Microbiology Fundamentals2 PDF

C HAPTER 4
Microbiology Fundamentals
15th International Biohydrometallurgy Symposium (IBS 2003)

September 14-19, Athens, Hellas
"Biohydrometallurgy: a sustainable technology in evolution"
A model for iron uptake in Acidithiobacillus ferrooxidans based

upon genome analysis
R. Quatrini1, F. Veloso1,3, E. Jedlicki2 and D.S. Holmes1,3*
1
Laboratory of Bioinformatics & Genome Biology, University of Santiago, Santiago,

Chile
2
Program of Cellular and Molecular Biology, ICBM, University of Chile, Santiago, Chile
3
Millenium Institute of Fundamental and Applied Biology, Santiago, Chile
Abstract
During growth at acid pH, Acidithiobacillus ferrooxidans is confronted with an
abundant supply of soluble iron. Therefore, it may have developed iron uptake
mechanisms quite distinct from those of neutrophilic organisms who are faced with the
problem of the uptake of insoluble forms of iron or of soluble iron at very low
concentrations. An analysis of the nearly complete genome sequence of A. ferrooxidans
reveals the presence of a number of potential genes and regulatory pathways involved in
iron uptake typical of those found in other organisms. However, in addition, bioinformatic
analysis suggests that A. ferrooxidans exhibits several unusual genetic features that may
reflect special requirements for iron uptake and homeostasis required for life at acidic pH.
Finally, an analysis of iron uptake mechanisms in A. ferrooxidans supports the growing
contention that the organism can live at a variety pHs between pH1 and, at least, pH 5.5.
Keywords: Acidithiobacillus ferrooxidans, iron uptake, iron regulation, FeoB, TonBdependent receptors, genome analysis
1.
INTRODUCTION
One of the critical problems confronting all organisms is the acquisition and uptake of
iron to fulfill metabolic requirements. Because iron is largely insoluble at neutral pH
under aerobic conditions, neutrophilic organisms have evolved a variety of sophisticated
iron uptake mechanisms. In addition, they have developed homeostatic controls to tightly
balance iron intracellular levels, guarding cell integrity against the deleterious effects of
excess iron (1,2).
The regulation of iron metabolism and its coupling with the regulation of defences
against oxidative stress must be strictly adjusted in response to environmental iron
bioavailability. Judging by the plethora of genes and response mechanisms involved, iron
Corresponding author: David Holmes: dsholmes2000@yahoo.com.

Work supported by Fondecyt No. 1010623 and the Millenium Institute of Fundamental and Applied
Biology, Santiago, Chile. We thank the Institute of Genome Research (TIGR) and Integrated Genomics, Inc.
(IG) for the use of their partial sequence of the Acidithiobacillus ferrooxidans genome.
989
metabolism and homeostasis appear to be a major concern for life in the presence of
oxygen.
We are interested in understanding how the bacterium Acidithiobacillus ferrooxidans
copes with iron uptake and homeostasis considering that it can grow aerobically at acidic
pHs in environmental situations where soluble iron is abundant. In addition, a special
feature of A. ferrooxidans, not found in the majority of organisms, is its use of iron as
energy and electron sources. This utilization must be balanced with the need for iron in
metabolism suggesting that A. ferrooxidans might exhibit novel iron regulatory
mechanisms not observed in neutrophilic heterotrophic organisms.
Contrary to what happens in most aerobic environments, where FeII rapidly
autooxidizes to FeIII which precipitates as ferric hydroxides, in acidic conditions FeII is
relatively stable and FeIII is much more soluble (>0.1M), exceeding by far the typical
requirements of bacteria of 10-8M (3,4). Given that A. ferrooxidans is usually confronted
with an abundant supply of both FeII and FeIII at acidic pHs, one can ask if it has evolved
novel mechanisms for Fe uptake that are not present or are substantially different from
those exhibited by neutrophilic organisms. One can also ask if this abundant supply of Fe
has necessitated the development of special or different mechanisms in A. ferrooxidans for
iron storage and for the avoidance of oxidative stress due to excess iron.
Studies concerning the strategies used by A. ferrooxidans for iron assimilation, iron
homeostasis and avoidance of the oxidative stress threat imposed by a unique iron-rich
and O2-plentiful environment are lacking and prompted the present investigation.
Knowledge of the physiology and metabolism of A. ferrooxidans has been impeded by the
paucity of classical genetic and molecular biological tools that have been so successfully
used to unravel the metabolism of other organisms. Due to its applied interest for
biomining and to circumvent this limitation, A. ferrooxidans was the first biomining
microorganism to have its genome almost completely sequenced (5) making it a candidate
for bioinformatic gene predictions and subsequent functional and metabolic model
building (6). In this paper, we present a partial bioinformatic survey of potential genes of
A. ferrooxidans involved in iron uptake. Candidate genes and pathways were first
identified by comparison to known iron uptake genes and pathways in other organisms.
Additional candidate genes were then identified by linkage analysis to the previously
known genes. Contributing to the elucidation of the physiological responses of A.
ferrooxidans is of fundamental interest and might, eventually, have a biotechnological
impact on the mining industry.
2.
METHODS
Known metabolic pathways of iron acquisition were obtained from BIOCYC
(http://biocyc.org:1555/META/server.html), KEGG (http://genome.ad.jp/kegg/kegg4.
html) and ERGO (http://wit.integratedgenomics.com/WIT2/). Amino acid sequences
derived from genes identified as being involved in iron acquisition were used as query
sequences to search the partial genome sequence of A. ferrooxidans ATCC 23270 in the
TIGR (http://www.tigr.org/) and ERGO data bases using TBlastN and BlastP respectively.
When a prospective candidate gene was identified in TIGR or ERGO its predicted amino
acid sequence was then used to formulate a BlastP (http://www.ncbi.nlm.nih.gov/BlastP/)
search of the non-redundant database at NCBI. Only bidirectional best hits were accepted
as evidence for putative homologs. Candidate genes and their translated proteins were
further characterized employing the following bioinformatic tools available in the web:
primary structure similarities (http://www.ebi.ac.uk/ClustalW/), secondary structure
990
predictions (HMM-based Protein Sequence Analysis http://www.cse.ucsc.edu/research/

compbio/HMM-apps/T99-query.html; JPred http://www.compbio.dundee.ac.uk/Software
/JPred/jpred.html), transmembrane predictions (http://www.ch.embnet.org/software/
TMPRED_form.html), motif predictions (http://www.blocks.fhcrc.org/, http://www.ebi.
ac.uk/interpro domain predictions (http://www.tolouse.inra.fr/prodom.htlm) and prediction
of protein localization sites (http://psort.nibb.ac.jp/).
3.
RESULTS AND DISCUSSION

Bioinformatic similarity searches for orthologous genes (evolutionarily homologous
gene with the same function in different species) in A. ferrooxidans involved in iron
uptake has revealed a number of candidate genes that can be assigned with a reasonable
level of confidence, due to their similarity with experimentally proven iron uptake genes
in other organisms (Table 1). Given the presence of such genes, metabolic models of iron
uptake in A. ferrooxidans can be inferred based upon known models in other organisms. A
summary of these results is shown in Figure 1. Additional work is in progress to identify
genes and pathways involved in iron sensing, homeostasis, storage and detoxification.
A. ferrooxidans encodes a putative FeII inner-membrane transport system termed
FeoAB (Figure 1A). FeoAB is present in most of the sequenced genomes but has been
observed to function only under anaerobic conditions where FeII is more stable and
soluble (7-9). As in other microorganisms, the feoB gene in A. ferrooxidans is adjacent to
a small open reading frame (ORF) termed feoA, of unknown function. Interestingly, in A.
ferrooxidans feoA and feoB appear to constitute an operon together with an ORF encoding
a putative porin (herein designated porA), which could facilitate FeII entrance in the
periplasm. Upstream of this porin is a putative Fur-box, suggesting typical iron-dependent
repression of FeII uptake mediated by a Fur-like repressor as has been demonstrated in
Escherichia coli (10). Bioinformatic analysis of the putative FeoB reveals the presence of
a highly conserved N-terminal GTPase motif and eight hydrophobic-rich transmembrane
segments supporting the contention that it is a membrane transport protein. In addition, the
second and third periplasmic loops have motifs similar to cytochrome CB5 and B6 which
could represent iron recognition or iron interacting portions of the protein.
Sequence comparison of the putative A. ferrooxidans FeoB ortholog and the other
three FeoB proteins so far characterized (GenBank accession #s:P33650, P74884,
D71909) reveals substantial amino acid sequence similarity except in the so-called
connector loop and in the C-terminal half of the protein. FeoB is an inner membrane
protein with several exposed loops facing the periplasm and it is these exposed loops that
harbour the major differences between the FeoB of A. ferrooxidans and the other three
orthologs of FeoB.
We speculate that these differences evolved as a response to the need for function and
structural integrity of the A. ferrooxidans FeoB in a low pH environment. In addition,
FeoB proteins of A. ferrooxidans and Helicobacter pylori have predicted periplasmic
loops with higher isoelectric points than E. coli. This could be a reflection of the
functional adaptation of these proteins to acidophilic or acid tolerant conditions. It has
recently been demonstrated that the FeoB of H. pylori plays an important role in iron
acquisition in the low-pH and low-oxygen environment of the stomach (7). We propose
that the FeoB of A. ferrooxidans may serve a similar function in iron uptake at low pH.
Inspection of the A. ferrooxidans genome also reveals the presence of several putative
outer membrane iron receptors (OMRs) (Figure 1B) with predicted affinities for either
hydroxamate-type siderophores or for FeIII-dicitrate. The degree of primary structure
991
similarity and the general accordance of secondary structure predictions between the FhuA
(FeIII-hydroxamate siderophores uptake) and FecA (FeIII-dicitrate uptake) of A.
ferrooxidans and E. coli highlights the conserved organization of the proteins. These
similarities include similar signal peptides, 2 to 4 N-terminal -helices, 22 -strands in the
putative -barrel domain and a TonB interaction C-terminal motif (11). Several of the
genes encoding OMRs are next to genes encoding TonB energy-transducing systems
(exbBDtonB) and some of them cluster together with all the necessary elements (OMR,
TonB system, ABC transporter components) to transport FeIII iron into the cytoplasm
(Figure 2) as observed in other organisms (12). The conservation of the siderophore
receptor proteins and organization of their genes into putative operons together with
relevant genes encoding other necessary components for iron uptake, allows us to propose
that mechanisms of siderophore uptake exist in A. ferrooxidans that are similar to those
found in other organisms.
Figure 1. Proposed model for iron uptake and storage A. ferrooxidans based upon a
bioinformatics analysis of its genome. (A) FeII uptake via an outer membrane porin
(black) and inner membrane FeoAB transporter complex (white). (B) FeIII uptake
via an outer membrane TonB dependent receptor OMR (black) and a periplasmic
siderophore binding and inner membrane uptake mechanism involving ABC
transporter Fec-like proteins (white) and a TonB energy transducing system (dark
gray). (C) Cytoplasmic FeII storage via a bacterioferritin-like protein Bfr.
Interestingly, the several OMRs of A. ferrooxidans exhibit a wide range of isoelectric
points (Table 1) that, by comparison to the OMRs of other organisms, would potentially
allow them to import iron at different pHs ranging from pH 1 to at least pH 5.5. This
supports the growing contention that A. ferrooxidans can live in various environments
with marked differences in pH (13, 14). It is not known how A. ferrooxidans might
regulate and coordinate the expression of these different OMRs to correspond to particular
environmental requirements for iron uptake.
An interesting aspect of the organization of the proposed FeIII-siderophore transport
operon in A. ferrooxidans is the presence of a small ORF potentially encoding a
hemoglobin-like protein (Figure 2). A bioinformatic analysis suggests that it belongs to a
992
family of genes known as truncated hemoglobins, identified in prokaryotes, protozoa,

algae and plants. The role of this hemoglobin-like protein is still unclear, although it has
been shown in plants to have low oxygen affinity and that the expression of its mRNA
decreases in microaerophilic conditions (15). The hemoglobin-like gene is associated with
other transporters annotated in genome database of ERGO in Burkholderia cepacia
(downstream of a putative molybdenum transport), Ralstonia eutropha (downstream of an
uncharacterized permease) and Sphingomonas aromaticivorans (JGI) (downstream of an
OMR). In addition to transporting and storing O2 and facilitating its diffusion, several
novel functions have emerged regarding hemoglobins, including control of nitric oxide
levels in microorganisms, binding and transport of sulfide in endosymbiont-harboring
species and protection against sulfide, scavenging of O2 in symbiotic leguminous plants,
O2 sensing in bacteria and archaebacteria, and dehaloperoxidase activity useful in
detoxification of chlorinated materials (16). In the case of A. ferrooxidans, its association
with a siderophore operon might represent a strategy to diminish the risks of O2 toxicity
caused by the entering iron until the metal has been targeted to its final destination.
Table 1. Identification and properties of candidate genes in A. ferrooxidans involved
in iron uptake. (A) name of genes involved in FeII uptake, FeIII-hydroxamate uptake
and FeIII-dicitrate uptake. (B) The % similarity of the predicted proteins from the
candidate genes to known proteins deposited in data banks. (C) lists the GenBank
accession number of the known proteins and (D) the organisms from which they are
derived. (E) lists the predicted location (loc) of the proteins where C: Cytoplasm,
OM: Outer Membrane and IM: Inner Membrane. (F) : predicted isoelectric point
(pI) of the mature proteins.
(A)
Gene
(B)
%S
(C)
Accesion #
(D)
Organism
(E)
Lo
(F)
pI
FeII
uptake
porA
feoA
feoB
42
55
53
ZP00053838
CAB49974
BAC09292
Magnetospirillum magnetotacticum
Pyrococcus abyssi
Thermosynechococcus elongatus
FeIII
hydroxamate
uptake
omr1
omr4
omr6
omr7
omr9
omr11
50
57
37
38
37
37
AAM72768
AAM43451
ZP00125990
ZP00003769
AAM35906
AAM42549
Chlorobium tepidum
Xanthomonas campestris
Pseudomonas syringae
Nitrosomonas europaea
Xanthomonas axonopodis
Xanthomonas campestris
OM
OM
OM
OM
OM
OM
6.98
8.18
9.02
9.15
6.53
8.63
FeIII
dicitrate
uptake
omr2
omr3
omr10
37
38
38
ZP00125403
ZP00125403
AAM35740
Xanthomonas axonopodis
OM
OM
OM
7.26
9.02
8.22
OM 5.85
C 11.14
IM 7.28
In accordance with what has been described for several siderophore transporting
genes in other bacteria this putative operon could be iron regulated, since a well conserved
Fur-box like sequence has been identified within the predicted promoter region of the
globin-like gene (unpublished data). The existence in A. ferrooxidans of an gene with 79%
similarity to the E. coli Fur ortholog (unpublished data) strengthens this interpretation and
promulgates further intriguing questions about iron regulation in this organism.
993
Figure 2. Gene organization of the proposed FeIII-siderophore uptake region of the

A. ferrooxidans genome. A: Globin-like gene; B: Outer Membrane TonB-dependent
Receptor I; C: TonB energy transducing system (exbB-exbD-tonB); D: ABC
transporter I Fhu-like proteins (perplasmic binding protein, permease, ATP-binding
protein); E: Outer Membrane TonB-dependent Receptor II and F: ABC transporter
II Fec-like proteins (perplasmic binding protein, permease, ATP-binding protein).
Also shown is the putative 19 nt Furbox. In the upper right corner is a scale bar in
base pairs (bp)
4.
CONCLUSIONS
A. ferrooxidans seems to have a complexity of iron uptake systems that rivals that
found in neutrophilic organisms. This argues against a simplistic view that A. ferrooxidans
has an easy time encountering and taking up readily available soluble iron in an acid
environment. It is possible that it requires this complexity despite the readily available
iron. However, a more reasonable argument is that it sometimes lives at higher pHs than
those typically associated with its growth in acid ferrous sulfate and sulfur, in which it
must scavenge less soluble forms of iron perhaps in strong competition with other
organisms. Notable is the plethora of siderophore uptake systems of A. ferrooxidans
characteristic of most organisms that have to scavenge and compete for poorly soluble
FeIII at neutral pHs. Demonstrations of the capacity of A. ferrooxidans to live at higher
pHs have appeared in the literature (13,14) and our findings, regarding iron acquisition
systems, support this point of view.
On the other hand, A. ferrooxidans iron uptake systems include components that seem
to be adapted for function at low pH, most obvious are the higher isoelectric points of the
periplasmic loops of membrane-associated iron uptake pumps. This characteristic is
expected for the maintenance of structural integrity and possible function of proteins
exposed to low pH.
The presence of a hemoglobin-like protein in A. ferrooxidans possibly associated with
a FeIII siderophore uptake operon is unusual, although descriptions of a similar protein are
now appearing in the literature. The best guess for its function is that it protects cells from
excess oxygen. Alternatively, and not mutually exclusive with this hypothesis, is the
possibility that it may measure oxygen levels and regulate genes in response to the
changes in oxygen concentration.
In this paper, we have focused on iron uptake mechanisms. Future studies will address
mechanisms to sense environmental levels of iron and the regulation of expression of iron
uptake and storage genes and how these might be coupled to genes involved in the
important capacity of A. ferrooxidans to oxidize iron for energy and electron gain.
Preliminary evidence points to the presence of a complex network of Fur regulated
operons and experimental work is underway to validate some of the bioinformatic models
we have developed.
994
REFERENCES
1. J.M. De Freitas and R. Meneghini. Mutat. Res., 475 (2001) 153.
2. D. Touati, M. Jacques, B. Tardat, L. Bouchard and S. Despied. J. Bacteriol., 177
(1995) 2305.
3. K. Hantke and V. Braun. In: Bacterial Stress Responses. (Chapter 19) G. Storz & R.
Hengge-Aronis. (Eds). ASM Press, Washington, D.C., 2000.
4. V. Braun and H. Killmann. Trends. Biochem. Sci., 24 (1999) 104.
5. E. Selkov, R. Overbeek, Y. Kogan, L. Chu, V. Vonstein, D.S. Holmes, S. Silver, R.
Haelkorn & M. Fonstein. Proc. Natl. Acad. Sci. USA, 97 (2000) 3509.
6. D.S. Holmes, J. Valdz, C. Dominguez, M. Barreto, C. Arriagada, S. Silver, S. Bueno
and E. Jedlicki. International Biohydrometallurgy Symposium (2001) 237.
7. J. Velayudhan, N.J. Hughes, A.A. McColm, J. Bagshaw, C.L. Clayton, C.S. Andrews
and D.J. Kelly. Mol. Microbiol., 37 (2000) 274.
8. R.M. Tsolis, A.J. Baumler, F. Heffron and I. Stojiljkovic. Infect. Immun., 64 (1996)
4549.
9. I. Stojiljkovic, M. Cobeljic and K. Hantke. FEMS Microbiol. Lett., 108 (1993) 111.
10. M. Kammler, C. Schn and K. Hantke. J. Bacteriol., 175 (1993) 6212.
11. R. Koebnik, K.P. Locher and P. van Gelder. Mol. Microbiol., 37 (2000) 239.
12. J.H. Crosa. Microbiol. Molec. Biol. Rev., 61(1997) 319.
13. N. Ohmura, K. Sasaki, N. Masumoto and H. Saiki. J. Bacteriol., 184 (2002) 2081.
14. M. Barreto, M. Rivas, E. Jedlicki and D.S. Holmes. International Biohydrometallurgy
Symposium (2003) (submitted, the Proceedings of this Congress).
15. R.A. Watts, P.W. Hunt, A.N. Hvitved, M.S. Hargrove, W.L. Peacock and E.S. Dennis
Proc. Natl. Acad. Sci. USA, 98 (2001) 10119.
16. R.E. Weber and S.N. Vinogradov. Physiol. Rev., 81(2001) 569.
995

Activity and occurrence of leaching bacteria in mine waste at

Cartagena, Spain, in the years 1991 until 2000
W. Sand, D. El Korchi-Buchert, T. Rohwerder*
Department of Microbiology, Institute for General Botany, University of Hamburg,
Ohnhorststr. 18, D-22609 Hamburg, Germany,
Tel/Fax: +49 40 428 16 423, e-mail: sand@mikrobiologie.uni-hamburg.de
Abstract
During the years 1991 until 2000 a total of 31 sampling campaigns have been
performed in the abandoned mining area between Portman and La Union near the city of
Cartagena in the province Murcia, Spain. Sampling was done at mine waste heaps near
and in the open pit Brunita, in the open pit Gloria, and at a rock formation with an
outcropping pyrite vein. More than 340 samples have been taken and analyzed for
numbers of leaching bacteria, microbial activity, and occurrence of sulfur compounds.
Besides, physicochemical factors such as pH, temperature, redox potential etc. were
measured. The statistical evaluation of all these data yielded the following results: (i) The
area is a typical habitat for acidophilic leaching bacteria. Consequently, a significant
positive correlation between the proton concentration and the numbers of leaching
bacteria, the concentrations of leaching products (iron ions, sulfate, and elemental sulfur),
and the microbial activity was obtained. (ii) The temporal development of leaching
activity indicates that in this climate biological leaching needs about 5 years to almost
fully oxidize all metal sulfides exposed at the surface (0-30 cm). (iii) Furthermore,
bioleaching is restricted to the humid season from November until April.
Keywords: AMD, open pit mining, mining waste, monitoring, microcalorimetry
1.
INTRODUCTION
Huge abandoned mining areas with waste heaps, open pits, and acidic lakes are the
unpleasant but often tolerated side effect of the metal winning activities of modern times.
Because these sites are often exposed to air and water and in most cases still hold
significant amounts of metal sulfides they tend to pollute rivers and groundwater with
acidic heavy-metal-containing solutions. This contamination, generally attributed to as
acid mine or rock drainage (AMD, ARD), is caused by the activity of acidophilc iron(II)and sulfur-oxidizing bacteria [1]. Therefore, risk assessment studies and rehabilitation
projects are only possible if information is provided on the parameters which determine
the bioleaching process. In particular, data on leaching kinetics and the overall temporal
development of the microbial flora in a special mining area have to be collected before
* Present address: UFZ Center for Environmental Research Leipzig-Halle, Permoserstr. 15, D-04318
Leipzig, Germany
997
starting rehabilitation activities. To meet these prerequisites we have studied for nearly 10
years the changes and interdependencies of leaching-relevant parameters in the abandoned
mining district between Portman and La Union in the southeast of Spain (province
Murcia). Sampling campaigns were performed at waste heaps near and in the open pit
mines Brunita and Gloria as well as at natural leaching sites (Fig. 1). The ore-rich region
near the city of Cartagena has suffered from human mining activities for thousands of
years. Thanks to the excellent natural harbor known as Portman (a contraction of the Latin
Portus Magnus) at the Mediterranean coast the richest ore deposits of the hills surrounding
this bay were already mined by the Romans for silver and lead. Intensive exploitation by
surface extraction technologies started in the late 1950s by S. M. M. Penarroya. During
the next 30 years waste materials from several open pit mines, ore processing, and
smelters were deposited as heaps or dumped in the bay of Portman [2]. The disposal in the
Mediterranean Sea was stopped in 1990, leaving the bay completely filled with 54 million
tons of heavy-metal-containing mining waste [2,3]. Finely, metal-winning activities
ceased with the closure of the open pit mine Los Blancos III in 1991 and the closure of the
Pb smelter near Cartagena in March 1992 [4].
Figure 1. Mining area between La Union and Portman in the Cartagena mountain
range (Autonomous Region of Murcia, southeast Spain). The dotted area marked the
abandoned mining district with the arrowhead pointing at the position of the open
pit mines Brunita and Gloria
2.
MATERIALS AND METHODS
2.1 Sample sites and sample analyses

The investigated area between La Union and Portman (Fig. 1) belongs to a geological
formation that contains Fe-, Pb-, Ag-, Cu-, Zn-, Mn-, and Sn-rich ore deposits [5].
Sampling campaigns were done at 4 sites: (a) A waste heap near La Esperanza to the north
of the open pit mine Brunita, which was piled up with material out of the mine as well as
998
with flotation residues; (b) the abandoned open pit mine Brunita; (c) a natural leaching site
at a rock formation with an outcropping pyrite vein, about 300 m to the south of Brunita;
(d) the abandoned open pit mine Gloria. Analyses of leaching parameters were performed
as described previously [6]. Briefly, sample humidity, temperature, pH, redox potential,
calorimetric activity, and conductivity were recorded. Furthermore, concentrations of
polythionates, sulfate, elemental sulfur, and iron ions were determined. For the assessment
of the aerobic microbial flora acidophilic iron(II) oxidizers, acidophilic sulfur oxidizers,
acidophilic chemoorganotrophs, neutrophilic sulfur oxidizers, and neutrophilic
chemoorgano-trophs were counted by MPN enumeration. In order to correlate the data of
sample analyses with the climatic situation in the province Murcia long-time mean values
of temperature and precipitation for San Javier were applied [7]. On the basis of this data
we were able to distinguish two seasons: winter covers the humid period from
November to April and summer covers the dry period from May to October.
2.2 Statistical methods
In a first statistical treatment the samples were compared with respect to all leaching
relevant parameters after dividing them in groups that had been classified with respect to
sample sites (see 2.1) and sample types (solid surface samples or from a depth of 20 to 30
cm, solid samples from the pyrite vein, lake sediment samples, and water samples). As the
collected data showed no normal distribution only non-parametric (i. e. distribution-free)
statistical tests could be applied. Multiple and bivariate group comparisons were
performed with the Kruskal-Wallis and Mann-Whitney U-test, respectively. In order to
elucidate the relationship between microbial activity and in situ conditions nonparametric, bivariate correlation analyses were performed (Kendall S-test). In this way
obtained results were tested with partial correlation analyses. Changes in the leaching
processes during the total campaign period were further tested by multiple linear
regressions. In this case, climatic influences were also considered by using the season at
which the samples were collected as a covariable. Generally, the significance level was set
for 5% and was adjusted to the number of statistical tests performed (Bonferronis
adjustment). The software SPSS 10.0.7 was used for the statistical analyses.
3.
3.1 Microbial activity

From previous studies it is already known that the microcalorimetrically measured
activity is one of the most reliable parameter in order to characterize a certain leaching
biotope [6,8]. A comparison of the microcalorimetric activity with the other parameters
recorded for the samples from the La Union mining district confirms this finding, too. The
mean activity of all samples was highest within the first 3 years of the campaigns (1991 to
1993) and then decreased continuously (Fig. 2). This temporal development of the
leaching activity can be explained with the closure of most open pit mines in the late
1980s and early 1990s. Obviously, in this mining area it takes about 5 years to almost
fully oxidize the metal sulfides exposed to the surface (0-30 cm). Samples which showed a
correlation between calorimetric heat production and cell numbers of leaching bacteria
had activity values strongly depending on the season at which the samples were collected.
As expected, the calorimetric activity was high in the humid winter season and low in the
dry summer time (Fig. 3A). In other words, significant biooxidation of metal sulfides at
the surface of the leaching sites (0-30 cm) only occurred between November and April. In
typical leaching biotopes the microbial processes are dominated by acidophilic bacteria.
999
Consequently, calorimetric values correlated negatively with the pH of the samples (Fig.
3B).
Figure 2. Temporal development of the microcalorimetrically measured activity of

samples collected from 4 different leaching sites at the La Union mining district
(sampling depth 0-30 cm)
Figure 3. Dependency of microbial activity of samples from the La Union mining

district (sampling depth 0-30 cm) on season (A) and pH (B)
3.2 Group comparisons and correlation analyses of other parameters
The interdependencies of the various leaching parameters were evaluated by group
comparisons and bivariate correlation analyses (see 2.2). Generally, it could be confirmed
that all sampling sites are typical acidophilic leaching biotopes (data not shown). Although
the total number of samples was comparatively low and the sampling sites consisted of
very heterogeneous materials a positive correlation was found between the proton
concentration and numbers of acidophilic leaching bacteria (iron(II)- and FeS2-oxidizing
bacteria). In contrast, but in conformity with the previous finding, the proton concentration
correlated negatively with cell numbers of neutrophilic bacteria (thiosulfate-oxidizing and
chemoorgano-heterotrophic microorganisms). As the leaching activity was highest in
1000
acidic samples (Fig. 3B) the amounts of leaching products (iron ions, sulfate, and
elemental sulfur) showed a highly significant correlation with the proton concentration.
4.
CONCLUSIONS
From the results of this study it can be concluded that rehabilitation projects should
start concomitantly with or immediately after mining, if climatic conditions are
comparable to the La Union district. In this region leaching bacteria oxidizes surfaceexposed metal sulfides (0-30 cm) almost completely within 3 to 5 years. Consequently,
AMD/ARD can only be stopped if mitigation measures accompany the piling of waste
heaps and the other depositing techniques. On the other hand, it is clearly shown that the
simple closure of open pit mines is not an adequate solution for the problem. The
consequences of this inactivity are acidic heavy-metal-contaminated soil and surface
water. This contamination is build up in a few years but will leave an abandoned mining
area unsuitable for many years with respect to farming and urban development as well as
for tourism.
ACKNOWLEDGMENTS
The help of J.I. Manteca (Universidad Politcnica de Cartagena) in providing detailed
information about the mining district La Union and the help of R. Buchert (University
Hospital Eppendorf, Hamburg) with the statistical analyses are greatly acknowledged.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
H. L. Ehrich, Geomicrobiology, Marcel Dekker, New York, 2002.

J. M. M. Orozco et al., Landscape Urban Plann., 23 (1993) 195.
C. Auernheimer and S. Chinchon, Environ. Geol., 29 (1997) 78.
S. Moreno-Grau et al., Atmos. Environ., 34 (2000) 5161.
I. S. Oen et al., Econom. Geol., 70 (1990) 1259.
A. Schippers et al., Appl. Environ. Microbiol., 61 (1995) 2930.
http://www.inm.es/wwc/html/dclimat/SAN_JA.html
W. Sand et al., Appl. Microbiol. Biotechnol., 40 (1993) 421.
1001

An AFM-study on the adhesion of Acidithiobacillus ferrooxidans

and Leptospirillum ferrooxidans to surfaces of pyrite
K. Kinzlera, W. Sanda*, J. Telegdib and E. Kalmanb
a
Institut fr Allgemeine Botanik, Mikrobiologie, Universitt Hamburg,

Ohnhorststrasse 18, D-22609 Hamburg, Germany
b
Hungarian Academy of Sciences, Chemical Research Center,
Pusztazeri ut 59-67, H 1025 Budapest, Hungary
Abstract
The attachment of Acidithiobacillus ferrooxidans and Leptospirillum ferrooxidans to
pyrite surfaces has been analyzed using atomic force microscopy (AFM). Cells of both
species attach preferentially to sites with visible surface defects. Although a part of the
cells from the inoculum was still planktonical, a complete coverage of the pyrite surface
never became detectable within 24 hours of attachment experiments. Only after the onset
of growth, the surface became fully EPS-covered and, thus, turned into an "organic"
surface. The extracellular polymeric substances (EPS) fill the space between the bacterial
cell (wall) and the substratum surface. Furthermore, the EPS extend beyond the cell body.
Especially the strain of L. ferrooxidans exhibited strong EPS-production and,
consequently, an enormous surface coverage. Due to this coverage, the cells have an
enlarged metabolic radius-of-action. Whereever EPS are in contact with the surface, the
metal sulfide dissolution will be enhanced, due to the EPS-complexed iron(III)ions.
1.
INTRODUCTION
Bioleaching is the bacterially mediated dissolution of metal sulfides to sulfuric acid
(sulfate) and metal cations. This process is known for decades and has been shown to be
of industrial importance as well as of environmental significance (acid rock drainage =
ARD) [1,2,3,4]. Since its discovery there has been a debate about the mechanism, by
which the microorganisms dissolve metal sulfides. Generally, a direct and an indirect
mechanism are proposed. The argument pro and contra continue still nowadays [5,6,7].
Recent work of our group indicated that bioleaching is a process, which combines
characteristics of both mechanisms [8,9,10]. This means that the dissolution of a metal
sulfide is a chemical process caused by oxidative attack of iron(III)ions and/or protons.
The iron(III)ions may be involved in any dissolution of metal sulfides, whereas protons
are restricted to those metal sulfides, which are amenable to a hydrolytic dissolution. The
group of metal sulfides, which obligately need iron(III)ions for a dissolution, comprise
pyrite (FeS2), molybdenite (MoS2), and tungstenite (WS2). Their characteristic is the fact
that the outer electron orbitals result from the metallic part of the compound, and not from
the metal-sulfur bond. Thus, the oxidative attack by iron(III)ions is a requirement. The
other metal sulfides have outer orbitals from the sulfur-metal bond and, thus, are
attackable by protons and by iron(III)ions. In the presence of the latter the rate of
1003
dissolution may become considerably enhanced, which often causes erroneously the
statement of a direct mechanism [11].
Furthermore, bioleaching is caused mainly by attached cells [8,10]. The planktonic
population plays a negligible role under most conditions. Only in case of a high
concentration of dissolved iron(III)ions planktonic cells could contribute. The iron(III)ions
are provided by the leaching bacteria complexed in their extracellular, polymeric
substances (EPS). The complex most likely consists of two moles of glucuronic acid and
one mole of iron(III)ions. Both, At. ferrooxidans and Leptospirillum ferrooxidans, have
these glucuronic acid - iron(III)ion complexes in their EPS. By an electrostatic interaction
between the negatively charged pyrite surface (pyrite assumes a negative charge at pH
below 3) and the net positive charge of the iron(III)ions-containing EPS, surface
attachment of the cells becomes possible as the primary event in adhesion. Whether in
later stages additional factors like pili etc. function, remains an open question. Some
indications exist that such interactions may contribute also to the formation of robust
surface attachment. Another open question is the site for attachment, whether it is a
randomly or a deliberately chosen place. Some considerations seem to suggest that in case
of bioleaching bacteria random attachment does not play the major role. Especially
chemotaxis seems in this context to be a very important surface site recognizing system
[12,13].
The present work was executed to visualize the attachment behavior of cells of two of
the main bioleaching bacteria, At. ferrooxidans and L. ferrooxidans. The idea was to find
criteria to judge the importance of EPS for attachment, to clarify the attachment
mechanism, and to obtain information about the characteristics of the attachment site. For
this purpose pure cultures of these species, which previously had been selected for good
attachment properties, and coupons of selected pyrite cubes were incubated jointly. The
coupons were analyzed for changes of the surface morphology in time-dependent adhesion
and biofilm formation experiments using atomic force microscopy (AFM). AFM, a variety
of tunneling scanning microscopy, allows to record the height profiles of plain surfaces by
a surface scanning mechanism, and to put this together to produce a topographical map
AFM-images of the investigated sample. Since this scanning action can be performed in
air or water, living cells may be investigated without any interference by vacuum etc.
Thus, the big advantages of this technique are the simple sample preparation and the
measuring conditions at atmospheric pressure and room temperature. Consequently, for an
investigation under natural conditions, AFM is the method of choice. The application of
AFM for monitoring of microorganisms has been described in several papers [14,15,16].
2.
2.1 Strains
For these investigations the strains At. ferrooxidans strain R7 and L. ferrooxidans
strain R3 were selected. They are kept in our culture collection in nutrient solution at pH
1.9 [17].
2.2 Pyrite sample preparation
Pyrite was obtained from mineral shops as cubes of at least 1 cm volume. To obtain
suitable samples for AFM investigations, the cubes were cut by a diamond saw into plates
of about 2 mm thickness. Afterwards, the plates were washed with ethanol and
hydrochloric acid to remove any organic contaminants and/or oxidation products like
iron(II/III)ions, elemental sulfur, or polythionates. This procedure was followed by
1004
sterilization in 70% ethanol for 12h. These plates were introduced into the flasks with the
bacterial cultures.
2.3 AFM experiments
For experimental purposes, 500 mL cultures were grown, harvested by centrifugation
after two-thirds of the iron(II)ions had been oxidized, and resuspended in 100 mL
nutrient-free solution ([17], without iron(II)ions). After a period of 24h for substrate
depletion (to allow the cells to oxidize any remaining substrate), pyrite samples were
introduced into the cultures and incubated for periods of up to 96h with stirring. At
different times pyrite samples were removed and analyzed using the AFM. The atomic
force microscope (NanoScope III, Digital Instruments, Veeco, USA) was operated in air in
contact mode. A silicium nitride tip (SiN4) was used for surface scanning. All figures
contain images, which have been obtained without any processing except autoflattening.
3.

A sterile pyrite surface prior to any attachment experiment is shown in Figure 1. The
various layers besides minor surface defects are clearly visible on the (typical) pyrite
surface.
Figure. 1. AFM-image of a sterile pyrite surface with a typical layered surface

structure and two inorganic crystals
2.1 Attachment of At. ferrooxidans strain R7 to pyrite
Attachment experiments were performed with At. ferrooxidans and immersed pyrite
plates. The following figures show the patterns of surface colonization by cells of At.
ferrooxidans R7. For evaluating the question for preferential attachment, especially
images in the time range of 2 to 8 h are useful. Such an image is given in Figure 2. Images
obtained lateron yielded only equivocal information due to growth effects. It becomes
obvious from Figure 2 and similar images (not shown) that the cells attached preferentially
to sites with visible surface defects. More than 80% of all cells were attached to
distortions of the pyrite surface (grooves, channels, edges of layers etc.) Images obtained
at a later stage of the surface colonization process prove the growth of the attached
bacteria (Fig. 3). Figure 4 demonstrates the existence of pits underneath bacteria. After
removal of the cells, pits with the shape and size of cells become visible. Besides the
bacteria footprints of cell-free EPS may remain on the surface (Fig. 5). Typical patterns of
time-dependent surface colonization are presented in Figure 6A-D. Within 6 h several
1005
cells have attached to the pyrite surface. The amount of attached cells increases up to 24 h.
Afterwards, until 48 h the amount of attached cells remains almost similar. One week
later, the surface morphology exhibits considerable alteration. Only a few single cells
remain visible. The majority is embedded in copious amounts of an amorphous mass, the
EPS (measurements in the friction mode of the AFM proved that the amorphous mass is
soft, whereas the free pyrite surface remained rough). The EPS cover a considerable part
of the total surface. In Figure 7 a scheme of a surface-attached, bioleaching cell of At.
ferrooxidans is shown.. It combines all our evidence collected from previous work [8] and
from AFM images. Most important is the finding that the EPS extends over the limits of
the cellbody and in this way considerably enlarge the metabolic radius-of-action.
Figure 2. AFM-image of a pyrite

surface after 8 h of incubation with
Acidithiobacillus ferrooxidans strain
R7. Cells attached preferentially to
sites with visible surface defects
Figure. 3. AFM-image of a pyrite

surface after 24 h of incubation with
R7. Pit formation is visible in the
centre of the image
Figure 4. AFM-image of a cleaned

pyrite surface after 24 h of
incubation
with
cells
of
R7
Figure 5. AFM-image of a single cell of

R7 attached to pyrite after 24 h of
incubation. Around the cell cell-free
exopolymers (footprints) are visible
1006
Figure 6A-D. Typical patterns of time dependent surface colonization by

Acidithiobacillus ferrooxidans strain R7 over the course of 8 d
Cell
A
cell of Acidithiobacillus
ferrooxidans
A
3+
S
2+
Fe S2O32-
attack
Figure 7. Schematical graph of a cell of Acidithiobacillus ferrooxidans attached to

pyrite. Included is a model for the mechanism of the indirect or contact-mode
leaching attack as catalyzed by A. ferrooxidans
1007
2.2 Attachment of Leptospirillum ferrooxidans strain R3 to pyrite

The same experiments, as shown for At. ferrooxidans, were performed with cells of L.
ferrooxidans strain R3. In Fig. 8 an image with a few cells of L. ferrooxidans attached to a
pyrite surface is shown. In the middle of the image a dividing cell or two recently divided
cells are visible, which are jointly embedded in their EPS. The cells are surrounded by
copious amounts of EPS. The amount of EPS is considerably higher than the amount of
the EPS of At. ferrooxidans strain R7 (Fig. 5). Furthermore, the EPS cover a considerably
larger part of the surface than the naked cells would be able to. The surface coverage is in
principle similar to the one shown in Figure 7 for At. ferrooxidans. Time-dependent
adhesion is shown in Fig. 9A-C. In the course of the colonization a serious change of the
surface morphology becomes visible. After 6 h of incubation (Fig. 9A) several cells have
attached to the pyrite surface and excreted some EPS around the cells. After 24 h (Fig.9B)
the morphology already has changed considerably. Single cells with their surrounding
EPS are still identifiable, although the EPS starts to merge together. After 48 h (Fig. 9C)
the surface of the pyrite is covered by an almost homogenous film of EPS (sea of EPS)
with some embedded cells. Consequently, the structure of the pyrite is not visible
anymore. The presence of the EPS again was proven by experiments using the frictionmode of the AFM (data not shown).
Figure 8. AFM-image of cells of Leptospirillum ferrooxidans attached to pyrite after

24 h of incubation
Figure 9A-C. AFM-image of cells of Leptospirillum ferrooxidans attached to pyrite

after 6 (A), 24 (B) or 48 (C) h of incubation
1008
3.
CONCLUSIONS
Summarizing, it is obvious that an attachment of these bioleaching bacteria takes
place at selected sites. These sites have a surface defect, which most likely causes a high
surface charge by electrical imbalance. This agrees well with the hypothesis of an
electrostatic interaction mechanism between EPS and pyrite surface. It also becomes clear
from these images that the EPS are always the compounds mediating the contact between
a bacterial cell and pyrite.
Once the cells are attached, they start to grow and, within a few days, may cover the
surface with large amounts of EPS. Consequently, the previously inorganic surface turns
to an organic one. It remains to be elucidated, what this means for further attachment.
Preliminary data indicate that this organic surface film hinders other bacteria to attach,
even those of the same strain. In any case, the EPS with their complexed iron(III)ions are
able to oxidatively attack the pyrite. Thus, the cells radius-of-action becomes considerably
larger than the dimensions of an EPS-free cell would indicate. Especially for L.
ferrooxidans, whose only substrate are the low-energy iron(II)ions, this enlarged radiusof-action might confer an improved environmental competitiveness.
REFERENCES
1. A. Schippers and K. Bosecker, Z. Angew. Geol., 48 (2002) 38.
2. D. Fortin, B. Davis and T.J. Beveridge, FEMS Microbiol. Ecol., 21 (1996) 11.
3. K.J. Edwards, T.M. Gihring and J.F. Banfield, Appl. Environ. Microbiol., 65 (1999)
3627.
4. N. Kuyucak, Min. Environ. Managem., 11 (2001) 12.
5. W.Sand, T. Gehrke, R. Hallmann and A. Schippers, Appl. Microbiol. Biotechnol., 43
(1995) 961.
6. M.P. Silverman, J. Bacteriol., 94 (1967) 1046.
7. F.K. Crundwell, In: V.S.T. Ciminelli and O. Jr. Garcia (eds.), Biohydrometallurgy:
Fundamentals, Technology and Sustainable Development, Elsevier, New York, 2001,
Part A, p. 149.
8. W.Sand, T.Gehrke, P.-G.Jozsa and A. Schippers, Hydromet., 59 (2001) 159.
9. T. Rohwerder, P.-G. Jozsa, T. Gehrke and W. Sand, In: G. Bitton (ed.) Encyclopedia
of Environmental Microbiology, Vol 2, Wiley, New York, 2002, p. 632.
10. D.E. Rawlings, Annu. Rev. Microbiol., 56 (2002) 65.
11. O.H. Tuovinen, In: H.L. Ehrlich and C.L. Brierley (eds.) Microbial Mineral Recovery,
McGraw-Hill, New York, 1990 p. 55.
12. J. Acuna, J. Rojas, A.M. Amaro, H. Toledo and C.A. Jerez, FEMS Microbiol. Lett., 96
(1986) 37.
13. G.Meyer, T. Schneider-Merck, S. Bhme and W. Sand, Acta Biotechnol., 22 (2002)
391.
14. Y.F. Dufrene, Ch.J.P. Boonaert, H.C. van der Mei, H.J. Busscher and P.G. Rouxhet,
Ultramicroscopy, 86 (2001) 113.
15. Y.F. Dufrene, Micron, 32 (2001) 153.
16. J. Telegdi, Zs. Keresztes, G. Plinks, E. Kalman and W. Sand, Applied Physics A, 66
(1998) S639.
17. M.E. Mackintosh, J. Gen. Microbiol., 105 (1978) 215.
1009

An X-ray photoelectron spectroscopy study of the mechanism

of microbially assisted dissolution of chalcopyrite
A. Parkera, C. Klauberb*, M. Stottb, H.R. Watlingb and W. van Bronswijka
a
A.J. Parker Cooperative Research Centre for Hydrometallurgy, School of Applied

Chemistry, Curtin University, P.O. Box U1987, Perth, WA 6845, Australia
b
A.J. Parker Cooperative Research Centre for Hydrometallurgy, CSIRO Minerals, P.O.
Box 90, Bentley, WA 6982, Australia
Abstract
The acidic oxidative dissolution of samples of p-type chalcopyrite (Mt Isa Mines
flotation concentrate and massive Mt Lyell) via the iron- and sulphur-oxidising microbes
Sulfobacillus thermosulfidooxidans (50C) and Sulfolobus metallicus (65C) has been
examined using X-ray photoelectron spectroscopy. This enabled surface speciation,
encompassing sulphides, disulphides, elemental sulphur, sulphate and various jarosites,
during the main dissolution stages. When compared to the case for the abiotic dissolution
of chalcopyrite conducted under similar chemical conditions, both the similarities and
differences can be highlighted. Principal differences between abiotic and microbially
assisted dissolution arise from the microbial tendency to oxidise elemental sulphur and the
effects of microbial nutrient media toward the type of jarosite formed. The role of the
surface sulphide dimer S22- in the sulphide oxidation has also been clarified.
1.
INTRODUCTION
The goal of complete copper recovery from chalcopyrite via low energy
hydrometallurgical processing is critically dependent upon the control of "passivating"
overlayer structures. The body of work previously undertaken by the authors (Klauber et
al., 2001; Klauber, 2003; Parker et al., 2003), concentrating on the abiotic chemistry of Mt
Isa Mines (MIM) and Mt. Lyell p-type chalcopyrites, for acidic sulphate oxidative
dissolution, shows related behaviour over the thermophile temperature range and describes
in detail a surface specific mechanism for abiotic chalcopyrite dissolution. The typical
behavioural traits are oxidative dissolution to produce iron and copper in solution, with the
sequential oxidation of the sulphide to disulphide, elemental sulphur and sulphate. The
latter three all contribute to the initial surface reaction phase. Eventually the sulphate
dominates, with a thin ferric sulphate precursor phase forming. With continued
dissolution, the ferric solution concentration reaches a point where massive hydronium
jarosite layers precipitate, hindering further chalcopyrite attack. Note that varied jarosites
(KFe3(SO4)2(OH)6) are possible with cation substitution of the K+ by H3O+, NH4+ etc. The
* Corresponding author. PO Box 90, Bentley, WA 6982, Australia, e-mail: craig.klauber@csiro.au
1011
initial ferric sulphate layer is observed to act as the crystallization precursor for jarosite
precipitation. Iron control appears to be the key to chalcopyrite hydrometallurgy.
Elemental sulphur plays a role but does not appear to be a significant candidate for
hindered dissolution. Despite thorough investigation, no spectroscopic evidence was found
for stable forms of the previously alluded to polysulphides (Parker et al., 2003). This work
places the abiotic chemistry into context with bioleaching in the presence of two microbial
strains. Particular aims were to examine the mechanism of sulphur oxidation further and
jarosite composition as a function of microbially assisted leaching conditions.
2.
EXPERIMENTAL
2.1 Microbial culture

Sulfolobus metallicus (DSM 6482T) and Sulfobacillus thermosulfidooxidans (DSM
9293T) were obtained from the German Collection for Microorganisms And Cell Cultures
(DSMZ). Sulfolobus metallicus was cultured in a nutrient medium containing (NH4)2SO4
(1.3 g/L), KH2PO4 (0.28 g/L), MgSO4.7H2O (0.25 g/L) and S. thermosulfidooxidans in a
nutrient medium containing (NH4)2SO4 (3.0 g/L), K2HPO4 (0.5 g/L), MgSO4.7H2O (0.5
g/L), KCl (0.1 g/L), Ca(NO3)2 (0.01 g/L). Both media were adjusted to pH 2.0 with
concentrated H2SO4. The media were autoclaved at 121C (2 atm) for 15 minutes.
Sulfolobus metallicus runs were also conducted in potassium free media where K2HPO4
was replaced with (NH4)2HPO4 (0.27 g/L) and (NH4)2SO4 was reduced to 1.0 g/L. Yeast
extract (0.02% (w/v)) was added aseptically to both media (via passage through a 0.22 m
pore size membrane). The chalcopyrite concentrate (CuFeS2 88%, FeS2 4% and gangue
minerals, mainly quartz, 8%) was wet sieved to a size fraction of +20 m 35 m, and
then dried in an oven at 30C. Residual xanthates were then washed from the concentrate
(Klauber et al., 2001). The cultures were grown at 65C (S. metallicus) and 50C (S.
thermosulfidooxidans) in an environmental incubator shaker operated at 180 rpm.
2.2 Leaching method
The microbially assisted leaches for the concentrate and the massive chalcopyrite
were conducted in different experimental apparatus. The concentrate was leached in
simple flask arrangements in an environmental incubator shaker. A 5 mL volume of
solution containing S. metallicus or S. thermosulfidooxidans (late exponential phase, ~72
hours post-inoculation) was inoculated into a series of 250 mL Erlenmeyer flasks (one for
each leach period), containing 95 mL of appropriate nutrient medium and 3.1 g
concentrate.
Samples of the massive chalcopyrite were silver dagged to sample stubs, dry polished
with 2000 grit carbide paper and dipped onto an overflowing leachant meniscus; surface
tension ensured contact only with the chalcopyrite surface. The leachant cell, containing
25 mL of inoculum and 0.03% (w/v) yeast extract in 500 mL of nutrient media, was run
for 12 hours with 16.0 g concentrate prior to exposure to the massive chalcopyrite.
2.3 Solution sampling and analysis
The pH and redox potential (Ag|AgCl) were measured periodically throughout the
experiments. For the concentrate leaches, concentrations of ferrous ion, total sulphur,
copper and iron in solution were determined after centrifugation to separate the solid
material. Ferrous ion concentrations in solution were determined by spectrophotometry
using the method of Wilson (1960). The total sulphur, iron and copper concentrations in
1012
solution were measured using Inductively Coupled Plasma - Atomic Emission

Spectrometry (ICP-AES).
2.4 X-ray photoelectron spectroscopy
At selected reaction intervals (12, 24, 36, 60 and 108 hours), the contents of a single
flask were filtered (0.22 m pore size membrane) and then washed with 10 mL of
acidified MilliQ water (H2SO4 to pH 2) to remove soluble metal species. A portion of the
solid material retained on the filter was then pressed into a sample stub and transferred
into the photoelectron spectrometer and cooled to ~150K to minimise sulphur
volatilisation. For the massive chalcopyrite, the surface was examined at shorter intervals
(0.5, 2, 24, 48 and 72 hours). These pre-mounted samples went directly into the
spectrometer after rinsing with acidified MilliQ water (H2SO4 to pH 2). XPS spectra were
obtained with either Mg K or Al K radiation (300 W) from a conventional source in a
VG ESCALAB Mk II fitted with 5-channeltron electron detection. Spectra were run at
base pressure of ~1x10-10 mbar, using 6 mm analyser slits and a fixed analyser pass energy
of 20 eV. The S 2p, Cu 2p, Fe 2p, O 1s, N 1s, K 2p, C 1s, K 2p, valence band, Fe 3s and
full survey regions were acquired in sequence. Methods of data reduction employed such
as X-ray line function removal (source function and satellites), spin-orbit splitting removal
(in the case of S 2p) and comments on the peak fitting are covered in detail in earlier
papers in the series (Klauber et al., 2001; Klauber et al., 2003; Parker et al., 2003).
3.
3.1 Solution conditions

Metal species solution concentration and redox potential for the concentrate cases are
shown in figure 1 and are typical for the microbially assisted ferric acid leach conditions
employed. As would be expected, Eh rises and solution ferrous concentration drops as the
leachant activity and microbial numbers increase. Also note the low concentrations of
soluble iron as compared to the leached copper (especially in the S. metallicus- assisted
leach) indicating substantial jarosite formation. The pH remained constant at ~2 until
extensive jarosite formation occurred. Note that a rate comparison to abiotic conditions is
not simple. In abiotic dissolution the initially high ferric level contrasts to a low value in
the microbial case and the microbial catalytic role of ferrous to ferric conversion is also
absent.
3.2 Photoelectron spectra of the inorganic phases
The matrix of four leach experiments (two microbes and two chalcopyrite types) over
a number of sampling periods and the acquisition of multiple spectral lines allows only a
few of the spectra to be shown. For clarity, the composite XPS spectra have arbitrary
shifts in the ordinate position to avoid overlap. Spectra at increasing time intervals are
arranged from top to bottom. For thin surface reaction layers, exhibiting negligible static
charge shifts, the binding energy (BE) scales are referenced to the underlying chalcopyrite
lattice sulphide species S2- at 161.15 eV (Klauber et al., 2001) to correct for spectrometer
drift. This is important for the S 2p analysis. Figures 2 and 3 show the typical
(unprocessed) S 2p data for the initial stages of the oxidation process for the concentrate.
Note that once deposition of thick non-conducting overlayers has occurred there is an
onset of static charging (typically 5-7 eV to higher apparent BE values). This has not been
corrected for in the figures so as to illustrate that critical layer formation. For example, the
raw sulphate peak in figure 2 can be seen to shift from 169.1 to 175.7 eV.
1013
Figure 1. Solution concentration of copper and iron and Eh variation during

concentrate leaching
Figure 2. Sulphur 2p spectra from concentrate leached in the presence of S.

thermosulfidooxidans at 50C for periods varying from 12 to 108 hours
1014
Figure 3. Sulphur 2p spectra from concentrate leached in the presence of S.

metallicus at 65C for periods varying from 12 to 108 hours in nutrient media with
and without K+
Due to the difficulty in reaction rate control for the microbially assisted leaches,
caution needs to be taken in comparing the observed times for surface phase formations.
Qualitatively, the same surface phase components form on both the MIM concentrate and
the Mt. Lyell massive samples, but they do not necessarily correlate in terms of time. This
is a consequence of differences in the experimental and sample set-ups (high surface area
concentrate versus low surface area massive). Even within individual repeat leach runs
exhibiting the same sequence of phases, the onset time for a given part of the sequence
could be at variance by hours. This is despite care in reproducing both inoculum activity
and the chalcopyrite sample. Other differences between the concentrate and massive relate
solely to particle size, e.g. the deposition of jarosite related layers. For the concentrate, the
static charge shifts are sudden as the particles are small and such layers reduce electrical
conductivity between particles and the particles and earth (i.e. multiple insulating layers).
By comparison the same relative jarosite levels on the massive material (in fixed contact
with earth) resulted in negligible shifts. This does not indicate a difference in chemistry.
As a diagnostic tool the Cu 2p spectrum is of limited value as all the spectra (e.g.
figure 4) are typically representative of Cu[I] for the unoxidized chalcopyrite lattice,
diminishing in intensity as the depositing overlayers attenuate the photoemission. The Fe
2p spectrum (figure 5) is more sensitive as iron species participate in the formation of
coherent overlayers, starting with pre-cursor iron sulphate acting as a nucleation site for
jarosite deposition. Analysis of the Fe 2p states via peak fitting is not practical, though a
least squares principal component analysis (Parker et al., 2003) can reliably indicate the
overlayer composition of iron containing phases.
1015
Figure 4. Copper 2p spectra from concentrate leached in the presence of S. metallicus

at 65C for periods varying from 12 to 108 hours in nutrient media with and without
K+
Figure 5. Iron 2p spectra from concentrate leached in the presence of S. metallicus at

65C for periods varying from 12 to 108 hours in nutrient media with and without K+
The S 2p peak remains the most useful means of understanding the oxidation process.
Figure 6 shows a detailed sulphur analysis of the initial oxidation in terms of states largely
already established (Klauber et al., 2001, Klauber 2003, Parker et al., 2003). Once the
state of reaction has progressed to significant jarosite formation, the reduced emission
1016
intensity from the mixed states region below the sulphate makes their analysis less
reliable. The component that evolved at a higher BE than the sulphate had not been
previously observed. It appeared to be characteristic of the high initial sulphate levels for
the concentrate, but not the massive. It is most likely to be sulphate emission from
nucleating jarosite crystals. This may be an illustration of differential static charging on
the surface or a genuine BE shift for conversion from a simple sulphate to one
characteristic of a jarosite.
Also noteworthy in figure 6, although the sulphide peak has been aligned to the lattice
value of 161.15 eV, as the elemental sulphur evolves the measured sulphide feature
actually shifts to a higher BE by ~0.13 eV whilst the sulphate BE remains approximately
constant. At the same time the loss feature (Klauber, 2003) diminishes in intensity. Both
this and the sulphide shift may be real and may point to a breakdown of the near surface
chalcopyrite lattice. There are two factors influencing observed BE. Based on the Muliken
charge population analysis (Hamajima et al., 1981) the sulphide donates charge toward the
iron, so a lattice breakdown to fully ionic S2- should cause a shift to lower BE. Countering
that, the simultaneous breakdown of the conduction band would mean less relaxation and
a shift to higher BE.
Figure 6. Detail of the sulphur chemical state analysis for processed spectra showing
the variation in states during the initial leaching. For figure clarity neither the final
fits nor the residuals are shown
1017
Comparison of the S 2p at 60 hours with the S. thermosulfidooxidans to that for

abiotic dissolution at 50C after 2 hours (Parker et al., 2003) shows a very similar
structure. Whilst the abiotic response appears "faster" (due to a higher initial ferric
concentration), the microbially assisted creates more surface sulphate and appears to allow
a more concerted attack on the sulphur dimer species by the ferric. The likely reason for
this is the direct microbial attack on the elemental sulphur, a pathway absent in the abiotic
case.
3.3 Photoelectron spectra of the entrained biomass
Any entrained or adherent biomass will contribute to the observed photoelectron
signal, principally for the C, N and O 1s peaks and to a lesser extent P 2p. In this study it
exhibited mainly in the C 1s and N 1s peaks; the O 1s was overwhelmed by the inorganic
phases. In terms of a microbial cell wall characterization, the use of XPS has generally
been limited to a cluster analysis based on nett N/C, O/C and P/C ratios (Van der Mei et
al., 2000). This is pragmatic as it circumvents applying detailed peak analysis that can be
especially difficult for complex organic systems. The current microbial studies do produce
some differences to the prior abiotic chemical investigations. This is due to the biomass
and more especially its nutrient media, e.g. ammonium ion presence. In this work, some
knowledge of the expected biomass XPS spectrum, especially the N 1s, is needed to
compare the leaching regimes, although the role of XPS in directly increasing microbe
knowledge is limited.
Based on the generalized chemical compositions of cells (Neidhardt et al., 1996) the
N 1s signal would be expected to primarily originate from proteins (~55%) with the RNA
component also a major contributor (~21%). As the backbone structures for proteins are
polypeptides, the basic N 1s signal for the protein should be similar to that observed for
amide linkages R-CO.NH-. Purely from a photoelectron BE viewpoint (i.e. chemical
environment), synthetic polymers such as nylon are polypeptide analogues. This provides
a guide to expected BE values (Beamson and Briggs, 1992):
polymer
R C1s
amide C 1s
amide N 1s
amide O 1s
nylon 6,6
285.00 - 286.02 eV 288.02 eV
399.81 eV
531.37 eV
nylon 6
285.00 - 286.00 eV 288.01 eV
399.77 eV
531.35 eV
nylon 12
285.00 - 285.90 eV 287.97 eV
399.84 eV
531.33 eV
Whilst the side groups on the polypeptide backbone define the protein type, it would
be expected that the amide nitrogen atom would dominate the overall N 1s signal. Hence a
BE of about 399.8 eV should be expected (confirmed by an examination of Escherichia
coli biomass). Carrying the synthetic polymer analogy further, the aromatic nitrogen
atoms in the heterocyclic RNA bases would be expected to have a BE value about 399.3
eV, i.e. 0.5 eV below that of the amide nitrogen atoms (Beamson and Briggs, 1992).
Hence two nitrogen peaks with a small separation would be expected for the biomass.
As can be seen from figure 7, two N 1s states are observed on the chalcopyrite
sample. Although the BE value of 1.9 eV is well above the expected amide-RNA
separation, it perfectly matches the separation between that of ammonium jarosite
standard and the biomass signature from E. coli. The jarosite was BE corrected for the
sulphate component to match that of jarosite on massive chalcopyrite and for E. coli the C
1s was fixed to 285.0 eV. The higher BE component was also noted to vary
disproportionately in relative intensity and found to dominate at long leach times with
significant jarosite deposition. This additionally indicated a non-biomass origin. Based on
the BE comparison, the low BE state can be identified as polypeptide amide and the
1018
higher BE state as the ammonium component in ammonium jarosite (the ammonium

coming from the nutrient media). Curiously, the ammonium also is present much earlier in
the oxidation. The signal-to-noise is insufficient to separate the protein-RNA components.
Note also that Al K is preferred over Mg K radiation as the N 1s region is otherwise
complicated by a series of copper Auger peaks.
Figure 7. Nitrogen 1s spectra from Mt. Lyell massive leached in the presence of S.
thermosulfidooxidans at 50C for 24 hours compared to E. coli and ammonium
jarosite. Both raw and processed spectra are shown
3.4 Peak in elemental sulphur production
Both the S. metallicus and S. thermosulfidooxidans assisted oxidative leaches exhibit
what appear to be peaks in the concentration of surface adsorbed elemental sulphur in the
36-60 hour range (figure 6 compared, for example, to 108 hours in figure 3). They are
both known to be elemental sulphur oxidisers (Huber and Stetter, 1991, Golovacheva and
Karavaiko, 1978) and also capable of oxidising ferrous so the nett behaviour will be
controlled by the availability of these species. As elemental sulphur forms, some of the
population oxidise elemental sulphur to sulphate. The specifics of the microbially assisted
and abiotic leaching mechanisms are too complex to reduce to consecutive, opposing and
concurrent reactions and also fit the limited rate data. We can, however, simplistically
view the paths involving elemental sulphur as two consecutive steps:
ferric
microbial
S2- S SO42For the classic case of irreversible first order consecutive reactions, the intermediate
species would first increase to a maximum and then fall to zero. Neither of these steps are
first order, but taking the concentrations of (solid) sulphide as approximately constant and
that of ferric as slowly varying, it will mimic first order. Similarly, for approximately
constant microbial activity, the oxidation of the elemental sulphur will mimic first order.
This would predict a peak in elemental sulphur concentration.
For the abiotic case sulphate is proposed to form via a thiosulphate intermediate
(Parker et al., 2003). Thiosulphate is known to be unstable at temperature in the presence
of sulphuric acid disproportionating to elemental sulphur and sulphate (Mizoguchi et al.,
1019
1976). By comparison with the microbial case the abiotic case does not have a
consumption route for elemental sulphur.
3.5 The disulphide intermediate in elemental sulphur production
Analysis of the sulphur 2p states for both the 12-60 hour S. thermosulfidooxidans and
12-36 hour S. metallicus leaches (figure 6) shows an interesting pattern in the relative
concentrations of the disulphide dimer S22- and elemental sulphur (assumed to be S8). We
have previously proposed the role of the pyritic-like dimer as an intermediate in the
chalcopyrite oxidative dissolution by ferric ions (Parker et al., 2003). From earlier work
the dimer appears to be a consequence of the fracturing of a chalcopyrite lattice and thus
systemic to the chalcopyrite system (Klauber, 2003). Interestingly, the detailed S 2p state
analysis for S. metallicus and S. thermosulfidooxidans (repeated for S. metallicus) shows a
clear interrelationship between S22- and S8. This is shown in figure 8. At the 36-hour point
when Eh and ferric concentration rise, there is a substantial decrease in the S22concentration and a stoichiometrically commensurate increase in the S8, i.e. the combined
sulphur content of the two states remains constant. It is the increased ferric activity that
converts S22- to S8. The concentrates small pyrite content complicates this interpretation,
but the changing levels of dimer and the subsequent quantity of elemental sulphur are well
in excess of what could be expected from the pyrite content. Moreover, preliminary acidic
ferric leaches on a pyrite concentrate showed no evidence for adsorbed elemental sulphur
being formed. Though the pyrite is reactive, oxidation is straight to sulphate and, unlike
the lower density dimer forming on chalcopyrite, appears not to be able to form elemental
sulphur under these conditions.
Figure 8. Decrease in surface dimer concentration is mirrored by a commensurate

increase in adsorbed elemental sulphur, indicating the dimer to be the intermediate
3.6 Cation role in jarosite formation
Prior abiotic work, in the absence of alkali cations, has indicated that eventual
deposition of hydronium jarosite, initiated by a ferric sulphate surface phase, hinders
dissolution (Parker et al., 2003). For the microbially assisted cases, the presence of
potassium enables potassium jarosite to form. Removal of the K+ content from the nutrient
1020
medium not only prevented potassium jarosite, but actually appeared to delay a jarosite
onset. This then occurred as the alternate ammonium jarosite (Stott et al., 2001) and is
qualitatively supported by the ammonium phase being observed in larger amounts, based
on the higher BE N 1s component, in the absence of K+.
Though only qualitatively examined, it would seem reasonable that the control of
singly charged cations may offer alternate avenues for jarosite control in addition to
employing ferric control. The relative sensitivity to cation type has not been explored,
though any mechanism that alters the final jarosite may also alter its structural coherence
and possibly encourage spontaneous nucleation rather than formation as an inhibiting
layer.
4.
CONCLUSIONS
In comparing earlier abiotic work on the acidic oxidative dissolution of samples of ptype chalcopyrite (Mt Isa Mines flotation concentrate and massive Mt Lyell) with that in
the presence of oxidising microbes S. thermosulfidooxidans and S. metallicus, at 50C and
65C we have been able to affirm that the interface chemistry is driven by the chemical
conditions. owever, the microbial activity and environment can have an impact upon that
chemistry. The microbial oxidation of the elemental sulphur is a key to the peak in
elemental sulphur concentration observed and appears to allow a more concerted attack,
by the ferric ion, on the surface sulphur dimer than in the abiotic case. Nutrient media
composition is also important. Due to the high concentrations of ferric, sulphate and
monovalent cations at elevated temperatures, microbial leach solutions strongly favour
jarosite formation. The potassium free nutrient media delayed the onset of potassium
jarosite, with the less stable ammonium jarosite forming at later reaction times.
Controlling the formation of jarosite and jarosite type remains the key in eliminating the
"passivation", or hindered dissolution, of chalcopyrite with respect to copper release, in
both abiotic and microbially assisted systems.
REFERENCES
1. Beamson, G., and Briggs, D., 1992. High Resolution XPS of Organic Polymers, The
Scienta ESCA300 Database. John Wiley & Sons, Chichester.
2. Golovacheva, R.S. and Karavaiko, G.I., 1978. A new genus of thermophilic sporeforming bacteria, Sulfobacillus. Mikrobiologiya, 47, 658-665.
3. Hamajima, T., Kambara, T., Ken, I. and Oguchi, T., 1981. Self-consistent electronic
structures of magnetic semiconductors by a discrete variational X calculation. III
Chalcopyrite CuFeS2, Phys. Rev. B: Condens. Matter 24, 3349-3353.
4. Huber, G., and Stetter, O., 1991. Sulfolobus metallicus, sp. nov., a novel strictly
chemolithoautotrophic thermophilic archaeal species of metal-mobilizers. System.
Appl. Microbiol., 14, 372-378.
5. Klauber, C., Parker, A., van Bronswijk, W. and Watling, H., 2001. Sulphur speciation
of leached chalcopyrite surfaces as determined by X-ray photoelectron spectroscopy.
Int. J. Min. Process., 62, 65-94.
6. Klauber, C., 2003. Fracture induced reconstruction of a chalcopyrite CuFeS2 surface.
Surf. Inter. Anal. (in press).
7. Mizoguchi, T., Takei, Y. and Okabe, T., 1976, The chemical behaviour of low valence
sulfur compounds. X. Disproportionation of thiosulfate, trithionate, tetrathionate and
sulfite under acidic conditions. Bull. Chem. Soc. Jpn., 49, 70-75.
1021
8. Neidhardt, F.C., Curtiss, III, R., Ingrahamn J.L., Lin, E.C.C., Low, K.B., Magasanik,
B., Reznikoff, W., Riley, M., Schaechter, M. and Umbarger, H.E., 1996. Escherichia
coli and Salmonella typhimurium: Cellular and Molecular Biology, 2nd Ed. American
Society of Microbiology, Washington D.C.
9. Parker, A., Klauber, C., Kougianos, A., Watling, H.R. and van Bronswijk, W., 2003.
An X-ray photoelectron spectroscopy study of the mechanism of oxidative dissolution
of chalcopyrite. Hydromet. (in press).
10. Stott, M.B., Watling, H.R., Franzmann, P.D. and Sutton, D.C., 2001. The effect of
solution chemistry on jarosite deposition during the leaching of chalcopyrite by the
thermophilic archaeon, Sulfolobus metallicus. In: Ciminelli, V.S.T. and Garcia Jr., O.
(eds.), IBS01: Biohydrometallurgy: Fundamentals, technology and sustainable
development: Proceedings of the International Biohydrometallurgy Symposium
IBS01 (Ouro Preto, Brazil), Elsevier Science B.V. Amsterdam, A, pp. 207-216.
11. Van der Mei, H.C., de Vries, J. and Busscher, H.J., 2000. X-ray photoelectron
spectroscopy for the study of microbial cell surfaces. Surf. Sci. Rep. 39, 1-24.
12. Wilson, A.D., 1960. The micro-determination of ferrous iron in silicate minerals by a
volumetric and a colorimetric method. Analyst (Lond.), 85, 823-827.
1022

Analysis of chalcopyrite (CuFeS2) electrodes utilizing galvanic

current in the presence of Acidithiobacillus ferrooxidans
D. Bevilaquaa*, A.V. Benedettib, C.S. Fugivarab, O. Garcia Jr.a
a
Department of Biochemistry and Chemical Technology, Institute of Chemistry, So

Paulo State University, P.O. Box 355, Araraquara,SP-14.801-970, Brazil
b
Department of Physical Chemistry, Institute of Chemistry, So Paulo State University,
P.O. Box 355, Araraquara, SP 14.801-970, Brazil
Abstract
The oxidative dissolution of chalcopyrite electrodes by Acidithiobacillus ferrooxidans
was studied utilizing a bacteriology battery, consisting of an electrochemical cell with two
identical compartments in two different conditions: (a) a solution in a natural condition
without any air or gas forced flux through the cell (non-air saturated) and (b) the solution
in the two compartments was saturated in air and a flux of air was passed over the solution
during the experiment. One compartment was sterile and the other inoculated with A.
ferrooxidans. During the experiment time course (around 10 days) the potential and
galvanic current was measured in a zero resistance ammeter. Results showed that the
forced aeration caused a progressive increasing in the values of potential and current due
the continuous bacterial activity, while a mineral dissolution limited by depletion of
oxygen in the solution was observed.
1.
INTRODUCTION
The microbiological leaching of chalcopyrite (CuFeS2) is of great interest because of
its potential application to many CuFeS2-rich ore materials. At present, copper bioleaching
processes are more efficient to other copper sulfides such as covellite, bornite, etc.,
because their leaching rates are much more favorable as compared to those of CuFeS2.
The slow dissolution rate reflects the recalcitrance of chalcopyrite to the chemical and
microbial oxidation [1, 2]. The slow kinetics of chalcopyrite dissolution is the result of the
intrinsic thermodynamic properties of this mineral. The electrochemical behavior of
chalcopyrite and a number of other sulfide ores have been investigated [3, 4, 5]. However
the biological mechanisms and the reactions that come into play are still poorly
understood.
The purpose of the present work was to study the oxidation of a research-grade
chalcopyrite electrode by A. ferrooxidans monitoring changes in the rest potential and in
the galvanic current during exposure of the sulfide to the bacterial attack. After the end of
* Financial support from the Fundao de Amparo Pesquisa do Estado de So Paulo (FAPESP) for the
project and for fellowship to D.B. is gratefully acknowledged. Acknowledgments are also due to the
Conselho Nacional de Desenvolvimento Cientfico e Tecnolgico (CNPq) (O.G. J. and A. V. B.)
Corresponding author (oswaldo@iq.unesp.br)
1023
experiments, the electrodes were also analyzed by scanning electronic microscopy in order
to verify the bacterial behavior regarding its colonization pattern on the sulfide. This study
is a part of a project designed to elucidate the fundamental basis and underlying
mechanisms involved in the microbiological leaching of sulfide minerals.
2.
MATERIAL AND METHODS
2.1 Bacterial strain and growth conditions

Acidithiobacillus ferrooxidans strain LR was used in this work [6]. The culture was
grown in a mineral salts medium [7] at pH 1.8 using ferrous sulfate as energy source. The
cell suspension for electrochemical experiment was obtained after growth for 48 hours in a
shaker (150 rpm and 30C) by successive washing and centrifugation (5000g) to eliminate
residual ferric ion from the medium [1]. The cells were then twice washed in Milli-Q
water (18 M.cm) and suspended in 10 mL of the mineral salts solution, medium T&K
[7], without ferrous iron in Milli-Q water. The cell suspension collected aseptically was
standardized by the modified Lowry protein determination method [8].
2.2 Mineral sample and electrode preparation
Research-grade chalcopyrite was obtained from Wards Natural Science
Establishment (Rochester, N.Y.). Samples were cut in pieces of approximately 1 cm2
using a diamond saw. One face was polished using silicon carbide emery paper with
different particle size and finally polished with an alumina suspension of 0.3 m. To
eliminate impurities the samples were rinsed with acetone, ethanol and Milli-Q water (15
min each one in an ultrasonic bath), and then dried with pure Argon and kept in
desiccators before using.
2.3 Electrochemical measurements
A modified Tait type electrochemical cell consisting of two compartments, each one
containing a massive electrode of chalcopyrite similarly prepared as possible was used in
the galvanic current and potential measurements. The working electrodes were fixed at the
bottom of each compartment of the cell with a Viton O-ring and the reference electrode
was fixed at the top of the cell. An area of 0.28 cm2 of each working electrode was
exposed to the solution, which was maintained at 30C. The electrodes and the
electrochemical cell were sterilized by UV exposure for 2 hours before each experiment.
The electrolytic solution (10 mL) composition was (g.L-1): 0.4 (NH4)2SO4, 0.4
K2HPO4, 0.4 MgSO4.7H20 at pH 1.8 adjusted with H2SO4 and sterilized by autoclaving (1
atm, 120C for 20 min). After a certain immersion time, one compartment was inoculated
with A. ferrooxidans-LR ( 8.5 109 cells.mL-1) and the other one was used as a control
(non-inoculated). Both compartments were connected via a glass tube containing the
electrolyte solution, and a cellophane membrane was fixed at the bottom of the glass tube
to impede the contamination of the control compartment with bacteria. The galvanic
current was measured at zero resistance ammeter by a short-circuiting the auxiliary and
reference electrodes outputs of the potentiostat, and imposing a potential difference of 0
0.3 mV between the two mineral samples used for recording the current fluctuations.
Ag/AgCl/KClsat electrode connected to the solution through a Luggin capillary was used
as reference. A quiescent solution was used in two conditions: (a) a solution in a natural
condition without any air or gas forced flux through the cell (non-air saturated) and (b) the
solution in the two compartments was saturated in air and a flux of air was passed over the
solution during the experiment (air saturated). The experiments were conducted for at least
1024
10 days, and then the electrodes were rinsed thoroughly with Milli-Q water to remove any
remaining unattached bacteria and examined using a scanning electron microscope (SEM)
Topcon SM-300.
3.

The Figure 1 shows the potential (E) and the galvanic current (Ig) vs. time curves
obtained for chalcopyrite electrode in the presence of A. ferrooxidans-LR in non-air
saturated condition. A decrease of the potential probably related to the dissolution of
oxides, naturally formed on the chalcopyrite surface was observed.
After addition of the cell suspension the potential increased up to 250 mV as a
consequence of biofilm formation and remained almost constant (around 250 mV) up to
the fifth day (region 1 of the figure). Afterward the potential decreased until 150 mV, the
minimum value, and increased again up to around 250 mV (region 2). The minimum
observed in the E vs. t curve can be associated with different concurrent processes: the
bacterial-assisted dissolution of iron from the mineral, and depletion of oxygen in the nonair saturated condition by the bacteria (cathodic reaction), leading to a decrease in the
bacterial activity with a consequent increase in the potential (region 3). This situation can
be illustrated by the global reaction of chalcopyrite oxidation by A. ferrooxidans, as shown
bellow:
2CuFeS2 + 8 O2 + H2SO4 2CuSO4 + Fe2(SO4)3 + H2O
In a previous work, relevant iron dissolution was also detected only after 4 days of
inoculation [1].
The galvanic current decreased initially and stabilized at around 1 nA up to the fifth
day and then increased reaching a maximum (at about 10 nA) which corresponds to
minimum potential (region 2), indicating an increasing in the mineral dissolution.
Increasing the immersion time the current decreased and stabilized at around 2 nA from
the eighth day to the end of the experiment. The same explanation given for the potentialtime curve is valid.
Figure 1. Potential and the galvanic current versus time curves obtained for
chalcopyrite electrode in non-air saturated condition. Arrows indicate the time of
bacterial addition into the electrochemical cell
1025
The experiment with air saturated solution (condition b), the E vs. t curve (Figure 2)
showed different behavior when compared with non-air saturated condition (Figure1). The
potential increased about 30 mV during the first hours of immersion and no significant
changes were observed until the end of the experiment, stabilizing around 340 mV. This
potential-time profile can be associated with the presence of oxidative atmosphere over the
solution during all the time course of the experiment, being the oxygen supplied to the
solution by a diffusion process. The minimum potential and the maximum current values
observed for non-aerated condition (Figure1) were not detected in the aerated experiment.
The oxygen diffusion controlled process was previously observed using electrochemical
impedance spectroscopy [9].
Figure 2 also depicts the galvanic current-time curve and a slow and continuous
increasing was observed, probably due to the progress of the bacterial activity, since there
was no restriction of oxygen supplying.
Figure 2. Potential and the galvanic current versus time (Eoc vs. t and Igalv vs. t) curves
obtained for chalcopyrite electrode in air-saturated condition. Arrows indicate the
time of bacterial addition into the electrochemical cell
Figure 3 (A and B) shows the adhesion of bacteria on chalcopyrite surface at the end
of experiment in air-saturated solution. As it can be seen mainly in Figure 3A a great
number of cells of A. ferrooxidans-LR attached on the chalcopyrite surface involved by a
biofilm, probably formed by exopolymeric substances. In Figure 3B (a zoom of Fig. 3A)
an unusual massive presence of diplobacilli can be observed, which is probably related
with an intense metabolic activity. In the non-air saturated experiment a similar pattern of
adhesion was observed (data not shown) despite of electrochemical differences detected
between both conditions.
The E and Ig curves of chalcopyrite electrodes in the presence of A. ferrooxidans-LR
showed significant differences depending on the test condition. For air-saturated
experiment the dissolved oxygen promoted a continuous activity of bacteria whereas in
non-air saturated test this activity seems to be limited, as revealed by the current peak after
around four days of incubation.
1026
()
()
Figure 3. Scanning electron microscopy of the A. ferrooxidans-LR on the surface of

chalcopyrite electrode in the experiment with air saturated solution at the end of
electrochemical measurements (12 days). The bars indicate the magnification
REFERENCES
1. D. Bevilaqua, A. L. L. C. Leite, O. Garcia Jr, O. H. Tuovinen, Process Biochem., 38
(2002) 587.
2. C.L. Brierley, Crit. Rev. Microbiol., 6 (1978) 207.
3. T. Cabral and I. Ignatiadis, Int. J. Miner. Process, 62 (2001) 41.
4. P. Velsquez, D. Leinen, J. Pascual, J. R. Ramos-Barrados, R. Cordova, H. Gmez, R.
Schrebler, J. Electroanalytical Chem., 494 (2000) 87.
5. P. Velsquez, H. Gmez, D. Leinen, J. R. Ramos-Barrados, Colloids and Surface A,
140 (1998) 177.
6. O.Garcia, Jr., Rev. Microbiol., 20 (1991) 1.
7. O.H. Tuovinen and D.P. Kelly, Arch. Mikrobiol., 88 (1973) 285.
8. E.F. Hartree, Anal. Biochem., 48 (1972) 422.
9. D. Bevilaqua, I. Diz-Perez, C. S. Fugivara, F. Sanz, A. V. Benedetti, O. Garcia Jr.
Proceedings of 15th International Corrosion Congress, Granada, Spain (2002) 063.
1027

Application of the bacterial weathering of silicate minerals in

the improvement of quality of non-metallics
I. tyriakova, I. tyriakb
a
Department of Biotechnology, Institute of Geotechnics of the Slovak Academy of

Sciences, Watsonova 45, 05343 Koice, Slovakia, E-mail: bacil@saske.sk
b
Department of Microbiology, Institute of Animal Physiology of the Slovak Academy of
Sciences, oltsovej 4-6, 040 01 Koice, Slovakia
Abstract
Silicate mineral weathering is one of the important biogeochemical processes in
nature. Heterotrophic bacteria of Bacillus spp. are often relatively active in these processes
because they are limited due to the lack of organic nutrients as well as other inhibitory
factors in environment. However, bacterial activity can be significantly increased by
forming of optimal conditions in bioleaching processes (pH, T, Eh, concentration of
biogenic elements). These processes can be used in the improvement of quality of several
non-metallics.
Our results indicate that removal of iron from non-metallics is the first positive
marker, which is visible after bioleaching. Except this, we observed the release of
contaminants from intergranular spaces of silicate grains, where Fe bearing minerals are
often impregnated. Moreover, the enrichment of non-metallics by fine-grained fraction is
the result of destruction of silicate minerals (mica, feldspars) and rounding-off of their
edges.
These findings suggest important applications of Bacillus strains in biotechnology
involving the quality improvement of kaolin, quartz sands, zeolite and feldspars.
Keywords: non-metallics, deferritization, silicate minerals, Bacillus spp.
1.
INTRODUCTION
Weathering of rocks is biogeochemical process occurring at earth's crust. While both
laboratory and field studies have demonstrated that microorganisms can play a role in
mineral weathering, the magnitude of this effect is still unknown.
Recent surveys of deep subsurface aquifers (hundreds of meters) show relatively high
microbial numbers (approximately 105 to 107 cells/cm3). Microbial abundance and
diversity is variable but does not decrease systematically with depth [1].
Acknowledgments: The authors are grateful to the Slovak Grant Agency for Science (Grant No.
2/2107/22).
1029
The identity of members of communities of microorganisms in natural environments,

their distribution, and the combined effects of their metabolisms on mineral dissolution are
currently poorly understood [2].
Weathering of minerals in soil releases major nutrients such as K, P, Fe, Ca, Mg, Si,
as well as trace ions which are necessary for microbial and plant growth [3]. Weathering
action may be the result of oxidative or reductive attack of appropriately reactive mineral
constituents (oxyhydroxides of Fe, Mn, sulphides) of a rock or mineral [4]. Silicate
bacteria are very active (at pH 7) in the release of impregnated sulphides from the silicate
matrix and sulphide minerals are accessible to destruction (at pH 4) by another bacterial
species of acidophilic sulfur-oxidizing bacteria [5]. In the case of metal sulphide ore
deposits, these activities can be and often are industrially exploited for metal recovery
without smelting in a process called biohydrometallurgy.
Feldspars, mica and quartz are the most abundant silicate minerals at the earth's crust.
While the geomicrobiology of silicate mineral weathering might seem an extremely arcane
subject, consider the following short list of processes affected by biogeochemically
mediated cycles: soil formation and plant nutrition, stability of architectural materials,
geochemistry of groundwater and movement of contaminants, long term stability of
geological nuclear repositories, effects of mineral weathering on climate on a geological
time scale [2]. However, in the case of silicate minerals weathering, this activity can be
also industrially used for improvement of quality of non-metallic raw materials and
treatment of wastes.
2.
2.1 Granitic eluvium

Granitic eluvium (GE), derived from granitic rocks (Slovakia). It is composed of
quartz (55-60%), feldspar (40-35%) and mica (10-20%).
2.2 Quartz sands
Quartz sands (QS) from Vyn Petrovec deposit composed of quartz (80-90%), mica
(5-10%), kaolinite (3-8%), rutil (2-5%) and siderite (4-6%).
2.3 Kaolin sample
The kaolin sample from Vyn Petrovec deposit composed of quartz (4-7%), mica (815%), kaolinite (60-75%), rutil (1-5%). The kaolin sample from Horn Prievrana deposit
composed of quartz (70-80%), mica (6-14%), kaolinite (10-20%).
The iron-bearing minerals decrease the quality of these raw materials.
2.4 Bacteria and media
Two bacterial strains of Bacillus spp. (Bacillus cereus and Bacillus pumilus) were
isolated from a kaolin quarry in Horn Prievrana. They were identified by means of the
Becton-Dickinson microbiology system (Becton Dickinson, USA). For the species
identification, the strains were cultivated on Columbia agar plates according to
recommendation of the system producer. This panel contains 29 enzymatic and
biochemical substrates and a fluorescence control on tips of plastic prongs. The resulting
pattern of the 29 reactions is converted into a ten-digit profile number that is used as the
basis for identification of a wide variety of microorganisms. For experiment, the bacterial
strains were grown in Nutrient broth No.2 (Imuna) at 37C for 18 hours. Bacterial cells were
1030
subsequently centrifuged at 4000 rpm for 15 min, subsequently washed twice with saline
solution (0.9% NaCl) and added in a concentration of 1010 cells per ml to modified
Bromfield liquid medium. Bioleaching of the samples was carried out in 3000 ml
Erlenmeyer flasks containing 2000 ml of modified Bromfield medium (NaH2PO4 0.5g/l,
MgSO4.7H2O - 0.5g/l, (NH4)2SO4 1.0g/l, NaCl 0.2g/l, molasses 0.3g/kg) inoculated
with a mixture of both Bacillus cereus and Bacillus pumilus strains. The flasks were
incubated statically for 28-120 days at 28C. The abiotic controls were cultivated under the
same conditions. After incubation, the culture solutions were separated from the biomass by
means of membrane filtration. The presence of vegetative bacterial cells in Erlenmeyer flasks
and their morphology were regularly examined by light microscopy after Gram staining.
2.5 Chemical analyses
Quantitative changes of samples (solid and liquid phases) investigated from view of
element composition stability were evaluated by standard analytical method atomic
absorption spectrometry on a VARIAN spectrometer AA - 30 apparatus (Varian, Australia)
after dissolution of the samples by standard procedure.
3.

The theses on weathering process as the most important geo-chemical process in the
hypergene zone that causes materialstructural changes of rocks and minerals, essentially
come out of the numerous studies, but there are still a lot of unclear issues concerning the
weathering stages, sequence of rock changes and participation of microorganisms in
genesis of minerals, including clay ones. We have started the study of biogenous
environmental transformation processes on the basis of similar considerations, which were
also formulated by Jn urlk [6], which concerned the fact that so far there is not any
precisely substantiating theory that could explain the genesis of eluvial processes and
creation of mineral zones of weathering crusts.
Microorganisms play an important role in the dissolution of silicate structure in the
rock weathering process and in the genesis of clay minerals. The activity of heterotrophic
bacteria and the release of organic compounds into circulating solutions in rock
environment are displayed in nature by zonal formation of a deposit. In the porcelain clay
deposit in Horn Prievrana, there is possible to observe the zones where the kaolin zone
and the Fe oxidation zone with its typical mineralization are evident. The relicts of the
bedrock of phyllites are also evident. The mire samples from Horn Prievrana kaolinite pit
are characteristic also by a high number of bacteria, especially of Bacillus species. There
are found bacteria of Bacillus genus with a wide range of species as B. cereus, B. pumilus
a B. megaterium and B. mycoides, which coexist together. They are relatively active under
aerobic and anaerobic conditions in environment with higher humidity [7].
The dissolution activity of selected Bacillus strains from these mire samples was
investigated with non-metallics materials contained feldspars, quartz, kaolinite, mica and
oxyhydroxides of iron. The chemical analyses as well as other methods indicated the
ability of bacteria to destruct these silicates and to release iron and other elements. This
fact suggests that these processes can be used in the removing of iron from many nonmetallics.
3.1 Granitic eluvium containing feldspars
Experimental studies of the effect of microorganisms on mineral weathering of
granitic eluvium with feldspars were performed in closed (batch) reaction vessels. In batch
1031
experiments, solutions were monitored over time, and solid phases were analysed in the
end of experiment.
Feldspars in granitic eluvium (GE) sample contain 1,44 wt.% iron. Iron was found
especially in the form of Fe oxyhydroxides. The presence of iron minerals was identified
also by visible brown-red colour of this material. The releasing of Fe was characteristic by
the change of visible brown-red colour of these non-metallic materials to brown-white.
The Bacillus strains used in this study markedly decreased Si content (from 73,3 to
53,5%) and partially decreased also content of potassium (from 2,48 to 2,15%) after 120
days bioleaching what resulted to a weak enrichment of Al content (from 11.51 to
11.73%) and destruction of feldspars. After 120 days of bioleaching, 31% Fe extraction
was detected in GE.
The release of Fe from GE suggests the removal of the cementing iron oxides from
the silicate surface. This partial Si and Fe removal can also improve the access of organic
acids to iron-bearing minerals and enable the increase of the feldspar raw material quality
by subsequent leaching with less concentrated organic acids. That is why bacterial
destruction of intergranular spaces of silicate grains, often impregnated with iron minerals,
should be used as pretreatment of raw material before chemical treatment of the sample by
oxalic acid.
Bobos and Gomes [8] investigated K-feldspars weathering at the kaolin deposit of Sao
Vicente de Pereira (Portugal). Their analyses indicated that the weathering products were
typical by high content of Al and small contents of Si, Fe, and K. Moreover, altered Kfeldspar grains exhibit "porous" surfaces. This fact indicates that the K-feldspar has been
rapidly dissolved or corroded by solutions with an acid pH, whereas no secondary mineral
phases were produced.
Mica and feldspars are the main constituents of rocks which were weathered into
secondary minerals, according to the environmental conditions. Feldspars (especially Kfeldspars) are often considered as relatively stable, and phyllosilicates (mica or chlorite)
are the main source of clay minerals, such as vermiculite or smectite. In humid tropical
climates, both micas and feldspars are completely transformed into kaolinite and/or
gibbsite, and iron oxides [9]. Hydrolysis is the process on which governs weathering. If
intense leaching of silicates by water occurs, it may lead to dissolution of the structure,
desilicification and precipitation of new secondary minerals (generally 1:1 clay minerals
or oxides and hydroxides). This process is the main source of clays in tropical areas.
Feldspars and micas experimentally weathered by water show pits in their structure in
addition to formation of hydroxides of Al and Fe [10]. Weathering is inextricably bound to
biological processes, for organisms inhabit a wide range of nichens in surface and
subsurface environments and influence various mineral transformation reactions [11,12].
3.2 Quartz sands
Quartz sands are largely used as a raw material in the glass industry, however, are
often associated with iron and titanium impurities that decrease their economic value and
hinder their application. Iron impurities lower an important technological factor related to
the degree of whiteness [13,14, 15].
In the first stage of bioleaching of the QS sample, there were removed visible Fe-bearing
minerals coated over quartz grains. Moreover, the extraction of elements was observed due to
the destruction of silicate minerals and intergranular spaces of brown polymineral grains with
releasing of ferrihydrite sealing siderite nodules. Chemical analyses of the bioleached QS
1032
sample showed the slight increase of Fe2O3 content from 0.15 to 0.19% and FeO content
decrease from 0.19 to 0.1% after 3 months bioleaching.
Bacterial destruction of these polycomponent grains (fraction size 0.1-0.8 mm) caused
the formation of a brown fine-grained fraction of mineral particles under 0.1 mm of size. This
brown fine-grained fraction with a high Fe-bearing mineral content was removed from QS
sample by simple elutriation process. Chemical analyses of QS show 47% removing of iron
by bioleaching and elutriation.
Upgrading of silica sand for the production of clear flat glass requires a partial
removal of iron, typically present in the form of iron oxide as discrete particles or coatings
on the surface of the sand particles. Traditional technology involves leaching with heated
sulphuric acid and producing a large quantity of acidic leachate, which is costly to treat
and discharge [16].
Other chemical methods consist of leaching with mineral acids, and treatment with
reducers, such as sodium dithionite plus aluminium sulfate or sulphur dioxide plus
aluminium powder. These bleaching methods are usually suitable for achieving a higher
degree of iron removal but are more expensive, have complex operating conditions, and
are environmentally hazardous [17].
Our experiments confirmed that heterotrophic bacteria of Bacillus spp. play active role at
biodestruction as well as at deferritization of quartz sands. Therefore, their activity should be
used in the improvement of quartz sands quality by a new progressive biotechnology.
3.3 Kaolin sample
The release of Fe from mica or from oxyhydroxides was used as indicator of mineral
dissolution and beneficiation of kaolin quality in our previous study [18].
Kaolin raw material composed of 0.92% - 1.43% iron, which was found especially in
the form of Fe oxyhydroxides (free Fe). The presence of Fe oxyhydroxides was identified
by visible brown-red colour of this material. However, a substantial part of iron in kaolin
is bound in the aluminosilicate lattice.
The Bacillus strains removed about 43% of free iron under anaerobic conditions from
kaolin sample of brown-red colour within 28 days. The removal of structural Fe ions from
kaolin of white colour under similar conditions was not so efficient because only about
15% of Fe was removed within 28 days. This fact can also suggest more difficult
availability of iron for bacterial removal because of probable binding of iron in mica.
Probable mechanisms involved in the biological removal of iron from elutriated kaolin
were examined on kaolin sample from Vyn Petrovec.
The prolongation of bioleaching time from 1 month to 3 months and thus longer
production of organic acids and metal-complexing substances by Bacillus strains caused
the increase of iron removal from mica. There was observed after 3 months the 52%
extraction of Fe atoms from octahedral position in mica when Al removal was only about
2%.
Kaolin mined at Vyn Petrovec deposit is economically interesting raw material but
its properties could be improved by removal of unsuitable minerals (mica,
oxyhydroxides). After such an improvement, it could be a suitable raw material for more
extensive use in ceramic industry.
Several strains of bacteria released by this way the cations from biotite (Si, Fe, Al),
plagioclase, and feldspar (Si, Al) much more than abiotic procedures [2]. Microbial
1033
production of organics by fermentation, or reductive dissolution of Fe Mn mineral

phases can greatly accelerate weathering rates of aluminosilicate minerals [19, 20].
4.
CONCLUSION
The iron content of industrial minerals can be reduced by a number of physical,
physicochemical and chemical methods. The appropriate method for the removal of iron
from an industrial mineral depends on the mineralogical forms and the distribution of iron
in the particular ore [21]. The red and yellow pigmentations noticed in many clay deposits
are mainly due to the associated oxides, hydroxides and hydrated oxides of ferric iron such
as hematite (red), maghemite (reddish brown), goethite (brownish yellow), lepidocrocite
(orange), ferrihydrite (brownish red), etc. These oxides and hydroxides occur either as
coatings on individual grains or as discrete fine particles throughout the clay mass.
Quantities as low as 0.4% of ferric iron may be sufficient enough to impart colour to the
deposit. Removal of these associated impurities improves the quality of the material. The
beneficiated clay finds use in paper coating, paper filling and as extenders in paints and
polymers [22].
Considerable efforts have been devoted to the problem of removing ferric
contaminants by physical [23] and chemical means. High intensity magnetic separation is
a standard method used, removing substantial quantities of iron and titanium as mineral
impurities, consequently improving the brightness of silicates. The chemical methods
consist of leaching with mineral acids, and treatment with reducers, such as sodium
dithionite plus aluminium sulphate, sulphur dioxide plus aluminium powder, and sulphur
dioxide plus zink powder. These bleaching methods are usually suitable for achieving a
higher degree of iron removal but they are more expensive, have complex operating
conditions, and are environmentally hazardous [17].
This paper also reports a possibility of microbial removal of iron from non metallics
and improvement of silicate properties via the action of bacterial destruction of Bacillus
spp. Our results indicate the removal of visible iron coatings from non-metallics as well as
iron contaminants from intergranular spaces of silicate grains. Moreover, the enrichment
of non-metallics by fine-grained fraction is the result of destruction of silicate minerals
(mica, feldspars) and rounding-off of their edges. It is also possible to accelerate
bioleaching process by continual monitoring of sugars consumption from molasses, which
was used as the energy source during samples bioleaching.
REFERENCES
1. D.J. Balkwill, Geomicrobiology J., 7 (1989) 33.
2. W.W. Barker, S.A.Welch, S. Chu and J.F. Banfield, In: Banfield JF, Nealson KH
(Eds) Geomicrobiology: Interactions between microbes and minerals, Reviews in
Mineralogy, Vol. 35, Washington, D.C., USA, 1997
3. P. Hinsinger, B. Jaillard and J.E. Dufey, Soil Soc Am J, 56 (1992) 977.
4. H.L. Ehrlich, Chem Geol, 132 (1996) 1.
5. I. tyriakov, I. tyriak and M. Kunierov, In: R. Amils, A. Ballester (eds) Process
Metallurgy 9A, Elsevier, 1999.
6. urlk,J., 1988: Geochmia geologickch procesov. Hypergnne procesy.
Prrodovedeck fakulta Univerzity Komenskho, Bratislava.
7. I. tyriakov, I. tyriak, Mineralia Slovaca, 34 (2002) 99.
8. I. Bobos and C. Gomes, Geologica Carpathica, 51 (2000) 49.
9. R. Romero, M. Robert, F. Elsass and C. Garcia, Clay Minerals 27, (1992) 21.
1034
10. M. Robert, D. Tessier, In: I.P. Martini and W. Chesworth (eds.): Weathering, Soils and
Paleosols. ELSEVIER, 1992.
11. J.F. Banfield and R.A. Eggleton, Clays and Clay Minerals, 36 (1988) 47.
12. P.M. Huang and M. Schnityer, Soil Science Society of America Special Publication
no.17 (1986) 606.
13. L.G. Yan, Y.J. Yu, S.S. Song, H.L Nan, Y.L. Cheng, X.M. Li, Q.W. Kong and Y.M
Dai, Proc. 2nd World Congress on Non-Metallic Minerals, Beijing, China, 1987.
14. F. Veglio and L. Toro, International Journal of Mineral Processing, 39 (1993) 87.
15. F. Veglio, Hydrometallurgy, 45 (1997) 181.
16. I.I.Tarasova, A.W.L. Dudeney and S. Pilurzu, Minerals Engineering, 14 (2001) 639.
17. L.M.S. Mesquita, T. Rodrigues and S.S. Gomes, Minerals Engineering, 9 (1996) 965.
18. I. tyriakov and I. tyriak, Ceramics Silikty, 44 (2000), 135.
19. S.A.Welch, W.J. Ullman, Geochim Cosmochim Acta 60 (1996) 2939.
20. P.C. Bennett, F.K. Hiebert and W.J. Choi, Chem Geol 132 (1996) 45.
21. M. Taxiarchou, D. Panias, I. Douni, I. Paspaliaris, and A. Kontopoulos,
Hydrometallurgy, 46 (1997) 215.
22. V.R. Ambikadevi and M. Lalithambika, Applied Clay Science 16 (2000) 133.
23. S.J.F. Guimares, N. De Oliveira, F.L. De Salles, Trans Inst. Min. metall. 51 (1987) 13.
1035

Assessment of acid production potential of sulphide minerals

using Acidithiobacillus ferrooxidans and microbial sulphate
reduction using Desulfotomaculum nigrificans
Evvie Chockalingam, S. Subramanian, K.A. Natarajan and J.J. Brauna
Department of Metallurgy and aIndo-French Centre for Water Sciences, Indian Institute of
Science, Bangalore 560 012, India
Abstract
The acid production potential of pyrite and chalcopyrite minerals has been assessed in
the presence of Acidithiobacillus ferrooxidans in iron-free Silverman and Lundgren
medium. The parameters such as pH, redox potential, bacterial cell number, ferrous, ferric
and copper concentrations were monitored as a function of time. The inoculated samples
show significant changes in the values of the parameters studied relative to the control
samples. The pH values decrease to 1.6 after 100 days for pyrite and to 2.5 after 150 days
for chalcopyrite respectively, from an initial value of about 7, after inoculation with
Acidithiobacillus ferrooxidans cells. The ferrous ion concentration is found to be higher
vis--vis that of ferric for both the mineral samples. The kinetics of copper dissolution
from chalcopyrite is found to be slower compared to the dissolution of iron. Significant
sulphate reduction and copper precipitation could be achieved using Desulfotomaculum
nigrificans, both at pH 7 and 5.5. An increase in the cell number enhances the kinetics of
sulphate reduction and copper precipitation.
Keywords: acid production potential, Acidithiobacillus ferrooxidans, pyrite, chalcopyrite,
Desulfotomaculum nigrificans
1.
INTRODUCTION
A major environmental concern to the mining industry, both during the operational
period and after closure is the generation of acidity, resulting from the oxidation of
sulphide mineral wastes, catalyzed by acidophilic microorganisms. The dearth of basic or
buffering materials to neutralise the acid, results in the leach water becoming acidic,
leading to dissolution and build up of metal species. This phenomenon known as acid
mine drainage, needs to be mitigated by suitable strategies. The technology employed for
the treatment of acid mine drainage includes pH adjustment, chemical precipitation,
aeration and settling. Several active and passive treatment methods, including the use of
soil and water covers have been investigated to curtail the oxidation of sulphidic wastes
[1]. Biotechnological approaches using sulphate-reducing bacteria (SRB) have also gained
importance in recent times [2-4]. The treatment of metal-contaminated wastes by sulphatereducing bacteria has been investigated by several workers [5-7]. Specific technologies
that have been successful in using SRB to remediate ground water contaminated with
heavy metals include the Paques (THIOPAQ) system [8,9] and the NTBC bio-sulphide
1037
process [10]. In an ongoing research programme, environmental impact studies have been
undertaken with reference to an abandoned copper mine. Keeping these objectives in
mind, acid production potential studies on pyrite and chalcopyrite minerals using
Acidithiobacillus ferrooxidans have been performed. Sulphate reduction studies have been
carried out in modified Barrs medium using Desulfotomaculum nigrificans, in the
simultaneous presence of copper ions, as a function of pH and bacterial cell number. The
variation in the sulphate and copper concentrations was monitored as a function of time.
2.
EXPERIMENTAL
2.1 Materials
2.1.1 Mineral samples
Pure mineral samples of pyrite and chalcopyrite were obtained from Alminrock
Indscer Fabriks, Bangalore, India. The samples were subjected to dry grinding in a
porcelain ball mill and the < 74 m size fraction was used for the experiments.
2.1.2 Bacterial strains
The strain of Acidithiobacillus ferrooxidans was isolated from the ore samples of the
Hutti gold mines, Karnataka, India, and used after purification. A pure strain of
Desulfotomaculum nigrificans was obtained from the National Collection of Industrial
Microorganisms, National Chemical Laboratory, Pune, India.
2.2 Methods
2.2.1 Medium and growth of Acidithiobacillus ferrooxidans
The bacteria Acidithiobacillus ferrooxidans were grown in 9K medium [11] which
has the following composition in grams per litre of solution: (NH4)2SO4, 3.0;
MgSO4.7H2O, 0.5; K2HPO4, 0.5; KCl, 0.1; Ca(NO3)2, 0.1 and FeSO4.7H2O, 44.2. The pH
was adjusted to 2 initially by the addition of sulphuric acid. The fully-grown culture of
Acidithiobacillus ferrooxidans from the 9K medium was centrifuged to harvest the cells.
The cells were then resuspended in pH 2 sulphuric acid to remove iron. This procedure
was repeated twice to get iron free cell suspension. The iron-free Acidithiobacillus
ferrooxidans cells thus obtained were used as inoculum for the acid production potential
studies.
2.2.2 Medium and growth of Desulfotomaculum nigrificans
The bacteria Desulfotomaculum nigrificans were grown in the modified Barrs
medium (MBM) [12] which contained 0.5 g K2HPO4, 1 g NH4Cl, 1 g CaSO4, 2 g
MgSO4.7H2O, 5 g sodium lactate and 1 g yeast extract in 1000 mL distilled water. The pH
of the Barrs medium was adjusted to 7.5 and sterilized for 3 consecutive days. The pH
was checked every time before sterilization and adjusted to the desired value. Before
inoculation, for each 100 ml of the medium, 0.05 g of FeSO4(NH4)2SO4.6H2O and 0.01 g
of sodium thioglycollate was added and filter sterilized. The media was stored in 250 ml
serum bottles stoppered with a tight fitting rubber cork. To each 90 ml of the abovementioned media, 10 ml of an active culture was added through a sterilized syringe.
Immediately after inoculation, nitrogen gas was flushed into the medium for 5 minutes
through a syringe and sealed tightly with parafilm to maintain anaerobic conditions. The
bottles were then incubated at 37C in a rotary shaker at 160 rpm. The fully-grown culture
1038
was centrifuged and the supernatant was discarded. The cells were resuspended in distilled
water. This process was repeated twice and the washed cells were used as inoculum for
the sulphate reduction studies.
2.2.3 Assessment of acid production potential
The acid production potential of pyrite and chalcopyrite minerals was evaluated
adopting the following procedure: 5 g of the mineral sample (< 74 m) was suspended in
100 ml of 9K- medium (without ferrous sulphate) containing 3 x 108 cells/ml under
natural pH condition, in a 250 ml Erlenmeyer flask. The flask was agitated at 240 rpm in a
Remi rotary shaker at 30 + 2C. The parameters such as pH, ESCE, bacterial cell number,
ferrous, ferric and copper concentrations were monitored periodically. A control sample
was also maintained in which 10% alcoholic thymol was added as a bactericide.
2.2.4 Sulphate reduction and copper precipitation kinetics using Desulfotomaculum
nigrificans
In these experiments, the variation in the sulphate and copper concentrations was
measured at pH 5.5 and 7 using different cell densities ( 107, 108 and 109 cells/ml),
during growth of Desulfotomaculum nigrificans in the modified Barrs medium. 10 ppm
of copper was added as cupric sulphate pentahydrate to the MBM for these studies.
2.2.5 Analytical procedures
The bacterial cell count was determined using a Petroff-Hauser counter attached to a
Leitz phase contrast microbiological microscope (Laborlux K Wild Mps12). The ferrous
and total iron concentrations were analysed using the orthophenanthroline method [13] in
a Shimadzu model UV260 uv-visible spectrophotometer. The total iron concentration was
determined by reducing the ferric iron in solution to ferrous by adding
hydroxylammonium chloride. The ferric concentration was then estimated as the
difference between total iron and ferrous concentrations. Copper was analysed using a
Video 11E Thermo-Jarell Ash atomic absorption spectrophotometer. Sulphate
concentration was determined by turbidimetric method using barium chloride at a
wavelength of 420nm [14] using a Shimadzu model UV260 uv-visible spectrophotometer.
3.
3.1 Acid production potential studies

The acid production potential of pyrite and chalcopyrite in the absence and presence
of Acidithiobacillus ferrooxidans was assessed by monitoring the variation in the pH,
redox potential, bacterial cell number, ferrous, ferric and copper concentrations as a
function of time. It must be emphasized that these experiments were carried out at a
natural pH of 6-7, in order to ascertain the propensity for acid generation, with or without
bacterial cells.
3.1.1 pH
The variation of pH as a function of time for pyrite and chalcopyrite suspensions in
9K medium for the inoculated and control samples is shown in Figure 1. In the case of the
pyrite sample, the pH steeply decreases from the initial value of 7 to 2 in about 50
days. The pH marginally decreases thereafter to 1.6 and remains more or less constant
upto about 220 days. However, in the case of the control sample for pyrite, the pH
marginally decreases to 6 in 50 days and is further lowered to about 4.5 in about 100
1039
days and remains at that value for the time period investigated. With respect to
chalcopyrite, the pH steadily decreases from 6 to 2.5 in about 150 days and remains
more or less unaltered thereafter. In the case of the control sample for chalcopyrite, the pH
remains close to the initial value of 6 upto 90 days and is subsequently decreased to 4.6
in 150 days and remains at that value upto 220 days. These results highlight that acidity is
generated in the case of both pyrite and chalcopyrite after inoculation with
Acidithiobacillus ferrooxidans. On a comparative basis, the acid production potential of
pyrite is higher than that of chalcopyrite after bacterial interaction.
3.1.2 Redox potential
Figure 2 depicts the redox potential changes as a function of time for pyrite and
chalcopyrite suspentions in 9K- medium. The ESCE value increases sharply from 100 mV
to 400 mV in about 40 days for pyrite after interaction with the bacterial cells. The redox
potential further increases to 425 mV in 50 days and remains unchanged upto about 220
days. In the case of the control sample for pyrite the ESCE increases from the initial value
of 100 mV to 180 mV in 50 days. Beyond that period, the redox potential marginally
increases and attains a value of 190 mV. The redox potential of the chalcopyrite sample
after inoculation with Acidithiobacillus ferrooxidans steadily increases from 100 mV to
250 mV in about 90 days. A sharp increase is observed to a value of 500 mV in 130 days
beyond which there is not much of a change. With respect to the control sample for
chalcopyrite, the redox potential is more or less unaltered at 100 mV upto about 90 days.
Thereafter, a steady increase is observed upto about 180 days and the redox potential
attains a value of 270 mV. As can be expected, the redox potential values are higher in the
case of the inoculated samples attesting to bacterial oxidation of the mineral samples. It is
pertinent to mention that the redox potential values observed in these tests are lower than
that commonly encountered in typical leaching systems. This is understandable as the
initial pH of the system was close to neutral (pH 6-7), in contrast to a typical leaching
system, wherein the pH is maintained close to 2-2.5.
900
Pyrite in 9K medium
Pyrite in 9K medium + Af cells
Chalcopyrite in 9K medium
Chalcopyrite in 9K medium +
Af cells
10
pH
600
450
300
150
0
0
50
100
150
200
Time (days)
Figure 1. pH as a function of time for

pyrite and chalcopyrite mineral
suspensions
1040
750
ESCE (mV)
Pyrite in 9K medium
Pyrite in 9K medium + Af cells
Chalcopyrite in 9K medium
Chalcopyrite in 9K medium +
Af cells
0
0
50
100
150
200
Time (days)
Figure 2. Redox potential as a function

of time for pyrite and chalcopyrite
mineral suspensions
3.1.3 Bacterial cell number

The changes in the bacterial cell number after interaction with pyrite and chalcopyrite
mineral samples as a function of time are portrayed in Figure 3. For both the samples, a
decrease in the cell number is observed initially, presumably due to the attachment of the
cells onto the mineral substrates. In the case of pyrite, the cell number steeply increases
from 2 x 107 cells/ml to 8 x 108 cells/ml in about 60 days and thereafter remains
unchanged. For chalcopyrite the cell number marginally increases from 7 x 107 to 2 x 108
cells/ml after 220 days. It is interesting to note that there is an increase in cell number after
interaction with pyrite while after interaction with chalcopyrite there is a marginal
decrease. The dissolution of copper ions from chalcopyrite as a function of time could
plausibly lead to toxicity impeding cell growth.
10
Number of cells/ml
10
Pyrite
Chalcopyrite
9
10
10
10
50
100
150
200
250
Time (days)
Figure 3. Bacterial cell number as a function of time for pyrite and chalcopyrite
mineral suspensions
3.1.4 Dissolution kinetics of ferrous and ferric ions
The changes in the ferrous and ferric concentrations as a function of time for pyrite
are shown in Figure 4a. The ferrous concentration steeply increases initially upto about 50
days and attains a value of 6g/l in the case of the inoculated sample. Around 170 days,
the ferrous concentration attains a value of 7g/l and remains at that value upto 220 days.
The ferric concentration steadily increases to 3g/l in about 50 days and thereafter
remains unchanged in the case of the inoculated sample. With respect to the control
samples, the ferric concentration slightly increases to 0.5g/l in about 75 days and
remains unaltered thereafter. The ferrous concentration in the control sample increases to
about 1g/l in 75 days and further increases to about 1.5g/l between 160-220 days. It is
evident that the ferrous concentration is higher in both the inoculated and control samples
compared to the ferric concentrations. These results are in agreement with those reported
earlier [15-17]. It is also pertinent to mention that the presence of the pyrite substrate
facilitates the preferential oxidation of sulphur in the mineral, rather than the ferrous ions
in solution, through the direct contact mechanism.
The dissolution of iron species from chalcopyrite is portrayed in Figure 4b. Almost a
similar trend as that observed for pyrite is revealed in this case also. In the case of the
inoculated sample, the ferrous concentration shows a marginal increase upto about 60 days
and subsequently a steep increase to 3.5 g/l is observed upto 160 days. Beyond that period,
1041
there is not much of a variation in the ferrous concentration. The ferric concentration
gradually increases to about 1 g/l in 150 days and remains at that value for the inoculated
sample. The ferrous concentration in the control sample marginally increases to about 0.75
g/l in 150 days and more or less remains unchanged thereafter. On the other hand, the
ferric concentration in the control sample remains more or less constant at 0.25 g/l in the
time frame investigated. A closer examination of Figures 4a and 4b highlights that the
dissolution rate of iron species from pyrite is higher vis--vis that of chalcopyrite. The
lower redox potential values observed in Figure 2 are consistent with the low ferric/ferrous
ratios observed (Figs. 4a and 4b).
8
6
2+
2+
3+
0
0
50
100
150
2+
Fe in 9K medium
2+
Fe in 9K medium + Af cells
3+
Fe in 9K medium
3+
6
5
3+
10
200
50
Time (days)
100
150
200
Time (days)
Figure 4a. Dissolution of ferrous and

ferric from pyrite as a function of time
Figure 4b. Dissolution of ferrous and

ferric from chalcopyrite as a function of
time
3.1.5 Dissolution kinetics of copper ions

Figure 5 depicts the kinetics of dissolution of copper from chalcopyrite in the case of
the inoculated and control samples. The amount of copper dissolved is very low in the first
50 days for both the samples. The copper concentration steadily increases to about 3 g/l in
the case of the inoculated sample after 220 days. However, the concentration of copper
remains unchanged upto about 170 days at 0.2 g/l and thereafter shows a marginal
increase to about 0.5 g/l in the case of the control sample. It becomes of interest to note
from Figures 4b and 5 that the dissolution rate of ferrous is higher compared to that of
copper from chalcopyrite. Similar results have been reported by other workers [15].
5
Chalcopyrite
Cu conc. (g/l)
9K medium
9K medium + Af cells
0
0
50
100
150
200
Time (days)
Figure 5. Dissolution of copper from chalcopyrite as a function of time

1042
Fe conc. (g/l)
Fe conc. (g/l)
2+
Chalcopyrite
Fe in 9K medium
2+
3+
Fe in 9K medium
3+
10
Fe conc. (g/l)
12
Pyrite
Fe conc. (g/l)
12
3.2 Studies on sulphate reduction and copper precipitation using Desulfotomaculum

nigrificans
Studies have been carried out to assess the efficacy of Desulfotomaculum nigrificans
to bring about the reduction of sulphate as well as the precipitation of copper. These
studies are of relevance to the mitigation of acid mine drainage. The effect of pH and cell
number on the reduction kinetics was assessed. The growth of the organism was carried
out in the modified Barrs medium [12] in the presence of copper ions ( 10 ppm). Figure
6 shows the change in sulphate concentration as a function of time and cell number. In
these tests, the initial pH was maintained at 7. It is evident from the figure that the
sulphate concentration decreases rapidly upto about 9 days in all cases and gradually
thereafter. Further, the rate of sulphate reduction increases with increase in the cell
number. For example, the sulphate reduction increases from 18.5% to 24.5% to 31.5% as
the cell number is increased from 107 to 109 cells/ml.
The change in copper concentration as a function of time and cell number was also
monitored. The results are depicted in Figure 7. It is noteworthy that, complete
precipitation of copper takes place by 3, 4 and 6 days respectively as the cell number is
increased from 107 to 109 cells/ml. It was found that the pH increases from the initial
value of 7 to about 7.8 after 4 days in all the samples and subsequently decreased back to
7. This may be attributed to the production of H2S by bacterial metabolism.
Initial pH = 7 + 0.2
1500
1350
1200
1050
Initial pH = 7 + 0.2
12
1.4 x 10 cells/ml
8
1.1 x 10 cells/ml
9
1.6 x 10 cells/ml
1.4 x 10 cells/ml
8
1.1 x 10 cells/ml
9
1.6 x 10 cells/ml
10
Cu conc. (ppm)
Sulphate conc. (ppm)
1650
8
6
4
2
900
0
0
10
12
14
16
Time (days)
Figure 6. Bioreduction of sulphate as a

function of time and bacterial cell
number at pH 7
Time (days)
Figure 7. Bioremoval of copper as a

number at pH 7
The feasibility of sulphate reduction at pH 5.5 was ascertained and the results are
portrayed in Figure 8. It is apparent that, beyond 7 days the amount of sulphate reduced,
increases with increase in cell number. A comparison of Figures 6 and 8 indicates that the
rate of sulphate reduction is lower when the pH is reduced to 5.5 from 7. Prolonged
experiments lasting for 57 days carried out at pH 7 and pH 5.5 respectively resulted in
95% and 91% sulphate reduction using 109 cells/ml.
The variation of copper concentration as a function of time for experiments carried
out at pH 5.5 at different cell densities is shown in Figure 9. It is evident that complete
precipitation of copper takes place by 6, 8 and 15 days respectively as the cell number is
increased from 107 to 109 cells/ml. As anticipated, the copper precipitation kinetics is
1043
slower at pH 5.5 when compared to pH 7 (Figures 7 and 9). The results presented in
Figures 6 to 9 unequivocally demonstrate that Desulfotomaculum nigrificans is capable of
significant sulphate reduction and complete precipitation of copper. The mechanisms of
sulphate reduction and metal ion precipitation are well documented [18, 19].
1.2 x 10 cells/ml
8
1.6 x 10 cells/ml
9
1.5 x 10 cells/ml
1650
1.2 x 10 cells/ml
8
1.6 x 10 cells/ml
9
1.5 x 10 cells/ml
12
10
1500
1350
1200
8
6
4
2
10
12
14
16
Time (days)
Figure 8. Bioreduction of sulphate as a

number at pH 5.5
4.
Initial pH = 5.5 + 0.2
Cu conc. (ppm)
Sulphate conc. (ppm)
14
Initial pH = 5.5 + 0.2
0
0
10
12
14
16
Time (days)
Figure 9. Bioremoval of copper as a

number at pH 5.5
CONCLUSIONS
From the results of the present investigation, the following conclusions can be arrived
at:
1. Acid production potential studies reveal that the pH values of pyrite and chalcopyrite
suspensions are decreased from 7 to about 1.6 and 2.5 respectively after inoculation
with Acidithiobacillus ferrooxidans cells.
2. An increase in the bacterial cell number is observed in the case of pyrite while for
chalcopyrite there is a marginal decrease in the cell number.
3. In the case of both pyrite and chalcopyrite minerals, the dissolution of ferrous ion is
substantially higher compared to ferric ion.
4. The amount of copper dissolved from chalcopyrite is lower than that of ferrous.
5. About 95% and 91% sulphate reduction could be achieved at pH 7 and 5.5
respectively, using 109 cells/ml of Desulfotomaculum nigrificans.
6. Complete precipitation of copper could be achieved both at pH 5.5 and 7 using
Desulfotomaculum nigrificans cells, though the kinetics of precipitation is slower at
pH 5.5.
7. The kinetics of sulphate reduction and copper precipitation is found to increase with
increase in the bacterial cell number.
ACKNOWLEDGEMENTS
The authors are grateful to the Indo-French Centre for Promotion of Advanced
Research (CEFIPRA) and the Institute for Research and Development (IRD) France, for
grant of research projects to carry out this investigation.
1044
REFERENCES
1. MEND, 1993, 2.11.2a: Literature Review Report: Possible Means of Evaluating the
Biological Effects of Sub-Aqueous Disposal of Mine Tailings, March 1993.
2. N. Kuyucak, Miner. Metall. Proc., No.2, 17 (2000) 85.
3. S. Foucher, F. Battaglia-Brunet, I. Ignatiadis and D. Morin, Chem. Eng. Sci., 56
(2001) 1639.
4. J.H. Tuttle, P.K. Dugan and C.I. Randles, Appl. Microbiol., No.2, 17 (1969) 297.
5. Kathy Jalali and S.A. Baldwin, Wat. Res., Vol. 34, 3 (2000) 797.
6. D.H. Dvorak, R.S. Hedin, H.M. Edenborn and P.E. McIntire, Biotechnol. Bioeng., 40
(1992) 609.
7. H. Kilty, J. Roussy, J.M. Tobin and J.R. Degorce-Dumas, Biohydrometallurgy:
Fundamentals, Technology and Sustainable Development, V.S.T. Ciminelli and O.
Garcia Jr. (eds.), Elsevier Science B.V, Part B (2001) 357.
8. L.J. Barnes, F.J. Janssen, J. Sherren, J.H. Versteegh, R.O. Koch and P.J.H. Scheeren,
Chem. Eng. Res. Des., 69 (1991) 184.
9. L.J. Barnes, F.J. Janssen, J. Sherren, J.H. Versteegh and R.O. Koch, Trans. Inst. Min.
Metall., 101 (1992) C183-C199.
10. M.V. Rowley, D.D. Warkentin and V. Sicotte, Proceedings of the Fourth International
Conference on Acid Rock Drainage, Vancouver, BC. Canada, 1997.
11. M.P. Silverman and D.G. Lundgren, J. Bacteriol., 77 (1959) 642.
12. E.S. Pankhurst, Isolation of Anaerobes, D.A. Shapton and R.G. Board (eds.), The
Society for Appl. Bacteriol., Technical series No.5, Academic Press, 1971.
13. A.I. Vogel, Vogels Text Book for Quantitative Chemical Analysis, 5th edn,
Longman, London, 1989.
14. APHA, AWWA and APCF, Standard Methods for the Examination of Water and
Waste Water, 14th edn, 1975.
15. Shrihari, R. Kumar, K.S. Gandhi and K.A. Natarajan, Appl. Microbiol. Biotechnol., 36
(1991) 278.
16. M.G. Monroy Fernandez, C. Mustin, Philippe de Donato, O. Barres, P. Marion and J.
Berthelin, Biotechnol. Bioeng., 46 (1995) 13.
17. K.J. Edwards, M.O. Schrenk, R. Hamers and J.F. Banfield, Am. Mineral., 83 (1998)
1444.
18. J.R. Postgate, The Sulphate Reducing Bacteria, Cambridge University Press, 2nd edn.,
1984.
19. C. Garcia, D.A. Moreno, A. Ballester, M.L. Blazquez and F. Gonzalez, Miner. Eng.,
Vol. 14, 9 (2001) 997.
1045

Comparative study on pit formation and interfacial chemistry

induced by Leptospirillum and Acidothiobacillus ferrooxidans
during FeS2 leaching
Helmut Tributsch and Jos Rojas-Chapana
Hahn-Meitner-Institut, Dept. Solare Energetik, 14109 Berlin, Germany
Abstract
After it had been assumed for a long time that the bacterium Leptospirillum
ferrooxidans is only capable of oxidizing Fe2+ which then oxidizes the sulfide, highresolution electron microscopy has recently revealed that the bacterium can also
selectively disintegrate a crystallized solid sulfide such as FeS2 in order to recover Fe2+.
Characteristic properties are the mineralization of its surface including the interacting
capsula, where sulfide nanoparticles can be identified, and localized pit formation, which
can result in a pronounced collective effort to dissolve the crystalline environment.
Acidothiobacillus ferrooxidans, on the other hand, profits from weakening and break
up of chemical bonds mediated by the formation of cysteine-pyrite complexes. This might
also be the case under natural conditions by the excretion of cysteine and cysteine-like
metabolites.
This contribution aims at a comparison of both Fe2+ oxidizing bacteria, to reveal their
adaptation to the leaching of the FeS2 interface. It is shown how Leptospirillum has
adapted electrochemical while Acidothiobacillus has adapted chemical mechanisms to
obtain an optimum output of energy supplying Fe2+ and sulfur species respectively.
1.
INTRODUCTION
Leaching of pyrite has received considerable attention especially due to its relevance
for the hydrometallurgical recovery of gold but also because pyrite, which is a component
in many sulfide ores, releases sulfuric acid during the oxidation process [1-12]. The
bacteria typically found in the presence of pyrite under leaching conditions are
Acidothiobacillus (formerly Thiobacillus) ferrooxidans (A.f.) and Leptospirillum
ferrooxidans (L.f.). As their names indicates, both are able to oxidize Fe2+ to Fe3+ in the
presence of oxygen, but while A.f. is also able to oxidize sulfide species, L.f. is apparently
only able to oxidize Fe2+. Numerous studies have been performed to investigate the
oxidation properties of these two bacterial species [13-14]. With respect to the bacterias
ability for Fe2+-oxidation it has been documented that L.f. is able to utilize Fe2+ as an
energy source at significantly, 200 mV, more positive oxidation potentials compared with
A.f. [9]. Under such conditions, correspondingly less electrochemical energy is available
for bacterial carbon dioxide fixation. In technical gold leaching installations with pyrite
containing ores, L.f. are the bacteria which are doing the main job especially when
1047
the solutions contain high concentrations of Fe3+. This contribution aims at a comparative
analysis of the two bacterial mechanisms for interacting with pyrite [13].
2.
EXPERIMENTAL
Bacterial strain and growth conditions: Leptospirillum ferrooxidans (isolated from
commercial biooxidation tanks in South Africa) was kindly provided by G. Hansford
(University of Cape Town, South Africa). L.f. was maintained on ferrous ions culture
solution using the medium described by Tuovinen and Kelly for Thiobacillus ferrooxidans
[15]. Bacterially mediated ferrous oxidation was followed by increasing coloration and
gradual precipitation due to the low solubility of ferric iron at pH 1.6.
For the culture of L. ferrooxidans on pyrite crystals (pyrite from Mogul-Turkey;
particle size ca. 5 mm and at 10% pulp density) the same medium was used but without
ferrous ions. Continuous agitation is required to maintain uniform pulp density.
The cultivation of A.f. on synthetic pyrite layers was performed following the
procedure described by Rojas-Chapana et al. [16-17].
Microscopy: High-resolution and analytical transmission electron microscopy (TEM)
and scanning electron microscopy (SEM) were used to study the bacterial morphology, the
stoichiometry and shape of the pyrite grains used as substrates as well as bacteria-pyrite
assemblages formed during the bacterial life cycle. All samples for TEM-examinations
were imaged in a Philips CM 12 electron microscope operating at 100 kV. High-resolution
SEM-examinations were imaged in a Gemini SEM-microscope (LEO1530). The SEM
microscope was operated in the 3-7 KeV range being configurated primarily for imaging
(using secondary electrons) and compositional analysis (using energy-dispersive x-ray
spectroscopy (EDS)).
2.1 Bacterial morphology in pyrite cultures
Neither A.f. nor L.f. show peculiarities in their morphology when cultivated on Fe2+ as
the only energy source. However, when cultivated on ground-up pyrite crystal material,
they both show a significant adaptation, in which they develop a pronounced capsule of
polysaccharides around their bacterial cells.
Fig. 1a and 1b show A.f. bacteria surrounded by such capsules, which are dotted with
nanoparticles. These nanoparticles turned out to be molecular sulfur [16] extracted from
the pyrite interface. Fig. 1b shows a A.f. cell which is splitting-up into a mother- and
daughter cell. It is seen, that also the capsule is separated during this process. The
mechanism by which A.f. is interacting with the pyrite surface, via the organic material of
the capsule has been discussed in some details in previous publications [11,16,17]. Since
the main energy source for this bacterium is a sulfide species that liberates significantly
more energy compared with Fe2+, the evolution of such a capsule is not surprising. The
bacterium needs an organic matrix to circle the catalyst species, which are disrupting
chemical bonds in the pyrite interface. Crystal-bound sulfur is extracted in the form of
polysulfide to yield sulfur colloids as intermediate energy storing elements. The
concentration, size and distribution of these colloids within the capsule may change during
the life cycle of the bacterium and in dependence of the sulfide energy supply.
It was a surprising discovery that L.f. are generating a capsule around their cells when
interacting with a pyrite energy source. Fig. 1c shows three bacterial cells characterized by
their spiral growth, partially overlapping each other in the transmission electron
microscope picture. The two bacterial cells on the right are shown in a slightly higher
magnification in Fig. 1d. The areas where the bacterial cells plus capsules overlap are
1048
clearly shown as areas of more intensive darkness. All over the capsule of L.f. dark spots
of nanoparticles are seen. They are typically unevenly distributed and vary in a certain
range with respect to their dimension. X-ray luminescence (EDAX) analysis has clearly
shown that one is dealing with FeS2 particles. The cubic crystal structure of pyrite could
be confirmed by electron diffraction analysis [18]. The question now arises: why does L.f.,
if it only oxidizes Fe2+, need a capsule? and, even more important; why do crystalline FeS2
particles accumulate in it? There is apparently a mechanism by which L.f. bacterial cells
are interacting with solid pyrite interfaces to disrupt them and to recover nanoparticles for
some metabolic purpose. In order to get a better insight into the interaction of L.f. with
pyrite interfaces it became necessary to perform studies of crystallized pyrite interfaces in
the presence of bacterial cells.
Figure 1a. Planktonic A.f. bacteria

grown on synthetic pyrite film
Figure 1b. A.f. bacteria during

cell division on synthetic pyrite
film
Figure 1c. Planktonic L.f. bacteria

grown on pyrite crystals
Figure 1d. Enlargement of 1c.

L.f. bacteria covered with pyrite
nanoparticles
2.2 Leaching patterns on crystallized pyrite interfaces

Our research group has a long-term experience with the physical chemistry of pyrite
interfaces and their electrochemical behavior. It was, therefore, quite easy to identify a
very unusual type of leaching patterns which develop in presence of L.f.. In otherwise
apparently untouched surfaces, pronounced 20 40 m large, deep etching structures
developed, which are shown in Fig. 2.
They partially show a hexagonal pattern (which can be explained for pyrite interfaces
[18]). The characteristic spiral-shaped cells of L.f. are rarely found on the exposed surface
of pyrite. However, interestingly, when looking into the etching craters, they can be
localized there in high concentrations. This is demonstrated with Fig. 2c. The bacterial
1049
concentration is so high, that the bacterial cells are touching each other. There appears to
be a correlation between the pronounced etching pit formation and the high concentration
of L.f. cells in them. This discovery raises new aspects in relation to L.f. activity. These
bacteria have apparently developed a technique for actively dissolving pyrite. If they can
do that, they presumably have an advantage from it regardless of the established notion
that they are unable to oxidize sulfur species.
Figure 2a-c. SEM micrographs of a pyrite surface leached by L.f. (pitting). The
bottom picture shows a high concentration of L.f. in the corrosion craters
It is reasonable now to confront this observed leaching behavior with the leaching
patterns observed with A.f. for pyrite. The pyrite leaching mechanism by these bacteria has
been examined before in some detail, especially by in-situ observation of bacterial activity
on 100 nm thin FeS2 synthetic layers [11,12,16]. A localized collective activity
comparable to that observed to L.f. was never observed. The bacterial cells moved freely
around on the surface interacting with it and leaching it away. Very homogeneous
interfaces were that way leached away quite homogeneously. When the substrate,
however, was inhomogeneous or when surface-active agents like Tween 80 were applied
in very low concentrations the bacterial cells developed localized activities, which yielded
pronounced bacterial etching pits.
Such pits could be observed in transmitted light microscopic studies as shown in Fig.
3a. Such pictures have been published before [10,16,17], but they are shown here again to
contrast the surface activity of L.f. The white spots are etching pits, which bacteria
created, making the thin pyrite layer more transparent for light. Chains of such etching pits
are clearly recognized and evidence a local multiplication of bacterial cells. Highresolution electron microscopical studies of such pits revealed that the nanocrystalline
pyrite substrate has been chemically leached as shown in Fig. 3b. Here an etching pit is
1050
seen on the left hand side and the shadow of a bacterial cell on the right. As explained in
preceding publications, such etching patterns arise when bacterial cells with their capsule
interact with the pyrite interface. Fe3+, H+, or the addition of a complex forming thiol such
as cysteine are cycled as catalysts to break-up interfacial chemical bonds of pyrite [19].
A.f. can use Fe3+ for oxidation of pyrite, however, this oxidation does not lead to an
essential disruption of chemical bonds because the energy bands from which electrons are
extracted have a non-bonding nature being d-states of iron. When, however, a complexing
agent like cysteine is added, then the surface can be disrupted. This has been shown in
complementary experiments in which pyrite layers were exposed to cysteine-containing
solution. It could be shown that pyrite is indeed dissolved by this complex forming thiol.
Figure 3. Optical transmission photographs of 100 nm thin synthetic pyrite layers

which are dotted with corrosion pits generated by bacterial activity. a) shows chains
of corrosion pits indicating cell multiplication on the pyrite surface; b) shows a
transmission electron micrograph of a corrosion pit with a bacterium nearby
It can, therefore, be understood why A.f. is able to dissolve pyrite. This bacterium is
specialized for the recovery of the sulfide species and is, therefore, able to disintegrate the
crystallized solid material. How can, on the other hand, a bacterium such as L.f., which is
only able to oxidize Fe2+, attack a pyrite interface where electron extraction does not
primarily break chemical bonds due to the non-bonding character of the energy states
involved (within 1 eV from the valence band edge)?
2.3 Comparative evaluation of leaching mechanisms
Since electrons extracted from the pyrite valence band by Fe3+ do not generate broken
bonds, a straightforward dissolution of pyrite by Fe3+ can, therefore, not be expected
[10,19]. Only when the pyrite electrode is electrochemically polarized beyond the standard
redox potential of Fe2+/3+
Eh = Eo
RT Fe 2 +
ln
zF Fe3+
(1)
an electrochemical dissolution of pyrite will occur. It is described by equation (2), which

shows that water molecules are needed for this dissolution:
FeS 2 + 8H 2O 14e + Fe 2+ + 2 SO42 + 16 H +
(2)
This has been verified by experiments in which small concentrations of water have
been added to organic electrolytes. It is known from literature that L.f. can oxidize Fe2+ as
an energy source at very high concentrations of Fe3+ corresponding to potentials, which
are approx. 200 mV more positive compared to the iron oxidation abilities of A.f. [9-13].
L.f., which is able to oxidize Fe2+, however not sulfur containing species, may therefore be
1051
able to oxidize pyrite by locally inducing a very positive Fe2+/3+ redox potential as shown
in Fig. 4a.
B
FeS2
Fe2+ + 2x +2SO42-
biofilm
FeS2
Fe2+2+
Fe
S2O 32-
Fe
3+
Fe3+
Fe
2+
Acidothiobacillus
ferrooxidans
Fe2+
2+
Fe
anodic
Fe
2+
e e cathodic
-
nanosized-FeS2
Fe2+
3+
Fe
H+
H
Fe2+
3+
Fe
FeS2 + 2x +4O2
(X= H+ +, Fe3+; Y (eg. cysteine))
dispersed sulfur
nanoparticles
ee-
FeS2-mineralized
Leptospirillum ferrooxidans
Figure 4. Schemes explaining the mechanisms for pyrite dissolution applied by L.f.
(a) and of A.f. (b)
The resulting local anodic dissolution of pyrite will liberate pyrite nanoparticles,
which are indeed found in the capsule of this bacterial species (Fig.1). The thus generated
14 electrons per pyrite molecule may be driven to a cathodic site on the pyrite surface,
where they may be engaged in the reduction of Fe3+ species according to equation (3).
14e + 14 Fe3+ 14 Fe 2+
(3)
When equation (2) and (3) are summed up, equation (4) results which corresponds to
the equation proposed by Boone et al. [1-2] for ferric leaching of pyrite ores:
FeS 2 + 8H 2O + 14 Fe3+ 15Fe 2+ + 2 SO42 + 16 H +
(4)
While the same global mechanism results, the strategy, which had to be adopted by
L.f. bacterial cells to accomplish this reaction turned out to be quite different from the
mechanism originally assumed for ferric leaching. We can use standard thermodynamic
considerations as applied to semiconductor electrochemistry [20] to understand the
process. When the anodic dissolution process (2), which is proposed to be driven by
bacterial activity is coupled with the reaction proceeding at the standard hydrogen
electrode,
14 H + + 14e 7 H 2
(5)
the following global reaction results which for simplicity is assumed to be reversible.
FeS 2 + 8 H 2O FeSO4 + H 2 SO4 + 7 H 2
(6)
Knowing the G value (125,6 kcal/M) of this reaction, one can deduce the
thermodynamic decomposition potential (against NHE). It is:
G 0
= +0.39V
(7)
+ zF
The decomposition may proceed if the redox potential of positive charges or holes is
more positive than the decomposition potential (7), provided there are no kinetic barriers.
Eo =
1052
Such kinetic barriers exist for pyrite. They arise because of the electronic d-character of its
valence band. The position of the band edges varies if the electric charge on the surface
varies, either by adsorbed ions, charged surface complexes or by high concentrations of
holes in the surface. The increase of free energy of holes by electron extraction via L.f.,
which determine the oxidative dissolution of pyrite is described by
p + p L
L
ln
E
E
kT
p F
V
N
V
(8)
(EV = position of valence band edge, NV = concentration of states in the valence band, PL
= concentration of holes generated by electron extraction through L.f. beyond the existing
concentration p). In this relation Fe3+ mediated electron extraction leading to an increase
in PL will pull the potential of holes towards positive values. Thus process will be
accompanied by a positive shift of the valence band position EV due to positive charging
of the pyrite surface. L.f. has obviously adapted to drive FeS2 beyond the decomposition
potential and its kinetic barrier.
A strategy was necessary which resembles the phenomenon of electrochemical
corrosion in which anodic and cathodic sites develop on the same micro heterogeneous
interface. Such a scenario is shown for L.f. activity and pyrite in Fig. 4a. Here high
concentrations of bacterial cells are locally inducing anodic electrochemical dissolution
processes while the thus liberated electrons are moving to cathodic sites where Fe3+ is
reduced. The Fe2+ then diffuses to the bacteria where the oxidation of this energy source
will occur. Experimental support for such a process are the pronounced large corrosion
craters produced by L.f. and the high concentration of bacterial cells sitting in these
craters. This special surface morphology created by bacterial activity may allow an
intensive circulation of Fe2+ for the bacterial energy supply needed for CO2-fixation.
While L.f. can sustain its metabolism by utilizing Fe2+ at very positive redox
potentials, approx. 200 mV more positive compared to A.f., the latter species operates in a
potential region where pyrite is typically not electrochemically dissolved. Since Fe3+ in
small concentrations have only a very limited effect on disruption of chemical bonds in
the pyrite interface, A.f. had to adopt a different strategy for energy recovery. This strategy
is shown in a simplified way in Fig. 4b (for more detail see [19]. Basically this bacterium
utilized protons, Fe3+, and complexing thiols (i.e extracellular polymeric substances (EPS)
containing cysteine or cysteine-containing metabolites) for a catalytic disruption of pyrite
interfaces. This works in such a way that protons interacting with the interface may
partially break bonds forming interfacial SH- groups. Fe3+ may extract electrons from thus
created interfacial states and cysteine may form complexes with the surface thus
disrupting it and extracting iron sulfur aggregates bonded to cysteine. Since protons, Fe3+,
species and cysteine are recycled this process may be considered catalytic. Viewed
globally, the process developed by L.f. is also catalytic (since Fe3+ is recycled, Fig. 4a),
however, the individual steps are electrochemical and require a separation of anodic and
cathodic sites as known from electrochemical corrosion mechanisms.
The effect of exogenous cysteine on the bacterial dissolution of pyrite has been
demonstrated [12], but so far no proof has been given that cysteine is actually present in
EPS secreted by A.f. in sufficient concentration. Though the EPS composition of some
determined A.f. strains have been published, it cannot be extrapolated to bacteria directly
interacting with the sulfide substrate. In-situ determinations of cysteine and cysteinecontaining metabolites in the presence of a biofilm on solid sulfides are required. In
addition, cysteine and methionine have recently been found In EPS produced by sulfatereducing bacteria during corrosion of mild steel [21].
1053
These distinct leaching strategies adopted by both pyrite oxidizing bacteria species
show the evolutionary adaptability of bacteria to their inorganic energy source.
Leptospirillum populations have apparently "discovered" that pyrite disintegrates under
very positive Fe2+/3+ potentials and have improved their ability to recover reducing power
for Fe3+ reduction by organizing themselves for cooperative leaching activities within
corrosion craters which apparently improve their ability to recover Fe2+ species.
Summarizing, the following conclusions can be drawn:
While A.f. takes advantage of thiol complex chemistry to brake up sulfur species, their
main energy source, from the sulfide interface, which are subsequently oxidized to sulfate
for energy recovery, the survival interest of L.f. is aimed at recovering a maximum amount
of reduced iron from the FeS2. Its interfacial activity aims at pushing the FeS2 towards
electrochemical oxidation (iron sulfate formation) so that the reducing power of electrons
can be harnessed for Fe3+ reduction elsewhere on the sulfide surface. L.f. bacteria
apparently induce electrochemical corrosion by locally polarizing the FeS2 interface with
Fe3+ immobilized in their capsule, thus oxidizing the sulfur species while generating
microscopic cathodic currents for Fe3+ reduction.
The specific leaching behavior of both species is not adapted for an efficient joint
leaching activity. L.f. aims at a more positive redox potential and at sacrificing the energy
rich sulfur species while utilizing electrochemically mobilized electrons for Fe3+ reduction
to recover a maximum amount of Fe2+. Since very little energy is generated in the oxidation of
ferrous to ferric ions, the bacteria must oxidize large amounts of iron in order to grow. A.f., on the
other hand, is adapted to extract the sulfur species for its sulfur metabolism, which
provides a high energy turnover and Fe2+ oxidation is just a parallel process which yields
significantly less energy but supports the interfacial leaching process by providing Fe3+.
ACKNOWLEDGEMENTS
The authors would like to thank Ms U. Bloeck and Mr. M. Wilhelm for their valuable
assistance in electron microscopy.
REFERENCES
1. M. Boon and J. J. Heijnen, Biohydrometallurgical Technologies, A. E. Torma, J. E.

Wey and V. L. Lakshmanan (eds.), The Minerals, Metals & Materials Society,
Warrendale Pennsylvania, (1993) 217.
2. M. Boon and J. J. Heijnen, Hydrometallurgy 48 (1998) 27.
3. W. Sand, K. K. Rohde, B. Sobotke and C. Zenneck, Appl. Environ. Microbiol., 58
(1992) 85.
4. T. Rohwerder, A. Schippers and W. Sand, Proceedings of the XX International
Mineral Processing Congress, Vol. 4, Solid-Liquid Separation, Hydro- und
Biohydrometallurgy, H. Hoberg and H. von Blottnitz (eds.), Gesellschaft fr Bergbau,
Metallurgy, Rohstoff- und Umwelttechnik (GMDB), Clausthal Zellerfeld, (1997) 475.
5. T. Rohwerder, A. Schippers, W. Sand, Thermochimica Acta, 309 (1998) 79.
6. P. R. Norris and D. P. Kelly, The use of mixed microbial cultures in metal recovery.
Microbial Interactions and Communities, A. T. Bull and J. H. Slater (eds.), Academic
Press London, (1982) 443.
7. P. R. Norri, Recent Progress in Biohydrometallurgy, G. Rossi and A. E. Torma (eds.),
Associazione Mieraria Sarda, Rom, Italy, (1983) 83.
1054
8. W. Sand, T. Gehrke, P.-G. Jozsa, A. Schippers, Proceedings of IBS 99.

Biohydrometallurgy and the Environment Towards Mining of the 21st Century,
Elsevier, (1999) 27.
9. G. S. Hansford and T. Vargas, Proceedings of IBS 99. Biohydrometallurgy and the
Environment Towards Mining of the 21st Century, Elsevier, (1999) 13.
10. H. Tributsch, Proceedings of IBS 99. Biohydrometallurgy and the Environment
Towards Mining of the 21st Century, Elsevier, (1999) 51.
11. J. Rojas-Chapana and H. Tributsch, Hydrometallurgy 59 (2001) 291-300.
12. J. A. Rojas-Chapana and H. Tributsch, Process Biochemistry 35 (8) (2000) 815.
13. D. Rawlings, H. Tributsch, G. Hansford, Microbiology 145 (1999) 5-13.
14. F. Battaglia-Brunet, P. dHughes, T. Cabral, P. Cezac, J. L. Garcia, D. Morin, Minerals
Engineering 11 (1998) 195.
15. O. Tuovinen, D. Kelly, Arch Mikrobiol 88(1973) 285.
16. J. Rojas-Chapana, M. Giersig, H. Tributsch, Arch. Microbiol 163 (1995) 352.
17. H. Tributsch, J. Rojas-Chapana, C. Baertels, A. Ennaoui, W. Hofmann, Corrosion 54
(1998) 216.
18. J. A. Rojas-Chapana and H. Tributsch, submitted to FEMS Microbiol.
19. H. Tributsch and J. A. Rojas-Chapana, Electrochimica Acta 45 (2000) 4705.
20. H. Gerischer, J. Vac. Sci. Technol.15 (1978) 1422.
21. K. Chan, L.. Xu, H. Fang. Environ. Sci. Technol. 36 (2002) 1720
1055

Composition of biofilm communities in acidic mine waters as

revealed by combined cultivation and biomolecular approaches
Sakurako Kimura, Kris Coupland, Kevin B. Hallberg and D. Barrie Johnson
School of Biological Sciences, University of Wales, Bangor, LL57 2UW, U.K.
Abstract
Waters draining a pyritic outcrop and a disused underground copper mine in north
Wales are both acidic and metal-rich. Both also contain copious amounts of microbial
biomass growing as biofilms ("acid streamers"). We have studied the microbial
communities in these streamers using a combination of cultivation-based and cultivationindependent molecular techniques. Isolation of acidophiles using a range of solid media
indicated that the predominant microbes at one site were iron-oxidizing microbes whereas
iron oxidizers and heterotrophic acidophiles were present in similar numbers in streamers
from the other site. Streamer communities were also analyzed by fluorescent in situ
hybridization (FISH) using a range of oligonucleotide probes of varying specificity. FISH
revealed that the streamers from both mine sites were similar in overall gross structure,
and were dominated (80-90% of the total bacterial counts) by -Proteobacteria. The next
most abundant group of microbes was -Proteobacteria, with Acidithiobacillus
ferrooxidans accounting for about 25-90% of these. Archaea, low G+C Gram-positive
bacteria and At. thiooxidans were not detected in streamers either by FISH and cultivation
techniques, though "Ferrimicrobium"-like bacteria were detected using both approaches.
These data show that the microbial communities of streamer biofilms from two different
mine sites are remarkably similar, but differ substantially from those found in an
abandoned pyrite mine at Iron Mountain, California.
Keywords: acid mine drainage; acid streamers; biodiversity; FISH

1.
INTRODUCTION
The diversity of prokaryotic microorganisms that may live in highly acidic, metal-rich
environments, such as acid mine drainage (AMD) and leachate liquors in biomining
operations is known to be extensive [1]. The most obvious manifestations of microbial life
in these environments are filamentous gelatinous growths, generally referred to as "acid
streamers". The earliest report of acid streamers was by Lackey [2] who observed them in
AMD streams in West Virginia. He originally believed that the long, colorless or light
brown structures were fungal, but microscopic examination showed that they were
"bacterial masses, presumably in a zoogloeal jelly". In contrast, Leathen [3] described acid
streamers as consisting of fibrous masses of sheath-like structures, and that bacterial cells
were absent. Temple and Koehlers [4] description of acid streamers consisting of nonfilamentous, non-orientated bacteria embedded within a tough slime, was more similar to
that of Lackey. They also noted that, inside mines, acid streamers were occasionally
1057
encrusted with orange-brown ferric precipitates and, in AMD streams outside mines,
colonized by phototrophic Euglena spp., giving them a green appearance. Dugan et al. [5]
isolated a neutrophilic Bacillus sp. from acid streamers, which produced copious amounts
of extracellular slime in liquid media; they concluded that this bacterium was a significant
member of the acid streamer microbial community.
Johnson et al. [6] also isolated a variety of neutrophilic bacteria (several of which
produced exopolysaccharides) from acid streamers growing extensively (estimated
biovolume >100 m3) in an abandoned pyrite mine in North Wales. In later work [7, 8],
acidophilic chemolithotrophic and heterotrophic bacteria were also isolated from the same
site. These authors concluded that these acid streamers were a mixed community of ironand sulfur-oxidizing bacteria, which were the primary producers in this pyrite mine, and
that the heterotrophic bacteria lived off exudates and lysates of the primary producers.
Interestingly, one of the isolates (coded "CCH7") was a ferrous iron-oxidizing filamentous
isolate that formed streamer-like growths in vitro. In contrast, Wakao et al. [9] considered
that it was highly unlikely that heterotrophic bacteria could form acid streamers in AMD
and concluded that such growths were composed of Acidithiobacillus ferrooxidans
embedded in a gelatinous matrix.
New insights into the microbial composition of acid streamer-like communities came
from the work of Bond et al. [10] who studied a 1 cm-thick slime biofilm that developed
on the surface of finely disseminated pyrite ore within an AMD stream at Iron Mountain,
California. Phylogenetic analysis of 16S rRNA genes cloned from this microbial
community indicated that the dominant sequences were of bacteria related to the ironoxidizer Leptospirillum. Archaeal (Thermoplasmales lineage), Acidimicrobium/
"Ferrimicrobium" and -proteobacterial gene sequences were also detected. On the basis
of these (and other) results, a variety of gene probes were designed and the slime
community further analyzed using fluorescent in situ hybridization (FISH). Whilst these
findings contrasted greatly with earlier results that were based chiefly on cultivation and
physiological criteria, the environment within Iron Mountain was also very different
(temperature 30-50C; pH typically 0.5-1.0) from the cooler, less acidic (pH 2-3) AMD
waters in which acid streamers have been reported elsewhere.
In this paper, we describe the microbial diversity in acid streamers from AMD at two
mine sites in north Wales, as revealed by using a combination of cultivation and molecular
approaches.
2.
2.1 Mine sites and water analysis

Acid streamers were sampled at two mine sites in north Wales: the former Parys
copper mine and Trefriw Spa, which is situated below the abandoned Cae Coch pyrite
mine [11]. Small streamer biofilm samples (ca. 1 cm3) were removed and taken to the
laboratory with one hour. On site analysis of pH, redox potential, temperature and
conductivity used a Whatman Water Tester (Whatman, U.K.); dissolved oxygen (DO) was
measured using a D400 DO meter (Whatman, U.K.). Ferrous iron was measured in filtered
(0.2 m pore-size) samples using the ferrozine assay [12]; other metals were analyzed by
atomic absorption spectrometry, and sulfate was measured turbidometrically as barium
sulfate (Hydrocheck, Cambridge, U.K.). Dissolved organic carbon was measured using a
Protoc Analyzer (Pollution & Process Monitoring Ltd., U.K.).
1058
2.2 Isolation of bacteria

Streamer samples were put into bottles containing basal salts solution [13], adjusted to
pH 3.5, and vortexed for 5 min. The suspension was then centrifuged at low speed to
remove undispersed streamer material, the supernatant was serially diluted, and samples
plated onto a variety of solid media. The majority of these were overlay media, which
facilitate the growth of a wide range of autotrophic and heterotrophic acidophiles. Full
details of the media can be found elsewhere [13, 14]. R2 agar [15] was used to enumerate
neutrophilic bacteria. Inoculated plates were incubated for up to 6 weeks at 20C under
aerobic and anaerobic conditions (the latter using AnaeroGen system; Oxoid, UK).
2.3 Phylogenetic analysis of isolates
The 16S rRNA genes of streamer isolates were amplified, using protocols described
elsewhere [16]. Purified PCR products were sequenced using a BeckmanCoulter dye
Terminator Cycle Sequencing kit and a CEQ8000 Genetic Analysis System (Beckman
Coulter, U.K.). The resulting gene sequences were compared with those available in
GenBank using BLAST [17].
2.4 Total bacterial counts
Bacteria in disrupted streamer suspensions (section 2.2) were harvested, washed with
phosphate buffered saline (PBS) and the re-suspended cells fixed in 3% (v/v)
paraformaldehyde in PBS. Fixed cells were washed with PBS, suspensions filtered
through 25 mm black polycarbonate membranes (0.2 m pore size) and stained with 4, 6diamidino-2-phenylindole (DAPI; 10 ml of a 1 g/ml solution). Membranes were placed
onto glass slides and viewed using a Nikon ECLIPSE E600 microscope. A minimum
number of 100 bacteria/membrane were counted.
2.5 Fluorescent in situ hybridization (FISH)
The fixed streamer cells (section 2.4) were also utilized for FISH analysis, using a
modified method of that described by Bond and Banfield [18]. The RNA probes used are
listed in Table 1. Where the probe used targeted Gram-positive bacteria, fixed cells were
first incubated (4C; 10 min) with lysozyme (1027.5 units of activity/ml) prior to
dehydration with ethanol. Hybridization was carried out using 25 ng of fluorescein-labeled
eubacterial probe (EUB338Fl) and 25 ng of a more specific, Cy3-labeled probe at the
same time. Mounting medium (70% glycerol in 100 mM sodium tetraborate, pH 9.2,
containing 3 mg/ml N-propyll gallate) was applied to reduce fading of the probe signal.
Various concentrations of formamide (0-50%, v/v) were tested to optimize specificity and
maximize signal response for each probe, using pure cultures of target and related
microorganisms (Table 1). To enumerate different groups of bacteria, Cy3-labeled cells
were counted relative to those stained with the EUB338Fl probe. For the eubacterial and
archaeal probes, counts of stained cells were made relative to those stained by DAPI.
1059
Table 1. List of 16S rRNA (except where noted) oligonucleotide probes used for FISH analysis and formamide concentrations
required for optimum specificity. All probes, except for EUB338Fl, were labeled with the fluorochrome Cy3.
Probe
EUB338Fl
ALF1B
Target
Eubacteria
-Proteobacteria, some Proteobacteria and most spirochetes
23S rRNA of most -Proteobacteria
23S rRNA of most -Proteobacteria
Thiomonas Group 1
Thiomonas Group 2
Acidithiobacillus ferrooxidans
Acidithiobacillus thiooxidans
Acidimicrobium and Ferrimicrobium
Acidimicrobium ferrooxidans
Low G+C Gram-positive bacteria
Leptospirillum groups 1, 2 and 3
Archaea
Probe sequence (5f3)

GCTGCCTCCCGTAGGAGT
CGTTCGYTCTGAGCCAG1
Formamide (%)
0 50
20
Reference
[19]
[20]
BET42a2
GCCTTCCCACTTCGTTT
35
[20]
2
GAM41a
GCCTTCCCACATCGTTT
35
[20]
TM1G138
GCAGTTATCCCCCATCAAT
40
This study
TM2G138
GTAGTTATCCCCCATCACA
40
This study
TF539
CAGACCTAACGTACCGCC
20
[21]
ATT223
AGACGTAGGCTCCTCTTC
40
This study
ACM732
GTACCGGCCCAGATCGCTG
35
[18]
ACM995
CTCTGCGGCTTTTCCCTCCATG
10
P. Norris, unpublished
LGC355
GGAAGATTCCCTACTGCTG
20
This study
LF655
CGCTTCCCTCTCCCAGCCT
35
[18]
ARCH915
GTGCTCCCCCGCCAATTCCT
40
[22]
1. Y = C or T
2. Hybridization using these two probes was carried out in the presence of a 10-fold excess of the other unlabeled oligonucleotide
3.
RESULTS
3.1 Mine water analysis

Physico-chemical data from the two mine sites from where streamer biofilms were
sampled are shown in Table 2. At both sites, conditions were reducing, with both
dissolved oxygen and redox potentials being relatively low, and all of the soluble iron
present was in the ferrous form.
Table 2. Physicochemical data from the two AMD sites from which streamer biofilms
were obtained. N.D. = not determined
pH
T (C)
Eh (mV)
Ec (S/cm)
DO2 (mg/L)
DOC (mg/L)
Fe2+ (mg/L)
SO42- (mg/L)
Al (mg/L)
Cu (mg/L)
Zn (mg/L)
Parys mine
2.70
8.8
+420
2994
0.55
6.0
473
1552
115
50
60
Trefriw spa
3.05
13.3
+374
1635
1.66
14.9
217
1093
N.D.
N.D.
N.D.
3.2 Plate and DAPI Counts

Colonies that grew on overlay media were differentiated on the basis of whether or
not they became encrusted with ferric iron (i.e. iron-oxidizing autotrophs and
heterotrophs), and on other morphological characteristics (color, size, etc.). Iron-oxidizing
isolates were further differentiated into moderate and extreme acidophiles, depending on
whether they grew on ferrous iron/thiosulfate (pH ~4) or ferrous iron (pH ~2.5) overlay
plates [14]. Enumeration of bacteria in dispersed biofilms determined by plate counts (on
aerobically-incubated plates) is shown in Table 3. Plating efficiencies, determined relative
to total (DAPI) cell counts were 1.8% for the Parys streamers and 0.96% of the Trefriw
streamers. Numbers of isolates obtained on solid media incubated under anaerobic
conditions were considerably, 102- to 103-fold, fewer than obtained with plates incubated
aerobically.
3.3 Phylogenetic analysis of streamer isolates
Identification of representative streamer isolates, differentiated on the basis of colony
morphologies and identified from analysis of amplified 16S rRNA genes, is shown in
Table 4. All of the iron-oxidizing autotrophic isolates examined in streamers at both sites
were At. ferrooxidans rather than Leptospirillum spp. In addition, "Ferrimicrobium"-like
isolates (Actinobacteria) were isolated from both streamers, as were Acidiphilium and
Acidocella spp. (-Proteobacteria). Parys streamer also included an isolate remotely
related to Acidobacterium capsulatum, a Frateuria-like isolate (-Proteobacteria), and a
Thiomonas-like isolate (-Proteobacteria). Interestingly, the latter was initially classified
as a heterotroph (being isolated on yeast extract solid medium) though it was later
confirmed that it was capable of oxidizing thiosulfate [23].
1061
Table 3. Plate counts (aerobic) and total microbes in the dispersed streamer samples
from the two sites (CFU per g wet weight dispersed streamer)
Parys streamers
Trefriw streamers
9.36 x 106
2.42 x 103
4.40 x 105
<102
9.80 x 106
5.44 x108
1.16 x 107
<102
4.80 x 106
<102
1.64 x 107
1.7 x109
Iron-oxidizing bacteria
- extreme acidophiles
- moderate acidophiles
Heterotrophic acidophiles
Neutrophiles
Total plate counts
Total counts (DAPI)
Table 4. Identification of acid streamer isolates based on analyses of amplified 16S

rRNA genes
(i)
Parys streamers
Isolate
Physiological class
Nearest relative (% identity)
KP1
Iron-oxidizing heterotroph
Uncultured eubacterium TRA2-10 (99.9)

Ferrimicrobium acidophilum (93.0)
PK11
Iron-oxidizing autotroph
At. ferrooxidans NO-37 (99.9)
PK51
Heterotroph
-proteobacterium WJ2 (100)
PK40
Heterotroph
Acidiphilium NO-17 (98.0)
PK46
Heterotroph
Acidiphilium organovorum (98.5)
PK48
Heterotroph
Acidiphilium acidophilum (97.5)
M21
Heterotroph
Acidocella NO-12 (99.4)
PK35
Heterotroph
Acidobacterium WJ7 (99.5)
KP3
Heterotroph
Uncultured eubacterium TRB82 (96.7)

Acidobacterium capsulatum (91.8)
PK44
Heterotroph
Thiomonas CO2 (97.4)
ii)
Trefriw streamers
Isolate
Physiological class
Nearest relative (% identity)
CS11
Iron-oxidizing heterotroph
Uncultured eubacterium TRA2-10 (99.7)

Ferrimicrobium acidophilum (93.0)
CCW10
At. ferrooxidans NO-37 (99.5)
CCW68
At. ferrooxidans ATCC 19859 (99.0)
CCW30
Heterotroph
Acidocella NO-12 (99.1)
CCP3
Heterotroph
Acidiphilium NO-17 (99.5)
3.4 FISH analysis of dispersed streamers

Analysis of microorganisms in the disrupted streamers by FISH produced data that
both confirmed and contrasted with results from cultivation techniques. The prokaryotes
in the streamers were shown to be exclusively (>99.9%) bacteria; no cells were detected
using the archaeal probe. Also in line with plating data, probes targeting Leptospirillum
spp., At. thiooxidans and low G+C Gram-positive bacteria also failed to detect any cells.
1062
The Gram-positive bacteria that were detected by FISH were exclusively Actinobacteria
detected by the Acidimicrobium/Ferrimicrobium probe but not by the Am.
ferrooxidans-specific probe.
The vast majority of cells in both streamer biofilms were Proteobacteria. Whilst -,
- and -Proteobacteria were all detected by FISH, the dominant group in both Parys and
Trefriw streamers were -Proteobacteria (Figure 1). The only -proteobacterium isolated
was a Thiomonas species from the Parys streamer. By FISH analysis with the
corresponding probe (TM1G138), however, this microbe was detected at less than 1 cell
per thousand in the Parys streamer and was not detected at all in the Trefriw streamer. A
second probe (TM2G138), targeting another group of Thiomonas species [23], also failed
to detect cells in either streamer.
Figure 1. Counts of microbes in the streamer samples targeted by a range of Cy3labeled oligonucleotide probes, given as percentages of the fluorescein-labeled
Eubacterial probe (EUB338Fl)
4.
DISCUSSION
The application of molecular techniques is increasingly becoming important in
microbial ecology [24, 25]. The power of these techniques lies in the ability to identify
and enumerate microorganisms without the reliance on cultivation. Cultivation-based
studies often lead to incorrect conclusions, such as the importance of Acidithiobacillus
ferrooxidans in commercial bioleaching operations, which have been shown by a
molecular approach to be dominated by Leptospirillum ferrooxidans and relatives [26].
In this study, we have taken the rarely used approach of combining cultivation with
molecular studies to characterize the microbial populations of streamer biofilms that
dominate mine water draining two metal mines in north Wales. Earlier cultivation-based
studies were carried out on both sites, and the identities of the microbes present were
determined solely on physiological characteristics. Since that time, improved solid media
for the cultivation of acidophiles have been developed, and were employed in the present
study along with 16S rRNA-based identification of the isolates. As would be expected
from earlier studies, the dominant cultivatable isolates from both sites were iron-oxidizing
microbes and the majority of these were found, by 16S rRNA gene sequence analysis, to
be Acidithiobacillus ferrooxidans.
1063
The heterotrophs isolated from both streamers proved to be more diverse than
previously thought, especially in the Parys streamer. The heterotrophs from both streamers
included isolates from the well-known genera of acidophiles Acidiphilium and Acidocella,
and amongst these were isolates that may represent new species of these genera. Some
previously unrecognized microbes also inhabit the streamer from Parys Mountain. One
isolate, PK51, is related to neutrophilic microbes of the genus Frateuria, PK51, and two
other isolates are related to Acidobacterium capsulatum. The homology of these isolates to
Ab. capsulatum and to each other is sufficiently low that they can be considered to belong
to different genera. It is interesting to note that few microbes of the phylogenetic group
that includes Ab. capsulatum have been cultivated, even though molecular studies show
that they are the dominant microbes in many habitats [e.g. 27]. The latter two groups of
microbes have also been isolated from a wetland constructed to treat acid mine drainage
[14].
As with the cultivation-based study, the populations of both of the streamers appeared
to be similar when analyzed with probes targeting ribosomal RNA. There was no
indication of archaea present in either streamer community, as indicated by FISH and also
by the absence of an amplification product following PCR with archaeal 16S rRNA genespecific primers (data not shown). In contrast to the cultivation study, both of the
streamers were dominated by -Proteobacteria and not -Proteobacteria, the probe for
which detects At. ferrooxidans. The only isolate obtained from either of the streamers,
PK44, that belongs to the -Proteobacteria is related to the genus Thiomonas. Using
probes to target the 16S rRNA gene of bacteria in this genus, however, it was shown that
they are not the dominant microbes in either streamer, in keeping with the cultivationbased studies. Further molecular characterization of these streamers, including the
generation of 16S rRNA gene libraries, is currently in progress and will reveal the identity
of the dominant microbes in the streamer biofilms. Identification of the unknown microbe,
or microbes, that dominate these streamers will be important to understanding the
biogeochemistry of these microbial growths that are very prominent in AMD.
Even though the streamers studied here were quite similar to each other, they differed
substantially from similar microbial growths found in AMD at Iron Mountain, California.
Analysis of 16S rRNA genes cloned from Iron Mountain streamer material showed that
those streamers were dominated by Leptospirillum-like microbes [28], which have only
been found there to date, and by archaea. Other microbes found were related to
Ferrimicrobium acidophilium and Acidimicrobium ferrooxidans. The Parys and Trefriw
isolates related to Ferrimicrobium were highly related (>99% 16S rRNA gene
homology) to the Iron Mountain clone TRA2-10. FISH analysis [10] confirmed that the
major microbes in the Iron Mountain streamers were the Leptospirillum-like microbes and
archaea, the latter namely Ferroplasma. Neither the Acidimicrobium/Ferrimicrobium
type bacteria, nor At. ferrooxidans, were detected by FISH, in contrast to the Parys and
Trefriw streamers.
While it is difficult to say what determines the differences in the microbial
constitution of these various streamers, certainly the physico-chemical parameters play a
central role, with temperature (~40C at Iron Mountain vs ~10C) and pH (<2 at Iron
Mountain vs ~3) having significant effects.
ACKNOWLEDGEMENTS
Kris Coupland is grateful to the NERC (U.K.; ref. NER/S/C/2001/06450) and Rio
Tinto Technologies for provision of a research studentship. K.B.H. and D.B.J. thank the
BBSRC and DTI (U.K.; ref. 5/BRM18412) for partial support of this work.
1064
REFERENCES
1. K. B. Hallberg and D. B. Johnson, Adv. Appl. Microbiol. 49 (2001) 37.

2. J. B. Lackey, Public Health Rep. 53 (1938) 1499.
3. W. W. Leathen (1952) Special Summary Report of Microbiological Studies of
Bituminous Coal MIne Drainage, Mellon Institute of Industrial Research, University
of Pittsburg, Pittsburg, Pa.
4. K. L. Temple and W. A. Kohler (1954) In Research Bulletin No. 25, Engineering Exp.
Station, University of West Virginia, Morgantown.
5. P. R. Dugan, C. B. Macmillan and R. M. Pfister, J. Bacteriol. 101 (1970) 982.
6. D. B. Johnson, W. I. Kelso and D. A. Jenkins, Environ. Pollut. 18 (1979) 107.
7. D. B. Johnson, M. A. Ghauri and M. F. Said, Appl. Environ. Microbiol. 58 (1992)
1423.
8. S. McGinness and D. B. Johnson, Sci. Total Environ. 132 (1993) 27.
9. N. Wakao, H. Tachibana, Y. Tanaka, Y. Sakurai and H. Shiota, J. Gen. Appl.
Microbiol. 31 (1985) 17.
10. P. L. Bond, G. K. Druschel and J. F. Banfield, Appl. Environ. Microbiol. 66 (2000)
4962.
11. D. A. Jenkins and D. B. Johnson, J. Russell Soc. 5 (1993) 40.
12. D. R. Lovley and E. J. P. Phillips, Appl. Environ. Microbiol. 53 (1987) 1536.
13. D. B. Johnson, J. Microbiol. Meth. 23 (1995) 205.
14. K. B. Hallberg and D. B. Johnson, Hydrometallurgy (2003) in press.
15. D. J. Reasoner and E. E. Geldreich, Appl. Environ. Microbiol. 49 (1985) 1.
16. N. Okibe, M. Gericke, K. B. Hallberg and D. B. Johnson, Appl. Environ. Microbiol. 69
(2003) 1936.
17. S. F. Altschul, T. L. Madden, A. A. Schaffer, J. Zhang, Z. Zhang, W. Miller and D. J.
Lipman, Nucleic Acids Res. 25 (1997) 3389.
18. P. L. Bond and J. F. Banfield, Microbial Ecol. 41 (2001) 149.
19. R. I. Amann, L. Krumholz and D. A. Stahl, J. Bacteriol. 172 (1990) 762.
20. W. Manz, R. Amann, W. Ludwig, M. Wagner and K. H. Schleifer, Syst. Appl.
Microbiol. 15 (1992) 593.
21. M. O. Schrenk, K. J. Edwards, R. M. Goodman, R. J. Hamers and J. F. Banfield,
Science 279 (1998) 1519.
22. D. A. Stahl and R. Amann, Nucleic Acid Techniques in Bacterial Systematics, E.
Stackebrandt and M. Goodfellow (eds.), Wiley, New York, (1991) 205.
23. K. Coupland, F. Battaglia-Brunet, K. B. Hallberg, M. C. Dictor, F. Garrido and D. B.
Johnson, This Volume
24. I. Amann, W. Ludwig and K. H. Schleifer, Microbiol. Rev. 59 (1995) 143.
25. I. M. Head, J. R. Saunders and R. W. Pickup, Microbial Ecol. 35 (1998) 1.
26. N. J. Coram and D. E. Rawlings, Appl. Environ. Microbiol. 68 (2002) 838.
27. P. Hugenholtz, B. M. Goebel and N. R. Pace, J. Bacteriol. 180 (1998) 4765.
28. P. L. Bond, S. P. Smriga and J. F. Banfield, Appl. Environ. Microbiol. 66 (2000) 3842.
1065

Computational fluid dynamics simulation of immobilized

Acidothiobacillus ferrooxidans
B. Metodieva, L. Lilovab and D. Karamanevb
a
Blue EdgeValley, 103 Gerald Cr., London, Ontario, Canada N5Z 5A4
Department of Chemical and Biochemical Engineering, The University of Western
Ontario, London, Ontario, Canada N6A 5B9
Abstract
This paper presents the results of the simulation, using computational fluid dynamics,
of the flow and ferric iron concentrations profiles around a single cell of A. ferrooxidans,
immobilized on the surface of a sulfide crystal. It has been shown that there are significant
concentration gradients of ferric iron concentration between the surface of the crystal and
the bulk of liquid. The gradient can reach several hundred mg/L. These results can explain
the difference between the chemical (sterile) and biological oxidation rates of metal
sulfides in liquid media containing iron ions.
Keywords: computational fluid dynamics, Acidithiobacillus ferrooxidans, bioleaching

1.
INTRODUCTION
The biological oxidation of metal sulfides by Acidothiobacillus ferrooxidans is of
great importance for the hydrometallurgical production of base, heavy and noble metals
such as copper, zinc, cobalt, uranium and gold. Because of the practical importance of
these processes, the mechanism of sulfide biooxidation has been a subject of intensive
studies in the recent decades. Two main mechanisms of the oxidation kinetics have been
proposed [1]:
the direct oxidation is based on the oxidation of sulfide minerals by an immobilized

on their surface microbial cell in the absence of intermediate oxidants such as ferric
iron. It is based on the following reaction (in the case of pyrite oxidation):
(1)
4FeS2 + 15O2 + 2H2O = 2Fe2(SO4)3 + 2H2SO4
the indirect oxidation is based on the ability of A. ferrooxidans and other ironoxidizing microorganisms to oxidize ferrous ions in aqueous solution:
(2)
2FeSO4 + H2SO4 + O2 = Fe2(SO4)3 + H2O
The ferric ions, produced as a result of this reaction, then can chemically oxidize
metal sulfides, such as pyrite:
7Fe2(SO4)3 + FeS2 + 8H2O = 15FeSO4 + 8H2SO4
(3)
The ferrous ions produced by reaction (3) are then oxidized by A. ferrooxidans to
ferric ions according to reaction (2). The ferric sulfate, produced by reaction (2) is further
1067
reused for the chemical oxidation of pyrite (reaction 3), which closes the cycle of iron ions
oxidation-reduction.
After many years of research, it seems that a consensus has been reached that the
indirect mechanism is if not the only [2-4], at least the predominant [5] mechanism in
sulfide biooxidation by A. ferrooxidans. However, the rate of purely chemical (sterile)
oxidation of pyrite by aqueous solutions of ferric sulfate is by an order of magnitude lower
than that in the presence of iron oxidizing microorganisms [6]. Different explanations
have been proposed to explain this discrepancy.
Boon and Heijnen [6] explained the higher rate of the biological sulfide oxidation by
the fact that the ferrous ions, produced as a result of the sulfide oxidation (3) are quickly
reoxidized by microorganisms, which keeps the redox potential high; in the sterile
oxidation of sulfides by ferric iron, the redox potential falls quickly because of the
decrease in the ratio of ferric to ferrous ions concentration. Alternatively, Gehrke et al. [7]
explained the acceleration of the bacterial process over the sterile one by the quick
reoxidation of iron, which is complexed with microbial extracellular polymeric
substances.
The purely hydrodynamic retention of ferric ions near the point of contact of
microorganisms with the sulfide surface has not been considered yet. In this work, we are
trying to explain the higher rate of bacterial sulfide oxidation, compared to the sterile one,
by the hydrodynamic retention of ferric ions. Computational fluid dynamics simulation is
used to determine the ferric ions concentration profiles in the vicinity of a microorganism,
attached physically to the surface of a sulfide crystal.
2.
NUMERICAL MODEL
The concentration and velocity profiles around a single microbial cell, immobilized
on the flat surface of a sulfide crystal were modeled numerically using the FLUENT
computational fluid dynamics software.
2.1 Assumptions
The following assumptions were made in order to develop the mathematical model:
Geometrically, the A. ferrooxidans cell was approximated as an ellipsoid of rotation
with a size along the axis of symmetry equal to 2 micrometers and a small axis,
perpendicular to it, of 1 micron (Fig. 1). While the size and shape slightly differ in
different strains, the one used here is a typical shape and size. The surface area of the
microbial cell was 5.34x10-12 m2.
The cell is a rigid solid particle, and therefore:
It does not deform as a result of the acting hydrodynamic forces.

The cell does not deform as a result of immobilization on the solid surface,
especially at the point of contact.
The cell is immobilized on a flat surface (Fig. 1).
The immobilization support is chemically inert.
No extacellular polymeric substances (EPS) were assumed to exist between the cell
and the sulfide surface. The available literature data shows that in some cases A.
ferrooxidans forms EPS when immobilized on a sulfide surface, while in other they
are absent.
The ferric ions concentration in the bulk of the liquid is equal to zero.
1068
The ferric ions are produced evenly over the entire surface of the cell and are
completely released into the liquid. The literature data show that the process of iron
oxidation takes place outside of the cell membrane and the iron ions do not permeate
the membrane [8]. The iron oxidation rate was assumed to be equal to 5.2x10-8 (kg
Fe3+)/(s.(m2 of cell surface)).
The approaching liquid flows in either laminar or turbulent mode, parallel to the solid
surface (Fig. 2), with two bulk velocities: 0.1 and 1.0 m/s. The case for 1.0 m/s was
calculated for two different cases laminar and turbulent liquid flow. For the
turbulent flow the turbulence intensity was 5% and the turbulent scale was chosen at
0.01 mm.
The real geometry of the microbial cell has been scaled up by a factor of 100 in order
to avoid round-off error of computing too small numbers (note that FLUENT
computes only in SI units). This required scaling down the liquid velocity by the same
factor in order to keep a constant Reynolds number. The computational mesh
consisted of 0.83 mill tetrahedral cells. The surface mesh of the bacteria is shown in
Fig. 3.
Advance turbulent model, namely the FLUENT Shear-Stress Transport (SST) k-
model was applied to resolve the low-Reynolds number turbulence around the
bacteria.
Figure 1. Geometry of the immobilized microbial cell
1069
Figure 2. Scheme of the modelling space
Figure 3. Bacteria surface mesh
2.2 Numerical results and discussion

The velocity and concentration profiles in the plane x-z (side view) were obtained
(Figs. 4-6) using the model, described above.
1070
The ferric iron concentration profiles in liquid, at the plane of contact with the sulfide
crystal surface, are shown in Fig. 7. Actually, Fig. 7 shows the top view (x-y plane) of
Figs. 4, 5 and 6, assuming that the microbial cell is "transparent".
Figure 4. Velocity and ferric ions concentration profiles around the immobilized cell.
Liquid velocity: 0.1 m/s (laminar flow)
1071
Liquid velocity: 1.0 m/s (laminar flow)
1072
Liquid velocity: 1.0 m/s (turbulent flow)
1073
Figure 7. Top view of the ferric ions concentrations in liquid, contacting the sulfide
crystal surface
The concentration profiles in Fig. 7 show that both the free stream velocity and flow
structure (laminar vs. turbulent) of liquid affect the concentration distribution only
slightly. This is due to the fact that the size of the microorganism is in the order of
micrometers, thus the entire microorganism is within the laminar viscous layer of the
liquid, regardless of the type of the flow laminar or turbulent. Within the liquid viscous
sublayer of 1 micrometer, the liquid velocity is very close to zero, which can be seen also
from the velocity plots in Figure 4. Therefore, the upstream velocity magnitude, as well as
1074
the upstream turbulence does have a significant effect on the concentration distribution
(Fig. 7). The tail of a relatively larger concentration downstream the bacteria is wider at
smaller upstream velocity. At larger velocity the larger amount of the ferric ions are
advected by the flow and the thinner tail of higher concentration is formed downstream of
the microbial cell.
3.
CONCLUSIONS
The results of the simulation show that there is significant spatial variation in ferric
ions concentration in the close surrounding of the microbial cell (Fig. 7). While the
concentration at the vicinity of the contact between the microbial cell and the crystal
surface is well over 100 mg/L, it drops to zero at a distance of 1 micrometer from the cell
surface (except for the tail behind the cell). Therefore, the concentration of ferric ions in
the bulk of liquid, far away from the cell, is different - much lower - than that on the
sulfide surface, surrounding the immobilized cell. The redox potential of the liquid, given
by the Nernst equation:
(4)
E = Eo + (RT/F)ln(Fe3+/Fe2+)
is a strong function of the ferric-to-ferrous ions concentration ratio. Therefore, since the
concentration of ferric ions in the immediate cell surroundings is higher than the bulk
concentration, the redox potential at the surface of the sulfide crystal will be higher that
that in the bulk of liquid. Since the ferric iron concentration at the cell-sulfide interface is
higher than the bulk concentration by approx. 150 mg/L, the redox potential at the cellsulfide interface will be significantly higher than the bulk potential mainly in the cases
when the bulk ferric iron concentration is below approx. 1 g/L.
ACKNOWLEDGEMENTS
This work was supported financially by the Natural Sciences and Engineering
Research Council of Canada and the Premiers Research Excellence Award (Ontario).
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
.E. Torma, Adv. Biochem. Eng., 6 (1977) 1.

W. Sand, T. Gehrke, P.-G. Jozsa and A. Schippers, Hydrometallurgy, 59 (2001)159.
H. Tributsch, Hydrometallurgy, 59 (2001) 177.
G.S. Hansford and T. Vargas, Proc. Metall., 9A (1999) 13.
K. Lilova and D. Karamanev, (paper to appear in this volume)
M. Boon and J. J. Heijnen, Hydrometallurgy, 48 (1998) 27.
T. Gehrke, R. Hallmann, K. Kinzler and W. Sand, Water Sci. Tech., 43 (2001) 159.
W. J. Ingledew, Biochim. Biophys. Acta, 683 (1982) 89.
1075

Contribution to the quantification of the Acidithiobacillus

ferrooxidans biomass concentration from the oxygen uptake
rate
D.S. Savi, V.B. Veljkovi, M.L. Lazi
Faculty of Technology, 16000 Leskovac, Bulevar oslobodjenja 124
Abstract
A simple, rapid and reproducible method for quantifying the A. ferrooxidans biomass
concentration from measuring the oxygen uptake rate is described. The specific oxygen
uptake rate, found to be 0.020 0.001 mg O2/mgd.w./min, for both concentrated cell
suspensions and during ferrous iron biooxidation, is used in the estimation of the viable
biomass concentration. The ferrous iron concentration in samples of the culture medium,
where the oxygen uptake is recorded, should be higher than 0.1 mg/L to avoid substrate
limitation.
Keywords: Acidithiobacillus ferrrooxidans, qantification of viable biomass concentration

1.
INTRODUCTION
The role of the chemolithotroph bacterium Acidithiobacillus ferrrooxidans in the
leaching of metals from sulfide ores has been well recognized. The bacterium obtains
energy for growth by the oxidation of ferrous to ferric iron using atmospheric oxygen, a
desirable reaction from the hydrometallurgical point of view. As in other cases, when
oxygen is closely related to the energetics of microbial growth under aerobic conditions
[1], the oxygen consumption of A. ferrooxidans could be correlated to its growth rate and,
therefore, used for indirect kinetic studies. A few earlier studies on oxidations by A.
ferrooxidans, such as the ferrous iron oxidation rate by cell envelopes [2], the competitive
inhibition of ferrous iron oxidation by cells and ferric iron [3, 4], and the relative
contributions of adhered and free suspended cells to pyrite biooxidation [5], are based on
the measurement of oxygen uptake. The oxygen uptake rate is also suggested for
evaluating bacterial activity for metal leaching [6].
A high degree of synchronization among processes of oxygen uptake, biomass growth
and ferrous iron utilization during ferrous iron oxidation by A. ferrooxidans in different
bioreactors, independently of pH (1.8 or 2.0), has been observed [7]. Thus, the dissolved
oxygen level could be used not only for studying the kinetics of the biooxidation process
but also for determining the viable biomass concentration. Such a method, based on the
measurement of the oxygen uptake rate, would be a simpler technique for quantifying the
A. ferrooxidans growth during the biooxidation of ferrous iron than the standard ones,
which have disadvantages in requiring considerable time, materials and manual labor
(serial end-point dilution, enumeration of colonies) or refers to dead and living cells
1077
(direct microscope counting, the measurement of the total dry weight and protein content).
The viable biomass concentration and the growth rate have been successfully determined
with sufficient accuracy by measuring the oxygen uptake rate during aerobic
fermentations, for instance by Saccharomyces cerevisiae, Streptomyces sp. and
Thermoactinomyces sp. [8] as well as Clostridium boldinii [9].
In the present work, the growth and oxygen uptake of A. ferrooxidans cells growing
on a ferrous iron medium were studied under different conditions with respect to the
ferrous iron concentration, the biomass concentration and the ferric/ferrous ratios. The
main aim was to reflect the relationship between microbial growth and oxygen uptake and
to establish a quantitative base for determining the viable biomass concentration by
measuring the oxygen uptake rate.
2.
2.1 Microorganism and cultivation conditions

A strain of A. ferrooxidans B5, isolated from the Bor Copper Mine (Serbia), was
grown on the following mineral salt medium (in g/L): (NH4)2SO4 3; KH2PO4 0.5;
MgSO4.7H2O 0.5; and KCl 0.1, the initial ferrous iron concentration being 9 g/L. The
culture was grown in an Erlenmeyer flask (2 L) containing 1 L of the medium at 28C on
a rotary shaker at 200 rpm.
2.2 Biomass
The biomass was determined by measuring the dry weight - mgd.w (80C, 24 h) of a
membrane filter (0.45 m pore diameter, Millipore, NVC 45218) after the sample (500
mL) was filtered under pressure (2 bar).
2.3 Preparation of a concentrated cell suspension
The biomass was harvested at the end of the exponential growth phase; it was
therefore assumed that the biomass contained a negligible number of dead cells. A volume
of the growing culture (500 mL; 25 mgd.w./L) was centrifuged at low speed (1000 min-1, 10
min) to remove the insoluble ferric iron, filtered through a membrane filter (0.45 m pore
diameter, Millipore, NVC 45218), washed twice with 0.05 M H2SO4 solution (100 mL)
and once with distilled water (100 mL), and concentrated by resuspending the biomass in
the iron-free salt medium (25 mL). According to the ratio of the oxygen uptake rate of the
original culture and that of the concentrated suspension, about tenfold concentration was
achieved (238 mgd.w./L).
2.4 Oxygen uptake rate
The oxygen uptake rate was calculated from the slope of the linear part of the
relationship between the dissolved oxygen level and time. The dissolved oxygen level was
measured by a Clarke-type oxygen probe (HANNA HI8043), and the oxygen solubility
(6.7 mg O2/L) in the mineral salt medium was estimated [10]. The reaction mixture (5 mL)
was prepared by mixing a certain volume of the concentrated cell suspension with
different volumes of ferrous iron and ferric iron solutions in the mineral salt medium to
achieve different ratios of ferric/ferrous iron. The oxygen uptake rate of the cell-free
mineral medium was found to be negligible (0.007 mg O2/L/min).
1078
2.5 Analytical methods

The ratio of ferrous and ferric ions was calculated from the redox potential measured
using a redox probe (HANNA HI 3932). The pH was monitored with a pH meter
(HANNA HI 9025).
3.

A summary of the effect of different biomass concentrations and ferric/ferrous ratios
on the specific oxygen uptake rate of the concentrated bacterial cells is given in Tables 1
and 2. When the biomass concentration is held constant at a level of 8.0 mgd.w./L and the
ferric/ferrous ratio is varied in the range from 1/10 to 1/0.1, which causes a change in the
pH of the reaction mixture from 1.9 to 1.4, the specific oxygen uptake rate does not
change. A mean value of 0.0185 0.001 mg O2/mgd.w./min (Table 1) was obtained. Also,
if the ferric/ferrous ratio of 1/10 is held constant and the biomass concentration in the
reaction mixture is varied from 4.3 to 47.7 mgd.w./L, the specific oxygen uptake rate is also
constant at a level of 0.020 0.003 mg O2/mgd.w./min (Table 2). At the significance level
of 0.05, as evaluated by the t-test (two populations), the two means are not significantly
different. As both the biomass concentration and the ferric/ferrous iron ratio do not have a
significant effect, the mean value of the specific oxygen uptake rate for all the
experimental conditions applied was calculated: 0,020 0.002 mg O2/mgd.w./min.
Table 1. Effect of the ferric/ferrous iron ratio on the specific oxygen uptake rate
(biomass concentration: 8.0 mgd.w./dm3)
Fe(III)/Fe(II)
1/10
1/7
1/5
1/2
1/0.3
1/0.1
Mean value
Standard deviation
qO2, mg O2/mgd.w./min
0.019
0.019
0.018
0.018
0.019
0.018
0.0185
0.001
Table 2. Effect of the biomass concentration on the specific oxygen uptake rate
(ferric/ferrous iron ratio: 1/10)
X, mgd.w./dm3
4.3
8.0
13.6
17.9
47.6
Mean value
Standard deviation
qO2, mg O2/mgd.w./min
0.017
0.019
0.025
0.021
0.022
0.020
0.003
The specific oxygen uptake rate found for the concentrated cells was practically the
same as that determined in our studies on ferrous iron biooxidation in a bubble column
without and with pH regulation at 1.8 and 2.0 and in a stirred tank with pH regulation at
2.0, as may be seen in Table 3. The mean value of the specific oxygen uptake rate was
calculated from all our data: 0.020 0.001 mg O2/mgd.w./min. This value agrees quite well
1079
with those measured for A. ferrooxidans grown in a batch reactor using the same type of
oxygen probe [11] or the manometric technique [12] while other published values are
twice [13, 14] or even tenfold [2] higher. The disagreement is probably due to different
strains, measuring methods or cultural conditions applied in the different studies.
Table 3. Specific oxygen uptake rate of A. ferrooxidans under different culture
conditions*
Culture condition
Concentrated cells, pH 1.4-2.4
Bubble column, 9 gFe/L, 28C without pH
regulation (pH 2.4-2.7)
Bubble column, 9 gFe/L, 28C
with pH regulation:
1.8
2.0
Stirred tank, 9 gFe/L, 28C
with pH regulation: 2.0
Fermentor, 9 gFe/L, 28C, pH 3.2
Batch reactor, 10 gFe/dm3, 30C, pH 1.6
Sample from uranium b
leaching + 0.20 gFe, 30C
Not defined a
qO2, mgO2/mgd.w./min
0.020 0.002
0.021 0.001
0.018 0.001
0.020 0.001
0.021 0.002
0.205
(230 mmol O2/mgpr./h)c
0.020
(630 nmol O2/mgd.w./min)
0.025
(700 mL O2/mgpr./h)c
Reference
This work
Unpublished data
[5]
[5]
[2]
[11]
[12]
0.045
[13]
(20840 mL O2/mgN/h)c
0.042
[14]
Shaken flasks, 0.56 gFe/dm3, 30C, pH 1.6
(785 nmol O2/mgpr./min)c
*
In all cases, a Clarke oxygen electrode was used with the following exceptions: (a) not defined
and (b) the manometric method
c
Assuming that the dry biomass of A. ferrooxidans contains about 60% of proteins [16, 17] and
12.4% of nitrogen [15].
The oxygen uptake rate of A. ferrooxidans, evaluated in experiments with the

concentrated cell suspensions and during ferrous iron biooxidation is shown in Figure 1 as
a function of the biomass concentration. All data fit a straight line, indicating that the
oxygen uptake rate of the bacterial population, OUR (in mg O2/L/min), during ferrous
iron biooxidation can be correlated to the biomass concentration, X (mgd.w./L), by the
well-known equation:
OUR = qO2 X
where qO2 is the specific oxygen uptake rate of the strain A. ferrooxidans B5 (0.020 mg
O2/mgd.w./min). This correlation allows the determination of the viable biomass
concentration of A. ferrooxidans from the oxygen uptake rate measured during the
biooxidation process. The only limitation for applying the above equation is the ferrous
iron concentration in the culture medium where the oxygen uptake rate is measured. From
the stoichiometry the ferrous iron oxidation and cell synthesis by A. ferrooxidans [15], and
taking the maximum oxygen rate and the usual duration time of the oxygen uptake rate
measurement as 0.70 mg O2/L/min [7] and 20 minutes, respectively, the critical ferrous
iron concentration in the medium, which limits the bacterial growth, was estimated to be
about 0.1 g/L. This value was verified by measuring the oxygen uptake rates in samples
1080
taken from the stationary growth phase, where the ferrous iron concentration was below
the critical value, and in the same samples after the addition of a ferrous iron solution to
adjust the ferrous iron concentration above the critical level. A summary of the oxygen
uptake rate of samples taken from the stationary growth phase with and without the
addition of ferrous iron is presented in Table 4. The oxygen uptake rate in the latter is
lower than that in the former, indicating the limitation of oxygen uptake by ferrous iron.
Figure 1. Oxygen uptake rate (OUR) as a function of dry biomass concentration

(experiments with concentrated biomass suspension: ; the biooxidation of ferrous iron in a
bubble column without pH regulation - U [unpublished data] and with pH regulation at 1.8 and 2.0 - ; and the biooxidation of ferrous iron in a stirred tank with pH regulation at
2.0: [7])
Table 4. The oxygen uptake rate of A. ferrooxidans in samples from the stationary
growth phase in a bubble column* before and after the addition of ferrous iron (up to
9 g/L)
Oxygen uptake rate, mg O2/L./min
Before the addition of
After the addition of
ferrous iron
ferrous iron
2.4-2.7
0.08
0.08
0.41
1.8
0.08
0.07
0.57
2.0
0.03
0.08
0.50
0.05
0.06
0.56
0.05
0.055
0.48
*
Data from the experiments described in [7], but not shown there.
pH
Ferrous iron
concentration in the
culture medium, g/L
1081
4.
CONCLUSIONS
Quantification of the A. ferrooxidans biomass concentration using usual methods,
such as serial end-point dilution and enumeration of colonies, is uncertain and timeconsuming. The measurement of dry biomass, although simpler and reproducible, can
only be used for estimating the biomass concentration in large bioreactors since a volume
of 500-750 mL [7] is needed as a sample.
This work describes a simple method for quantifying the A. ferrooxidans biomass
concentration indirectly by measuring the oxygen uptake rate by a cell population. The
advantages of this method are that the test is rapid, small samples are needed, and the
results are reproducible. The method could also be appropriate for enumerating aerobic
microorganisms in bioleaching experiments, since many cells are attached to the ore and
could not be detected by standard methods. The following assumptions were made when
the indirect method was used. Firstly, all the oxygen consumed in the bioprocess was only
used for ferrous iron oxidation, the oxygen, which became part of the biomass was derived
from combined oxygen found in CO2, and the oxygen required for maintenance was
negligible. Secondly, autooxidation of the ferrous iron did not occur, which was verified
expermentally.
REFERENCES
1. M.L. Shuler and F. Kargi, Bioprocess Engineering, Prentice Hall, Englewood Cliff,
New Jersey, 1992.
2. C. Bodo and D.G. Lundgren, Can. J. Microbiol., 20 (1974) 1647
3. H.M. Lizama and I. Suzuki, Appl. Environ. Microbiol., 55 (1989) 2588
4. I. Suzuki, H.M. Lizama, P.D. Tackaberry, Appl. Environ. Microbiol., 55 (1989) 1117
5. D. Savi, V. Veljkovi, M. Lazi and M. Vrvi, in: R. Amils and A. Ballester (eds.),
Biohidrometallurgy and the Environment toward the Mining of the 21st Century,
Elsevier, Amsterdam, 1999, Part A, pp 625-630.
6. G.I. Karavaiko, G. Rossi, A.D. Agate, S.N. Groudev and Z.A. Avakyan, (eds.)
Biogeotechnology of Metals, CIP GKNT, Moscow, 1988
7. V. Veljkovi, D. Savi, M. Lazi and M. Vrvi, in: R. Amilis and A. Ballester (eds.),
Biohidrometallurgy and the Environment toward the Mining of the 21st Century,
Elsevier, Amsterdam, 1999, Part A, pp 617-623.
8. D.W. Zabriskie and A.E. Humphrey, AIChE J., 24 (1978) 138
9. M. Reu, R.P. Jefferis and J. Lehman, GBR Monogr., 3 (1976) 107
10. A. Schumpe, I. Adler and W. Deckwer, Biotechnol. Bioeng., 20 (1978) 145
11. J.T. Pronk, W.M. Meijer, W. Hazeu, J.P. vanDungen, P. Bos and G. Kuenen, Appl.
Environ. Microbiol., 57 (1991), 2057
12. C.L. Brierley, in: L.E. Murr, A.E. Torma and J.A. Brierley (eds.) Metallurgical
Applications of Bacterial Leaching and Related Microbiological Phenomena,
Academic, New York, 1978, pp. 345-364
13. S.N. Groudev, F.N. Genchev and S.S. Gaidarjev, in: L.E. Murr, A.E. Torma and J.A.
Brierley (eds.) Metallurgical Applications of Bacterial Leaching and Related
Microbiological Phenomena, Academic, New York, 1978, pp. 253-272.
14. D.P. Kelly and C.A. Jones, in: L.E. Murr, A.E. Torma and J.A. Brierley (eds.)
Metallurgical Applications of Bacterial Leaching and Related Microbiological
Phenomena, Academic, New York, 1978, pp. 19-44.
15. J.R. Smith, R.G. Luthy and A.C. Middleton, J. WPCF, 60 (1988) 518
1082
16. C.A. Jones and D.P. Kelly, J. Chem. Tech. Biotechnol., 33B (1983) 241
17. A.H. Basaran and O.H. Tuovinen, Coal Prep., 5 (1987) 39
1083

Electrochemical and microbiological characterization of

mercury in contact with mud
F. Cruza, A. Welzelb, C. Sampaioc, G.E. Englertc, I.L. Mllerc
a
Material Department - UFRGS - Brazil

b
PPGEM - UFRGS - Brazil
Metallurgical Department - UFRGS - Brazil
Abstract
HgCl2 poured onto liquid nutritive broth was used as culture media for aerobic
bacteria, Bacillus sp. This media was used to evaluate the bacterial activity resistance to
toxic metal and to activate the bacteria protective mechanism. Two interconnected flasks
were projected to evaluate: i) growth of the bacteria and ii) the electrochemical behaviour
of gaseous mercury trapped in acid solution.
Bacteria were cultivated and maintained in plate count agar (PCA), a solid nutritive
broth. Electrochemical experiments were done in a liquid culture medium using a
Potentiostat PAR EG&G 273A. Potentiodynamic curves were obtained over a platinum
surface during 10 days to follow the oxygen depletion by the bacteria. Electrochemical
experiments were conducted in nutritive broth with and without mercuric chlorides and
bacteria. At the end of the experiments bacteria remained viable despite the contact with
toxic metal while the broth pH value changed of 7 through 10.
1.
INTRODUCTION
It is well known that mercury is a toxic metal and any contaminated with it cannot be
buried due to its hazardous effect. This kind of residues is catalogued as belonging to the
Class I. Gaseous monatomic mercury Hg may be produced by several natural sources and
then it is delivered to the atmosphere. Almost 95% of atmospheric mercury is found in the
Hg state. However the fraction value that is free as a monatomic gas or adsorbed onto the
solid particles, which are suspended in the air, may change considerably (1). On the other
hand, in aqueous media, Hg can be oxidized to HgII as a result of the ozone activity (1). In
this case bacteria with an active methylating mechanism seem to regulate environmental
conditions that control the bioavailability of HgII. It was proposed in the literature that the
accumulation rate of mercury in water depends on pH, selenium concentration, presence
of sulfate, dissolved organic carbon in seepage lakes and decomposable organic matter
(2). One of the ways of determining the rate of oxidation of mercury is by measuring the
concentration of HgII. Some researchers have studied the detection of mercury through
bio-indicators but they are not specific to a given form of HgII. It is also possible to detect
the presence of mercury by electrochemical methods using platinum electrodes (3).
Universidade Federal do Rio Grande do Sul - Av. Osvaldo Aranha 99/615-D.90035-090 Porto Alegre,
RS_Brasil. genglert@vortex.ufrgs.br Phone numbers 55 51 3316 3344/3212 4547
1085
Many researchers have been looking for a technical solution to eliminate mercury
from waters using microorganisms (1, 4, 5). Since microorganisms live in environments
that can be reproduced artificially, it is possible to isolate them easily. Hence, it is possible
to determine the effects that environmental factors play on the availability of inorganic
mercury in the presence of bacteria in aquatic environments. Such knowledge is very
important from an ecological point of view (3).
2.
EXPERIMENTAL PROCEDURE
Bacterial strain. A portion of mud was collected from an industrial area in order to
isolate one microorganism adapted to the environmental conditions. They were placed in
contact with a liquid broth appropriated for aerobic bacteria. The liquid broth consisted of
a solution of peptone, yeast extract and sodium chloride in water at pH 7. During one
week the culture was kept at 30C to produce a quick growth. After this stage, a liquid
portion was poured in a dish plate in order to start to the bacterial isolation. The culture
was then maintained at the same temperature indicated before. Bacterial identification was
carried out by morphological aspects, temperature resistance (after the broth liquid with
bacteria was heated up to 80C for 15 minutes) and biochemical tests (Table 1). The
temperature resistance indicates conditions for spore formation and the biochemical tests
pointed out to Bacillus sp.
Table 1. Biochemical tests to identify bacteria
Motility
Urease
Citrate
Gelatin
Nitrate reduction
Glucose
Catalase
Oxidase
+
+
+
+
+
+
+
An aqueous solution with bacteria characterized previously was taken after 48 hours
of incubation and poured into a fresh portion of modified liquid broth. The modified broth
consisted of 1 mL of bacteria solution plus 99 mL of sterilized water. (In order to improve
the bacterial growing, oxygen was continuously poured onto the flask). The resistance of
the microorganisms to mercury was evaluated adding 1mL of 0.5M HgCl2 and HgCl2
which is a toxic solution to 0.2 L of broth liquid. The later solution pH was reduced from
7 to 4 by adding some drops of 0.1M nitric acid. The viability of the microorganisms was
evaluated by counting them at Neusbauer camera. Some other tests were also carried out
with the same solution mentioned before at pH 7, with and without mercury.
Electrochemical parameters. Anodic and cathodic potentiodynamic curves, in broth
culture with and without mercury, in the presence and in the absence of bacteria, were
obtained using a potentiostat PAR G&G 273A.The saturated calomel electrode (SCE) was
used as a reference and a graphite was an auxiliary electrode. The working electrode
(where electrochemical reactions happened) consisted of a mini-electrode of platinum with
an area of 0.00035 cm2. All the experiments started at -1.0V sce and went until +1.0V sce
at a scan rate of 2 mV/s. All experiments were done at sterilized and quiescent conditions.
1086
3.

Biochemical tests to characterize the bacteria indicate Bacillus sp. as the isolated
bacteria. They are considered resistant to mercury according to literature (1,4,5). In this
work it was observed that those species cannot grow as well in modified media as they do
in the original media, although both were contaminated by Hg. In the former media the
organic concentration was low, since the solution was diluted so the bacteria were in
closer contact with mercury. The morphological aspects after experimentation in modified
media with Hg are shown in Figure 1, in which the stressed structure is clearly observed.
Bacteria that survived in a complete media without Hg were not stressed as can be clearly
seen in Figure 2. It was observed that bacteria species in contact with modified broth with
Hg survived despite the oxygen concentration, which was reduced progressively after
electrochemical experiments (Figure 3).
Figure 1. Bacteria after 30 days in modified media without Hg
Figure 2. Bacteria after 30 days in modified media with Hg
1087
1,5
st
th
1 day
after 5 day
th
1,0
th
after 10 day
10 day
E (VSCE)
0,5
0,0
-0,5
-1,0
1E-7 1E-6 1E-5 1E-4 1E-3 0,01
0,1
10
100
i (A/cm )
Figure 3. Electrochemical curves in modified media with bacteria and without

mercury
It was observed that, even reducing the oxygen concentration that was measured by
the reduction of the current density, the bacteria were kept alive. The turbidity of the
liquid (which remained for more than 10 days without any oxygen reposition) was an
evidence that bacteria were still alive and was confirmed when bacteria grew in PCA.
From electrochemical point of view it is possible to observe in Figure 3 that at the 5th day
of the experimentation, there was a dislocation of the anodic curve to the right hand when,
apparently, all oxygen had been removed by bacteria. The shape of the polarization curve
is an indicative that the hydrogen cathodic reaction was the cathodic reactive, especially
because pH was low. It is well known that the main cathodic reactions from the
electrochemical point of view are: the reduction of oxygen dissolved in water and the
reduction of the hydrogen from the water or from the acid media. The reintroduction of
oxygen to the system brought the curve to the original behaviour with lower current
density values. In order to compare the experiments with and without bacteria (data not
shown), the same experiment was done with the liquid broth media alone and it was
possible to observe that the cathodic current values did not suffer any change. This is a
strong indicative that, despite oxygen is by far continuously removed by its
electrochemical reduction, remotion is not enough to dislocate the curves as can be seen in
Figure 3. Finally, a tiny anodic peak was observed and it may be an indicative that some
change happened when comparing the experiments at the first day and the 15th day. As
microorganisms grew, the production of their by-products is high enough to interact with
the platinum surface. It is known that factors such as low concentration of organic carbon
dissolved, the presence of sulphate, low alkalinity and low pH, may interfere in the
mechanism by which bacteria retain mercury (2). However, a high concentration of total
organic carbon may interfere in the mechanism that bacteria use to reduce mercury. From
the anodic curve (data not shown) obtained in the broth liquid with mud, it was not
possible to confirm the presence of Hg probably due to the interference to the high organic
matter concentration. It is important to highlight that previous analysis through atomic
absorption confirmed the presence of Hg in the mud in a concentration of a few ppm. The
1088
oxygen uptake is partially responsible to the high pH value observed at the end of all
experiments since the initial value was 7 and the final one was 10. The participation of the
by-products may also contribute to the high pH value observed.
4.
CONCLUSION
Tested bacteria Bacillus sp. can survive without oxygen. They create self-mechanisms
to produce a new by-product noticed by the changes in cathodic potentiodynamic curves
and also in pH values. This observation was confirmed by pH 10 registered at the end of
experimentation. The electrochemical tools may be jointly used with the microbiological
ones to characterize the growth of bacteria during mercury remotion from contaminated
liquids.
ACKNOWLEDGEMENTS
To FAPERGS and CNPq for the financial support in the form of scholarship as well
to Prof. Washington Aliaga by his help to structure this work.
REFERENCES
1. T. Smith, K. Pitts, J.A. McGarvey, A.O. Summers. Applied and environmental

microbiology, Apr. 1998, vol. 64, No 4, p. 1328-1332.
2. T. Barkay, M. Gillman, R.R. Turner. Applied and environmental microbiology, Nov.
1997, vol. 63, No 11, p. 4267-4271.
3. Ionashiro, E.Y., Souza, G.R., Milar, E., Bebedetti, A.V., Ionashiro, M. and Fertonani,
F.L. XII Simpsio Brasileiro de Eletroqumica e eltroanaltica, April, 2001, CDrom,
www.ufrgs.br/sibee.
4. A.O. Ogunseitan. Applied and environmental microbiology, Fev. 1998, vol. 64, No 2,
p. 695-702.
5. M.D. Mullen, D.C. Wolf, F.G. Ferris. Applied and environmental microbiology, Dec.
1989, vol. 55, No 12, p. 3143-3149.
1089

Evaluating the growth of free and attached cells during the

bioleaching of chalcopyrite with Sulfolobus metallicus
B. Escobar, M.J. Hevia and T. Vargas
Centro de Estudios en Hidro/Electrometalurgia, Departamento de Ingeniera de
Minas/Departamento de Ingeniera Qumica, Universidad de Chile,
Tupper 2069, Santiago, Chile. Fax 56-2-6991084
Abstract
Chalcopyrite can be dissolved at covenient rates in the presence of thermophilic
microorganisms such as Sulfolobus metallicus, a chemolitothrophic archaea that can grow
autotrophically at temperatures between 65 and 80C. The catalytic influence of this
microorganism is partially related to the microbial oxidation of ferrous iron and the
oxidation of residual sulfur produced during chalcopyrite dissolution. Chalcopyrite
bioleaching studies, however, normally report only the growth of the planktonic
population of Sulfolobus metallicus which makes difficult to assess the pattern of
utilization of these two subtrates by the microorganism during the process.
The present work reports experimental studies on the bioleaching of a chalcopyrite
concentrate with Sulfolobus metallicus in shake flasks at 70C and pH 1.8. The growth of
the planktonic population was determined by cell counting, MPN and protein
determination by Lowry. The population of microorganisms attached to the solids was
evaluated by protein determination using a modified Lowry method.
Results showed that about 68% of the Sulfolobus metallicus present in the flask grows
attached to chalcopyrite particles, while the remaining 32% grows in the solution. The
percentage associated to attached growth coincides with the fraction of microorganisms
which initially attaches to chalcopyrite after addition of the inoculum. Cell growth during
the first 72 hours of both attached and free microorganisms is mainly based on the
oxidation of ferrous iron. Bacterial oxidation of sulfur, which is conducted only by
attached microorganisms, becomes relevant only after 72 hours of leaching. After 120
hours of leaching there is a net reduction of both the planktonic and attached population
which evidences an important process of cell death and breakage.
Keywords: S. metallicus, bioleaching, chalcopyrite, biomass estimation, attached cells

1.
INTRODUCTION
Primary copper sulfides such as chalcopyrite has been demonstrated to be refractory
to bacterial leaching with mesophiles such as At. ferrooxidans in ambient conditions (1).
During the last years, however, several researchers have demonstrated that chalcopyrite
can be dissolved at covenient rates in the presence of thermophilic microorganisms such
1091
as Sulfolobus metallicus, Acidianus brierley or Metallosphera sedula (2, 3). The use of
Sulfolobus metallicus, a chemolitothrophic archaea that can grow autotrophically at
temperatures between 65 and 80C, has been largely investigated in the bioleaching of
chalcopyrite (4-6). The catalytic influence of S. metallicus in this process is partially
related to the microbial oxidation of ferrous iron, an activity which has been evidenced by
several authors (1, 7). Other works have also indicated that S. metallicus presence have
also a catalytic influence on the oxidation of residual elemental sulfur to sulfate (7,8). In
chalcopyrite bioleaching studies only the growth of the planktonic S. metallicus is
normally monitored, which makes difficult to assess the relative importance of these two
substrates.
The present work reports experimental studies on the bioleaching of a chalcopyrite
concentrate with Sulfolobus metallicus in shake flasks at 70C and pH 1.8. An important
methodological aspect, the growth of both the population of microorganisms attached to
the sulfide particles and microorganisms free in solution were simultaneously monitored
during these experiments. The relative importance of sulfur and ferrous iron as substrates
for microbial growth was evaluated from relating bacterial growth to the evolution of
dissolved metals and pH during the experiment.
2.
2.1 Microorganism
A strain of Sulfolobus metallicus was used in this study. The cells were grown in
basal medium containing 1% chalcopyrite concentrate as the energy source. The
composition of the basal medium was 0.4 (NH4)2SO4, 0.5 MgSO4.7H2O and 0.2 KH2PO4
g/l, adjusted to pH 1.6 with sulfuric acid.
2.2 Mineral
A -325 +400 mesh (-0.045mm +0.038mm) sample of a high purity chalcopyrite
sample prepared from an Andina Copper concentrate was used in the experiments.
2.3 Experimental procedure
Attachment experiments were conducted in 125 ml Erlenmeyer flasks containing 50
ml basal medium. One g of chalcopyrite was contacted with an initial concentration of
1.1x108 cells of S. metallicus/ml. The flasks were incubated in an environmental shaker at
70C and 100 rpm. The rate of cell attachment was determined by periodic sampling and
determination of free cells by direct counting onto the microscope.
Leaching experiments were conducted in 250 ml Erlenmeyer flasks containing 95 ml
of the basal medium, 1% (w/w) chalcopyrite and 5 ml of active inoculum of Sulfolobus
metallicus. The flasks, 8 in total, were incubated in an environmental shaker at 100 rpm
and 70C. Two flasks were removed at different leaching times to determine total iron,
ferrous iron, copper and pH in solution, and bacterial growth in solution and solids. Total
and ferrous iron in the solution were determined by o-phenanthroline colorimetric method
(9); for solutions with high ferric:ferrous ratio ferrous iron was determined by the
modified o-phenanthroline colorimetric method (10). The number of free cells in solution
was determined by direct counting onto the microscope using a Petroff-Hausser chamber
and by the MPN of iron oxidizing microorganisms, in which cells were grown at 70C in
basal medium containing 3.6 mM ferrous sulfate during 10 days. Also free cells were
estimated by protein determination using the Lowry methodology (11). Attached
1092
microorganisms were also estimated by protein assay using the modified Lowry method
proposed by Karan et al., (12). In order to correlate the amount of proteins determined on
the solids to number of cells, a calibration curve obtained from measurements of
planktonic cells with both methods, direct counting and modified Lowry was used (see
Figure 1).
3.
3.1 Adherence of S. metallicus to chalcopyrite

Figure 2 shows the evolution of the population of free bateria in solution during
attachment of Sulfolobus metallicus to chalcopyrite. The attachment of S. metallicus onto
chalcopyrite concentrate was found to reach a stationary value within the first 100 minutes
of contacting the cells with mineral, when 68% of the initial cells were attached to the
mineral.
8.0E +09
Total Cell/50 ml
Cell/ml
6 .0 E + 0 8
4 .0 E + 0 8
2 .0 E + 0 8
0 .0 E + 0 0
6.0E +09
4.0E +09
2.0E +09
0.0E +00
10
20
30
P r o te in ( m g /m l)
Figure 1. Correlation between cell

number of S. metallicus and
measured protein
50
100
150
T im e (m in)
Figure 2. Attachment of S.
metallicus onto 1 g chalcopyrite
concentrate
3.2 Bioleaching of chalcopyrite

Figure 3 shows the rate of copper and iron dissolution, determined during the
bioleaching of chalcopyrite with Sulfolobus metallicus. The total amount of dissolved iron,
8.86 mg, was much smaller than the total amount of dissolved copper, 134 mg.
Ferrous/ferric determinations indicated that iron released during chalcopyrite dissolution
was quantitatively oxidized to ferric iron from the start of the experiment (results not
included), which indicated an effective microbial catalytic activity. Very likely, most of
the iron released during chalcopyrite dissolution was eventually precipitated, which
showed in the presence of large amounts of jarosite-like compounds either in the
bioleaching solution or adhered to the flasks wall. The amount of ferric iron precipitated at
each time corresponds to the difference between the effective and measured iron
concentration, the theoretical iron dissolution expected according to the stoichiometry:
(1)
CuFeS2 + 4Fe+3 Cu+2 + 5Fe+2 + 2S
The steady increase of this difference indicates that ferric iron precipitation occurred
continuously since the start of the experiment.
The evolution of the number of cells and biomass of planktonik cells of S. metallicus
during the experiment is shown in Figure 4. Figure 5 shows the evolution of
1093
microorganisms attached to the solids presents in the flasks, solids which includes residual
chalcopyrite together with the ferric iron precipitates which were simultaneously separated
from the solution by filtration. These results show that planktonic cells gives only account
of 30% of cell growth occurring during the process and, consequently, about 70% of
bacterial growth occurs in the cells attached to the solids. The relative importance of these
two populations is similar to the distribution of cells obtained during the initial attachment
stage, shown in Figure 2.
6.E+08
6.E+07
4.E+08
4.E+07
2.E+08
2.E+07
0.E+00
0.E+00
1400
1000
800
600
Protein/ml
1200
Cell/ml
Cu and Total Fe (mg/l)
1600
400
200
0
0
100
200
300
T im e (h o u rs )
200
300
Tim e (hours)
Figure 3. Copper , total iron

measured and effective total iron
disolution
in
bioleaching
of
chalcopyrite concentrate
Figure 4. Cells of S. metallicus in

bioleaching solution determined by
direct count , MPN and protein
concentration
160
4.E+10
140
120
3.E+10
Sulfur (mg)
Total Cell and Adhered

Cells
100
2.E+10
1.E+10
100
80
60
40
20
0.E+00
0
100
200
300
Time (hours)
Figure 5. Total and adhered cells
of S. metallicus in the bioleaching
chalcopyrite concentrate
0
0
100
200
300
Time (hours)
Figure 6. Sulfur generated

(curve a) and oxidized (curve b)
estimated from copper and acidity
of solution
The cell population of planktonic and attached cells initially increases reaching a
maximum at 72 and 120 hours, respectively. After this initial stage there is a rapid decay
in the population of both free and attached microorganisms.
In the case of planktonic cells, the decrease in cellular activity, determined by MPN,
is paralelled by a decrease of the number of cells, determined by counting. This indicates
that the decrease in cellular activity is not due to some type of inhibition, but rather to a
process of cell death and/or cell breakage. This phenomenon can be linked to the
sensitivity which these microorganisms show to the presence of solids, as it is widely
known (3,13). Elimination of cells by incorporation into ferric iron precipitates can be
1094
ruled out as a main cause of cell number reduction, as this would show in a simultaneous
increase in the biomass of attached bacteria, which was not observed.
It is interesting to outline that the decrease in cell number and activity of planktonic
cells is paralelled by a decrease in protein concentration. This indicates that the constituent
proteins of thermophilic microorganisms, once the cell breaks, are not stable at the
temperature and acidity at which bioleaching solutions operate. This fact was confirmed in
a complementary experiment in which Sulfolobus metallicus digested cell proteins were
maintained in a shaker in an iron-free basal solution at pH 1.8 and at 70C. The protein
concentration in the solution after 72 hours in these conditions, determined by Lowry,
decreased 28% with respect to the initial value.
The decay in protein concentration observed on solids indicates that in the attached
cell population also a stage is reached in which the rate of cell death and/or breakage
predominates over the rate of net cell growth.
3.3 Bacterial utilization of ferrous iron and sulfur
Ferric iron was not initially added in the reported experiments, hence the scarce ferric
iron present in the solutions resulted from the microbial oxidation of ferrous iron released
during chalcopyrite dissolution. Bioleaching solutions presented a high Eh (data not
shown), indicating a high ferrous iron oxidation activity of S. metallicus. The efficiency of
this oxidative activity was also evidenced from the fact that MNP cell number
determinations, which were based on the reponse of microorganisms to ferrous iron
oxidation, gave similar values to those determined by direct cell counting. Attached cells,
however, are exposed to two substrates, ferrous iron and surface elemental sulfur formed
after reaction (1). It is interesting to assess the pattern of utilization of these two substrates
during bacterial growth and see how this relates to the chalcopyrite bioleaching process.
The pH of the bioleaching solution is the result of dinamic balance linked to a series
of chemical reactions which consume or produce protons. Protons are consumed during
the oxidation of ferrous iron which occurrs according to:
(2)
Fe+2 + 1/2O2 + 2H+ Fe+3 + H2O
On the other hand, protons are produced through: a) hydrolysis of the Fe(III) with
jarosite formation according to the reaction:
(3)
M+ + 3Fe+3 + 6H2O + 2(SO4)-2 MFe3(SO4)2(OH)6 + 6H+
+
+
+
+
Where M is either K , NH4 or H3O ; and through b) oxidation of sulfur to sulfate
according to the reaction:
(4)
2S + 3O2 + 2H2O 2H2SO4
In order to assess the amount of elemental sulfur oxidized by Sulfolobus metallicus
to sulfate, the acidity resulting from this reaction was evaluated at different times during
chalcopyrite bioleaching with the expression:
4 = Ct - Ci - 3 + 2
(5)
where 4 is the acid generated by reaction 4, Ct is the acid concentration determined in the
solution from pH measurement, Ci is the initial acid concentration, 3 is the acid produced
in reaction 3, and 2 the acid consumed in reaction 2. The amount of ferrous released at
each time was calculated from the amount of copper dissolved from chalcopyrite,
assuming that this follows the stoichiometry of reaction (1). Ferrous iron was assumed to
be quantitatively converted into ferric according to reaction (2). The amount of ferric iron
precipitated was calculated as the difference between the theoretical amount of ferrous
1095
iron dissolved and the iron experimentally determined in solution (Figure 3). The acid
generation associated to ferric iron precipitation was calculated assuming formation of
potassium jarosites which, according to XRD analysis, is the compound we have found
that predominate in concentrates bioleached with thermophiles.
Results of this calculation are shown in Figure 6. Curve a represents the amount of
elemental sulfur generated during chalcopyrite dissolution according to reaction (1), and
was calculated from the amount of copper dissolved at different leaching times. Curve b
represents the amount of elemental sulfur oxidized to sulfate calculated from reaction (4)
using the 4 values as generated acid. Comparison of data in Figure 6 with the evolution
of cell growth in Figures 4 and 5 indicate that the fast cell growth observed during the
initial 72 hours in both the planktonic and attached microorganisms, is mainly based on
the oxidation of ferrous iron. The growth is faster on the solids as the attached initial
population is larger. In addition, cells attached to chalcopyrite surface should have access
to larger concentrations of ferrous ions which is released in that vicinity.
The oxidation of elemental sulfur becomes significant only after the first 72 hours
when already 50 mg of this compound have been accumulated on the chalcopyrite surface.
At this stage the growth of the attached population continues fast for another 50 hours
while the planktonic population started to decrease. It is very likely that the continued
growth trend observed in the attached microorganisms is partially linked to triggering of
sulfur oxidation, as these microorganisms are the only ones with can have direct access to
that alternative solid substrate. On the other hand, the decrease in the planktonic
population after the first 72 hours can be due to depletion of ferrous iron in the solution
bulk. This can be related to the large population of attached microorganisms reached at
this stage, which can consume most of the ferrous iron released from the sulfide.
Studies done by several authors about the bioleaching of pyrite and chalcopyrite with
thermophilic microorganisms show differences on the cells adsorption to the mineral; in
some cases, in spite of the fact that the cellular attachment was not detected by
microscope, they showed that pyrite leaching required the direct contact of Acidianus
brierley with the mineral (14). Nemati and Harrison (15), working with S. metallicus and
pyrite, showed a low cell adsorption (5-10%) and considered the concentration of the cells
in the liquid phase as indicative of total microbial growth in the reactor. On the other
hand, Konishi et al., (16) showed a high attachment of cells to chalcopyrite and found that
the chalcopyrite leaching with A. brierley at 65C takes place with a direct attack by
adsorbed cells on the mineral surface, the ferric leaching being insignificant. The results
presented in this work show that most of the population of S. metallicus present in a
bioreactor remains attached to the solids, where they grow out of ferrous iron and sulfur
oxidation. Any attempt to assess in depth the performance of a bioleaching reactor should
include an adequate monitoring of this population.
4.
CONCLUSIONS
Experimental results demonstrated that the growth of attached Sulfolobus metallicus
can be efficiently monitored using the modified Lowry method of Karan et al.,(16).
During bioleaching of chalcopyrite about 68% of the Sulfolobus metallicus present in
the flask grows attached to chalcopyrite particles, while the remaining 32% grows in the
solution. The percentage associated to attached growth coincides with the fraction of
microorganisms which initially attaches to chalcopyrite after addition of the inoculum.
Cell growth during the first 72 hours of both attached and free microorganisms is
mainly based on the oxidation of ferrous iron. Bacterial oxidation of sulfur, which is
1096
conducted only by attached microorganisms, becomes relevant only after 72 hours of

leaching.
After 120 hours of leaching there is a a net reduction of both the planktonic and
attached population which evidences an important process of cell death and breakage.
ACKNOWLEDGEMENTS
This work was funded by BHP Billiton and Conicyt Chile under Fondef Project D 00
I 1050.
REFERENCES
1. P.R. Norris, Biomining: Theory, Microbes and Industrial Processes, D.E. Rawlings
(ed), Landes, Bioscience, Austin, TX USA. (1997) 281.
2. P.R. Norris, L. Parrot and R.M. Marsh, Workshop on Biotechnology for the Mining,
Metal-refining and Fossil Fuel Processing Industries, Biotechnol. Bioeng., Symposium
N16, H.L. Ehrlich and D.S Holmes (eds), Wiley, New York (1986) 253.
3. D.A. Clark and P. Norris, Microbiology 142 (1996) 785.
4. P. dHugues, D. Morin and S. Foucher, Biohydrometallurgy: Fundamentals,
Technology and Sustainable Development, Part A, V.S.T. Ciminelli and O. Garcia
(eds), Elsevier, Amsterdam (2001) 75.
5. M.L. Blzquez, A. Alvarez, A. Ballester, F. Gonzlez and J.A. Muoz,
Biohydrometallurgy and the Environment toward the mining of the 21st Century, A.
Ballester and R. Amils (eds.) Elsevier, Amsterdam (1999) 137.
6. Y. Rodriguez, A. Ballester, M. L. Blzquez, F. Gonzlez and J.A. Muoz,
Biohydrometallurgy: Fundamentals, Technology and Sustainable Development, Part
A, V.S.T. Ciminelli and O. Garcia (eds), Elsevier, Amsterdam (2001) 125.
7. R. M. Marsh, P.R. Norris and N. W. Le Roux, Progress in Biohydrometallurgy, G.
Rossi and A.E. Torma (eds.) Cagliari, Italy (1983) 71.
8. D.W. Shivvers and T. D. Brock, J. Bacteriol., 114 (1973) 706.
9. M. K. Muir and T. Anderson, Metall. Trans. 8B (1977) 517.
10. L. Herrera, P. Ruiz, J.C. Aguilln and A. Fehrmann, J. Chem. Tech. Biotechnol 44
(1989) 171.
11. O.H. Lowry, N.J. Rosenborough, A.L. Farr and R.J. Randall, J. Biol. Chem., 193
(1951) 265.
12. G. Karan, K.A. Natarajan and J.M. Modak, Hydrometallurgy 42 (1996) 169.
13. B. Escobar, J.M. Casas, J. Mamani and R. Badilla-Ohlbaum Biohydrometallurgical
Technologies, A. E. Torma, J.E. Wey and V.I. Lakshmanan (eds.) The Minerals,
Metals & Materials Society (1993) 195.
14. L. Larsson, G. Olsson, O. Holst and H. T. Karlsson, Biotechnol. Lett. 15 (1993) 1, 99.
15. M. Nemati and S.T.L. Harrison, Biohydrometallurgy and the Environment toward the
mining of the 21st Century, Part A, A. Ballester and R. Amils (eds) Elsevier,
Amsterdam (1999) 473.
16. Y. Konishi, M. Tokushige and S. Asai, Biohydrometallurgy and the Environment
toward the mining of the 21st Century, Part A, A. Ballester and R. Amils (eds)
Elsevier, Amsterdam (1999) 367.
1097

Experimental and modeling studies on inhibition effect of

solution conditions on activity of Acidithiobacillus ferrooxidans
during biooxidation of mixed sulphidic concentrates
M.N. Chandraprabhaa, Jayant M. Modakb and K.A. Natarajana
a
Department of Metallurgy, Indian Institute of Science, India

Department of Chemical Engineering, Indian Institute of Science, India
Abstract
In this paper, we report experimental and modeling results aimed at characterizing the
effect of solution conditions on the growth and ferrous iron oxidation ability of
Acidithiobacillus ferrooxidans. The conditions we have chosen pertain to the bioleaching
of mixed sulphidic (pyrite and arsenopyrite) gold ore concentrates resulting in the
accumulation of ferric and arsenic ions in the solution. Apart from the dissolved metal
species, the pH and CO2 concentrations in the solution are also important parameters in
bioleaching processes.
The iron oxidizing efficiency of the strain increases with increasing carbon dioxide
supplementation of the sparged air upto 1% (V/V), beyond which the bacterial activity
reduces. Maintaining lower growth pH (1.6) reduced the formation of precipitates to a
large extent.
Ferric and arsenic ions are found to be toxic to the bacterium and strains tolerant to
higher concentrations of these metal ions could be obtained by adaptation. The growth
kinetics of the bacterium has been modeled incorporating the ferric and arsenic inhibition.
The adaptation of Acidithiobacillus ferrooxidans to higher metal concentrations is
quantified by gradual reduction in inhibition constants. The simulation results were found
to match well with the experimental data obtained. The applicability of the inhibition
constants obtained to the batch bioleaching model of a mixed sulphidic concentrate has
been analysed.
Keywords: Acidithiobacillus ferrooxidans, biooxidation, inhibition, adaptation, modeling

1.
INTRODUCTION
Biooxidation rate depends to a large extent on the efficiency of the bacterial strain
used. The regeneration of ferric iron by bacterial oxidation of ferrous sulphate in the
solution phase is an essential step in the biooxidation process. The oxidation rate is,
therefore, directly related to the activity of the bacteria in solution phase, which in turn is
determined and driven by the solution conditions.
Bioleaching experiments results in the formation of ferric and arsenic hydroxides that
precipitate at pH 2. One way of minimizing the formation of these precipitates is by
maintaining a low process pH. Studies by several workers (Jose et al, 1997; Kelly, 1978;
1099
Nagpal et al, 1992) have clearly indicated the reduction in precipitate formation by
lowering pH. Since the bacterium employed has an optimum pH of 2, it has to be suitably
adapted to the low pH required in order to retain its optimal activity. Studies done by
Toxiarchou et al, (1994) on the extent of arsenopyrite and pyrite oxidation at various pH
values showed that although the selective oxidation of arsenopyrite is enhanced at low pH,
pH values lower than 1.6 adversely affects the overall oxidation kinetics since under these
conditions the oxidation of pyrite is hindered. Hence a pH of 1.6 would be optimal for
reducing the precipitate formation without hindering the oxidation kinetics.
Oxygen and carbon dioxide are essential nutrients for bacterially catalyzed mineral
oxidation and bacterial growth. Reduction in oxygen concentration in the aqueous phase
below optimum levels leads to reduced growth rates. Similarly, reduction of aqueous
phase carbon dioxide also sharply reduces the bacterial growth. Hence, the effect of CO2
enrichment on the growth of bacteria needs to be studied.
Various toxic substances, often inherently present in the biooxidation of a sulphide
mineral, have been found to inhibit the oxidation ability of bacteria (Bailey and Hansford,
1993). During biooxidation process, the concentration of the released metal ions in the
solution continually increases and hence there will be build-up of ferric and arsenic ions.
Both ferric and arsenic ions are toxic to the bacterial activity and hence the growth of
bacteria in solution phase in presence of these metal ions needs to be examined. This
detrimental effect can be overcome by adapting the bacteria to higher concentrations of
these metal ions (Modak and Natarajan, 1995).
The objective of our present work was to analyse the effect of solution conditions on
the activity of the bacteria, model the growth kinetics of the bacterium incorporating the
inhibition effects of the toxic metal ions present in the solution and to analyse the
applicability of this to the batch bioleaching model of a mixed sulphidic concentrate.
2.
EXPERIMENTAL
2.1 Materials
The bacterial culture used was a pure strain of Acidithiobacillus ferrooxidans that was
isolated from Hutti Gold Mines (HGML) and is referred to as TfH6. The purity was
ascertained by the procedure outlined by Karavaiko (1988). The bacteria were cultured in
sterile 9K medium developed by Silverman and Lundgren (1959). The refractory sulphidic
concentrate obtained from HGML was in finely ground form with an average particle
diameter of 70 m. The chemical composition was 31.95 wt% Fe, 29.25 wt% S and 22.5
g/MT gold.
2.2 Methods
2.2.1 Growth of bacteria
The growth experiments were carried out by inoculating 10%(V/V) of active
inoculum to sterilized 9K medium. The flasks were incubated at 30C on a rotary shaker
at 240 rpm. The ferrous and the total iron concentration were monitored at regular
intervals by spectroscopic method. The solution cell number was determined by
microscopic counting with Petroff-Hausser counter using a Leitz phase contrast
microscope (Laborlux K Wild MPS12). pH changes were also monitored simultaneously
using a Systronics digital pH meter.
1100
2.2.2 Effect of carbon dioxide on growth

The effect of carbon dioxide availability on the activity of bacteria was monitored by
measuring the iron oxidization rate of bacteria at different levels of carbon dioxide
supplementation. These experiments were carried out in 17L bioreactors with the
operating volume of 6L. The aeration rate and the carbon dioxide flow rate were
monitored using suitably calibrated rotameters.
2.2.3 Growth inhibition and adaptation studies
The inhibition effect of ferric iron and arsenic on the growth of bacteria was studied
by adding ferric sulphate or sodium arsenite at different concentrations to the flasks
containing 9K medium solution, prior to inoculation. The iron oxidation rate of these
cultures were monitored and compared with that of the control without inhibition.
Adaptation to lower pH was done by serial subculturing of the bacterial strain in 9K
medium maintained at the required pH. Adaptation of the bacteria to the toxic ferric and
arsenic ions was carried out by serial subculturing of bacteria in 9K medium in presence
of externally added ferric sulphate and sodium arsenite as the sources of ferric and arsenic
ions respectively.
2.2.4 Batch bioleaching experiments
The bioleaching experiments were performed in batch mode in well-agitated and
aerated 17L bioreactor with an operating volume of 6L. The reactor was aerated at the rate
of 0.5 v/v/min with 1% carbon dioxide enrichment. The experiments were started with a
low pulp density of 1%(W/V), which was then increased in stepwise with periodic
addition of fresh concentrate when the concentrate in the reactor was completely oxidised.
Complete oxidation was considered to be achieved when the total iron in the solution
approached the theoretical stoichiometric value estimated.
3.
3.1 Growth of Acidithiobacillus ferrooxidans

The growth of a typical culture of Acidithiobacillus ferrooxidans (TfH6) on ferrous
iron with time under optimum conditions of temperature (30C) and pH (2.0) is shown in
Figure 1a. No significant increase in cell density was observed during the initial lag period
of 20hrs, beyond which the growth followed exponential kinetics. The growth remained in
log phase till the 50th hr after which the cells entered the stationary phase. The maximum
cell number was reached in 48-50 hrs.
3.2 Effect of pH
Lowering of pH did not have any significant effect on the activity of the bacterial
strain studied and just after one subculturing the activity of the adapted strain matched
with that of the wild strain as seen from Figure 1b. Comparing the growth curves at pH 2
and pH 1.6 (Figures 1a and 1b), we find that at the end of exponential phase when all the
ferrous iron is oxidized, the ferric iron concentration in the flask maintained at low pH is
8.9 g/L, which is higher than that observed at pH 2 (6.3 g/L). This clearly indicates that
the precipitation of ferric hydroxide found at pH 2 is reduced to a large extent by lowering
the pH.
1101
(a)
10
2+
Fe
3+
Fe
Iron concentration (g/L)
Iron concentration (g/L)
10
6
4
2
0
(b)
8
6
2+
Fe
3+
Fe
4
2
0
10
20
30
40
50
60
70
10
20
Time (h)
30
40
50
60
70
Time (h)
Figure 1. Growth of Acidithiobacillus ferrooxidans at (a) pH 2 and (b) pH 1.6

3.2 Effect of CO2 on growth of organism
In order to study the effect of CO2 on the growth of bacteria, the growth experiments
were conducted in 17L bucket bioreactor at different levels of CO2 supplementation (vol
CO2/vol air). The results obtained are plotted in Figure 2. It is clear from the figure that
the supplementation of CO2 to the sparged air resulted in enhancement of growth rate. In
the absence of CO2 supplementation, the bacteria were able to oxidize the ferrous iron in
9K medium in 52 hours. With 0.5% CO2 supplementation, the time taken for complete
growth was just 42 hours, which reduced to 38 hours at 1% CO2 supplementation. The cell
yield also showed slight increase with increase in CO2 supplementation. However beyond
1% CO2 supplementation, there was no much benefit since the oxidation rate was less at
2% CO2 supplementation as compared to 1% CO2 supplementation.
25
(a)
Cell number X 10 (cells/L)
0% CO2
0.5% CO2
1% CO2
2% CO2
6
4
10
Ferrous iron concentration (g/L)
10
2
0
0
10
20
30
40
50
60
Time (h)
(a) Ferrous iron concentration
70
(b)
20
15
10
0% CO2
0.5% CO2
1% CO2
2% CO2
5
0
0
10
20
30
40
50
60
Time (h)
(b) Cell density
Figure 2: Effect of CO2 supplementation on growth of Acidithiobacillus ferrooxidans

1102
70
Nagpal et al., (1992) have shown that bacteria grown on a sulphide ore concentrate
utilize oxygen and carbon dioxide in a molar ratio of approximately 20 mol oxygen per
mole CO2. This suggests that air supplemented with 1% CO2 would provide the
appropriate molar ratio of the two nutrients. During their tests at 16% solids, carbon
dioxide concentrations in the liquor in excess of 10mg/L were found to be inhibitory to
bacterial growth. The obtained results are thereby, in good agreement with that reported.
3.4 Effect of ferric and arsenic inhibition
In order to study the effect of Fe3+ on ferrous oxidation, the growth of TfH6 was
monitored with varying concentrations of Fe3+ added initially to the nutrient medium. The
results of the experiments with 2, 5 and 8 g/L of initial Fe3+ is presented in Figure 3a.
While 2 g/L initial Fe3+ had no significant effect on growth, an extended lag phase was
observed with 5 g/L and 8 g/L initial Fe3+ concentration. This is accompanied by reduced
rates of ferrous oxidation and cell growth.
The arsenic ions coming to the solution due to solubilization of arsenopyrite is in the
form of arsenite [As3+] and arsenate [As5+] ions. Since arsenate ions are less toxic
compared to arsenite ions, the inhibition effect of only arsenite ions is considered in our
studies. Figure 3b, shows the cell growth and ferrous iron oxidation by TfH6 in presence
of different initial concentrations of arsenite ions. From the figure we can observe that the
wild strain is naturally tolerant upto 400 ppm of arsenite concentration. Further increase in
arsenite concentration resulted in extended lag phase and decrease in cell yield. The
extended lag phase was significant when the initial arsenite concentration maintained was
1000 ppm and the wild strain took more than 58hrs to oxidize 9 g/L ferrous. It was
observed that the total cell yield was less at arsenite concentrations of 500 and 1000 ppm.
10
(a)
3+
Fe concentration
0 g/L
2 g/L
5 g/L
8 g/L
6
4
2
0
0
10
20
30
40
Time (h)
50
60
70
10
(b)
8
6
3+
As concentration
4
0 ppm
100 ppm
200 ppm
2
500 ppm
1000 ppm
10
20
30
40
50
60
70
Time (h)
Figure 3: Effect of (a) ferric and (b) arsenite ions on growth of Acidithiobacillus
ferrooxidans
3.5 Bacterial adaptation studies
From Figures 3, it is clear that the growth of TfH6 is inhibited in presence of high
initial ferric and arsenic concentration. However, repeated subculturing in presence of
these toxic ions results in reduced lag period and increased growth rate due to its
adaptation. The bacterial strain is said to be adapted when the cell growth and iron
oxidation rate of the strain in presence of the externally added toxic metal ion is
1103
comparable with that of the control, which is the growth rate of the strain in the absence of
any inhibition. The results obtained in presence of 8 g/L Fe3+ and 1000 ppm As3+ is shown
in Figure 4. The growth rate of the wild bacterial strain in presence of the externally added
toxic metal ion is represented as its natural tolerance. When grown in presence of 8g/L
Fe3+ growth rate after two subcultures matched with that of the wild strain, thereby
indicating that the strain is now adapted to 8 g/L Fe3+. Hence, with just two subcultures it
was possible to obtain a strain tolerant to 8 g/L Fe3+. As evident from Figure 4b, it was
possible to obtain a strain tolerant to 1000 ppm As3+ by subculturing the wild strain thrice
in presence of 1000 ppm As3+.
25
25
(b)
Cell number X 10 cells/L
20
15
10
10
Cell number X 10 cells/L
(a)
10
5
Natural tolerance
1 subculture
2 subculture
No inhibition (control)
0
0
10
20
30
40
Time (h)
50
60
70
20
15
10
Natural tolerance
1 subculture
2 subculture
3 subculture
No inhibition
5
0
0
10
20
30
40
50
60
70
Time (h)
Figure 4. Adaptation of Acidithiobacillus ferrooxidans to (a) 8 g/L ferric; (b) 1000 ppm
arsenite
3.6 Modeling of bacterial growth
The model for the growth of bacteria on ferrous iron should be able to predict the
rates at which ferrous is being oxidised under the changing conditions of ferrous and ferric
concentrations and the cell densities. The growth rate of bacteria in liquid is assumed to be
described adequately by Monod's model (Boon et al, 1999; Jensen and Webb, 1995) with
ferrous iron as the substrate, that is, the rate of substrate utilization is given by:
m Fe 2+
dFe 2+
=
dt
Y ( K S + Fe 2+ )
(1)
where m is the maximum specific growth rate in hr-1 and KS is the saturation constant for
ferrous iron in g/L respectively.
Y is the yield coefficient given by,
dX
dFe 2+
= Y
(2)
dt
dt
where X represents cell density in solution in cells/L.
Both ferric and arsenic ions inhibit the growth of Acidithiobacillus ferrooxidans
competitively without affecting the maximum specific growth rate. The rate is thus given
by the below equation (Bailey and Olis, 1986),
1104
m Fe 2+
dFe 2+
=
dt
Y ( K S (1 + K I Fe3+ + K a As 3+ ) + Fe2+ )
(3)
where KI is the inhibition constant for ferric iron and Ka for arsenic ion in L/g.
Experimental data of growth in absence of any inhibition was used to fit the
parameters of the above equation and the simulation result obtained is shown in Figure 5.
The values of m and KS obtained are 0.07 hr-1 and 1.6 g/L respectively.
(a) Ferrous iron oxidation
(b) Cell growth
Figure 5. Growth of Acidithiobacillus ferrooxidans at 30C and pH 1.5
The experimental data obtained at different levels of CO2 supplementation were used
in fitting the parameters of the Monod equation and it was found that the supplementation
of CO2 affects the maximum specific growth rate. The m obtained at different levels of
supplementation are given in Table 1, from which it is clear that the optimum level of
supplementation is 1% V/V of CO2.
The growth rate data obtained in presence of 8 g/L Fe3+ and 1000 ppm As3+ was used
to obtain the inhibition parameters in equation 3. Since both Fe3+ and As3+ ions inhibit
competitively, m remains unaffected. A comparison of the model parameters obtained
with that reported in literature is given in Table 2. As see from the table, the values
obtained are in close agreement with the reported values (Jensen and Webb, 1995; Lizama
and Suzuki, 1989).
Table 1. Variation in m with CO2
supplementation
Table 2. Model parameters of bacterial

growth
Percent CO2
supplementation
0
0.5
1
2
m, hr-1
Parameter
0.07
0.0825
0.0925
0.0785
m (hr-1)
Ks (g/L)
Y (cells/mol Fe2+)
KI (L/g)
Ka (L/g)
Estimated
value
0.09
1.6
1.36x1012
0.14
1.15
Reported
value
0.05 - 1.3
1-2
9x1013 - 5x1014
0.5 - 2.3
0.45 - 0.6
The comparison between model simulation using the parameters listed in Table 2 and
the experimental data is shown in Figure 6. The bacteria, on continuous exposure, develop
tolerance to ferric and arsenic ions as discussed earlier. The effect of bacterial adaptation
1105
to ferric and arsenic ions on the inhibition constants KI and Ka is shown in Figure 6. From
the figure we observe that with repeated subculture, the value of the constants approaches
zero.
Figure 6. Variation of inhibition constants with adaptation of Acidithiobacillus

ferrooxidans to (a) 8 g/L ferric; (b) 1000 ppm arsenic
3.7 Application to batch leaching model
A complete batch bioleaching model for a mixed sulphidic concentrate was developed
earlier by considering the various subprocesses independently to obtain the model
parameters (Chandraprabha et al, 2001). In this model growth of bacteria was described by
equation 1, neglecting the inhibition effect of ferric and arsenic ions coming into the
solution. Here the effect of the inhibition constants on the model prediction of the batch
biooxidation of 1% pd pyrite-arsenopyrite concentrate was tested and the results obtained
are shown in Figure 7. The bacterial growth in solution was described by equation 3 and
the parameters listed in Table 2 were used for simulation. As seen from the figure, the
model overpredicts the ferric iron concentration when the inhibition effect of ferric and
arsenic ions is neglected (dotted lines).
The presence of ferric and arsenic ions in the solution thereby inhibits the bacterial
activity. After 5 days of leaching, when the concentrate was almost completely oxidized,
the pulp density was increased from 1% to 2% by addition of fresh concentrate.
Simulation done at 2% pd also showed similar behaviour (data not shown). The slurry was
maintained at 2% pd for 5 days after which it was increased to 3% by addition of fresh
concentrate. The results for 3% pulp density are shown in Figure 7b. During the course of
biooxidation the bacteria are subjected to increasing concentrations of the toxic ferric and
arsenic ions coming from the concentrate and grow in presence of these stressed
conditions. When solid pulp density reached 3%, the bacteria were exposed to total of
about 240 h. As revealed by our adaptation studies, the bacteria attain tolerance at 3
subcultures (approximately 200 h). Hence the later generations of bacteria are adapted
strains that are tolerant to these toxic ions. This is clear from Figure 7b, which shows the
simulation results obtained at 3% pd concentrate with KI and Ka values set to zero. Here
we find that the prediction obtained by neglecting the inhibition effects of ferric and
arsenic fits well with the experimental data points.
1106
Figure 7. Simulation of batch bioleaching at (a) 1% pd and (b) 3% pd concentrate. Solid

lines represent model curves with ferric and arsenic inhibition and dashed lines without
any inhibition. Symbols represent experimental data points
4.
CONCLUSIONS
The effect of solution conditions on the activity of Acidithiobacillus ferrooxidans is
reported and the following conclusions can be arrived at:
Lowering of process pH resulted in reduction of the precipitate formation, thereby
enhancing bacterial oxidation rates.
Supply of air supplemented with CO2 enhanced the growth kinetics of the strain used.
The iron oxidation rate increased with increase in CO2 supplementation upto 1%
(V/V), beyond which the cell yield decreased.
Presence of both ferric and arsenic ions inhibited the bacterial growth. Serial
subculturing in presence of these metal ions resulted in adapted strains that were
tolerant to higher concentrations of the metal ions.
The growth kinetics of bacteria was modeled incorporating the inhibition constants
obtained for ferric and arsenic inhibition and it was found that with repeated
subculturing the value of the inhibition constant approached zero which clearly
indicated the adaptation of bacteria to the toxic metal ions due to continuous
exposure.
The model developed for the batch bioleaching of pyrite-arsenopyrite concentrate was
found to overpredict the ferric iron concentration in the solution at 1% pd when the
effects of ferric and arsenic inhibition were neglected. This was overcome by
incorporating the inhibition constants in the growth equation of bacteria.
The predictions obtained at later stages at a higher pulp density were found to fit well
with the experimental data even when the inhibition effects were neglected. This was
due to the adaptation of the bacteria to the released metal ions in solution that was
achieved due to their continuous exposure to these ions.
REFERENCES
1. Bailey, J.E. and Hansford, G.S., 1993. Factors affecting bio-oxidation of sulphide
minerals at high concentrations of solids: A Review. Biotechnol. Bioengg., 42, 11641174.
1107
2. Bailey, J.E. and Olis, D.F., 1986. Biochemical engineering fundamentals, second ed.
McGraw Hill, Singapore.
3. Boon, M., Ras, C., Heijnen, J. J., 1999. The ferrous iron oxidation kinetics of
Thiobacillus ferrooxidans in batch cultures. Appl. Microbiol. Biotechnol., 51, 813819.
4. Chandraprabha, M.N., Modak, J.M., Raichur, A.M. and Natarajan, K.A., 2001.
Modeling of biooxidation of gold bearing pyrite-arsenopyrite concentrates by
Acidithiobacilli ferrooxidans. In Ciminelli, V.S.T. and Garcia, O., (Eds.),
Biohydrometallurgy: Fundamentals, Technology and Sustainable Development,
Process Metallurgy 11A, Elsevier, Amsterdam, pp.43-54.
5. Jensen, A.B. and Webb, C., 1995. Ferrous sulphate oxidation using Thiobacillus
ferrooxidans: A review. Process Biochem., 30, 225-236.
6. Jose, M.G., Domingo, C. and Barrie, J., 1997. Comparison of the effects of
temperature and pH on iron oxidation and survival of Thiobacillus ferrooxidans (type
strain) and Leptospirillum ferrooxidans. Biotechnol. Bioengg., 42, 581-589.
7. Karavaiko, G.I., 1988 Methods of isolation, evaluation and studying of
microorganisms. In Karavaiko, G.I., Rossi, G., Agate, A.D., Groudev, S.N., Avakyan,
Z.A., (Eds.), Biotechnology of Metals Manual, Center for International Projects
GKNT, Moscow, pp.47-79.
8. Kelly, D.P. and Jones, C.A., 1978. Factors affecting metabolism and ferrous iron
oxidation in suspensions and batch cultures of Thiobacillus ferrooxidans. In: Murr,
L.E., Torma, A.E. and Brierly, J.A., (Eds), Metallurgical applications of bacterial
leaching and related microbiological phenomenon, Academic, New York, pp. 19-44.
9. Lizama, H.M. and Suzuki, I., 1989. Synergitic competitive inhibition of ferrous iron
oxidation by Thiobacillus ferrooxidans by increasing concentrations of ferric iron and
cells. Appl. Environ. Microbiol., 55, 2588-2591.
10. Modak, J.M. and Natarajan, K.A., 1995. Development of special strains of
Thiobacillus ferrooxidans for enhanced bioleaching of sulphide minerals. In: Jerez,
C.A., Vargas, T., Toledo, H. and Wiertz, J.V., (Eds), Biohydrometallurgical
processing, Vol I, University of Chile, Santiago, Chile, pp. 34-46.
11. Nagpal, S. and Dahlstrom, D., 1992. Effect of carbon dioxide concentration on the
bioleaching of a pyrite-arsenopyrite ore concentrate. Biotechnol. Bioengg., 41, 459464.
12. Silverman, M.P. and Lundgren, D.G., 1959. Studies on chemoautotrophic iron
bacterium Thiobacillus ferrooxidans. J. Bacteriol., 77, 642-646.
13. Taxiarchou, M., Adam, K. and Kontopoulos, A., 1994. Bacterial oxidation conditions
for gold extraction from Olympias refractory arsenical pyrite concentrate.
Hydrometallurgy 36, 169-185.
1108

Ferrous ion oxidation by an activated carbon cloth modified

with Acidithiobacillus ferrooxidans
F. de J. Cerino-Cordova, J.P. Magnin, N. Gondrexon, P. Ozil
Laboratoire dElectrochimie et de Physico-chimie des Matriaux et des Interfaces,
UMR 5631 INPG-CNRS- UJF, Equipe Gnie des Procds, ENSEEG BP75,
38402 St Martin dHres, France
Abstract
The formation of Acidithiobacillus ferrooxidans biofilm on an Activated Carbon
Cloth (ACC) was performed during subcultures of a biologically modified ACC in sterile
ferrous iron medium. Two ways of modification were investigated: the classical culture of
free bacteria on ACC, an enhanced way using a preliminary sorption of concentrated
bacteria on ACC surface.
In the absence of bacteria, the ACC exhibited a low ferrous-iron oxidation capacity,
23 mg Fe2+ L-1/h. The presence of free (initial concentration 0.07 mg protein L-1) or
immobilised bacteria on ACC WWP3 and WKL20 (4-8 mg protein/g carbon) increased
the ferrous ion oxidation rate respectively up to 60-90 and 600 mg Fe2+ L-1/h.
The preliminary sorption step allowed to decrease the time required for biofilm
stabilisation of about 30 percent, depending on the initial concentration of bacteria
immobilised on ACC. No difference was observed concerning the biofilm formation and
activity for the two types of ACC, excepted the final fragility of bacterial modified KL20
ACC.
Keywords: Acidithiobacillus ferrooxidans, biofilm, activated carbon cloth

1.
INTRODUCTION
Acidithiobacillus ferrooxidans is a chemoautotrophic acidophilic microorganism,
which gets the energy indispensable for its growth from oxidation of ions ferrous,
elemental sulphur and reduced sulphur compounds including metal sulphides [1]. This
ability has been used from a long time in different processes such as bioleaching [2-6],
H2S removal [7] and acid mine drainage treatment [8-9].
It is well known that the free growth of A. ferrooxidans is characterised by a low rate,
thus requiring long residence times and consequently large reactor volumes resulting in
high investment and operating costs [10]. Therefore many research works have been
focused on increasing ferrous iron oxidation rate using immobilised biomass. Various
configurations were investigated using different bioreactor types (packed bed [9-13],
fluidised bed [13-15] trickle bed reactor [16] and rotating biological contactor [17,18,19])
with different non-metabolisable supports (glass beads [13-14]; ion exchange particles
1109
[13-14], diatomeous earth [20-21], polyurethane foam [22], silica gel [23], sand [8], nickel
alloy fibre [24], and activated carbon [13-14, 23]).
Activated carbon exists under various forms such as granule, fibre or cloth. Activated
Carbon Clothes (ACC) offer faster kinetics than Granular Activated Carbon to treat
micropolluants [26] and to sorb heavy metals [27].
Our previous studies were devoted to fixing A. ferrooxidans on Granular Activated
Carbon in order to use this biological modified support as the biocathode of a
bioelectroreactor [28]. Activated Carbon Cloth showed an improved biomass capacity for
A. ferrooxidans immobilisation (respectively 25 mg protein /g Carbon for WWP3 and 14,9
mg protein /g Carbon for WKL20) [unpublished data]. Therefore the present paper focuses
on developing a A. ferrooxidans biofilm on ACC and characterising it by its iron-oxidising
activity and from Scanning Electron Microscopy observations.
2
2.1 Culture and medium

The bacterial strain Acidithiobacillis ferrooxidans DSM 583 was routinely maintained
in 9K liquid medium (33 g.L-1 FeSO4, 0.4 g L-1 MgSO4.7H2O, 0.4 g.L-1 K2HPO4, 0.4 g.L-1
(NH4)2SO4) at pH=1.4 (H2SO4)
After bacterial growth at 30C in a 5-L bioreactor, the biomass was concentrated by
cross-flow filtration before being harvested at 5000 g for 10 min. The resulting biomass
was washed four times with H2SO4 (pH 1.4) and harvested at 1000 rpm during 10 min to
remove jarosite precipitate. Finally, the biomass was suspended again in H2SO4 (pH 1.4)
and kept at 5C during 24 h before utilisation.
2.2 Activated Carbon Cloths
The physical characteristics of ACC WWP3 and WKL20 (purchased from Actitex,
France) are recapitulated in table 1.
Table 1. Industrial characteristics of Activated Carbon Cloths WWP3 and WKL20
Characteristics
Specific Surface (MEB)
Weight
Mechanical resistance
Median Pore diameter
Weaving
Pore < 20
Pore < 500
Microporosity
Thickness
WWP3
1000-1300 m2/g
110-130 g/m2
5.0 DaN/cm
8.4
Woven fabric
0.52 cc/g
0.65 cc/g
>80%
0.5 m
WKL20
1000 m2/g
100 g/m2
7
Knitted fabric
0.31 cc/g
0.33 cc/g
>90%
0.5 m
2.3 Analytical procedures

The concentration ratio of ferrous ion to total (ferrous + ferric) ions was measured by
the o-phenantroline method [29]. The bacterial concentration in solution correlated with
protein concentration was estimated by the Lowry method [30].
1110
2.4 Modification of ACC with bacterial biomass

The ACC biological modification was performed via preliminary adsorption batch
experiments at 30C and 100 rpm. A piece of ACC (0.1345 g) was introduced in a 0.1 L
flask containing 0.02 L H2SO4 (pH=1.4) and 0.2 g protein L-1. After 2h, this piece was
rinsed with sterile demineralised water to eliminate non-attached bacteria. It was next used
in the ferrous ion oxidation experiments.
2.5 Bacterial subcultures in the presence of ACC cloths
The A. ferrooxidans bacterial cultures with modified or unmodified ACC (WWP3 or
WKL20) were realised in 0.1 L flask containing 0.2 L of culture medium at 30C under
rotary shaker agitation (150 rpm).
In the case of non-biologically modified ACC, biomass (0.07 mg prot. L-1) was
introduced in the medium. During incubation, samples were regularly collected and
analysed to determine both ferrous ion oxidation activity (ratio Fe2+/ (Fe2+ + Fe3+)) and
protein concentration in the medium. When 90% of ferrous ion concentration were
oxidised (corresponding to a time noted T90), the ACC samples were picked up under
sterile condition, rinsed with sterile demineralised and introduced into a 20-ml sterile
culture medium. This procedure was repeated several times until reaching constant T90
and bacterial concentrations in solution.
2.6 Scanning Electronic Microscopic
ACC samples were rinsed with sterile demineralised water and put into a Nacacodylate solution (4% cacodilyc acid, sodium salt trihydrate 98%) dissolved in a 25%
glutaraldehyde solution (pH 7). After 4-h incubation, the samples were transferred firstly
into a water solution for 1h and then successively into 25%, 50%, 75% and 100%
xylene/acetone solutions for 10 minutes each. Finally, the samples were air-dried and
prepared before SEM observation (Jeol 6400).
3.

The WKL20 and WWP3 ACC under study exhibited a capacity to oxidise ferrous ion
in solution as shown in Fig. 1. Without ACC and in an uninoculated medium, no ferrous
iron was oxidised. In the absence of bacteria, 90% of ferrous ions were thus oxidised after
170 and 340 h respectively with WWP3 and WKL20. This Fe2+ oxidation capacity of
ferrous ion oxidation was previously observed on activated carbon [11] and carbon fibre
[31].
Figure 1. Ferrous oxidation by Activated Carbon Cloth WWP3 or WKL20 under

different conditions: with and without bacteria
1111
The presence of bacteria in solution did not change the Fe2+ oxidation rate during the
first 40h for of WKL20 and during 70 h for WWP3. Ferrous ions were rapidly oxidised
either by ACC samples than by A. ferrooxidans biomass. After this period, bacterial Fe2+
oxidation was more important than that observed with ACC alone. 90% of ferrous ions
were thus biologically oxidised within a short time (respectively 50 h and 90 h in the
presence of WKL20 and WWP3 ACC). This oxidation activity was directly correlated to
the free or immobilised biomass production. The protein concentrations (5 mg.L-1 with of
WKL20 and 3 mg.L-1 with WWP3) were lower than that obtained without ACC (7 mg.L1
). As a conclusion ACC supports present an inhibition effect for iron-oxidising activity
such as observed for activated carbon [23], flowers of sulphur, fluorapatite, glass bead and
pyrite [5,32]
This bacterial diminution in solution was linked to the development of a biofilm onto
the ACC surface associated with a jarosite precipitation (Fig. 2). No jarosite or biofilm
structures were observed onto initial ACC samples (data not shown here). A. ferrooxidans
modified ACC materials show a catalytic effect in ferrous ion oxidation as observed in the
case of Biomass Support Particles with Activated Carbon coating [11].
Figure. 2. SEM observations of A. ferrooxidans biofilm obtained (a) onto WKL20

after 60h, (b) WWP3 after 90h
The development of a stable biofilm onto ACC was studied during subcultures of A.
ferrooxidans modified ACC in Fe2+ growth medium (see Materials and Methods) (Fig. 3).
The amount of biomass present within the biofilm was indirectly measured from the
amount of biomass present in solution (via protein concentration) and associated to SEM
observation of the biofilm. The total Fe2+ oxidation activity was estimated by determining
the time required to oxidise 90% of ferrous ions (T90), resulting to the biological activity
due to the biomass eventually present in solution or immobilised into the biofilm.
The evolution of protein concentration or T90 appeared to be similar for both
biologically modified ACC. The ferrous activity curves involved three distinct steps.
During step I, the bacterial concentration in solution after oxidation of 90% Fe2+ was
greater than that obtained during a batch culture without AC Cloths (5 mg.L-1). These
freely dispersed bacteria resulted from two actions:
either cells were divided in the biofilm, left the biofilm, and were suspended in
medium,
or / and the loss of some bacteria from the biofilm due to friction between the ACC
support and glass flask. [15]
In both cases, Fe2+ oxidation rate was increased as demonstrated by the fast T90
decreasing from 85h to 10-18 h depending on the ACC biofilm. The biofilm was more
developed that that obtained after the first culture and consisted of jarosite precipitates and
1112
bacteria (Fig 4). Bacterial cells become embedded within iron precipitates and were
adsorbed through complex multilayers patterns.
Figure 3. Biological characteristics (protein concentration and time necessary to

oxidise 90% of ferrous ions, T90) of subcultures of ACC WKL20 (A) and WWP3 (B)
in 9K medium. A. ferrooxidans biomass (0.07 mg.L-1) being only introduced during
the first subculture). The vertical arrows with dotted lines indicate the different steps
in the biofilm formation (see text)
During step II, T90 was increased from 10-18 h to 40 h, thus indicating a decrease of
the overall activity for ferrous ions oxidation. Such diminution of biological activity was
correlated to a large bacterial concentration in solution (6-11 mg protein L-1). The biomass
present in solution exhibited a very low or a non-existent ferrous ion oxidation activity.
However more jarosite deposits and bacteria were observed onto biofilm (Fig. 4).
Step III led to the formation of a biofilm which is stable for its structure and its
metabolic activity: no variation in the biomass concentration present in solution and in the
time necessary to oxidise 90% of ferrous ions initially present in solution were observed.
A weak bacterial concentration in solution (1-2 mg protein L-1) was reached after 13
subcultures. The final biofilm recovered the whole ACC surface and involved a significant
jarosite deposit (Fig. 4).
Finally, a stable and active A. ferrooxidans biofilm was obtained onto both ACC
under study after about 300-350 h subcultures. This biofilm oxidised five times faster the
substrate (ferrous ions) than a culture with free biomass.
In order to decrease the time required to obtain a stable biofilm onto ACC, the clothes
used latter were previously modified by A. ferrooxidans biomass. This modification
resulted of an initial sorption of concentrated biomass onto ACC (see Materials and
Methods). A 2 h contact period allowed to reach the equilibrium between non immobilised
1113
and immobilised cells. A better bacterial sorption capacity was obtained with WWP3 (7.8
mg protein L1) than with WKL20 (4.6 mg protein L1) as confirmed by SEM observations
(Fig. 5).
Figure 4. SEM observations of A. ferrooxidans biofilm onto Activated Carbon Cloths

WKL20 and WWP3 during subcultures in 9K medium (arrows =bacterial cell)
Figure 5. SEM observations of Activated Carbon Cloths modified with bacteria after
2h of bacterial fixation (pH 1.4, 30C)
The development of A. ferrooxidans biofilm from these initially fixed biomasses was
studied during subcultures of biologically modified ACC in 9K medium (Fig. 6A, B).
No shape difference was observed on the evolution of biological characteristics
whatever the ACC nature may be. However these curves were simpler than those observed
when the cloth is directly introduced into the bacterial culture (Fig. 3): only two steps
were observed instead of three.
1114
During the first culture, the ferrous ions were rapidly oxidised by bacteria as shown
by the weak T90 value (12 h). This oxidation was correlated to a desorption process
confirmed by measuring protein concentration in solution and by comparing SEM
observations (Fig. 7 and Fig. 4 for WWP3).
Figure 6. Biological characteristics (protein concentration and T90) of subcultures

for ACC WKL20 (A) and WWP3 (B) modified with A. ferrooxidans biomass in 9K
medium. The vertical arrows with dotted lines indicate the different steps in the
biofilm formation (see text)
Protein concentration in solution differed according to the AA cloth nature: 12 mg.L-1

for WKL20 and 4 mg.L-1 for WWP3. During the five first subcultures, the protein
concentration in solution increased because of bacterial desorption. This desorption was
correlated to an increase of the time required to biologically oxidise ferrous ions: 12h for
WKL20 and 16 h for WWP3. However, the biofilm was developing as shown by SEM
observations on Fig. 7.
From the sixth subculture, protein concentration in solution continuously decreased
down to 1-2 mg protein L-1. T90 simultaneously decreased indicating a high ferrous ion
oxidising capacity. This capacity was mainly due to biofilm because of the very low
bacteria concentration in solution.
1115
After 232-265h of total subculture, the biologically active biofilm oxidised the ferrous
ion substrate within 10-12h, with a very little biomass desorption (1-2 mg protein L-1).
Figure 7. SEM observations of the formation of an A. ferrooxidans biofilm on

Activated Carbon Cloth WKL20 and WWP3
4.
CONCLUSION
Activated Carbon Cloth has been selected as a supporting material because of its large
specific volume and high ability to adsorb A. ferrooxidans. The ACC here under study
showed their capacity to oxidise ferrous ions. However, this capacity is improved by
developing a A. ferrooxidans biofilm.
As observed by SEM, the biofilm consisted in jarosite deposits wherein bacterial cells
are able to develop. The development of a stable iron-oxidising biofilm may be explained
by the model of "biofilm with adsorbed cells" previously described by Karamanev [15]. A
very high iron-oxidising activity (600 mg Fe2+ oxidised L-1/h) was obtained for
biologically modified ACC. A preliminary bacterial immobilisation onto ACC allowed an
enhancement in developing and stabilising the biofilm. The formation of a stable biofilm
and long time incubation in acidic solution led to a loss of mechanical resistance.
Therefore ACC WKL20 rapidly breaks in contrast to ACC WWP3.
ACKNOWLEDGEMENTS
This work was carried out owing to a grant from "Consejo Nacional de Ciencia y
Tecnologia" (CONACYT, Mexico) accorded to Mr Felipe Jesus Cerino Cordova.
1116
REFERENCES
1. A. Das and A.K. Mishra, App. Microbiol. Biotechnol., 45 (1996) 377.

2. J. Haddadin, C. Dagot, M. Fick, Enzyme Microbial Technol., 17 (1995) 290.
3. G. Mercier, M. Chartier, D. Couillard, J.F. Blais, Environ. Manag., 4 (1999) 517.
4. E. Donati, C. Pogliani, J.L. Boiardi, Appl. Microbiol. Biotechnol., 47 (1997) 636.
5. N. Wakao, M. Mishina, Y. Sakurai, H. Shiota, J. Gen. Appl. Microbiol., 30 (1984) 63.
6. T.A. Fowler, P.R Holmes, F.K. Crudndwell, Appl. Environ. Microbiol.,.7 (1999) 2987.
7. A. Pagella and D.M. De Faveri, Chem. Eng. Sci., 55 (1999) 2185.
8. T.A. Wood, K.R. Murray, J.G. Burgess, Appl. Microbiol. Biotechnol., 56 (2001) 560.
9. H.R. Diz and J.T. Novak, J. Environm. Eng., 2 (1999) 109.
10. M. Nemati and C. Webb, Appl. Microbiol. Biotechnol., 46 (1996) 250.
11. M. Nemati and C. Webb, J.Chem. Technol. Biotechnol., 74 (1999) 562.
12. A. B. Jensen and C. Webb, Process Bioch., 2 (1995) 225.
13. S. I. Grishi and O.H. Tuoniven, Appl. Environm. Microbiol., 12 (1988) 3092.
14. S. I. Grishi and O.H. Tuoniven, Appl. Environm. Microbiol., 31 (1989) 31.
15. D.G. Karamanev, J. Biotechnol., 20 (1991) 51.
16. A. B. Jensen and C. Webb, Biotechn. Techn., 2 (1994) 87.
17. L. Nikolov, D. Karamanev, V. Mamatarkova, D. Mehochev, D. Dimitrov, Biochem.
Eng. J., 12 (2002) 43.
18. H. Olen and R.F. Unz, J. Water Pollut. Control Fed, 52 (1980) 257.
19. H. Olen and R.F. Unz, Biotechnol. Bioeng, 19 (1977) 1475.
20. T. Imaizumi, Biotechn. Bioeng Symp., 16 (1986) 363.
21. H. Magota, Y. Shiratori, C. Inoue, E. Yanagisawa, JP patent No.01215399 (1989).
22. H. Armentia and C. Webb, Appl. Microbiol. Biotechnol., 36 (1992) 697.
23. T. Kai, T. Takahashi, Y. Shirakawa, Y. Kawabata, Biotechnol. Bioeng., 36 (1990)
1105.
24. S. Sandhya and R.A. Pandey, J. Environ. Sci. Health, 2 (1992) 445.
25. J.M. Gomez, D. Cantero, C. Webb, Appl. Microbiol. Biotechnol., 54 (2000) 335
26. C. Brasquet, E. Subrenat, P. Le Cloirec, Wat. Sci. Tech. 39 (1999) 201.
27. K. Kadirvelu, C. Faur-Brasquet, P. Le Cloirec, Langmuir, 16 (2000) 8404.
28. A. Boyer, J.P. Magnin, P. Ozil, Int. Biohyd. Symp., Madrid, Spain, (1999) 211.
29. O.H. Tuoniven and D. P. Kelly, Arch. Microbiol., 88 (1973). 285.
30. M.K. Muir and T.N. Anderson, Metall. Trans. B., 88 (1977) 517.
31. M. Uchida, O. Shinohara, S. Ito, N. Kawasaki, T. Nakamura, S. Tanada, J. Colloid
Interf. Sci., 224 (2000) 347.
32. A.A. Dispirito, P.R. Dungan, O.H. Tuoniven, Biotechnol. Bioeng., 23 (1981) 2761.
1117

Heavy metal precipitation by off-gases from aerobic culture of

Klebsiella pneumoniae M426
A.M.M. Essa, L.E. Macaskie and N.L. Brown
School of Biosciences, The University of Birmingham,
Edgbaston, Birmingham B15 2TT UK
Abstract
In this study we have identified a new method for the removal of heavy metals from
solution by uses the off-gas produced from aerobic cultures of Klebsiella pneumoniae
M426. GC-Mass Spectroscopic analysis showed that the outlet gases contain
dimethyldisulfide (DMDS), which precipitates some metals from their solutions. The
outlet gases precipitated mercury, cadmium, palladium, and lead as metallo-thiol
complexes but did not precipitate other metals such as nickel, cobalt, zinc and arsenite.
Energy dispersive X-ray microanalysis showed that the metal precipitate consisted of the
metal plus carbon, sulfur and oxygen. Formation of metal carbonates was discounted since
precipitation was similar at pH 2.0; the precipitate did not have the physical characteristics
of metal sulfides. Bio-mineralization of heavy metals by using DMDS is a novel approach
because it can be used with concentrated or physiologically incompatible solutions, and
since the metal precipitate is kept separate from the bacterial biomass, the biological and
chemical end products can be managed independently.
1.
INTRODUCTION
The discharge of heavy metals into the environment due to agricultural, industrial and
military operations and the effect of this pollution on the ecosystem and human health are
growing concerns. Since the natural mineralization of metals is a slow process, pollution
by metals constitutes one of the most important environmental problems. Some metals
seem to serve no biologically relevant function e.g. Cd, Pb, Sn, Hg instead they cause
damage due to their avidity for the sulfhydryl groups of proteins which they block and
inactivate [1-6]. Different procedures for the removal of toxic metal species from
contaminated environments have been developed; most of them are based on ionexchange technologies and/or chemical precipitation of cations in an inert form. These
methods are expensive and also require additional products for desorption of metals and
for cleaning up of the inorganic matrix and lead to formation of concentrated secondary
wastes. Remediation technologies using microorganisms are feasible alternatives to the
physical cleaning of soils or the concentration of metals in polluted waters by physical or
chemical means [7]. For example, sulphate-reducing bacteria have been used to precipitate
metals as metal sulfides [8] but this requires control of emission of highly toxic H2S. As
To whom correspondence should be addressed (Fax: +44(0)121-414-5925;
e-mail: l.e.macaskie@bham.ac.uk)
1119
an alternative S-based precipitation process organosulfur compounds have been

considered, but these are too expensive for routine use [9].
Dimethyldisulfide (DMDS) has been reported to be produced from different
organisms: Pseudomonas sp. [10], Streptomyces sp. [11], Aspergillus sp. [12], Clostridium
sporogenes [13], Alteromonas putrefaciens, Pseudomonas fluorescens and Achromobacter
sp. [14], Pseudomonas putida [15], Proteus sp. [16] and Pseudomonas fluorescens,
Proteus vulgaris, and Serratia marcescens [17]. These bacteria were isolated from soil,
wastewater, and fish. Some bacterial strains isolated from activated sludge [18] as
Lactobacillus sp., Corynebacterium sp., Alcaligenes sp., and Pseudomonas sp. showed an
ability to produce DMDS from activated sludge.
The formation of methanethiol from methionine by cell extracts of Lactococcus lactis
subsp. cremoris B78 was proposed by Engels et al. [19]. The first step is the
transamination during which 4-methylthio-2-ketobutyric acid (KMBA) is produced. The
intermediate KMBA is probably, converted to methanethiol, which under aerobic
conditions is converted rapidly to dimethyldisulfide and/or dimethyltrisulfide. In this study
we describe a novel approach for the removal of some heavy metals by using the outlet
gases produced during the aerobic culture of Klebsiella pneumoniae M426. These gases
contain a volatile sulfur compound dimethyldisulfide (DMDS), which is proposed to
precipitate heavy metals with high efficiency.
2.
2.1 Bacterial strains and bioreactor operation

Klebsiella pneumoniae M426 strain was originally obtained from the NCTC (National
Culture Type Collection) Colindale, London. Bacterial cells were grown in Luria Broth
medium (LB; 20) in a shaker agitated at 200 rpm and incubated at 37C. A batch
bioreactor was constructed for the metal precipitation (Fig. 1) consisting of two chambers,
one for the bacterial growth (about 1L in volume, maintained under aerobic conditions by
pumping in a sterilised air, 37C) and the other chamber (50ml) for the precipitation of
metal via the pumping of the headspace gases from the culture through the metal solution.
Cultures were inoculated (100 ml) using cells in an exponential growth phase (6hrs) and
bacterial growth was monitored by measuring the optical density at 600nm.
2.2.1 Analysis of the soluble metals
Heavy metals (Hg2+, Cd2+, Pb2+, Pd2+) were prepared as aqueous solutions, and used
as the following concentrations: mercuric chloride, 0.02-10.0 mg/ml; cadmium chloride,
0.7 mg/ml; lead nitrate, 0.6 mg/ml; sodium tetrachloropalladate, 0.9 mg/ml. The metal
precipitates from the precipitation chamber were removed by centrifugation at 1600g for
15 minutes. Aliquots (10 ml) of the supernatant were filtered through 0.2 m (Nalgene
Syringe filters). The filtrates (Cd2+, Pb2+, Pd2+) were analysed using a 1999 Duo HR Iris
advanced inductive coupled plasma (ICP) spectrophotometer [21]. Mercuric ion
concentrations in the solutions were measured with a mercury analyser (Buck Scientific,
model 400A) applying cold vapour atomic absorption spectrometry [22].
1120
Figure 1. A bioreactor used for metal precipitation via the culture off-gas consisting
of two chambers one is the bacterial growth chamber and the second is the metal
precipitation chamber
2.2.2 Analytical electron microscopy
Samples from the precipitation chamber were centrifuged at 1600g for 15 minutes.
The supernatant was removed and precipitates were washed in 20ml deionised water
followed by centrifugation as before. This step was repeated 3 times. The precipitate was
collected and dried at 30C. The metal precipitates were examined with a JEOL JSM 5900
SEM. The elemental composition of the mineral phases was determined by energy
dispersive X-ray microanalysis (EDX) using an Oxford Link ISIS System.
2.2.3 GCmass spectral analysis
The outlet gases produced from the bioreactor were collected in serum bottles (50ml)
with a rubber caps. The samples were immediately analysed by GCMS [23], using a
Hewlett-Packard 5890/5970A system, with a HP1 column (50 mm x 0.2 mm fused silica
capillary column, film thickness, 0.5 mm). GC oven initial temperature was 60C and was
programmed to 220C at a rate of 2C/min, and finally held at 220C during 40 min.
Operating conditions of the GC were as follows: helium was used as carrier gas (0.8
ml/min); the temperature of injector and detector was 250C. The mass spectra were
determined at 70 eV in the mass range of 35-400.
1121
3.
3.1 Precipitation of mercury as HgS by growing cultures

During the aerobic growth of K. pneumoniae M426 a black precipitate was visible in
the extracellular matrix (Fig. 2a). EDX analysis data showed that the precipitate contained
mercury and sulfur (Fig. 2b). Accordingly, it was assumed that K. pneumoniae M426 has
the ability to produce hydrogen sulfide (H2S) under aerobic conditions, which can
precipitate Hg2+ as insoluble HgS [24]. These results were confirmed when a suspended
filter paper saturated with HgCl2 turned black after exposure to the culture outlet gases.
This phenomenon was recognised previously by Aiking et al. [25], who recorded an
aerobic mechanism for mercury bioremediation, in which Klebsiella aerogenes formed
HgS when grown in continuous aerobic culture in the presence of HgCl2 and at the same
time there was no bacterial mercury volatilization. This strain showed mercuric ion
sensitivity under sulfate-limited conditions and elevated levels of total cellular sulfide
were detected in cells grown in the presence of mercuric ions. It was suggested that this
was due to the formation of HgS. A similar phenomenon was recorded by Wang et al. [26]
who showed the ability of Pseudomonas aeruginosa to remove more than 99% of the
cadmium from 5mM cadmium solution by precipitation on the cell wall as cadmium
sulfide under aerobic conditions.
Figure 2. (a) Electron micrograph of aerobic culture of K. pneumoniae M426

showing an extracellular dark precipitate, (b) EDX spectrum of the dark precipitate
in (a)
3.2 Mercury and cadmium bio-mineralization using culture off-gases
A bioreactor for mercury bioremediation was designed according to the ability of K.
pneumoniae M426 to precipitate the soluble Hg2+ as HgS. The bacterial cells were used as
a gas generator in a separate chamber and the headspace gases produced during aerobic
growth were pumped into the metal solutions in the precipitation chamber through a
bacterial membrane filter to avoid any bacterial contamination.
A black HgS precipitate was anticipated, but more than 99% of the soluble Hg2+ was
recovered as a yellowish white precipitate within 2h (Fig. 3). A similar result was obtained
1122
with cadmium; within 1h there was more than 99% of Cd2+ removal from CdCl2 solution.
EDX analysis showed that the precipitates consisted of mercury or cadmium plus carbon,
sulfur and oxygen and there was no phosphorus in the precipitate, which confirms that
there was no bacterial contamination (Fig. 4a & 4b, 5a & 5b).
1
900
Bacterial growth
0.9
Residual soluble mercury

800
0.8
700
0.7
600
0.6
500
Bacterial
growth
0.5
400
0.4
(O.D600)
300
0.3
Residual
soluble
Hg2+
(mg/l)
200
0.2
100
0.1
0
0
0
20
40
60
Time (min)
80
100
120
Figure 3. A time course for bacterial growth and mercury bio-precipitation by using
the outlet gases produced from the aerobic culture of K. pneumoniae M426. The data
shown are for an initial Hg2+ concentration of 0.8 mg/ml, similar results were
obtained using Hg2+ up to 10.0 mg/ml (Error bars represent means standard
errors)
Figure 4. Mercury precipitation by using the volatile thiols produced by K.

pneumoniae M426, (a) an electron micrograph of the mercury precipitate, (b) EDX
spectrum of the mercury precipitate
1123
Figure 5. Cadmium precipitation by using the volatile thiols produced by K.

pneumoniae M426, (a) an electron micrograph of the cadmium precipitate, (b) EDX
spectrum of the cadmium precipitate
The EDX spectrum for the Hg precipitate cannot easily distinguish between Hg and
sulfur due to peak overlap, but the peaks were clearly differentiated in the corresponding
Cd-precipitate. It could be suggested that metal carbonates were precipitated. However
similar results were obtained at pH 2.0, at which pH little HCO3- would be present in the
solution and no CO32- [27]. GC-Mass Spectroscopic analysis showed that the outlet gases
from the bioreactor contained volatile dimethyldisulfide (DMDS). According to these
results, we suggest a novel mechanism for mercury bioremediation by the production of
DMDS as a metabolic by-product during the aerobic growth of K. pneumoniae M426.
This volatile species may form insoluble mercurithiol-compounds. There was more
than 99% mercury precipitation by this mechanism with different concentrations of HgCl2
(from 20 g/ml to 10 mg/ml) within 2 hrs. This bioprocess also gave high metal removal
efficiency over a wide pH range (acidic, neutral or alkaline) and also at salinity levels up
to 1M NaCl.
DMDS for bio-mineralization of heavy metals
The head gas produced by aerobic culture of K. pneumoniae M426 did not precipitate
some metals e.g. nickel, cobalt, zinc and arsenite but showed a high efficiency of
precipitation of cadmium, lead and palladium from solution with about 99% metal
removal efficiency within 1h. EDX microanalysis showed that the metal precipitates
consisted of metal plus carbon, sulfur and oxygen, while GC-Mass Spectrum analysis
showed the presence of dimethyldisulfide (DMDS) in the culture off-gas. We attributed
the precipitation to the high affinity between DMDS and some metal ions to produce
metallo-thiol complexes. The formation of metal sulfides seems unlikely since mercury
and cadmium deposits were white instead of black and yellow respectively and the
characteristic X-ray powder diffraction spectra of metal sulfides could not be obtained
(data not shown). There are some reports of the use of organosulfur compounds [2,3Dimercaptopropanol, meso-2,3-Dimercaptosuccinic acid (DMSA), Sodium 2,3dimercapto-1-propanesulfonate (DMPS)] as therapeutic chelating agents for the removal
of some heavy metals (arsenic, cadmium, lead and mercury) from living cells. The
formation of complexes with metal ions can be attributed to the presence of sulfhydryl
group in these chelators, which show a great affinity for metal ions producing the metal
1124
complexes [28-35]. According to our knowledge this is the first study using a naturally
produced organo-volatile sulfur gases as a chelating agent to precipitate heavy metals with
such high metal removal efficiency under aerobic conditions. Current studies are aiming to
quantify DMDS production and its contribution to metal precipitation.
4.
CONCLUSIONS
There have been attempts to engineer aerobic sulfate or thiosulfate reducing systems
for metal removal to produce more sulfide [36, 37]. This study describes a novel natural
process not only for mercury bioremediation but also for some other heavy metals e.g.
cadmium, lead and palladium via aerobic bio-precipitation of metal ions as a metal-thiol
complex. This mechanism provides an excellent approach for bioremediation of some
heavy metal and has some advantages. Firstly, since culture off-gas is used, it is applicable
for concentrated or non-physiologically compatible solutions under a wide range of
environmental factors such as pH and salinity. Secondly, with this bioprocess one
organism, K. pneumoniae, was used to precipitate different metals from their solutions and
could be used with mixed metal solution. Thirdly, and an important advantage, during
metal bio-mineralization there is no direct contact between the bacterial biomass and the
metal precipitate; thus keeping the biomass away from the original contaminated waste
water, ensuring that there is no cross-contamination between the biomass and the final
metal precipitate.
REFERENCES
1. G. M. Gadd, Accumulation of metals by microorganisms and algae. In: Biotechnology,

Ed. H. J. Rehm. VCHV, Weinheim, 1988.
2. G. M. Gadd, Microbial control of heavy metal pollution. In: Microbial Control of
Pollution, Eds J. C. Fry, G. M. Gadd, R. A. Herbert, C. W. Jones and I. Watson-Craik.
Cambridge University Press, Cambridge, 1992.
3. G. M. Gadd, FEMS Microbiol. Rev. 11 (1993) 297.
4. M. D.Mullen, D. C. Wolf, F. G. Ferris, T. J. Beveridge, C. A. Flemming and G. W.
Bailey, Appl. Environ. Microbiol. 55 (1989) 3143.
5. S. G.Walker, C. A. Flemming, F. G. Ferris, T. J. Beveridge and G. W. Bailey, Appl.
Microbiol. Biotechnol. 55 (1989) 2967.
6. T. J. Beveridge, Rev. Microbiol. 43 (1989) 147.
7. M. M. Valls and V. de Lorenzo, FEMS Microbiol. Rev. 26 (2002) 327.
8. G. M. Gadd and C. White, Trends Biotechnol. 11 (1993) 353.
9. R. M. Austrain, Ph.D Thesis, Institut fr Thermische Verfahrenstechnik und
Umwelttechnik, Germany, 1994.
10. R. E. Kallio and A. D. Larson, Methionine degradation by a species of Pseudomonas.
In W. D. McElroy and H. B. Glass (ed.), A symposium on amino acid metabolism.
Johns Hopkins University Press, Baltimore, 1955.
11. W. Segal and R. L. Starkey, J. Bacteriol. 98 (1969) 908.
12. J. Ruiz-Herrera, and R. L. Starkey, J. Bacteriol. 99 (1969) 544.
13. W. Kreis and C. Hession, Cancer Res. 33 (1973) 1862.
14. A. Miller, R. A. Scanlan, J. S. Lee and L. M. Libbey, Appl. Microbiol. 26 (1971)18.
15. S. Ito, T. Nakamura and Y. Eguchi, J. Biochem. 79 (1976) 1263
16. N. J. Hayward, T. H. Jeavons, A. J. C. Nicholson and A. G. Thornton, J. Clin.
Microbiol. 6 (1977) 187.
17. M. Pohl, E. Bock, M. Rinken, M. Avvdin and W. A. Konig. Z, Naturforsch. Teil C. 39
(1984) 240.
1125
18. B. Tomita, N. Hmamura and Y. Ose. Eisei Kagaku, 27 (1981) 98.

19. W. J. M. Engels, A. C. Alting, M. M. T. G. Arntz, H. Gruppen, A. G. J.Voragen, G.
Smit and S. Visser, Int. Dairy J. 10 (2000) 443.
20. J. Sambrook, E. F. Fritsch and T. Maniatis, Molecular Cloning: a Laboratory Manual,
2nd ed., vol. 3. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y. 1989.
21. M. M. Matlock, B. S. Howerton and D. A. Atwood. J. Hazardous Materials 84 (2001)
73.
22. J. S. Chang, J. Hong, Q. A. Ogunseitan and B. H. Olson, Biotechnol. Prog. 9 (1993)
526.
23. M. Mestres, O. Busto and J. Guasch, J. Chromatography 945 (2002) 211.
24. A. M. M. Essa, L. E. Macaskie and N. L. Brown, Biochem. Soc. Trans. 4 (2002) 672.
25. H. Aiking, H. Govers and J. 't Riet, Appl. Environ. Microbiol. 50 (1985) 1262.
26. C. L. Wang, P. C. Michels, S. C. Dawson, S. Kitisakkul, J. A. Bross, J. D. Keasling
and D. S. Clark, Appl. Environ. Microbiol. 63 (1997) 4075.
27. C. N. Mulligan, R. N. Yong and B. F. Gibbs, J. Hazardous Materials, 85 (2001) 111.
28. H. V. Aposhian and M. M. Aposhian, Ann. Rev. Pharmacol. Toxicol. 30 (1990) 279.
29. H. V. Aposhian, D. C. Bruce, W. Alter, R. C. Dart, K. M. Hurbult and M. M.
Aposhian, FASEB J. 6 (1992) 2472.
30. M. M. Aposhian, R. M. Maiorino, Z. Xu and H. V. Aposhian, Toxicol. 109 (1996) 49.
31. P. Asiedu, T. Moulton, C. B. Blum, E. Roldan, N. J. Lolacono and J. H. Graziano,
Environ. Health Perspect. 103 (1995) 734.
32. J. Chisolm, J. Clin. Toxicol. 30 (1992) 493.
33. J. H. Graziano, N. J. Lolacono, T. Moulton, M. E. Mitchell, V. Slavkovich and C.
Zarate, J. Pediatr. 120 (1992) 133.
34. M. M. Jones, Crit. Rev. Toxicol. 21 (1991) 209.
35. R. A. Peters, L. A. Stocken and R. H. S. Thomson, Nature. 156 (1945) 616.
36. S. W. Bang, D. S. Clark and J. D. Keasling, Appl. Environ. Microbiol. 66 (2000) 3939.
37. C. L.Wang, P. D. Maratukulam, A. M. Lum, D. S. Clark and J. D. Keasling, Appl.
Environ. Microbiol. 66 (2000) 4497.
1126

Influence of pH, Mg2+ and Mn2+ on the bioleaching of nickel

laterite ore using the fungus Aspergillus niger O5
O. Cotoa, D. Gutierreza, L. Abna J. Marreroa and K. Boseckerb
a
Departamento de Microbiologa. Facultad de Biologa. Universidad de la Habana.

Calle 25 entre J e I. Plaza. Vedado. Ciudad Habana. Cuba
b
Federal Institute for Geosciences and Natural Resources (BGR)
Stilleweg 2, D 30655 Hannover, Germany
Abstract
The main global nickel deposits are laterite deposits. Their nickel contents vary from
0.3-1.5% and are associated with a non-soluble phase that cannot be processed completely
by conventional methods. The Cuban nickel deposits are recognized as one of the largest
deposits in the world. Besides nickel they contain other valuable metals, such as cobalt
and manganese, which in spite of low concentrations of valuable metals, are of important
commercial interest.
The aim of this investigation was to study the influence of pH and Mg2+ and Mn2+ ion
concentrations on the dissolution of Ni, Co, Mn, and Mg from serpentinite in the Punta
Gorda deposit in Moa (Cuba). For leaching, Aspergillus niger O5 was chosen because of
its ability to dissolve a high percentage of Co, Ni and Mn from the mineral substrate and
for its production of citric acid. The strain produced higher concentration of citric acid
(25-27 g/L) when grown in the presence of the ore than in the absence of ore (14 g/L). The
production of organic acid by the fungus varied with the conditions of incubation. The
initial pH of the medium was a very important variable in this process; its change during
the course of the time affected the bioleaching efficiency and gave rise to biosorption
processes.
Keywords: bioleaching, heterotrophic microorganisms, Aspergillus niger, Penicillium

simplicissimum, citric acid, organic acids, laterite ore, oxide ore
1.
INTRODUCTION
Cuba has an estimated world reserve of Ni of 37% and ranks first in the world
reserves of Ni and second in world reserves of Co. The main nickel deposits in Cuba are
on the east side of the island in the Moa-Nicaro area of Holgun province. In this area
there are three commercial plants where Ni+Co oxides and Ni+Co sulfides are obtained by
hydrometallurgy and pyrometallurgy, respectively. Both technologies require high
temperature, special equipment and constitute a high environmental risk. For this reason,
bioleaching with organic acid produced by fungi (1-9) might be an attractive alternative
for the Cuban nickel industry in order to develop a sustainable technology.
The number of scientific papers and patents on bioleaching of sulfide ores is quite
large and exceeds several times the number of papers related to bioleaching of non-sulfide
1127
ores. Bioleaching of non-sulfide ores is based on the ability of fungi to produce organic
acids such as citric, oxalic, and fumaric acid, which act as leaching agents (19). They are
excreted into the environment and dissolve heavy metals by direct displacement of metal
ions from the ore matrix by hydrogen ions and by the formation of soluble metal
complexes and chelates (2-3). Fungi most active in leaching are Penicillium
simplicissimum and Aspergillus niger. Studies of nickel leaching have revealed that the
best recovery has been obtained with citric acid (2, 3, 6, 8).
The main objective of this paper was to study the influence of pH and Mg2+ and Mn2+
ion concentrations on the dissolution of Ni, Co, Mn and Mg present in the serpentinite
from the Punta Gorda deposit at Moa, Cuba.
2.
2.1 Laterite ore

Nickeliferous ore from the Punta Gorda, Moa (Cuba) deposit was used in all
experiments.
Table 1. Chemical and mineralogical composition of Cuban laterite ore
Concentration (%)
Phases
%
Ni
1.54
Al2O3
1.64
Co
0.27
Fe2O3
17.53
Fe
14.8
MnO
0.253
Element
Mg
19.86
MgO
29.56
Mn
0.49
SiO2
35.87
Al
1.95
Cr
1.02
2.2 Screening of microorganisms for bioleaching

Sixteen strains of fungi were used in this work, 15 of them identified as Penicillium
simplicissimum (strains 2, 3, 14, 16, 22, 23, 25, 26, 42, 43, 47, 50, 56, 60, 61). They were
obtained from the Federal Institute for Geosciences and Natural Resources, Hannover,
Germany. The other fungus was Aspergillus niger strain O5 (6). All strains were grown on
solid Saboraud medium and incubated for one week at 30C. Culture medium 1a (100 mL)
was inoculated with 2 mL containing 107-108 spores/mL in a 500 mL Erlenmeyer flask.
Culture medium 1a was prepared as follows: 140 g/L sucrose; 0.3 g/L yeast extract; 0.5
g/L KH2PO4; 0.250 g/L MgSO4 and 0.125 g/L MnSO4. The pH was adjusted to 6.5. The
spores were counted using a Thoma counting chamber.
All leaching experiments were carried out at 5% pulp density of laterite ore. The ore
was milled to a mean particle size of 0.250 mm. The culture medium was incubated at 100
rpm at 30C. Each time a sample was taken from the solution the same volume was
replaced with sterile culture medium. One Erlenmeyer flask without inoculum was used as
a control.
2.3 Culture conditions for bioleaching
To study the effect of oxygen on bioleaching, one Erlenmeyer flask was placed on a
shaker for the duration of the experiment, a second one was not shaken.
The influence of the culture medium was investigated using seven variants of the
growth medium. Culture medium 3a was used only for the Ni leaching experiment.
1128
Table 2. Variants of culture media
Mg2+
Mn2+
pH
Culture media (with 140 g/L sucrose and 0.3 g/L yeast extract)
1
1a
2
2a
3
3a
3b
+
+
+
+
+
+
+
+
6.5
6.5
5.5
3.0
5.5
5.0
3.0
2.4 Bioleaching in two-stages

In the first step, culture medium 3a (100 mL in a 500 mL Erlenmeyer flask) was
inoculated with A. niger O5 and incubated at 150 rpm/min for 8 days at 30C. In the
second step, the biomass was removed by filtration; the filtrate was autoclaved and used to
leach Ni ore with a 5% pulp density.
At regular intervals, samples were withdrawn for pH measurement and element
analysis by ICP-AES. The type and concentration of the organic acids produced were
determined using a densitometer after separation by thin layer chromatography (TLC).
The spots became apparent under ultraviolet light (366 nm) and after staining at visible
light.
The mineralogical composition of the Cuban laterite ore and some solid residues from
bioleaching was determined by X-ray diffraction (XRD). This was done using a Philips
X'pert system PW1730 equipped with a Cu tube, primary and secondary Soller slits and a
secondary monochromator. The main phases are chrysotile, magnesioferrite and goethite.
3.
3.1 Screening for microbial leaching

In a preliminary experiment, we tested the ability of 16 strains of fungi to leach nickel
and other metals from Cuban laterite ore. Most of of the strains were able to dissolve Ni,
Co, Mn, Al, Fe, and Cr to some extent, whereas strains 42, 56, 60 and 61 dissolved less
than 10% (data not shown).
The amount of metal leached by four strains of the 16 strains studied is shown in
Figure 1 as a function of time. The trend is rather similar with all four strains.
In the majority of strains, the initial pH of 6.2 decreased, indicating production of
organic acid by the fungus, and the most of the change in pH occurred within the first
seven days.
We took Ni solubilization as the main criterion for screening programs, because Ni is
the metal of the most economic importance in Cuban laterite ore. Table 3 shows the
percentage of nickel recovered after one and two weeks of bioleaching, indicating that the
strains can be separated into three groups: the first group with the strain 23 and O5,
recovered up to 60% Ni within 14 days. Strains 2, 3, 14, 22, 25, 26 and 50 form the second
group, dissolving more than 20% and less than 50% Ni. The fungi of the third group,
including strains 16, 42, 43, 47, 56, 60 and 61, extracted less than 15% Ni. The best strain
was A. niger O5.
1129
Figure 1. Laterite bioleaching in medium 1a: A) strain 50, B) strain 3, C) strain 2 and
D) strain 42. Pulp density: 5%; carbon source: sucrose 140 g/L; temperature: 30C;
reciprocal agitation: 100 rpm.min-1. (: Mn, : Mg, : pH, : Co, : Ni, : Cr, : Al
and : Fe)
Table 3. Solubilization of Ni laterite ore by different strains
Fungal strain
P. simplicissimum
2
3
14
16
22
23
25
26
42
43
47
50
56
60
61
A. niger O5
1130
% recovery after
7 days
14 days
0.27
14.73
12.64
9.56
15.61
30.15
16.10
21.51
4.93
6.92
11.42
17.52
1.60
10.60
9.36
34.15
34.30
25.89
21.02
14.91
24.42
56.23
24.92
35.43
6.57
11.18
13.72
31.71
2.09
12.35
10.01
57.8
The selection of the best strain is a very important factor in the bioleaching of Ni
laterite ore and the leaching efficiency depends on the amount of acid produced by the
fungus (1, 2, 6, 8).
3.2 Effect of incubation conditions on bioleaching of metals
The results of bioleaching and change of pH with strain O5 under shaking and static
conditions are illustrated in Figure 2. Under both conditions the pH decreased to 3.8
within 7 days and remained below 4.0 thereafter.
Manganese was leached best and within 10 days about 70% Mn was dissolved.
However, the pattern of bioleaching changed in the course of time depending on the
incubation conditions. Aspergillus niger O5 was the only strain that showed loss of Ni,
Mn and Co from dissolution when cultivated under shaking conditions (Figures 1 and 2).
This was observed after 14 days of experiment (Figure 2A) and was due to the uptake of
Ni, Co, and Mn by Aspergillus niger, which is known for nickel accumulation in the
biomass (2, 10). Some abiotic factors, such as pH, allow the biosorption of heavy metals
by fungi (10); precipitation of metals in the presence of oxalic acid (1) or electrosorption
(11) may occur as well. Besides manganese and nickel, up to 40% of cobalt and
magnesium was removed from the laterite ore (Figures 1 and 2). Previous mineralogical
studies of this Cuban laterite ore revealed that nickel and cobalt are associated with
goethite and manganese mineral phases, respectively.
In general, there are two phases in the bioleaching of nickel laterite ore, similar to the
bioleaching of chalcopyrite (12). The first phase is characterized by an increasingly high
rate of metal dissolution, due to a strong decrease in the pH, which reflects a high
metabolic activity of the microorganisms (7). In the second phase, the rate of metal
dissolution decreases or remains almost constant.
Figure 2. Laterite bioleaching with A. niger O5 in medium 1a. A) shaking conditions

(reciprocal agitation: 100 rpm.min-1) and B) static conditions. Pulp density: 5%,
carbon source: sucrose 140g/L, temperature: 30C. (: Mn, : Mg, : pH, : Co, :
Ni, : Cr, : Al and : Fe)
3.3 Influence of pH and Mg2+ and Mn2+ ion concentrations on the bioleaching
Considering that some cations, such as Mn2+, Mg2+, Fe2+, and Fe3+, can affect cell
morphology, enzymatic activity and citric acid production by Aspergillus, they are
essential micronutrients in the metabolism. We used some variations of the culture
1131
medium to study the effect of these cations on the release of Ni from Cuban laterite and on
the type and concentration of organic acid produced by the fungus.
The composition of the culture medium considerably influenced metal removal as
well as type and amount of organic acids produced by strain O5 (Figures 3 and 4).
The highest amount of Ni was leached using medium 3a. After 18 days of incubation,
60% Ni was dissolved, and 86.6% Ni was extracted at the end of experiment (Figure 3C).
Medium 3b was the worst for A. niger O5. A different behaviour was observed with
Penicillium simplicissimum strain 2 (data not shown).
Regardless of the initial pH in the growth medium, all pH values were around 5 after
7 days of bioleaching, after 14 days pH values of 3.8-4.8 were measured. The pH
behaviour in medium 3a was different from that of the other media. The concentration of
dissolved Mn, Ni and Co increased when media 1a, 2, 2a and 3 were used. Later, a loss of
dissolved metals was observed (Figure 3).
A)
B)
Ni Dissolved (%)
Mn Dissolved (%)
120
100
80
60
40
20
100
80
60
40
20
0
0
0
10
15
20
Time (days)
25
30
10
20
30
Time (days)
D)
30
25
20
15
10
5
0
0
10
20
Mg Dissolved (%)
Co Dissolved (%)
C)
30
20
15
10
5
0
0
Time (days)
10
20
Time (days)
30
E)
8
pH
6
4
2
0
0
10
20
Time (days)
30
Figure 3. Influence of initial pH, Mg and Mn on bioleaching of Cuban laterite ore by

A. niger O5 (A, B, C and D) and change of pH (E). Pulp density: 5%, carbon source:
140g/L sucrose, temperature: 30C, reciprocal agitation: 100 rpm.min-1. Media with
Mg and Mn: : 1a, : 3, : 3a, : 3b and media without Mg and Mn: : 2, : 2a, x: 1
1132
Oxalic acid (g/L)
60
50
40
30
20
10
Citric acid (g/L)
As shown in Figure 4, the initial pH of the media and the concentrations of Mn2+ and
Mg2+ cations influenced the amount and type of organic acid produced by fungal activity.
When the strain was cultivated in medium 3a (pH 5.0), the yield of citric acid was
considerably enhanced and oxalic acid production was inhibited. This is of considerable
advantage, since oxalic acid inhibits Ni laterite bioleaching (1). This strain produced a
higher concentration of citric acid (25-27 g/L) when grown in the presence of the ore than
in the absence of ore (14 g/L) (data not shown).
30
25
20
15
10
5
0
2a
3b
3
1
1a
2
3a
Figure 4. Influence of different culture media on oxalic acid (A) and citric acid (B)
production by A. niger strain O5 to10 days of Cuban laterite bioleaching. Pulp
density: 5%, carbon source: sucrose 140 g/L, temperature: 30C, reciprocal
agitation: 100 rpm.min-1. Medium: 1, 2, and 2a (without Mg and Mn) and 1a, 3, 3a
and 3b (with Mg and Mn)
Some released metal cations inhibit citric acid synthesis by A. niger O5. Mg2+ is a
cofactor of the enzyme isocitrate dehydrogenase, and some concentrations of Mn2+ act in a
negative way on the citric acid cycle, and citric acid is the best leaching agent for Cuban
laterite nickel ore (6). Therefore, we need to define the optimum concentration of metal
ions for maximum yield of citric acid.
The X-ray diffraction analysis of some residue samples from this set of experiment
demonstrated a new phase in the solid residue after bioleaching. This was identified as
magnesium oxalate hydrate (Figure 5).
Figure 5. X-ray diffraction: A) Cuban laterite ore before bioleaching and B) solid
residues from bioleaching with A. niger O5 in medium 3 (pH 5.5)
1133
3.4 Two-stage bioleaching

The leaching agent produced in the first step was used for the leaching of the ore in
the second step. The results of the leaching of laterite ore are shown in Figure 6. The only
organic acid metabolite accumulated in the filtrate was citric acid (data not shown). The
percentage of Ni recovered using this method (Figure 6A) was similar to that obtained
using a single stage method (Figure 3B). pH during the leaching phase is shown in Figure
6B as a function of time. Note that the increase in pH with time (Figure 6B). This behavior
is due to the release of 30% Mg, which is a neutralizing agent. In the single-stage
bioleaching method, the pH decreased with time (Figure 3E) as consequence of metabolite
activity of the fungus.
Figure 6. (A) Chemical leaching of Cuban laterite ore for 12 days by metabolic
products of A. niger O5 cultivated in medium 3a (pH 5) and (B) pH without ore. Pulp
density: 5%, temperature: 30C and reciprocal agitation: 100 rpm.min-1
4.
CONCLUSIONS
Cuban laterite ore was more amenable to bioleaching with Aspergillus niger O5 than
with Penicillium simplicissimum.
A. niger is capable of producing oxalic and citric acid or only citric acid depending on
the culture conditions. The best leaching agent for Cuban laterite nickel ore was citric
acid. The initial pH of culture medium and the composition of growth medium is an
important variable for the bioleaching of non-sulfide ore. Cuban sugar cane can be a
substitute for sucrose in the culture medium.
There was a high solubilization of nickel (89% after 28 days of bioleaching) in
medium 3a with an initial pH of 5.0. The single-stage and two-stage bioleaching methods
yielded similar results. On the basis of these results, we recommend that a preliminary
engineering and economic study be made.
REFERENCES
1. K. Alibhai, D. Leak, A.W.L. Dudeney, S. Agatzini and P. Tzeferis, in: R.W. Smith and
M. Misra (eds.), Mineral Bioprocessing, The Minerals, Metals and Materials Society,
TMS (1991) 191.
2. K. Bosecker, in: R.L. Lawrence, R.M.R. Branion and H.H Ebner (eds.), Fundamental
and Applied Biohydrometallurgy, Elsevier, Amsterdam (1986) 367.
3. K. Bosecker, in: J. Salley, R.G.L. McCready, and P.L. Wichlaz (eds.),
Biohydrometallurgy, CANMET SP89-10 (1989) 15.
1134
4. W. Burgstaller and F. Schinner, in: A.E. Torma, J.E. Wey and V.I. Lakshmanan (eds.),
Biohydrometallurgical Technologies, Vol. 1, The Minerals, Metals & Materials
Society, TMS (1993) 325.
5. W. Burgstaller and F. Schinner, J. Biotechnol. 26 (1993), 340.
6. O. Coto, N. Bruguera, L. Abn, J. Gamboa and Y. Gmez, in: V.S.T Ciminelli and O.
Garca Jr. (eds.), Biohydrometallurgy: Fundamentals, Technology and Sustainable
Development, Part A, Elsevier, Amsterdam (2001) 175.
7. L. Abin, O. Coto, B. Marrero and J. Marrero, Ciencias Biolgicas CENIC No 3 (2002)
In press.
8. P. Tzeferis, Int. Journal of Mineral Processing, 42 (1994) 267.
9. L.B. Sukla and V.V. Panchanadikar, in: A.E. Torma, J.E. Wey and V.I. Lakshmanan
(eds.), Biohydrometallurgical Technologies, Vol. 1, The Minerals, Metals & Materials
Society, TMS (1993) 373.
10. B. Volesky, in: R. Amils and A. Ballester (eds.), Biohydrometallurgy and the
Environment towards the Mining of the 21st Century, Part B, Elsevier, Amsterdam
(1999) 161.
11. M. Valix, J.Y. Tang and W.H. Cheung, Minerals Engineering., 14 (12) (2001) 1629.
12. O. Coto, A. Ballester, M.L. Blzquez and F. Gonzlez, Biorecovery, 2 (3) (1993) 121.
1135

Mercury tolerance of thermophilic Bacillus sp. and

Ureibacillus sp.
K.J. Glendinning and N.L. Brown
Edgbaston, Birmingham, B15 2TT, U.K.
Abstract
Although resistance of microorganisms to Hg(II) salts has been widely investigated
and resistant strains have been reported from many eubacterial genera, there are few
reports of mercuric ion resistance in extremophilic microorganisms. Thermophilic
mercury resistant bacteria were selected by growth at 62C on Luria agar containing
HgCl2. Sequence analysis of 16S rRNA genes of two isolates showed the closest matches
to be with Bacillus pallidus and Ureibacillus thermosphaericus. Minimum inhibitory
concentration (MIC) values for HgCl2 were found to be 30 g/ml and 80g/ml for these
isolates respectively, compared to 10 g/ml for B. pallidus H12 DSM3670, a mercurysensitive control. The best-characterised mercury-resistant Bacillus strain, B. cereus
RC607, had an MIC of 60 g/ml.
Growth of the thermophilic isolates in media containing HgCl2 resulted in the
formation of a black precipitate. X-Ray Diffraction (XRD) analysis of this precipitate
showed it to be HgS. Growth in the presence of radioactive 203HgCl2 at 45C or 62C
showed progressive removal of mercury from the medium, but the mechanism by which
this occurs is uncertain. Biochemical assays for mercuric reductase, which reduces 203Hg2+
to volatile 203Hg, did not detect this enzyme. It is possible that volatilisation of mercury is
occurring as a by-product of normal metabolism, through complexation into volatile
compounds.
Keywords: Ureibacillus sp., Bacillus sp., thermophilic, mercury resistance

1.
INTRODUCTION
Compounds of mercury, such as mercuric chloride and organomercurials, have long
been known to be toxic to both eukaryotic and prokaryotic cells. These compounds can
pass through biological membranes [1] and bind with high affinity to thiol (SH) groups in
proteins [2], thus causing damage to membranes and inactivating enzymes. Mercury is
also genotoxic; inorganic Hg(II) is capable of strong reversible interactions with the
nitrogens in purines and pyrimidines, and organic mercury compounds, e.g.
methylmercury, also produce irreversible damage to nucleic acids [3]. Environmental
contamination with mercury compounds can have devastating effects, as mercury toxicity
Corresponding author. Current address: Birmingham Women's Health Care, Metchley Park Road,
Birmingham B15 2TG, UK. Email: kj_glendinning@yahoo.co.uk; Fax: +44-121-627-2711.
1137
is cumulative, with the highest levels of mercury compounds being found in consumers at
the top of the food chain.
A number of microorganisms have evolved resistance mechanisms to deal with
mercury compounds. Mercury resistance was first reported in Staphylococcus aureus [4]
and since then has been described in a number of bacterial species.
One of the best-defined mercury resistance determinants is the mer operon encoded
by transposon Tn501, found in Gram-negative bacteria. The functions of the minimal
number of proteins required to confer full resistance are as follows [5]:
MerR is the mer regulatory protein which controls the expression of all the other
proteins in the operon in response to the presence of Hg(II) salts. MerP is a periplasmic
protein which binds mercuric ions entering the periplasm through the porins OmpC and
OmpF, and transfers Hg(II) to the MerT transport protein located in the cytoplasmic
membrane. The mercuric ions are then passed from cysteine residues in MerT which are
exposed to the periplasmic compartment to cysteine residues on the cytoplasmic face of
the membrane, and thence to the cytoplasmic mercuric reductase (MR, the merA gene
product). Hg(II) is complexed with the active site cysteine residues of MR and reduced
from Hg(II) to Hg(0) using NADPH as the reductant. Hg(0) is then lost from the cell in
the gas phase.
Mercuric ion resistance has also been characterised in Gram-positive genera. The
resistance determinants from Bacillus sp. RC607 [6], Streptomyces lividans [7] and
Staphylococcus aureus [8] have been characterised. The mechanisms of induction,
mercury transport and mercury reduction in Gram-positive genera are similar to those of
the Gram-negative systems, but the transport functions differ, as might be expected from
the differences in the surfaces of Gram-positive and Gram-negative cells [5].
Although resistance to Hg(II) has been widely investigated, there are few reports of
mercuric ion resistance in extremophilic microorganisms [9, 10]. Given the wide spread of
mercuric ion resistance/tolerance throughout the bacterial kingdom, it seems likely that
many extremophilic microorganisms will encode mercury resistance. This paper describes
the investigation of such resistances.
Here we report the isolation of thermophilic Bacillus sp. [11] and Ureibacillus sp.
[12] and characterisation of their observed mercuric ion resistance.
2.
2.1 Isolation and identification of organisms from compost

Suspensions of garden compost were plated onto LB agar plates supplemented with 0,
15 and 20 g/ml HgCl2, and the plates were incubated at 62C for 48 hours. Colonies were
picked from the LB agar plates containing mercury and repeatedly subcultured on LB agar
containing 15 g/ml HgCl2 to obtain two axenic cultures. Approximately 700bp of 16S
rDNA gene was amplified from each organism using primers pA [13] and 530r [14]. The
PCR products obtained were sequenced and subjected to BLAST analysis.
2.2 HgCl2 resistance testing
Minimum Inhibitory Concentration (MIC) assays for HgCl2 were performed on
Ureibacillus sp. and Bacillus sp. The controls used were a mercuric ion sensitive strain,
Bacillus pallidus H12 DSM3670, and a known mer operon-containing mercury resistant
Bacillus cereus RC6067 [6]. MIC assays were performed on LB agar plates containing
1138
HgCl2 in 10 g/ml increments from 0g/ml to 100g/ml. Plates were incubated at 37C
and 62C.
2.3 Identification of culture precipitate
It was noted that when Bacillus sp. and Ureibacillus sp. were grown in culture with
HgCl2 a black precipitate was formed. Bacillus sp. cells were grown with and without
HgCl2 followed by transmission electron microscopy (TEM) to localise the precipitate. In
order to test whether the precipitate could be HgS formed by the reaction between Hg(II)
and microbially produced H2S [15], [16], H2S production was determined. The positive
control used was Proteus vulgaris, a known aerobic H2S producer, with Escherichia coli
as the negative control. P. vulgaris, E. coli and B. cereus RC607 were grown overnight at
37C in LB broth, and B. pallidus H12 DSM 3670, Ureibacillus sp. and Bacillus sp. were
grown overnight at 62C in LB broth. Overnight cultures were diluted 1:100 into sterile
glass universal bottles containing 5ml volumes of fresh LB broth with 0g/ml and 10
g/ml HgCl2 (only used for mercury resistant strains). Lead acetate strips (BDH) were
folded over the neck of the bottles (care was taken not to wet the strips) and the caps were
screwed on. These were then incubated at 37C or 62C (as above) and examined after
24h and 48h.
In order to identify the precipitate, Bacillus sp. cells were grown in the presence of 50
g/ml HgCl2 for 48 hours. Cells and precipitate were harvested by centrifugation and
separated by careful washing with sterile distilled water. The precipitate was dried and
examined by X-ray diffraction (XRD) analysis.
2.4 Mercury volatilisation assays
Volatilisation assays were performed essentially by the method of [17] with the
exception that LB broth was used instead of M9 medium and induction was performed
with 2.5 g/ml HgCl2 for mercury resistant isolates and 0.25 g/ml for sensitive isolates.
Cultures were grown to OD600 = 0.4-0.5, then preinduced with HgCl2 for one hour.
Cultures were adjusted to precisely OD600 = 0.4 followed by washing and resuspension of
cells in fresh LB broth. Three 1ml aliquots of the resuspended cells were dispensed into
separate 50ml conical flasks. Three control flasks each containing 1ml uninoculated LB
broth were prepared. Flasks were placed in a shaking waterbath set to the desired
temperature and to each flask was added 4ml pre-warmed LB broth supplemented with
1.25M HgCl2 and 1:2000 diluted 203HgCl2 (at a concentration of 2.55mg/ml Hg;
Amersham Pharmacia plc). Timing commenced upon the removal of the first sample. 100
l samples were taken from each flask at 0, 20, 40, 60, 120, 180 and 240 minutes. These
samples were immediately added to 5ml volumes of scintillation cocktail (OptiPhase
"HiSafe" 2) and counted in a Packard TR1-CARB 2700 TR liquid scintillation analyser
with the window set at 0.0-220 MeV.
2.4.1 Mercuric reductase assays
Mercuric reductase assays were performed by the method of Lund and Brown [18].
Cells were grown overnight, then inoculated 1:25 into fresh LB broth without
selection. Cells were grown at the appropriate temperature (45C or 62C) with shaking to
OD600 = 0.5-0.6, induced with 10 g/ml HgCl2 for one hour, then a further 10 g/ml HgCl2
for a further hour (mercury sensitive B. pallidus H12 DSM 3670 cells were induced with 1
g/ml HgCl2 for one hour followed by a further 1 g/ml HgCl2 for a further hour). Cells
were harvested by centrifugation, then washed in 20 ml wash buffer (30mM Tris-HCl
1139
pH7.6, 20% sucrose, 1mM DTT, 1mM EDTA). The cell suspension was re-centrifuged
and the supernatant discarded. Cell pellets were washed twice with 20ml resuspension
buffer (100mM sodium phosphate pH7.0, 1mM -mercaptoethanol, 0.5mM EDTA,
0.05mM PMSF, 0.1mM benzamidine) followed by resuspension in 5ml fresh buffer. The
cells were lysed by sonication and cell debris was pelleted by centifugation. Supernatants
were retained and the pellets discarded. The supernatant was clarified by two further
rounds of centrifugation in order to remove all traces of cellular debris. Lysates were
stored at 0C overnight. The following day, 7ml aliquots of assay buffer (0.1M sodium
phosphate buffer pH 7.0, 1mM -mercaptoethanol, 0.5mM EDTA, 0.1mM NADPH) were
dispensed into sterile glass universal bottles and stored on ice until use. For the assay, 70l
of cell extract was added to 7ml assay buffer, mixed well and the mixture was preincubated at the desired temperature for 5 minutes. 1ml was removed and an A340 reading
was taken. This was immediately followed by addition of 60l of 10mM HgCl2 stock,
which gave a final concentration in the assay mix of 100M HgCl2; 60l sterile distilled
water was added to the control assay mix. Samples were mixed well and incubated at the
desired temperature for 1 minute followed by an A340 reading. Further readings were taken
at 3 and 5 minutes after addition of HgCl2/sterile distilled water.
3.
3.1 Identification of microorganisms

The 16S rDNA PCR products obtained from the two cultures were sequenced and
BLAST analysis of the sequences obtained revealed 97% homology with Bacillus pallidus
H12 DSM3670 [11] and >99% homology with Ureibacillus thermosphaericus [12].
3.2 HgCl2 resistance
Table 1. HgCl2 MIC values (g/ml) for Bacillus sp. and Ureibacillus sp.
Organism
HgCl2MIC (mg/ml) at 37C
HgCl2MIC (mg/ml) at 62C
Bacillus cereus RC607

Bacillus pallidus DSM 3670
Bacillus sp.
Ureibacillus sp.
60
10
60
30
10
80
30
No data are shown for Bacillus cereus at 62C as this temperature was outside the
growth range for this organism.
The results obtained from MIC testing showed that mercury resistance of the
thermophilic isolates was not temperature dependent. The growth of Bacillus sp. was
reduced on normal LB agar at 37C; this temperature is near to the bottom end of the
organism's growth range, which may help explain why mercury resistance was reduced,
but not abolished.
3.3 TEM of Bacillus sp.
The results of Transmission Electron Microscopy (TEM) (Figure 1) show a clear
difference in the appearance of bacterial cultures grown with and without mercury. There
is a large amount of black precipitate visible when Hg(II) is present, which was thought to
be HgS. The appearance of the cells in the TEM suggests that the precipitate may be
forming within the cells, possibly causing cells to rupture.
1140
Figure 1(a). TEM (10,000 X magnification) of Bacillus sp. grown with 50g/ml HgCl2
Figure 1(b): TEM (10,000X magnification) of Bacillus sp. grown without HgCl2
3.4 Identification of the precipitate

XRD analysis was performed upon a sample of dried crystalline precipitate collected
from a culture of Bacillus sp. grown with HgCl2. The thick black lines on the XRD profile
(Figure 2) show the peaks obtained for the experimental sample, and the fainter, vertical
lines show the peaks obtained from a database for an ultra-pure reference sample of HgS. There was sufficiently good correlation between the two sets of information to make
a positive identification of the experimental sample as -HgS. The profile was very similar
to one obtained for HgS precipitated by Desulfovibrio desulfuricans [19].
1141
Figure 2. X-Ray Dispersion Analysis of the precipitate from Bacillus sp. grown in the
presence of Hg(II)
3.5 Mercury volatilisation assays
The results of mercury volatilisation experiments (Figure 3) showed that Ureibacillus
sp. removed Hg(II) from the media more efficiently than Bacillus sp. at 45C than at
62C. This may reflect the growth temperature optima for these organisms. From 16S
rRNA data, the most closely related species to Bacillus sp. was Bacillus pallidus, which is
reported as having a maximum growth temperature of 70C with a temperature optimum
of 60-65C. The closest species to Ureibacillus sp. was Ureibacillus thermosphaericus
which has a maximum growth temperature of 64C.
Mercury Volatilisation at 45oC
Mercury Volatilisation at 62oC
100
100
%Hg remaining
120
%Hg remaining
120
80
60
40
B.cereus RC607
80
Bacillus sp.
Ureibacillus sp.
60
Bacillus pallidus
H12 DSM 3670
Luria broth
40
20
20
0
0
50
100
150
Time (min)
200
50
100
150
Time (min)
200
Figure 3: Results of volatilisation assays on Bacillus spp. and Ureibacillus sp. at 62C
and 45C. Results are shown as the means of triplicate readings with error bars
shown as sample standard deviations (n-1)
This suggests that Ureibacillus sp. probably has a lower optimum growth temperature
than Bacillus sp., hence Ureibacillus sp. may remove 203Hg(II) more efficiently at 45C.
At 45C, 203Hg(II) was being removed from the media at a high rate by B. cereus RC607.
1142
This rate was reduced at 62C, probably due to the fact that 62C was outside the growth
range of this organism. There appeared to be no appreciable difference in the rates of
removal in LB broth alone and in the presence of B. pallidus H12 DSM3670 at 45C.
There did appear to be some removal of 203Hg(II) from the media by B. pallidus H12
DSM3670 at 62C, the reason for which is not known.
3.6 Mercuric reductase assays
The results of mercuric reductase assays are shown in Figure 4. Mercuric reductase
activity was observed in B. cereus RC607 at 37, 45 and 62C. There appeared to be no
specific NADPH-dependent mercuric reductase activity in B. pallidus H12 DSM3670,
Ureibacillus sp. or Bacillus sp. This result was surprising in view of the mercuric ion
resistance and results from volatilisation assays, which showed Ureibacillus sp. and
Bacillus sp. as capable of removing 203Hg(II) from the medium. An alternative, non-mer
mechanism of mercury detoxification and volatilisation may operate within Ureibacillus
sp. and Bacillus sp., possibly as a by-product of normal metabolism.
90
o
37 C with HgCl2
nmoles NADPH oxidised/min/mg of protein
80
45 C with HgCl2
o
62 C with HgCl2
70
37 C without HgCl2
o
45 C without HgCl2
60
62 C without HgCl2
50
40
30
20
10
0
B.cereus
RC607
Bacillus
sp.
Ureibacillus
sp.
B.pallidus
H12
DSM3670
Figure 4. Mercuric reductase activity of Bacillus spp. and Ureibacillus sp. at 37C,
45C and 62C. Errors bars show one standard deviation (n-1) of triplicate readings
3.7 Testing for H2S production
The precipitation of HgS by microorganisms is often due to a reaction of Hg(II) with
H2S gas. The strains in this study were tested for H2S production as described in section
2.3. Growth occurred in all bottles; however, blackening of the lead acetate strip only
occurred in the bottle containing P. vulgaris, indicating that only this isolate produced
detectable amounts of H2S. It seems likely, therefore, that any production of HgS by
Ureibacillus sp. and Bacillus sp. proceeded due to complexation of Hg(II) with cellular
sulphides and not with H2S.
1143
4.
CONCLUSIONS
Mercuric ion tolerant isolates of thermophilic Bacillus sp. and Ureibacillus sp. were
isolated from compost. Mercury tolerance in these isolates was apparently not due to
classical mer operon-mediated mercury reduction, but by some other mechanism. Mercury
tolerance was not temperature dependent, as both isolates showed elevated tolerance to
mercuric ions at 37C and 62C.
When considering features of mercuric ion resistance in these isolates it is important
to make the distinction between mercury resistance and mercury tolerance. Mercuric ion
resistance can be thought of as a genetically encoded detoxification mechanism, which is
specifically induced in response to mercurials. Mercury tolerance can be thought of as a
detoxification mechanism, which is a by-product of normal metabolism and is not
specifically induced [19]. From the results obtained it seems possible that Ureibacillus sp.
and Bacillus sp. are exhibiting mercury tolerance mechanisms rather than a specific mer
operon-encoded mercury resistance. The evidence to support this is the lack of detectable
NADPH-dependent mercuric reductase activity, and precipitation of HgS in the medium.
Several organisms have been reported as being mercury tolerant due to HgS precipitation
[15, 16, 20, 21]. This mechanism has not been shown to be specifically induced in
response to mercurials, suggesting that it is indeed a by-product of normal metabolism.
HgS can be formed by various mechanisms in microbial mercury cycling. The growth
of Desulfovibrio desulfuricans under sulphate-reducing conditions resulted in the
precipitation of HgS when the medium was spiked with Hg(II) [15, 19]. Hg(II) reacts
directly with H2S produced by D. desulfuricans [15] and Clostridium cochlearium [16] to
form HgS. This is thought to provide a means of mercury tolerance in these organisms.
The formation of HgS from monomethylmercury (MMHg) by D. desulfuricans may occur
via the reaction of MMHg with H2S to produce unstable dimethylmercurysulphide
(DMHgS), which breaks down to HgS and volatile dimethylmercury (DMHg) [15].
Precipitation of HgS by aerobic microorganisms has also been suggested as a tolerance
mechanism. Klebsiella aerogenes NCTC418 was thought to form HgS when grown in
continuous (aerobic) culture with the addition of 2ppm (equivalent to 2g/ml) HgCl2. No
bacterial mercury volatilisation could be detected, and it was claimed mercuric ion
sensitivity was increased under sulphate-limited conditions [20]. K. aerogenes is unable to
produce H2S gas, therefore any precipitation of HgS presumably occurred via an
alternative mechanism.
There are few reports of precipitation of HgS by bacterial cultures, which makes it of
unknown importance in microbial mercury detoxification. At least part of the mercury
tolerance shown by Ureibacillus sp. and Bacillus sp. may be due to precipitation of HgS.
This probably occurs by reaction of Hg(II) with cellular sulphides, as H2S was not
produced by these microorganisms when grown with or without HgCl2.
The results of mercuric reductase assays showed that no NADPH-dependent mercuric
reductase activity was present in Ureibacillus sp. and Bacillus sp., therefore mercury
removal was not due to reduction of Hg(II) to Hg(0) by mercuric reductase, unless it
uniquely uses another cofacter. It is possible that Ureibacillus sp. and Bacillus sp. may
produce high levels of non-specific reductases which are capable of reducing Hg(II) to
Hg(0). Alternatively, there may be a different, as yet unknown, mechanism for mercury
removal operating in these organisms.
A possible explanation for removal of mercury from the media by Ureibacillus sp.
and Bacillus sp. may be the methylation of mercury to either MMHg or DMHg. MMHg
1144
and DMHg are volatile, and could be lost from the medium in a similar fashion to Hg(0)
The formation of DMHg by microorganisms may represent a mercury detoxification
method due to the volatility of the compound [19]. Methylation of mercury by anaerobic
microorganisms has been reported [22]. At present there are no reliable reports of mercury
methylation by aerobic cultures.
The mechanisms of mercury tolerance in Ureibacillus sp. and Bacillus sp. are not
fully understood. Whilst precipitation of HgS in culture is one tolerance mechanism, the
means by which mercury volatilisation occurs in the absence of specific mercuric
reductase is unclear.
ACKNOWLEDGEMENTS
This work was supported by a BBSRC CASE studentship to KJG. Thanks are due to
Dr Jon L. Hobman and Dr Ching Leang for discussion and advice in setting up assay
systems.
REFERENCES
1.
2.
3.
4.
5.
6.
Gutknecht, J. J. Memb. Biol. (1981). 61:61.

Leach, S.J. J. Aust. Chem. (1960). 13:520.
Sletten, E. and Nerdal, W. Metal Ions In Biological Systems. (1997). 34:479.
Moore, B. Lancet. (1960). II:453.
Hobman, J.L. and Brown, N.L. Metal Ions In Biological Systems. (1997). 34:527.
Wang, Y., Moore, M.,Levinson, H.S., Silver, S., Walsh, C. and Mahler, I. J. Bacteriol.
(1989). 171:83.
7. Sedlmeier, R. and Altenbuchner, J. Mol. Gen. Genet. (1992). 236:76.
8. Laddaga, R.A., Chu, L., Misra, T.K. and Silver, S. Proc. Natl. Acad. Sci. (1987).
84:5106.
9. Olson, G.J., Porter, F.D., Rubinstein, J. and Silver, S. J. Bacteriol. (1982). 151:1230.
10. Bogdanova, E.S., Mindlin, S.Z., Kalyaeva, E.S. and Nikiforov, V.G. FEBS Lett.
(1988). 234:280.
11. Scholz, T., Demharter, W., Hensell, R. and Kandler, O. Syst. Appl. Microbiol. (1987).
9:91.
12. Fortina, M.G., Rdiger, P., Schumann, P., Mora, D., Parini, C., Manachini, L.P. and
Stackebrandt, E. Int. J. Syst. Evol. Microbiol. (2001). 51:447
13. Edwards, U., Rogall, T., Blocker, H., Emde, M. and Bottger, E.C. Nucleic Acids Res.
(1989). 17:7843.
14. Lane, D.J., Pace, B., Olsen, G.J., Stahl, D.A., Sogin, M.L. and Pace, N.R. Proc.
Natl.Acad. Sci. (1985). 82:6955.
15. Baldi, F., Pepi, M. and Filippelli, M. Appl. Environ. Microbiol. (1993). 59:2479.
16. Pan-Hou, H. and Imura, N. Arch. Microbiol. (1981). 129:49.
17. Leang, C., PhD thesis (1999), University of Birmingham: Birmingham, UK.
18. Lund, P.A. and Brown, N.L. Gene. (1987). 52:207.
19. Baldi, F. Metal Ions In Biological Systems. (1997). 34:213.
20. Aiking, H., Govers, H. and van 't Riet, J. Appl. Environ. Microbiol. (1985). 50:1262.
21. Belliveau, B.H., Starodub, M.E. and Trevors, J.T. Can. J. Microbiol. (1991). 37:513.
22. Choi, S.C. and Bartha, R. Appl. Environ. Microbiol. (1993). 59:290.
1145

Reduction of Pd(II) with Desulfovibrio fructosovorans,

its [Fe]-only hydrogenase negative mutant and the activity of
the obtained hybrid bioinorganic catalysts
I.P. Mikheenkoa, V.S. Baxter-Planta, M. Roussetb, S. Dementinb,
G. Adryanczyk-Perrierb and L.E. Macaskiea
a
School of Biosciences, the University of Birmingham,

Edgbaston, Birmingham, B15 2TT, England, United Kingdom
b
Bioenrgtique et Ingnierie des Protines, CNRS,
31 Chemin Cedex 20, Joseph Aiguier, 13402 Marseille France
Abstract
A novel biocatalyst was obtained via reduction of Pd(II) to Pd(0), at the expense of
hydrogen, onto the cell surface of Desulfovibrio fructosovorans and a mutant of D.
fructosovorans with deactivated [Fe]-only hydrogenase. The ability of the palladiumcoated biomass to reduce chromium (Cr+6 to Cr+3), to release hydrogen from sodium
hypophosphite and reductively dehalogenate chlorophenol and polychlorinated biphenyl
congeners was demonstrated. Dried, palladised cells of D. fructosovorans wild type and its
mutant were more effective catalysts than Pd(II) reduced chemically under hydrogen or
commercially available sub-micron size Pd(0) powder. Differences were observed in the
catalytic activity of the wild type and the mutant strain of D. fructosovorans when
compared with each other. Negligible chloride release occurred from the chlorophenol and
polychlorinated biphenyl species using biomass alone. The structure of the bioinorganic
catalyst was investigated using transmission electron microscopy.
1.
INTRODUCTION
Palladium is one of the worlds most expensive metals. Catalyst systems based on
palladium are widely used in chemical manufacturing and processing (1). The highest
palladium consumer is the auto motive industry, where this metal, together with other
platinum group metals, is used in automobile catalytic converters to reduce the toxicity of
exhaust gases. Palladium is also extensively used in the electronic industry. Since
palladium is a highly valuable metal with only limited world resources (2), developing
new methods of Pd recovery and reprocessing of scrap is necessary since existing hydroand pyrometallurgical routes are energy demanding and/or not environmentally friendly.
In the work of Lloyd et al. (3) it was shown that palladium could be effectively
recovered from solution by resting cells of the sulphate-reducing bacterium Desulfovibrio
desulfuricans at ambient temperature. It was concluded that Pd(II) was reduced by the
activity of hydrogenase and possibly cytochrome c3.
Yong et al. (4) demonstrated that palladised D. desulfuricans biomass could be used
as a catalyst without any further processing. It has also been reported (5, 6) that in the
1147
process of biorecovery metallic Pd with unusual magnetic properties could be obtained:

ferromagnetic Pd nanocrystals were formed on the biomass. Subsequent studies showed
(7) that biologically reduced palladium (Bio-Pd) can be used as a catalyst for partial
dehalogenation of polychlorinated biphenyls, which are very stable recalcitrant
xenobiotics which are not generally susceptible to aerobic or anaerobic microbial
metabolization.
The catalytic properties of biomineralized palladium (Bio-Pd) depend on the size and
the distribution pattern of metal deposits on the cell surface. To obtain a Bio-Pd catalyst
with certain catalytic properties it is important to understand the process of Pd(II)
reduction and Pd(0) grain formation which take place within the periplasm of sulphate
reducing bacteria (5). According to Lloyd et al. (3) hydrogenase is one of the important
components of the Pd(II) reducing system. Sulphate-reducers have more than one
hydrogenase system, with Desulfovibrio sp. possessing up to four different hydrogenases,
which differ by structure (enzymes with Fe, Fe-Ni or Fe-Ni(Se) in their active centres) and
localisation within the cell (periplasmic and cytoplasmic, membrane bound and nonbound) (8). Several hydrogenase mutants of Desulfovibrio sp. have been constructed
lacking different hydrogenases (9, 10). The aim of this work is to investigate the catalytic
properties of Bio-Pd obtained via Pd(II) reduction by the sulphate reducer D.
fructosovorans and its [Fe]-only hydrogenase mutant. The hydrogenase and cytochrome
complex of the Desulfovibrio sp. has been intensively investigated (8, 11, 12 ).
2.
EXPERIMENTAL
2.1 Preparation of cells

D. fructosovorans and its mutant (10) were maintained anaerobically (13). Growth
medium was inoculated with 10% of culture and cells were incubated at 37 C for three
days before harvesting. The medium for D.fructosovorans culture was as in (13).
For Bio-Pd production the cells were harvested by centrifugation at 6000 rpm for 15
min at 4C, washed three times in degassed 20 mM MOPS buffer at pH 7.0. The cells
were stored refrigerated.
2.2 Preparation of Bio-Pd
Aliquots of the washed cells were resuspended as a concentrated suspension (diluted
to 4-6 mg dry wt/ml for use) in MOPS buffer and transferred under N2 to 100 ml of Pd(II)
(2 mM sodium tetrachloropalladate(II) Na2PdCl4 (Aldrich Chemical Company) in 0.01 M
HNO3, pH 2.3) solution in 100 ml serum bottles sealed with butyl rubber stoppers and preequilibrated with N2. The suspension was allowed to stand for 1 h at 30C to form
nucleation sites on the biomass. The electron donor used was H2 gas, which was bubbled
through the solution for 20 min. After the black precipitate had formed, the solution was
centrifuged at 6000 rpm for 10 min and the precipitate was washed three times with
distilled water, dried in air and then washed with acetone. Chemical palladium was
prepared identically except that no biomass was added and reduction took approximately
60 min under H2 in the absence of cells.
2.3 Pd(II) and Cr(VI) assay
The kinetics of Pd(II) reduction were studied by recording the A420 of the residual
Na2PdCl4 solution, withdrawn (1.5 ml) from the reaction mixture each 20 min, (Unicam
5600 UV-VIS spectrophotometer: Cambridge, UK). The maximum height of the specific
1148
absorption peak of [PdCl4]2- at 420 nm depends linearly on the palladium salt

concentration. The concentration of Pd(II) in solution was also determined
spectrophotometrically with Sn(II) (14). For the Pd(II) assay 200 l of sample was added
to 800 l of SnCl2 solution. Readings were taken after 60 min at a wavelength of 463 nm.
The residual concentration of Cr(VI) (where Pd(0) was used as a catalyst for the
reduction of Cr(VI) (below)) was measured using the diphenylcarbazide method (5). Test
solution (100 l) was added to 1.4 ml of the diphenylcarbazide solution, mixed well, left
to stabilise for 5 min and read at 540 nm.
2.4 Examination of Pd-loaded cells by electron microscopy
Pellets of Pd-loaded bacteria were harvested using a Heraeus Sepatech Biofuge 13
microfuge (13000 rpm, 5 min) and fixed in 2.5% (wt/vol) glutaraldehyde in 0.1 M Nacacodylate buffer, pH 6.8. After centrifugation the pellet was resuspended in 1.5 ml of 0.1
M Na-cacodylate buffer, pH 6.8. The preparation was stained in 1% osmium tetroxide in
0.1 M phosphate buffer, pH 7.0, dehydrated and embedded in epoxy resin. Sections (100 150 nm thick) were cut from the resin block and placed onto a copper grid prior to
analyses. Sections were viewed with a JEOL 120CX2 transmission electron microscope.
2.5 Catalytic activity measurement
Three test reactions were used to characterise the catalytic activity of the Bio-Pd:
hydrogen release from sodium hypophosphite, Cr(VI) reduction and reductive
dehalogenation (RD) of polychlorinated aromatic compounds.
2.5.1 Hydrogen release from sodium hypophosphite
The reaction was initiated via the addition of 4 mg of Bio-Pd (Pd:dry wt biomass 1:3)
to 20 ml of a 2% solution of hypophosphite (NaH2PO2) at 30C. The volume of liberated
hydrogen was recorded at 5 min intervals starting 20 mins after the initiation of the
reaction (reaction onset time). The reaction rate was calculated as the rate of H2 release
from 5 replicate experiments. Chemical palladium (4 mg) was used as a control test
reaction.
2.5.2 Cr(VI) reduction
Tests were performed under oxygen free nitrogen (OFN) in 12 ml sealed serum
bottles. Bio-Pd (2 mg) was added to 5 ml of 7 mM Na2CrO4 in 20 mM MOPS-NaOH
buffer at pH 7.0. The solution was equilibrated at 30C for 10 min and the reaction was
initiated by addition of sodium formate (electron donor) to a final concentration of 25
mM. Samples were taken from the bottle via a rubber septum and centrifuged for 5 min at
13000 rpm to remove Pd. The supernatant was analysed for residual Cr (VI). The catalytic
activity was expressed as the percent of reduced Cr(VI). Chemical palladium (2 mg) was
used as a control test reaction.
2.5.3 Reductive dehalogenation of chlorinated aromatic compounds
The Bio-Pd (2 mg, 1:3 Pd:dry wt biomass) was placed in 12 ml serum bottles and the
chlorinated aromatic compound added to the required concentration in a carrier of 20 mM
MOPS NaOH buffer pH 7.0. Experiments were initiated via the addition of formate (10
mM sodium formate, pH 7.3). The final volume was 10ml. Dehalogenation of the
chlorinated aromatic substance was estimated in periodically withdrawn samples as
release of chloride ion assayed spectrophotometrically by the mercury (II) thiocyanate
method (15) versus sodium chloride as standard. For these tests the washing procedures
1149
were carried out during preparation of the Bio-Pd with omission of Cl- from all processing
steps to ensure a low background level of Cl-, against which the Cl- release was measured.
The limit of sensitivity was in the range 0.5-100 g/ml Cl-.
2.5.4 Chlorinated aromatic compounds
Chlorophenols were used as aqueous solutions diluted to the required concentration
(usually 5 mM) in the media. Polychlorinated biphenyls (PCBs) were dissolved in hexane
and diluted to the required nominal concentrations (80 g/ml as chloride; shaken to mix)
as hexane suspensions in the media.
3.

The bioreduction of Pd(II) by D. fructosovorans and its [Fe]-only hydrogenase mutant
was compared. Solutions (2 mM) of sodium tetrachloropalladate (Na2PdCl4) in 0.01 M
HNO3 or tetraamminepalladium chloride ([Pd(NH3)4]Cl2) in 20 mM MOPS buffer, pH 7.0
were poured into 10 ml sealed vials and degassed under OFN. The resting cell suspensions
of each strain were added to the Pd(II) solution (3 parts dry weight of biomass to 1 one
part of Pd(II) and left for biosorption at 30C. After 1 h samples of supernatant were
analysed for remaining Pd(II). This concentration of Pd(II) was regarded as the time-zero
concentration.
Following the one hour of biosorption the Pd(II) bioreduction was initiated via the
addition of electron donor (20 mM sodium formate solution, pH 7.0). The reaction was
carried out anaerobically at 30C. The residual Pd concentration in the supernatant is
shown in Fig. 1.
Figure 1. Reduction of Pd(II) by D. fructosovorans, wild type and [Fe]-only

hydrogenase negative mutant. Mass ratio Pd(II) dry biomass is was 1:3. Solution: 2
mM Na2PdCl4, in 0.1 M HNO3. Electron donor 0.25 mM sodium formate
The rate of Pd(II) reduction was monitored via the decreasing concentration of the
coloured [PdCl4]-2 complex ion (A420), and confirmed using SnCl2. Both methods gave
identical results. The reduction of Pd by sodium formate alone was used as a control. Dead
cells (killed by autoclaving at 121C for 15 mins) did not reduce Pd (not shown). Fig 1
shows that both wild type and mutant reduce Pd(II) at approximately the same rate hence
the absence of [Fe]-only hydrogenase did not significantly effect the Pd(II) reduction
process. It should be noted that [Fe]-only periplasmic hydrogenase is less abundant
1150
mM reduced Cr(VI)/mg Pd.h
mM reduced Cr(VI)/mg Pd.h
compared to the [Fe-Ni] enzyme, which constitutes a larger part of hydrogenase activity in
the periplasm (8).
The catalytic activity of Bio-Pd samples obtained using D. fructosovorans and its
mutant with respect to Cr(VI) reduction was substantially higher when compared to
palladium powder of 39 m particles size (Fig. 2). The Bio-Pd obtained under acidic
conditions (pH 2.3) using Na2PdCl4 salt (Fig.2a) proved to be a better catalyst when
compared to the catalyst obtained at pH 7.0 using [Pd(NH3)4]Cl2 salt (Fig.2b).
Interestingly, Bio-Pd obtained with [Fe]-only hydrogenase negative mutant was
significantly more catalytically active when compared to Bio-Pd obtained from the wild
type regardless of which palladium salt or pH was used in the preparation of Bio-Pd (Fig.
2).
2,5
2,0
1,5
1,0
0,5
0,0
0,6
0,5
0,4
0,3
0,2
0,1
0,0
1 - Control: commercial Pd powder ; Bio-Pd: 2 - wild type; 3 - [Fe]-only hydrogenase negative mutant.
Figure 2. The reduction of Cr(VI) to Cr(III) using D. fructosovorans and its [Fe]-only
hydrogenase negative mutant Bio-Pd produced with Na2PdCl4 salt, pH 2.3 (a); and
[Pd(NH3)4]Cl2 salt, pH 7.0 (b)
The Bio-Pd was also a more active catalyst than its chemical counterpart in the release
of hydrogen from sodium hypophosphite (Fig. 3). The rate of hydrogen release was higher
in the presence of mutant-based Bio-Pd catalyst obtained with Na2PdCl4 salt (made at pH
2.3) (Fig. 3a) when compared to wild type-based Bio-Pd. However, in the case of Bio-Pd
obtained using [Pd(NH3)4]Cl2 salt made at pH 7.0 (Fig. 3b) the difference between the two
types of biocatalyst was not significant.
0,14
0,12
ml H2/mg Pd.min
.
ml H2/mg Pd min
0,30
0,25
0,20
0,15
0,10
0,05
0,10
0,08
0,06
0,04
0,02
0,00
0,00
1 - Control: commercial Pd powder ; Bio-Pd: 2 - wild type; 3 - [Fe]-only hydrogenase negative mutant.
Figure 3. The release of hydrogen from sodium hypophosphite using D.

fructosovorans and its [Fe]-only hydrogenae negative mutants produced Bio-Pd with
Na2PdCl4 salt, pH 2.3 (a); and [Pd(NH3)4]Cl2 salt, pH 7.0 (b) bars. Bars indicate
standard error
1151
The reductive dehalogenation of chlorinated aromatic compounds was investigated.

Chloride release from 4-chlorobiphenyl, 2,3,4,5-tetrachlorobiphenyl and 2,2,4,4,6,6hexachlorobiphenyl using Desulfovibrio fructosovorans and its hydrogenase negative
mutant Bio-Pd produced with Na2PdCl4 salt (pH 2.3) was tested. Throughout, no Clliberation was promoted by non-palladised bacteria, chemically prepared Pd(0) or
commercially available finely divided Pd(0) (not shown) although the Bio-Pd promoted
Cl- release from all the PCBs (Figs. 4a-c). However only small differences were seen
between the parent and the mutant Bio-Pds in each case.
16
a
Chloride, g/ml
Chloride, g/ml
16
12
8
monoPCB wild type
monoPCB [Fe]-negative
hydrogenase mutant
12
8
TetraBCB wild type
TetraPCB [Fe]-negative
hydrogenase mutant
Time, hours
20
16
12
8
HexaPCB wild type
HexaPCB [Fe]-negative
hydrogenase mutant
4
1
Time, hours
Chloride, g/ml
16
Chloride, g/ml
Time, hours
12
8
2-chlorophenol wild type
2-chlorophenol [Fe]-negative
hydrogenase mutant
Time, hours
Figure 4. Reductive dehalogenation of 4-chlorobiphenyl (a) and 2,3,4,5tetrachlorobiphenyl (b) of 2,2,4,4,6,6-hexachlorobiphenyl (c) and 2-chlorophenol
(d) using D.fructosovorans and its hydrogenase negative mutants Bio-Pd
The rate of chloride release from 5 mM 2-chlorophenol was also studied under similar
conditions. In this case the rate of Cl- release for the [Fe]-only hydrogenase mutant was
lower than for the wild type (Fig. 4d). The controls included resting cells of the bacteria
without bound Pd(0), chemically prepared Pd(0) and commercially available finely
divided Pd(0). In confirmation of a previous study the chemically prepared Pd(0) and
commercially available finely divided Pd(0) gave a similar rate of Cl- release to D.
fructosovorans, whereas no Cl- liberation was promoted by the non-palladised bacteria
(not shown).
Thus, the release of Cl- from 2-chlorophenol was achieved using both Bio-Pd
preparations, with no advantage over "chemical" Pd(0). However, the reductive
dehalogenation of polychlorinated biphenyls was achieved using the Bio-Pd obtained from
D. fructosovorans and its hydrogenase negative mutant under conditions where the
chemically produced Pd counterpart is largely ineffective, but no advantage was
demonstrated using the mutant strain.
The bioreduction tests and studies using X-ray diffraction analysis (not shown)
showed that D. fructosovorans and its [Fe]-only hydrogenase negative mutant reduced
1152
palladium to its metallic state on their surface. Electron microscopic investigation of BioPd in this study showed that for D. fructosovorans and its mutant the initial reduction of
Pd takes place in the same way as that previously described for in D. desulfuricans (5). It
begins within the periplasmic space (the site of localisation of hydrogenases and
cytochromes) and then the growing crystals protrude through the outer cell membrane and
form larger crystals on the cell surface (Fig. 5). The pattern of Pd crystal deposition was
very similar in both the wild type and mutant stain. The catalytic activity of Bio-Pd
obtained using the [Fe]-only hydrogenase negative mutant was higher when compared to
the Bio-Pd obtained using wild type strain in the reduction of Cr(VI) and in the other tests
the difference varied according to the Pd(II) salt used and pH of its preparation (hydrogen
release tests) or chlorinated aromatic substrate (RD tests). We assume that the reason for
differences in catalytic activity is the size and, consequently, the surface area of palladium
crystals formed in the process of Pd(II) reduction. It may be possible that the lack of one
of the hydrogenases leads to formation of a larger quantity of Pd(0) nanoclusters via the
activity of other redox-enzymes within the periplasm. These nanoparticles are not visible
on the electron micrographs but the presence within the preparation can be detected
through their enhanced catalytic activity. The presence of such Pd nanoclusters was shown
in the Bio-Pd preparations of D. desulfuricans via the measurement of the magnetic
moment of Bio-Pd powder in a varying external magnetic field (5, 6); magnetic
measurements to determine the Pd-cluster size on the wild type and mutant D.
fructosovorans are in progress.
Figure 5. Electron micrographs of Pd deposition on the surface of D. fructosovorans

(a) and its mutant (b) cell surface. Scale bars indicate 500 nm
4
CONCLUSIONS
The catalytic activity of Bio-Pd obtained under different conditions using D.
fructosovorans and its hydrogenase negative mutant vary in absolute values although
shares a general trend. It was significantly higher in all tests (except RD of chlorophenol)
when compared to chemical palladium. The [Fe]-only hydrogenase negative mutant
showed better catalytic activity compared to the wild type in most reactions. This provides
evidence that, firstly, hydrogenase is not the only enzyme involved in Pd(II) reduction,
and, secondly, it shows that the mutation has an additional effect, which causes variations
in Bio-Pd catalytic activity.
AKNOWLEDGEMENTS
We acknowledge the financial support of the EU BIO-CAT (Grant number GRD12001-40424) and the BBSRC (Grant number E13817) and EPSRC (Grant
NoGR/NO8445).
1153
REFERENCES
1.
2.
3.
4.
Platinum 2002 Interim Review. Johnson Matthey. November (2002) 28.

P. Crowson. Minerals Handbook. Macmillan Ltd., Surrey, 1982.
J.R. Lloyd, P. Yong, L.E. Macaskie. Appl. Environ. Microbiol., 64 (1998) 4608.
P.Yong, N.A. Rowson, J.P. Farr, I.R. Harris, L.E. Macaskie. Biotechnol. Bioeng., 80
(2002) 369.
5. I. Mikheenko, P. Mikheenko, C.N.W. Darlington, C.M. Muirhead, L.E. Macaskie. In
Biohydrometallurgy: Fundamentals, Technology and Sustainable Development Eds
V.S.T. Ciminelli & O Garcia, Elsevier (2001) 525.
6. I.P. Mikheenko, L.E. Macaskie, P.M. Mikheenko, C.M. Muirhead. Pd. 19th General
Conference of the EPS Condensed Matter Division held jointly with CMMP 2002
Condensed Matter and Materials Physics. Brighton. Europhysics Conference
Abstracts. 26A (2002) 93.
7. V.S Baxter-Plant, I.P. Mikheenko, L.E. Macaskie, Biodegradation (2002) in press.
8. R.Cammack, M. Frey, R. Robson (eds.), Hydrogen as a Fuel. Learning from Nature.
Taylor & Francis. London, New York, 2001.
9. L. Casalot, De Luca G, Dermoun Z, Rousset M, de Philip P. J. Bacteriol.,184 (2002)
853.
10. B.K. Pohorelic, J.K. Voordouw, E. Lojou, A. Dolla, J. Harder, G. Voordouw. J.
Bacteriol., 184 (2002) 679.
11. C. Aubert, M. Brugna, A. Dolla, M. Bruschi, M-T. Giudici-Orticoni. Biochim.
Biophys. Acta - Protein Structure & Molecular Enzymology, 1476 (2000) 85.
12. C. Wawer, G Muyzer. Appl. Environ. Microbiol., 61 (1995) 2203.
13. M. Rousset, L. Casalot, B. J. Rapp-Giles, Z. Dermoun, P. de Philip, J. P. Belaich, J.
D.Wall. Plasmid, 39 (1998) 114.
14. G.C. Dasages. Absorptiometriques des elements mineraux. Ed. Masson, Paris, 1978.
15. G.H. Jeffrey, J. Bassett, J. Mendham, R.C. Denny. Vogels textbook of quantitative
chemical analysis, Fifth Edition, Bath Press, Avon, UK, 1989.
1154

Removal of cobalt, strontium and caesium from aqueous

solutions using native biofilm of Serratia sp. and biofilm
pre-coated with hydrogen uranyl phosphate
M. Paterson-Beedle and L.E. Macaskie
Edgbaston, Birmingham, B15 2TT, UK
Abstract
Heavy metals are removed by Citrobacter sp. (NCIMB 40259) now identified as a
Serratia sp. via the activity of an acid-type phosphatase enzyme, which liberates HPO42from a suitable organic phosphate donor with the stoichiometric precipitation of heavy
metal cations (M2+) as insoluble MHPO4 at the cell surface. Previous studies have shown
that it is possible to remove Ni2+ effectively via its intercalation into pre-formed or
growing crystals of hydrogen uranyl phosphate (HUP), with the "guest" metal species
intercalating within the HUP matrix. Also, the effective biomineralisation of HUP from
uranium mine waters has been demonstrated. Nuclear wastes contain not only uranium but
also fission products like 90Sr, 137Cs, and 60Co. We have shown that in the presence of
uranyl ion and glycerol 2-phosphate the deposited HUP is able to remove "cold"
surrogates of the fission products via intercalation using either continuous co-crystal
growth or by using cells pre-coated with HUP as inorganic ion-exchangers. Using Serratia
sp. biofilm immobilised onto polyurethane reticulated foam continuous removal of fission
products surrogates was obtained by intercalation techniques, but not using the
phosphatase-biomineralisation route alone, in the absence of the HUP host crystal.
1. INTRODUCTION
Fission of 235U yields radioactive fission products such as the isotopes 60Co, 137Cs and
90
Sr. Urgency is prompted by the nuclear industry to treat radionuclide-loaded liquid
wastes. Biological methods for removal of metals from nuclear wastes can succeed where
traditional physico-chemical treatments fail [1, 2]. A model for phosphatase-mediated
uranyl ion accumulation, with exocellular deposition of polycrystalline (i.e., multitude of
small crystals [3]) hydrogen uranyl phosphate (HUO2PO4.4H2O, HUP) that is identical to
the structure of HUP prepared by inorganic crystal growth has been well published [4].
"Chemical" HUP is an established ion-exchange material, with the intralamellar mobile
protons freely exchangeable for other ions like Na+, Ni2+ [5] and Co2+ [6, 7]. Previous
studies have demonstrated two approaches for the bioremediation of Ni2+ from dilute
aqueous flows using Serratia sp. cells [8-12]. The first approach was based on a two step
process: (i) enzymatically-mediated generation of a polycrystalline host lattice (priming
layer), i.e. HUP bound to the surface of the Serratia sp. cells and (ii) cation-exchange
intercalation of Ni2+ into the interlamellar spaces of HUP [8-12]. The second approach [8]
was to co-challenge immobilised Serratia cells in a packed-bed reactor with a Ni2+ and
1155
UO22+ solution in the presence of glycerol-2-phosphate (phosphate donor for phosphate

release and metal precipitation) giving a sustained removal of both metals. Since Co and
Ni are in the same group of the periodic table of the elements they should behave
similarly. Therefore, the purpose of this study was to evaluate the methods described
above using Serratia sp. biofilm immobilised onto polyurethane foam for the removal of
Co2+ from solution. Previous studies showed the uranyl phosphate precipitate to be, more
likely, NaUO2PO4.4H2O [13], therefore, Cs+ should be similarly removed as CsUO2PO4
and, in the presence of both fission products surrogates, bio-driven formation of mixed
metal uranyl phosphate is distinctly feasible. Previous studies showed removal of Sr2+ as
the phosphate precipitate [14] and removal of Sr2+ as a co-crystal was also explored.
2.
2.1 Support, organism and biofilm production

Citrobacter sp. (NCIMB 40259) now identified as Serratia sp. [15] was grown as
biofilm onto polyurethane reticulated foam Filtren TM30 (supplied by Recticel, Belgium)
in an air-lift fermenter [16] and the phosphatase specific activity of the cells from the
fermenter outflow was determined as described previously [16]; the steady state specific
activity was ~ 3000 units (nmol p-nitrophenol released from p-nitrophenyl phosphate min1
mg protein-1).
2.2 Preparation of packed-bed reactor systems for metal bioaccumulation
Cubes of polyurethane foam (88, 125 mm3) coated with Serratia sp. biofilm were
packed in a cylindrical glass column (length 9.0 cm and internal diameter 1.5 cm) of
working volume of ca. 13 ml. The total amount of protein immobilised onto the foam was
ca. 95 mg per reactor, measured by the bicinchoninic acid/CuSO4 method (Sigma protein
test kit TPRO-562) using bovine serum albumin as standard [17]. Considering that the
amount of protein is ca. 50% of the dry weight of biomass [unpublished], the total biomass
per reactor was, therefore, ca. 190 mg. The reactors were challenged with an upward flow
(ca. 10 ml h-1) via an external peristaltic pump (Watson-Marlow). All tests were done at
ambient temperature. Challenge solutions comprised sodium citrate buffer (2 mM), pH 6.0
and sodium glycerol-2-phosphate (5 mM) supplemented with Co(NO3)2.6H2O (1 mM),
CsCl (1 mM), Sr(NO3)2 (1 mM) or UO2(NO3)2.6H2O (1 mM) as specified in individual
experiments.
2.3 Metal biosorption by packed-bed reactors
Reactor systems were prepared, similar to those described in section 2.2, and
challenged with metal (1 mM) in the presence of citrate buffer (2 mM), pH 6.0, but in the
absence of glycerol-2-phosphate.
2.4 Co-crystallisation of metals by packed-bed reactors not previously primed with
HUP
Reactor systems were prepared, similar to those described in section 2.2, and
challenged with solutions comprising UO2(NO3)2.6H2O (1 mM), sodium citrate buffer (2
mM), pH 6.0 and sodium glycerol-2-phosphate (5 mM) supplemented with
Co(NO3)2.6H2O (1 mM), CsCl (1 mM) or Sr(NO3)2 (1 mM) as specified in individual
experiments.
1156
2.5 Use of reactor packed-bed systems previously primed with HUP

Reactor systems were "primed" (deposition of HUP onto the biomass) using a
solution of UO2(NO3)2.6H2O (1 mM), sodium glycerol-2-phosphate (5 mM) and sodium
citrate buffer (2 mM), pH 6.0. First step: four reactors were primed for 24.3 h at a flow
rate of 24 ml.h-1 giving a removal of ca. 65% of the input uranyl ion (total uranium loaded
was ca 139 mg). Second step: two reactors were challenged with a solution comprising
Co(NO3)2.6H2O (1 mM) in the presence of glycerol-2-phosphate (5 mM), pH 6.0 and the
other two were challenged with a solution comprising Co(NO3)2.6H2O (1 mM) in the
presence of glycerol-2-phosphate (5 mM) and sodium citrate buffer (2 mM), pH 6.0.
2.6 Spectrophotometric analysis of Co2+
The Co2+ contents of reactor outflows were estimated using the method described by
Onishi [18], with a slight modification. The sample or standard containing Co2+ (30 l, 110 g) was transferred to a test tube, citric acid solution (0.2 M, 250 l) and phosphateboric acid buffer (6.2 g of boric acid, 35.6 g of disodium phosphate dehydrated, and 500
ml of 1 M sodium hydroxide in a total volume of 1 l, 300 l) were added. The pH of the
solution should be close to 8.0. Nitroso-R salt (supplied by Fluka, UK) solution (0.2%,
125 l) was added while stirring. The test tubes containing the samples were covered,
transferred to a water bath at 100C and left for 1 min. Concentrated HNO3 (250 l) was
added and the samples left for a further 1 min. Samples were cooled in the dark and then
deionised water was added (325 l). The transmittance of the solution was measured at
420 nm.
2.7 Spectrophotometric analyses of Sr2+ and/or UO22+
A method was developed, similar to that described by Yong et al. [19], for the
simultaneous determination of uranium and strontium in mixed solutions using arsenazo
III. Michaylova and Kouleva [20] showed that the complex formation of strontium with
arsenazo III is maximum at pH 9-10. To each test tube containing 30 l of a target metal
(or metal mixture) was added 1.97 ml of one of the two solutions, as appropriate to the
metal under test (0.1 M HCl for uranium and 0.1 M borate buffer, pH 9.0 for strontium).
Metal was visualised by the addition of 0.1 ml of 0.15% (w/v) arsenazo III, with
estimation of the blue-violet complex at 652 nm and 649 nm for uranium and strontium,
respectively. The blue-violet complex develops in 25 min with strontium [20].
2.8 Spectrophotometric analysis of Cs+
The Cs+ contents of reactor outflows were estimated using the method described by
Huey and Hargis [21] with modifications. Sample or standard containing Cs+ (60 l, 1-50
g) was transferred to an Eppendorf tube (1.5 ml), perchloric acid solution (6 M, 84 l)
was added and diluted with deionised water (192 l). Phosphomolybdic acid (supplied by
Riedel-deHan) solution (14%, w/v, 60 l) was added, mixed and allowed to stand for 30
min. The suspension was centrifuged at 16,060 g for 30 min, the supernatant discarded
using a glass Pasteur pipette and the precipitate washed with perchloric acid (1.2 M, 300
l). The suspension was centrifuged at 16,060 g for 30 min, the supernatant discarded and
the precipitate was dissolved in borate buffer pH 9.0 (0.05 M, 360 l) during 30 min.
Borate buffer (840 l) was added to give a final volume of 1.2 ml. The absorbance was
measured at 226 nm.
1157
3.
3.1 Biosorption and bioaccumulation of Co2+ by Serratia sp. biofilm

The first experiment was designed to test the background adsorption/biosorption of
Co2+ from aqueous flows by polyurethane foam coated with Serratia sp. biofilm (Fig. 1).
Approximately 11% of the Co2+ was removed initially (up to 21 ml, or ca. 2 column
volumes), decreasing rapidly thereafter. Biosorption of Co2+ by the biomass was
negligible (less than 1% of the biomass dry weight). Bioaccumulation of Co2+ by Serratia
sp. in the presence of glycerol 2-phosphate was also negligible (Fig. 1) even though excess
phosphate was produced (not shown). These results are similar to those obtained using
immobilised Serratia sp. cells for the removal of Ni2+ [8, 11], and show that phosphatasemediated metal bioprecipitation is not appropriate for these metals.
Figure 1. Co2+ removal by reactors containing Serratia sp. biofilm immobilised onto
polyurethane foam and challenged with: , Co(NO3)2.6H2O (1 mM) in the presence
of citrate buffer (2 mM), pH 6.0 (biosorption); , Co(NO3)2.6H2O (1 mM) in the
presence of glycerol-2-phosphate (5 mM) and citrate buffer (2 mM), pH 6.0
(bioaccumulation)
3.2 Co-crystallisation of Co2+ and HUP
Previous experiments [8] showed that Serratia sp. cells, when challenged with a
solution of uranyl nitrate (0.5 mM) and nickel nitrate (0.5 mM) in the presence of
glycerol-2-phosphate and citrate buffer, accumulated both metals. The flow rate was set to
give a removal of UO22+ of 56.3% and the corresponding removal of Ni2+ from solution
was maintained at 27.3% (during 42 h, i.e. 25 column volumes). The proportion of metal
removed suggest the formation of Ni(UO2PO4)2.7H2O; the identity of the material
accumulated by the cells was confirmed by X-ray powder diffraction [8]. We tested cells
immobilised onto polyurethane foam challenged with a solution of UO2(NO3)2.6H2O (1
mM) and Co(NO3)2.6H2O (1 mM) in the presence of glycerol-2-phosphate with a flow rate
set at 10 ml.h-1. The removal of UO22+ was ca. 99% but the corresponding removal of Co2+
was less than 10% (Fig. 2).
3.3 Uptake of Co2+ onto/into HUP previously accumulated onto the biomass
In this study the reactors were primed with HUP as described in section 2.5 which
resulted in ca. 90 mg uranium deposited per reactor (ca. 47% of bacterial dry weight). The
initial challenge with Co2+ (Fig. 3) resulted in a slowly decreasing removal of Co2+ over
226 ml, with no sharp breakthrough after saturation of the HUP host. This pattern is
similar to that obtained using Serratia biofilm immobilised onto ceramic support, primed
1158
with HUP, and initially challenged with Ni2+ solution [11]. The theoretical capacity of a
reactor containing 90 mg uranium (in the form of HUP) would be 11 mg of cobalt at a
molar ratio of Co:U of 1:2, i.e. for the formation of Co(UO2PO4)2. A loss of Co-removing
capacity after 226 ml corresponds to ca. 10 mg of deposited cobalt which, therefore, is
attributable to column saturation. As expected the presence of citrate in the challenge
solution reduced the amount of Co2+ deposited (Fig. 3), since the metal would be
presented to the cells as the citrate complex; citrate was shown previously to remove Ni2+
from its position within the HUP host crystal. [12].
Figure 2. Metal removal by reactors containing Serratia sp. biofilm immobilised onto
polyurethane foam and co-challenged with a solution of Co(NO3)2.6H2O (1 mM) and
UO2(NO3)2.6H2O (1 mM) in the presence of glycerol-2-phosphate (5 mM) and citrate
buffer (2 mM), pH 6.0. , UO22+ removed. , Co2+ removed
Figure. 3. Co2+ removal by reactors containing Serratia sp. biofilm immobilised onto
polyurethane foam, primed with HUP and challenged with: , a solution of
Co(NO3)2.6H2O (1 mM) in the presence of glycerol-2-phosphate (5 mM), pH 6.0; , a
solution of Co(NO3)2.6H2O (1 mM) in the presence of glycerol-2-phosphate (5 mM)
and citrate buffer (2 mM), pH 6.0.
3.4 Co-crystallisation of Cs+ and HUP
Serratia sp. cells immobilised onto polyurethane foam were co-challenged with
UO2(NO3)2.6H2O (1 mM) and CsCl (1 mM) in the presence of glycerol-2-phosphate at a
flow rate of ca. 10 ml.h-1. The removal of UO22+ was ca. 96% and the corresponding
removal of Cs+ was ca. 58% (Fig. 4). The removal of Cs+ using the phosphatasebiomineralisation route alone, in the absence of the HUP host crystal, was less than 20%.
Thus it was concluded that although co-crystallization did not promote removal of Co2+
(above) this approach has potential for the removal of Cs+. It should be noted that since
glycerol 2-phosphate was provided as the sodium salt (5 mM) Na+ would be present to a
five-fold excess over Cs+ and in this respect it can be suggested that co-crystal formation
1159
favours formation of the CsUO2PO4 precipitate. However, it is not easy to distinguish

between HUO2PO4 and NaUO2PO4 by X-ray powder diffraction because they both have
very similar crystal lattice [13].
polyurethane foam and co-challenged with a solution of CsCl (1 mM) and
buffer (2 mM), pH 6.0. , UO22+ removed. , Cs+ removed.
3.5 Co-crystallisation of Sr2+ and HUP
Serratia sp. cells immobilised onto polyurethane foam were co-challenged with
UO2(NO3)2.6H2O (1 mM) and Sr(NO3)2 (1 mM) in the presence of glycerol-2-phosphate at
a flow rate of ca. 10 ml.h-1. The steady-state removal of UO22+ was more than 92% and the
corresponding removal of Sr2+ was in the range of 50-56% (Fig. 5). The removal of Sr2+
using the phosphatase-biomineralisation route alone, in the absence of the HUP host
crystal, was negligible. The molar ratio accumulated was in accordance with the
deposition of Sr(UO2PO4)2.
polyurethane foam and co-challenged with a solution of Sr(NO3)2 (1 mM) and
buffer (2 mM), pH 6.0. , UO22+ removed. , Sr2+ removed
4.
CONCLUSIONS
We have shown that using the continuous co-challenge system, i.e. in the presence of
uranyl ion and glycerol 2-phosphate, the deposited HUP is able to promote removal of ca.
58% of Cs+ and 50-56% of Sr2+ (during 53 column volumes) but was not able to remove
the Co2+ (less than 10%). However, using cells pre-coated with HUP as inorganic ionexchangers it was possible to remove the Co2+ until the column reached saturation
corresponding to a molar ratio of U:Co of ~2:1 in accordance with a crystal of
1160
Co(UO2PO4)2. Further work to be reported will confirm the identity of the crystals formed,
using X ray powder diffraction analysis and EXAFS. However, the preliminary data
reported here show that removal of Co2+, Sr2+ and Cs+ as, respectively, Co(UO2PO4)2,
Sr(UO2PO4)2 and CsUO2PO4 is feasible since no removal of the surrogate fission products
occurred in the absence of the uranyl ion. Such uranyl phosphate-driven co-deposited or
intercalation is entirely feasible for the treatment of real nuclear wastes since the fission
products residues often occur with an excess of residual uranium [1, 2] and, furthermore,
the metal precipitating phosphatase is highly radioresistent [22].
ACKNOWLEDGMENTS
This work is supported by the Biotechnology and Biological Sciences Research
Council. The authors thank Recticel (Belgium) for the polyurethane reticulated foam.
REFERENCES
1. L. E. Macaskie, CRC Crit. Rev. Biotechnol. 11 (1991) 41-112.

2. J. R. Lloyd and L. E. Macaskie, In: Environmental Microbe-Metal Interactions, (D. R.
Lovely, ed.), ASM Press, Washington, (2000) 277-327.
3. B. K. Vainshtein, Fundamentals of Crystals, Modern Crystallography, Vol. 1, 2nd edn.,
Springer, Berlin, 1994.
4. L. E. Macaskie, R. M. Empson, A. K. Cheetham, C. P. Grey and J. Skarnulis, Science
257 (1992) 782-784.
5. A. Clearfield, Chem. Rev. 88 (1988) 125-148.
6. R. Pozas-Tormo, L. Moreno-Real, M. Martinez-Lara and S. Bruque-Gamez, Can. J.
Chem. 64 (1986) 30-34.
7. R. Pozas-Tormo, S. Bruque-Gamez, M. Martinez-Lara and L. Moreno-Real, Can. J.
Chem. 66 (1988) 2849-2854.
8. K. M. Bonthrone, G. Basnakova, F. Lin and L. E. Macaskie, Nature Biotechnol. 14
(1996) 635-638.
9. G. Basnakova and L. E. Macaskie, Biotechnol. Bioeng. 54 (1997) 319-328.
10. G. Basnakova, A. J. Spencer, E. Palsgard, G. Grime and L. E. Macaskie, Environ. Sci.
Technol., 32 (1998) 760-765.
11. G. Basnakova, J. A. Finlay and L. E. Macaskie, Biotechnol. Lett., 20 (1998) 949-952.
12. G. Basnakova and L. E. Macaskie, Biotechnol. Lett., 23 (2001) 67-70.
13. P. Yong and L. Macaskie, J. Chem. Tech. Biotechnol. 63 (1995) 101-108.
14. L. E. Macaskie and A. C. R. Dean, Biotechnol. Lett., 7 (1985) 627-630.
15. P. Pattanapipitpaisal, A. N. Mabbett, J. A. Finlay, A. J. Beswick, M. Paterson-Beedle,
A. Essa, J. Wright, M. R. Tolley, U. Badar, N. Ahmed, J. L. Hobman, N. L. Brown
and L. E. Macaskie, Environ. Technol., 23(7) (2002) 731-45.
16. J. A. Finlay, V. J. M. Allan, A. Conner, M.E. Callow, G. Basnakova and L. E.
Macaskie, Biotechnol. Bioeng., 63 (1999) 87-97.
17. K. P. Nott, M. Paterson-Beedle, L. E. Macaskie and L. D. Hall, Biotechnol. Lett., 23
(2001) 1749-1757.
18. H. Onishi, Photometric Determination of Traces of Metals, Part IIA: Individual
Metals, Aluminum to Lithium, 4th edn., John Wiley, New York (1986) 454-459.
19. P. Yong, H. Eccles and L. E. Macaskie, Anal. Chim. Acta, 329 (1996) 173-179.
20. V. Michaylova and N. Kouleva, Talanta, 21 (1974) 523-532.
21. F. Huey and L. G. Hargis, Anal. Chem., 39(1) (1967) 125-127.
22. B. C. Jeong, Studies on the atypical phosphatase of a heavy metal accumulating
Citrobacter sp., D. Phil. Thesis, University of Oxford, UK (1992).
1161

Removal of soluble manganese from mine waters using a

fixed-bed column bioreactor
D. Barrie Johnson, Helen Miller, Sandra Ukermann and Kevin B. Hallberg
Abstract
Acid mine drainage waters vary considerably in the range and concentration of heavy
metals they contain. Besides iron, aluminum and manganese are frequently present at
elevated concentrations in waters draining both coal and metal mines. Passive treatment
systems (aerobic wetlands and compost bioreactors) are designed to remove iron by
biologically induced oxidation/precipitation, and aluminum by precipitation as Al(OH)3 as
a result of biologically-generated alkalinity. Manganese, however, is problematic as it
does not readily form sulfidic minerals (in compost bioreactors) and requires elevated pH
(>8) for oxidation of Mn(II) to insoluble Mn(IV). As a result, manganese removal is often
less effective than iron and aluminum removal, such as at the pilot passive plant treating
water draining the former Wheal Jane tin mine in Cornwall, U.K. We have sought to
devise a novel microbiological approach for effective removal of manganese from mine
waters at pH 5-7. Ferromanganese nodules (about 2 cm diameter) were collected from an
abandoned mine adit in the Gwydyr forest, north Wales, and used to inoculate a fixed bed
bioreactor (working volume ca. 700 ml). Pumice-like porous beads, made from inert recycled glass, were used as carrier materials for microbial biofilms. Following colonization
of the beads, the aerated reactor was tested for removal of soluble manganese in synthetic
and actual mine waters, using a continuous plug-flow mode of operation. Data from
preliminary experiments show that the bioreactor is highly efficient at catalyzing the
removal of manganese from the mine waters via oxidation of soluble Mn (II) and
precipitation of the resultant Mn (IV) in the bioreactor. Such an approach appears to be a
suitable alternative to current bioremediation strategies employed for manganese removal
from mine waters.
Keywords: acid mine drainage; bioreactors; bioremediation; manganese

1.
INTRODUCTION
Soluble manganese (Mn (II)) is often found in considerably greater amounts in AMD
than in unpolluted streams and groundwater [1]. Even though there are uncertainties
regarding the toxicity of manganese, the removal of this metal from surface- and
groundwaters is desirable for several reasons. As with iron and aluminum, manganese also
contributes to the total mineral acidity of mine waters when it is oxidized. Additionally,
the presence of manganese in water sources for human consumption is undesirable
because it can impart a metallic taste to water, it will stain laundry and water fixtures and,
as it precipitates readily as Mn (IV), manganese tends to block water distribution
1163
networks. Due to these concerns, the U.S. Environmental Protection Agency (EPA) has
set a secondary maximum contaminant level for Mn of 0.05 mg/l. The EPA has also
established guidelines limiting the concentration of Mn in acidic waters discharging from
mines at maximum of 4 mg/l, as long as average discharges for a 30-day period to not
exceed 2 mg/l.
To address such concerns, and to meet the potentially more stringent limits that are
being discussed in Europe and in the U.S., effective means of removal of Mn from mine
waters are required. Although manganese readily precipitates as Mn(IV), little oxidation
of Mn(II) occurs in solutions below pH 8, in contrast to iron, and the kinetics for
manganese oxidation is slow, relative to that of ferrous iron. In addition, biological
oxidation of Mn(II) does not proceed rapidly in the presence of Fe(II), and thus it is not
removed significantly in wetlands where the concentration of iron exceeds 1 mg/l [2]. The
lack of affinity of manganese for sulfide prohibits any significant removal as a sulfide in
compost wetlands and bioreactors. Therefore, alternative approaches are needed for the
removal of manganese from mine drainage waters.
The treatment systems currently used are typically applied at the end of any treatment
process scheme, to ensure the all iron has been removed first. An abiotic Mn-removal
system consisting of columns filled with limestone (to significantly raise the pH for Mn
oxidation and subsequent precipitation) has been described [3]. Biological approaches that
have been proposed for removing manganese from solution, by causing solution pH to rise
above 8 via oxygenic photosynthesis, include columns of immobilized cyanobacteria [4]
and small pools filled with rocks that have been colonized by algae [5]. The latter
approach has been applied at the Wheal Jane passive treatment plant. This study site was
built in 1993 following a catastrophic spill of acid mine drainage from the Wheal Jane tin
mine in Cornwall, England. However, the algal ponds have proved highly ineffective at
removing soluble manganese (as described below) and an alternative approach, using a
fixed-bed column bioreactor, has been initiated as an alternative strategy, which would be
equally applicable for treating other mine waters.
2.
MANGANESE REMOVAL AT THE WHEAL JANE PASSIVE TREATMENT

PLANT
A composite, pilot-scale passive treatment plant ("PPTP") was established at the
former Wheal Jane tin mine in Cornwall, England in 1995, as a large-scale experimental
facility to examine the efficacy of using constructed "wetlands" to remediate acidic, metalrich minewaters. A more complete description of the design and operation of the treatment
system can be found elsewhere [6]. In brief, a small fraction (<5%) of the acid (pH ~ 3.5)
mine drainage (AMD) water flows into the pilot passive treatment plant, where it is fed
into three similar composite passive systems (Figure 1). The main variations are the pretreatment that the AMD is subjected to ahead of entering the "wetlands". These are: (i)
addition of lime (CaO) to adjust the AMD to a pre-determined pH value (the "lime-dosed"
system); (ii) passage through an anaerobic (manure-based) cell to remove dissolved
oxygen and then through an anoxic limestone drain (to passively add alkalinity; the
"ALD" system); (iii) no pretreatment (the "lime-free" system). In each system, AMD then
flows (by gravity) through a series of five aerobic cells (which are considerably larger in
the "lime-free" system), then through a single "anaerobic cell" (containing sawdust and
cow manure; the "compost bioreactor") and finally through a series of ten smaller "rock
filters" (algal ponds). The key roles of the components in the passive systems are:
(a) aerobic cells: oxidation and precipitation of iron, and co-precipitation of arsenic;
1164
(b) anaerobic cell: sulfidogenesis and alkalinity genesis, causing precipitation of

chalcophilic metals as sulfides (e.g. copper, zinc and cadmium) and other metals as
hydroxides (e.g. aluminum);
(c) rock filters: (i) oxygenic photosynthesis causing aeration and further increase in
pH and precipitation of manganese (as carbonate and/or oxide) and (ii) polishing of
processed AMD (e.g. removal of organic carbon compounds draining the compost
bioreactors).
Fe2+ oxidation and Fe/As removal
MeS
Mn
removal removal
Rock Filters (10)
Compost Bioreactor
Aerobic Cell 5
Aerobic Cell 3
Aerobic Cell 1
Inflowing
AMD
Outflow
Figure 1. Schematic of the pilot passive treatment plant for remediating AMD at the
former Wheal Jane tin mine, Cornwall, England
The performance of the PPTP has been monitored by a group of geochemists and
microbiologists since 1998. In general, the aerobic cells have proved highly effective in
fulfilling their roles, but in only one of the three systems (the "lime-free" system) has the
performance of the compost bioreactors and algal ponds been satisfactory [7]. Figure 2
shows the concentrations of manganese in waters sampled throughout the three systems in
the PPTP in April 2002, and is representative of the general trends that have been
observed. Data in Figure 2 show that changes in soluble manganese do occur within the
aerobic cells (which are mostly highly acidic, pH <3, due to ferric iron hydrolysis); this is,
in part, due to dilution of the AMD by rainwater. However, significant concentrations (~25 mg/l) are present in the water draining the compost bioreactors and, as shown in Figure
2, is removed effectively only in the "lime-free" system. Indeed there was evidence of remobilization of manganese in the rock filters of the "lime-dosed" system on this sampling
occasion.
A major reason for the contrasting trends observed in Mn geochemistry found within
the algal ponds is that the necessary elevation in pH resulting from oxygenic
photosynthesis only occurred in the "lime free" system, as illustrated in Figure 3. The
discrepancy in Mn removal is due, in the main, to the variation in the performance of the
compost bioreactors in each of the three systems. Water draining these bioreactors in the
"ALD" and "lime-free" systems contains higher concentrations of both ferrous iron and
sulfide. Microbiological (and abiotic) oxidation of these species in the algal ponds causes
the pH of the water to decline, which is counter-productive to the overall objective [7]. As
the treatment plant is built in a temperate climate that receives considerable rainfall,
photosynthetic activity appears to be limited such that the rock filters are unable to
overcome the additional acid burden from the poorly performing compost bioreactors.
1165
Figure 2. Manganese concentrations in various components of the pilot passive

treatment plant at the former Wheal Jane tin mine; sampling date April 2002
Figure 3. Manganese (bars) and pH (lines) in the compost bioreactor influent (CB in)
and effluent (CB out) and in each of the 10 rock filter pools (RF1-RF10). The
samples were taken in April, 2002, from the ALD (A) and LF (B) systems. The
straight lines at the bottom of each graph represent the lower limit of manganese
detection, which was 0.25 mg/l
1166
3.
REMOVAL OF SOLUBLE MANGANESE BY FRESHWATER FE-MN

NODULES
To overcome the limitations of the rock filters, we have proposed an alternative
treatment strategy based on bioreactors containing immobilized manganese-oxidizing
bacteria. Ferromanganese nodules were collected from a trial mine adit located in the
Gwydyr forest, north Wales. The irregular, spherical nodules, ca. 1-2 cm in diameter, were
found in a pool within the adit (pH 6.9, soluble [Mn] 30 mg/l, conductivity 149 S/cm).
These were transferred to the laboratory and stored at 4C. Two of the nodules were
placed into replicate 250 ml conical flasks containing 100 ml of filter-sterilized "MMS"
medium [8], which had been developed to promote the growth of manganese
oxidizing/depositing bacteria. MMS medium contains sodium pyruvate (2.4 g/l) and basal
salts (pH 7); 50 l of a 100 mM manganese sulfate solution (filter-sterilized) was added
just prior to inoculation with the nodules, and the flasks were incubated, shaken (100 rpm)
at 20C. Samples were withdrawn at regular intervals and concentrations of soluble Mn
analyzed colorimetrically by complexing with formaldoxime (Hydrocheck, Cambridge,
U.K.). Results, given in Figure 4, show that concentrations of soluble manganese declined
steadily during incubation; this corresponded with the appearance of brown-black
precipitates (and associated bacteria) within cultures.
Figure 4. Removal of soluble manganese in shake flask cultures containing

freshwater ferro-manganese nodules. The error bars (occluded in most cases by
symbols) show the range of soluble [Mn] at each sampling occasion
4.
COMMISIONING AND PERFORMANCE OF A FIXED-BED BIOREACTOR

A fixed-bed bioreactor for removing soluble manganese from mine waters was set up
by mixing freshwater ferromanganese nodules (ca. 10%, v/v) with porous, acid-washed
glass beads (8-16 mm) manufactured by Poraver GmbH, Germany [9]. The nodules/beads
mixture was packed (to a depth of 14.5 cm) into a 25 by 10 cm perspex column, between
3.5 cm layers of acid-washed pea gravel (Figure 5a). The base and cap of the column
contained ports for aeration (applied at ca. 2 l/min), medium circulation and sampling. The
column was filled to the top of the upper gravel layer surface with MMS medium, which
was then re-circulated through the column using a peristaltic pump until the soluble
1167
manganese had declined to <10% of the initial concentration, at which point the column
was drained and fresh medium introduced. This sequence was repeated several times, and
the column was observed to accumulate increasing amounts of brown-black (Mn (IV))
precipitates (Figure 5b). The external plumbing on the bioreactor was then re-configured
to allow the system to be tested in continuous plug-flow mode.
Representative data showing removal of soluble manganese from MMS medium with
the fixed-bed bioreactor run in recirculating and plug-flow modes are shown in Figure 6.
In both cases, >90% of soluble Mn was removed within 4 hours. When run in plug-flow
mode, 50% of soluble manganese was removed with a flow rate of 400 ml/h, and 68%
when the flow rate was lowered to 195 ml/h.
Figure 5. Fixed-bed bioreactor for removing soluble Mn from solution: (a) after
initial construction; (b) several weeks after continuous operation
Figure 6. Performance of the fixed-bed bioreactor in removing soluble manganese

from MMS medium when run in recirculating () and plug-flow ( ) modes
1168
5.
DISCUSSION
Manganese is the second most abundant heavy metal in the lithosphere, occurring at
between 0.002 and 10% (w/w) in terrestrial ecosystems, ca. 0.2 g/l in marine waters, and
an average of 8 g/l in fresh waters [10]. As with many other metals, greatly enhanced
concentrations of soluble manganese can occur in waters draining coal and metal mines.
The major oxidation states of manganese are +2 and +4: Mn (II) can occur as free ions in
solution (Mn2+, and Mn(OH)+ in circum-neutral pH waters) or complexed with a variety of
organic and inorganic ligands. In contrast, Mn (IV) forms very insoluble oxide phases;
MnO2 has a point of zero charge at pH 2.8 which, together with its large specific surface
area, gives rise to a marked capacity to adsorb metal cations, including Mn (II).
Thermodynamically, Mn (II) should oxidize spontaneously to Mn (IV) in aerated,
neutral pH waters, but the activation energy required is relatively high, and this greatly
slows down the process, causing Mn (II) to be far more stable in non-acidic waters than is
ferrous iron. In contrast, the adsorption of Mn (II) onto MnO2 is much more rapid. The
"activation energy barrier" for Mn oxidation can be overcome biologically. Many
microorganisms are known to be able to catalyze the oxidation of Mn (II), and at least
some of these are known utilize the energy available from the reaction [10]. Manganese
(II) may also form a highly insoluble carbonate (MnCO3; rhodocrosite). Since, however,
the concentration of CO32- in solution is dictated by pH, formation of this mineral is only
significant in neutral-alkaline water bodies.
The existence of marine "manganese nodules" has been known for some time, and
these are mined as an ore of the metal. Iron/manganese-rich nodules may also occur in
freshwaters, though these tend to be much smaller than their marine equivalents. Stein et
al. [11] analyzed prokaryotic populations associated with freshwater ferromanganese
micronodules (mm scale) and sediments by extracting nucleic acids, amplification of 16S
rRNA genes and phylogenetic analysis of cloned products. Two groups of known Mn (II)oxidizing genera were identified, Leptothrix and Hyphomicrobium, as well as a group
related to metal-reducing bacteria.
To date, the microorganisms associated with the freshwater ferromanganese nodules
used in the present study have not been subjected to rigorous analysis, though Mn(IV)depositing bacteria have been isolated on solid media inoculated with crushed nodules and
bioreactor liquor. The increasingly-efficient performance of the bioreactor (in terms of
removing soluble Mn(II)) as Mn(IV) deposits accumulate may be due both to the activities
of planktonic-phase Mn(II)-oxidizers and to the abiotic uptake of Mn(II) onto MnO2
precipitates, and subsequent oxidation of the metal by sessile bacteria. What is clear is that
a simple fixed-bed bioreactor containing Mn(II)-oxidizing/depositing microorganisms can
be highly effective in catalyzing the removal of the metal from a synthetic, metal-rich
waste water. So far, limited tests of the technology for bioremediating actual Mn-rich
mine waters has shown considerable promise (data not shown). Fixed-bed bioreactors for
oxidizing and precipitating ferrous iron from AMD have been described by a number of
research groups [e.g. 12].
A major advantage offered by metal-removing bioreactors is that they occupy a far
smaller "footprint" than, for example, the rock filters/algal ponds at the Wheal Jane
passive treatment plant. Land availability frequently restricts the application of passive
mine water treatment systems, and low-maintenance biological systems are attractive
alternatives in such circumstances.
1169
ACKNOWLEDGEMENTS
We wish to thank DTI (ref. # BTL/20/71), MIRO (U.K.) and the Environment
Agency (U.K.) for financial support. We also acknowledge the contributions made by our
colleagues at the University of Reading, Imperial College, London, the Centre for Ecology
and Hydrology, the Camborne School of Mines and the Knight Pisold Consulting Group
(U.K.) to the "Wheal Jane Project".
REFERENCES
1. D. Banks, P. L. Younger, R.-T. Arnesen, E. R. Iversen and S. B. Banks, Environ. Geol.
32 (1997) 157.
2. B. Nairn and R. S. Hedin, Constructed wetlands for water quality improvement, G. A.
Moshiri (eds.), Lewis Publishers, Boca Raton, Fl, (1993) 187.
3. F. C. Thornton, Ecol. Eng. 4 (1995) 11.
4. J. Bender, J. P. Gould, Y. Vatcharapijarn, J. S. Young and P. Phillips, Water Environ.
Res. 66 (1994) 679.
5. P. Phillips, J. Bender, R. Simms, S. Rodriguezeaton and C. Britt, Water Sci. Technol.
31 (1995) 161.
6. Q. U. I. Hamilton, H. M. Lamb, C. Hallett and J. A. Proctor, J. Chart. Inst. Water.
Environ. Manage. 13 (1999) 93.
7. K. B. Hallberg and D. B. Johnson, Mine Water Treatment:A Decade of Progress, C. A.
Nuttall (eds.), Newcastle upon Tyne, U.K., (2002) 143.
8. Y. M. Nelson, L. W. Lion, W. C. Ghiorse and M. L. Shuler, Appl. Environ. Microbiol.
65 (1999) 175.
9. . Kolmert and D. B. Johnson, J. Chem. Technol. Biotechnol. 76 (2001) 836.
10. H. L. Ehrlich, Geomicrobiology, 3rd edition. Marcel Dekker, New York (1996).
11. L. Y. Stein, M. T. La Duc, T. J. Grundl and K. H. Nealson, Environ. Microbiol. 3
(2001) 10.
12. A. Mazuelos, F. Carranza, I. Palencia and R. Romero, Hydrometallurgy 58 (2000) 269.
1170

Sulfane sulfur of persulfides is the actual substrate of the

sulfur-oxidizing enzymes from Acidithiobacillus and
Acidiphilium spp.
T. Rohwerder and W. Sand
Department of Microbiology, Institute for General Botany, University of Hamburg,
Ohnhorststr. 18, D-22609 Hamburg, Germany
Tel/Fax: +49 40 428 16 423, e-mail: sand@mikrobiologie.uni-hamburg.de
Abstract
The sulfur chemistry of sulfur dioxygenase (EC 1.13.11.18) was analyzed in various
strains of the genera Acidithiobacillus and Acidiphilium. Cell-free extracts oxidized
elemental sulfur well, however only in the presence of glutathione (GSH). Sulfur had to be
merged with GSH to form the organic persulfide or higher homologues (GSnH, n > 1)
prior to the enzymic oxidation to sulfite. The latter reacted non-enzymically further either
to sulfate, thiosulfate, or glutathione S-sulfonate (GSSO3-). In contrast, oxidation of
hydrogen sulfide by the dioxygenase system did not require the reduced but the oxidized
form of glutathione (GSSG). The latter formed with sulfide non-enzymically GSSH prior
to enzymic oxidation. Persulfide species as sulfur donors could not be replaced by other
sulfane-sulfur-containing compounds such as thiosulfate, polythionates, aryl
thiosulfonates, or bisorganyl-polysulfanes. In conclusion, persulfides and higher
monoorganylpolysulfanes are exclusively the substrates for the sulfur dioxygenases of
meso-acidophilic sulfur bacteria. On the basis of these findings a biochemical model for
both elemental sulfur and sulfide oxidation in Acidithiobacillus and Acidiphilium spp. is
proposed.
Keywords: sulfur dioxygenase, sulfur oxidation, sulfide oxidation, Acidithiobacillus,

Acidiphilium
1.
INTRODUCTION
For the oxidation of elemental sulfur a conclusive biochemical pathway has not yet
been identified. Diverse enzymic systems like a reverse sulfite reductase [1,2], a
thiosulfate-oxidizing enzyme complex [3,4], and dioxygenases or oxygenase-reductases
[5,6,7,8] are considered to be involved. The meso- and moderately thermophilic leaching
bacteria seem to have a pathway depending on low-molecular thiols. For in vitro assays
glutathione (GSH) is the preferentially used thiol compound (reviewed in [9]). Due to its
capability to incorporate molecular oxygen into the sulfur substrate and to produce sulfite
as the primary reaction product the system has been named sulfur dioxygenase (EC
Present address: UFZ Center for Environmental Research Leipzig-Halle, Permoserstr. 15, D04318 Leipzig, Germany
1171
1.13.11.18; [5]). This GSH-dependent activity has been detected in Acidithiobacillus

thiooxidans and Acidithiobacillus ferrooxidans [10,11,12] as well as in Sulfobacillus
thermosulfidooxidans [13]. Although the known sulfur dioxygenases are very similar in
molecular size and cofactor content, the role of GSH is under dispute. The work of Suzuki
[5,10] and Silver & Lundgren [11] suggest a catalytic role for GSH, where organic
polysulfanes are formed and sulfur is oxidized at the zero valence state to sulfite (equation
1). In contrast, Sugio and coworkers [12,14] described an enzyme system that probably
used dissolved sulfide as substrate. Consequently, GSH would be consumed
stoichiometrically in the course of sulfur oxidation due to the preceding non-enzymic
reduction of elemental sulfur to sulfide.
GSnH + O2 + H2O GSn-1H + SO32- + 2 H+
n>1
(1)
So far, the sulfur chemistry and especially the fate of GSH has not been analyzed
leaving the actual activation mechanism for elemental sulfur unresolved. Therefore, we
reinvestigated the GSH-dependent sulfur-oxidizing system. Special analytical attention
was given to the organic sulfur species that occurred in the in vitro assay. Besides
elemental sulfur other sulfane compounds and hydrogen sulfide were tested to elucidate
the actual substrate used by the sulfur dioxygenase. For this purpose the enzymic activity
of cell-free extracts of At. thiooxidans, At. ferrooxidans, Acidiphilium acidophilum, and
Acidiphilium cryptum was investigated.
2.
2.1 Bacterial strains, growth conditions, and cell-free preparations

The bacterial strains of this study are listed in Table 1. Growth on elemental sulfur
powder (5 g/L) was performed in a salt solution modified from Mackintosh [15] with 2
mM NH4Cl and an initial pH of 3.0 adjusted with HCl. In addition, a lithotrophic iron
medium [15] with 5 g/L iron(II) ions and a glucose-based medium [16] were applied for
the growth of iron-oxidizers and facultative or obligate heterotrophs, respectively. All
strains were cultivated aerobically at 28C. Cell disruption was performed at 4C under
anaerobic conditions by treatment with glass beads [17]. Suspensions of disrupted cells
were centrifuged (20 min, 25,000 x g, twice) to remove intact cells and cell residues. The
supernatant, referred to as crude or cell-free extract, usually contained 1 to 2 g/L protein
and was used either directly for activity assays or was stored under an anaerobic
atmosphere at 25C (less than 3 months).
2.2 Enzyme assays
All assays were performed aerobically at 30C in phosphate buffer (10 mM, pH 6.5)
with stirring at 300 rpm. If needed, the pH was maintained by titration with 50 mM KOH
or 50 mM HCl. In order to determine non-enzymic reactions, assays without protein, with
0.2 g/L BSA, or with heat-inactivated crude extracts (90C for 30 min) were used. For
measuring elemental sulfur oxidation a system of dispersed sulfur in water was developed.
To 50 mL of deionized water an equal volume of a saturated acetonic sulfur solution was
added and dialyzed against 5 L deionized water to remove the acetone. The whitish
dialysis product contained sulfur droplets of 2 to 10 m in diameter and was used at a
concentration of 4 mM. When sulfide, thiosulfate, tetrathionate, and p-toluolthiosulfonate
were tested stock solutions of the respective potassium or sodium salts in deionized water
were prepared. A mixture of GSSG and higher homologues was obtained by an incubation
of 500 mM elemental sulfur (powder) with 100 mM GSH at pH 7.5 (adjusted with KOH)
under stirring and anaerobic conditions until the solution got lemon-colored. At this point
1172
the pH was lowered to 5.0 by addition of HCl. The resulting hydrogen sulfide was
removed by evacuation of the gas phase. Finally, the pH was adjusted to 6.5.
2.3 Analyses of sulfur compounds
Thiosulfate, polythionates, GSH, GSH-derivatives, and p-toluolthiosulfonate were
analyzed by ion pair chromatography using tetrabutylammonium as counter ion [17]. A
HPLC system from Kontron/BIO-TEK Instruments was applied. Chromatograms were
recorded at 205, 215, 265, and 300 nm concomitantly with spectra from 190 to 320 nm.
Elemental sulfur was analyzed by reversed-phase chromatography followed by UVdetection [18]. Sulfite and sulfate were quantified by ion exchange chromatography and
conductivity detection [18] applying a Dionex DX 500 system. Sulfite was fixed with
methanal prior to analysis [19]. Dissolved sulfide was determined by the methylene blue
method [20]. With the exception of elemental sulfur quantification, all samples had been
filtered (nylon filter, 0.2 m) prior to analysis in order to remove elemental sulfur.
3.
3.1 Oxidation of elemental sulfur via glutathione persulfide

The enzymic oxidation of elemental sulfur by cell-free extracts from various mesoacidophilic sulfur-oxidizing bacteria was performed mainly according to the sulfur
dioxygenase assay of Suzuki [5,10]. As has been demonstrated previously for similar
enzyme preparations (reviewed in [9]) enzymic oxidation of elemental sulfur could not be
observed in the absence of GSH. With GSH, however, elemental sulfur was readily
oxidized by crude extracts derived from sulfur oxidizers. In contrast to data of a previous
study [21], preparations from the obligate iron(II) oxidizer Leptospirillum ferrooxidans
did not show any sulfur oxidation activity (Table 1). The main reaction product of positive
assays was thiosulfate (99%). Besides, traces of sulfite, sulfate, and glutathione Ssulfonate were detectable (data not shown). Obviously, sulfite as the first oxidation
product rapidly reacted further with the finely dispersed sulfur to thiosulfate (equation 2).
Only trace amounts were consecutively oxidized to sulfate or reacted with GSSG to
glutathione S-sulfonate (equation 3).
SO32- + 1/8 S8 S2O32(2)
GSSG + SO32- + H+ GSSO3- + GSH
(3)
GSSG was regularly detected in the assay solutions. It was formed in the course of the
non-enzymic reduction of elemental sulfur with GSH (equation 4 and 5). This reaction
occurred in parallel to the enzymic activity and, besides GSSG, GSnG species with n
ranging from 3 to 5 and hydrogen sulfide were formed (data not shown).
S8 + GSH (GS9H) GSnH + (9-n)/8 S8
(4)
x 2, y 1
(5)
GSxH + GSyH GSx+y-1G + H2S
Comparing sulfur oxidation and reduction activities, it turned out that GSH was not
consumed stoichiometrically in the course of elemental sulfur oxidation. In other words,
free sulfide was not a substrate for the sulfur-oxidizing enzyme under the experimental
conditions. Interestingly, the non-enzymic sulfur reduction by GSH (equation 4 plus 5)
and the GSH-dependent enzymic sulfur oxidation were connected in another way: the
higher the rates of the enzymic sulfur oxidation were, the lower were the GSH oxidation
rates (Fig. 1). Furthermore, for a certain GSH concentration a maximal sulfur dioxygenase
activity was obtainable, irrespective of the amount of enzyme, at which no further GSH
oxidation occurred. For example, with 100 M GSH an upper limit of about 350 M/h
1173
was achieved (Fig. 1). The explanation for this phenomenon can be deduced from
equations 1 and 5. Both reactions compete for the highly unstable
monoorganylpolysulfane derivatives (GSnH, n > 1) of GSH. Summarizing, from the
analyses of all relevant sulfur compounds in these sulfur dioxygenase assays we can fully
confirm Suzukis proposal [5] of persulfides and their higher homologues being the actual
substrates for the sulfur-oxidizing enzyme system of meso-acidophilic sulfur oxidizers
(equation 1). This finding is valid for the genera Acidithiobacillus and Acidiphilium.
Table 1. Specific activities of sulphur dioxygenase in crude extracts of different
strains of meso-acidophilic bacteria
strain
substrate for
cell growth*
activity
mol.h-1.mg-1
sulfur
3.8 1.4
Acidithiobacillus
ferrooxidans R1
iron(II)
3.5 0.6
Acidithiobacillus thiooxidans DSM 504
sulfur
2.4 0.8
Acidithiobacillus thiooxidans K6
sulfur
1.5 0.5
sulfur
22.4 5.4
Acidiphilium
acidophilum DSM 700
glucose
7.8 2.5
Acidiphilium cryptum DSM 2389
glucose
0.3 0.1
Leptospirillum ferrooxidans DSM 2705
iron(II)
ND
*Strains were subcultured at least for 2 years on the indicated substrates.
Specific activity is expressed as the amount of sulfur atoms oxidized to
the valence state of sulfite within 1 hour by 1 mg protein.
ND not detectable.
Figure 1. Sulfur dioxygenase assay with 100 M GSH. Relationship between initial
GSH consumption rate V and enzymic activity
1174
3.2 Oxidation of hydrogen sulfide

Although the free sulfide formed during the incubation of elemental sulfur with GSH
was not oxidized enzymically, hydrogen sulfide was tested separately for enzymic
oxidation. As expected, cell-free preparations incubated with or without GSH did not
oxidize the sulfide. Surprisingly, low but significant enzymic activity was observed after
GSSG addition to the assays (at 1 mM). In this case, the formation of thiosulfate and
glutathione S-sulfonate was observed (Fig. 2). The latter was the main oxidation product.
Furthermore, GSSG was reduced to GSH. From the stoichiometry of all these compounds
(Fig. 2) the following reaction mechanism can be deduced. GSSG reacted with hydrogen
sulfide to GSH and GSSH (equation 6). The persulfide was consecutively oxidized by the
sulfur dioxygenase to sulfite (equation 1) and then formed with excess GSSG the
glutathione S-sulfonate (equation 3).
GSSG + H2S GSSH + GSH
(6)
+
2 GSSG + H2S + O2 + H2O 3 GSH + GSSO3 + H
(7)
Summing up, GSH and glutathione S-sulfonate should be produced in a ratio of 3:1
(equation 7) which agrees quite well with the observed values (Fig. 2). Consequently,
sulfide was not oxidized directly, as proposed by Sugio and coworkers [14], but via the
formation of persulfide species.
concentration (M)
3.5
15
3.0
10
2.5
2.0
1.5
ratio GSH/GSSO3-
4.0
20
1.0
0.0
GSH
thiosulfate
0.4
0.8
time (h)
1.2
GSSO3ratio GSH/GSSO3-
Figure 2. Oxidation of sulfide by cell-free extracts in the presence of 1 mM GSSG

3.3 Oxidation of other sulfane compounds
From the results of the oxidation tests performed with elemental sulfur and sulfide, it
can be clearly deduced that the active enzyme preparations contained no enzymic activity
for the direct oxidation of these sulfur species. Only the sulfane sulfur of persulfides was
oxidizable. In additional experiments the specificity of this enzymic activity was tested
with various inorganic and organic sulfane compounds (see 2.2.). Briefly, with none of
these compounds an oxidation activity was observed. However, when a mixture of GSSG
and higher GSnG species (2<n<6) was incubated with GSH the higher
bisorganylpolysulfanes were readily oxidized by active crude extracts (Fig. 3). This
indicates that the GSnG species were also degraded via GSnH compounds according to
equation 8. The sulfane sulfur of the polysulfanes was oxidized to thiosulfate and
glutathione S-sulfonate. The latter compound was formed in parallel with GSH (Fig. 3 A).
1175
Consequently, it was highly likely formed from sulfite and GSSG because this reaction
produced equimolar amounts of GSSO3- and GSH (equation 3).
500
300
200
50
100
0
0
0.0
GSnG (mAbs min)
400
100
0.5
1.0 1.5
time (h)
thiosulfate
2.0
GSH
GSSO3-
150
1100
100
1000
50
900
800
0.0
0.5
1.0
1.5
GSSG (M)
thiosulfate (M)
150
GSH, GSSO3 (M)
(8)
GSnG + GSH GSxH + GSyGn 3, x 2, y 2
Summarizing, only the sulfane sulfur of persulfide compounds can be oxidized by the
sulfur dioxygenase of meso-acidophilic bacteria. Neither polythionates nor
bisorganylpolysulfanes are oxidized directly.
2.0
time (h)
GS3 G
GS4 G
GS5 G
G SSG
Figure 3 A+B. Oxidation of GSnG species (2 < n < 6) by cell-free extracts in the
presence of 200 M GSH. Amounts of GSnG are expressed as peak areas recorded at
215 nm
3.4 Proposal of a biochemical model for sulfur and sulfide oxidation
Compiling all data available, a model for the oxidation of sulfide and elemental sulfur
can now be proposed (Fig. 4). Extracellular elemental sulfur (S8) reacts with thiol groups
of special outer-membrane proteins (OMP-SH) and, thus, can be transported as persulfide
sulfur (OMP-SSH) into the periplasmic space. Here sulfur dioxygenase (SDO) takes over
the sulfane sulfur and oxidizes it to sulfite. Sulfite is oxidized further to sulfate by a
sulfite:acceptor oxidoreductase (SOR). This enzyme most probably uses cytochromes
(Cyt) as electron acceptors [22,23,24]. Free sulfide is oxidized by a separate
sulfide:quinone oxidoreductase (SQR), which is located at the periplasmic site of the
cytoplasmic membrane [25,26]. For this reaction polysulfides have recently been proposed
to be the oxidation products [27]. However, at least for acidophilic bacteria we believe
that membrane-bound persulfides (OMP-SSH) as products of SQR activity are more likely
(Fig. 4). In fact, the membrane binding of sulfur at zero valence state was demonstrated in
sulfide oxidation experiments with cell-free extracts from strains of At. thiooxidans
[25,28]. Candidates for the not yet identified thiol-bearing OMP are the sulfide-binding
1176
protein isolated from At. ferrooxidans AP 19-3 [29] and several outer membrane proteins
that have been associated with sulfur oxidation in strains of At. ferrooxidans [30,31,32].
Figure 4. Proposed biochemical pathway for sulfide and elemental sulfur oxidation in
leaching bacteria of the Gram-negative genera Acidithiobacillus and Acidiphilium
(see text).
ACKNOWLEDGEMENTS
This work was supported by grants to W. S. from BMBF via UBA (1490954), DBU
(05333), and the Max-Buchner-Forschungsstiftung (DECHEMA e.V.).
REFERENCES
1. M. Schedel and H. G. Trper, Biochim. Biophys. Acta, 568 (1979) 454.
2. A. S. Pott and C. Dahl, Microbiology, 144 (1998) 1881.
3. C. G. Friedrich et al., Appl. Environ. Microbiol., 67 (2001) 2873.
4. D. Rother et al., J. Bacteriol., 183 (2001) 4499.
5. I. Suzuki, Biochim. Biophys. Acta, 110 (1965) 97.
6. T. Emmel et al., J. Gen. Microbiol., 132 (1986) 3415.
7. A. Kletzin, J. Bacteriol., 171 (1989) 1638.
8. A. Kletzin, J. Bacteriol., 174 (1992) 5854.
9. I. Suzuki, Methods Enzymol., 243 (1994) 455.
10. I. Suzuki, Biochim. Biophys. Acta, 104 (1965) 359.
11. M. Silver and D. G. Lundgren, Can. J. Biochem., 46 (1968) 457.
12. T. Sugio et al., J. Bacteriol., 169 (1987) 4916.
13. E. N. Krasil`nikova et al., Mikrobiologiya, 67 (1998) 156.
14. T. Sugio et al., Biochim. Biophys. Acta, 973 (1989) 250.
15. M. E. Mackintosh, J. Gen. Microbiol., 105 (1978) 215.
1177
16. A. P. Harrison Jr., Int. J. Syst. Bacteriol., 31 (1981) 327.

17. T. Rohwerder, Untersuchungen zur enzymatischen Oxidation von Elementarschwefel
bei acidophilen Laugungsbakterien, PhD thesis, University of Hamburg, 2002
(www.sub.uni-hamburg.de/disse/849/dissertation.pdf).
18. A. Schippers and B. B. Jrgensen, Geochim. Cosmochim. Acta, 65 (2001) 915.
19. J. Wei, Ionenchromatographie, VCH, Weinheim, 1991.
20. Anonymous, Deutsche Einheitsverfahren zur Wasser-, Abwasser- und Schlammuntersuchung (DEV), VCH, Weinheim, 1984.
21. T. Sugio et al., Appl. Environ. Microbiol., 58 (1992) 431.
22. J. R. Vestal and D. G. Lundgren, Can. J. Biochem., 49 (1971) 1125.
23. K. Nakamura et al., Biosci. Biotech. Biochem., 59 (1995) 11.
24. G. A. H. de Jong et al., J. Mol. Catal. B, 8 (2000) 61.
25. D. J. W. Moriarty and D. J. D. Nicholas, Biochim. Biophys. Acta, 184 (1969) 114.
26. T. Nbel et al., Arch. Microbiol., 173 (2000) 233.
27. C. Griesbeck et al., Biochemistry, 41 (2002) 11552.
28. D. J. W. Moriarty and D. J. D. Nicholas, Biochim. Biophys. Acta, 197 (1970) 143.
29. T. Sugio et al., Agric. Biol. Chem., 55 (1991) 2091.
30. V. Buonfiglio et al., FEMS Microbiol. Rev., 11 (1993) 43.
31. V. Buonfiglio et al., J. Biotechnol., 72 (1999) 85.
32. N. Ohmura et al., J. Bacteriol., 178 (1996) 5776.
1178

Sulfate reduction at low pH by mixed cultures of acidophilic

bacteria
Anna M. Sen, Sakurako Kimura, Kevin B. Hallberg and D. Barrie Johnson
Abstract
Mixed cultures of acid-tolerant sulfate reducing bacteria (SRB) and other acidophilic
bacteria were shown to form stable microbial communities in sulfidogenic bioreactors.
Cultures were set up in pH-controlled (pH 2.5-5.5) bioreactors, and run in batch and
continuous mode using glycerol or ethanol as electron donor. Reduction of sulfate (and
corresponding formation of hydrogen sulfide), acid-consumption and changes in the
indigenous microflora (SRB and other bacteria) were monitored. Sparging of the
bioreactors with nitrogen effectively removed the H2S, which was used to precipitate
copper as CuS (off-line) or zinc as ZnS (within the bioreactor). Following earlier
experiments with undefined mixed populations of anaerobic acidophiles, we studied
sulfidogenesis at low pH by mixed cultures of Desulfosporosinus spp. (SRB) and an
Acidocella-like acidophilic heterotroph. In these cultures, accumulation of acetic acid (the
product of incomplete glycerol oxidation by the SRB) was much less than in the undefined
mixed cultures, leading to more efficient metal (zinc) immobilization in. Syntrophy
between acid-tolerant SRB and acidophilic heterotrophic bacteria appears to be important
in mediating sulfidogenesis in acidic liquors.
Keywords: acid mine drainage; acidophiles; bioremediation; metal recovery; sulfate

reducing bacteria
1.
INTRODUCTION
Studies of the microbiology in extremely acidic liquors, such as acid mine drainage
(AMD) and mineral leachates, have focused primarily on aerobic processes, chiefly
because the major reactions responsible for mineral dissolution are oxidative. Some
acidophiles are able to grow in anoxic environments [1]. Iron respiration (where ferric iron
acts as electron acceptor) seems particularly widespread amongst acidophiles. Ironoxidizing moderate thermophiles and obligately heterotrophic Acidiphilium spp. can
couple the oxidation of organic substrates to the reduction of ferric iron [2-4].
Acidithiobacillus ferrooxidans can also grow anaerobically by iron respiration, using
either reduced sulfur or hydrogen as electron donor [5, 6].
Sulfate is found in mineral leaching environments often at concentrations of many
grams per liter. Dissimilatory sulfate reduction might therefore be anticipated to be
widespread in acidic anaerobic zones (sediments etc.), and there are a number of reports in
the literature documenting sulfidogenesis in these situations [e.g. 7, 8, 9]. However,
attempts to isolate and cultivate acid-tolerant/acidophilic sulfate reducing bacteria (aSRB)
1179
from mine waters have mostly been unsuccessful. Most mine water isolates appear to be
neutrophilic, and are not active below pH ~ 5 [e.g. 7, 10]. There are a number of reasons
for this, including the use of inappropriate media (substrates, such as lactate, used
routinely to enrich neutrophilic species may be toxic to acidophiles). Sen and Johnson [11]
obtained enrichment cultures of aSRB from sediment samples from a copper mine and
from geothermal areas on Montserrat. Sulfidogenesis by these mixed cultures occurred at
pH <3, but was more rapid between pH 3 and 5.
This paper describes sulfidogenesis by undefined and defined cultures of aSRB, and
the immobilization of heavy metals at low pH due to biogenic sulfide. Syntrophy between
aSRB and a non-sulfidogenic heterotrophic acidophile is discussed.
2.
2.1 Establishment and maintenance of enrichment cultures

Enrichment cultures containing acidophilic/acid-tolerant strains of SRB were obtained
from sediments from the Afon Goch, Parys copper mine (north Wales) and Montserrat, as
described elsewhere [11]. These were subcultured in ~ 500 ml of ethanol-containing liquid
medium (below) containing ~ 300 ml of porous glass beads (2-4 mm diameter; [12]). The
beads were included to act as a surface for bacterial colonization. These cultures were
used to inoculate 2 l bench-scale bioreactors (P350; Electrolab, U.K.), which were run in
either batch or continuous mode. A schematic diagram of the continuous mode set up is
shown in Figure 1. Cultures were maintained at 30C, and pH control was maintained via
addition of either 0.1 M HCl or KOH.
Impeller
Sample
point
(input)
MEDIUM
PUMPED IN
ORP/pH electrodes
Acid/
alkali
Filter
OFN
in
WASTE
PUMPED
OUT
(to sample
point output
and waste vessel)
2L bioreactor
vessel
Biofilm of aSRB
on Poraver beads
Gas collection bottles

each containing 500 ml
copper acetate (10 mM)
Figure 1. Bioreactor system set-up for continuous mode experiments. OFN = oxygenfree nitrogen
1180
2.2 Batch and continuous cultures of mixed (undefined) SRB cultures

The growth medium for batch culture experiments contained 10 mM ethanol, 0.002%
yeast extract, 10 mM FeSO4, together with basal salts and trace elements [13]. The
medium was heat-sterilized prior to addition of filter-sterilized ethanol and FeSO4. The
cultures (working volumes, 1.8 l) were sparged daily for 5-10 minutes with oxygen-free
nitrogen (OFN) to remove H2S, which might otherwise inhibit the SRB. Samples were
removed at regular intervals and analyzed for sulfate, ferrous iron, ethanol, acetate and
protein. Bacteria were isolated and enumerated on overlay solid media, as described
elsewhere [11].
The bioreactor was run in continuous mode for the Parys mixed culture only. In this
case, a modified growth medium was used, with glycerol (at 5 mM) replacing ethanol as
electron donor. In addition, the concentration of FeSO4 was lowered to 1 mM, with sulfate
concentrations being maintained as earlier by the inclusion of 9 mM K2SO4. The sterile
feed medium was adjusted to pH 3.0 with H2SO4, and was added to the culture vessel at
25 ml/h, while that in the bioreactor was maintained at pH 2.5, 3.0, 3.5 or 4.0. In addition,
inflow rates of 50 and 100 ml/h were applied in the pH 4.0-maintained culture.
Throughout the working day, the reactor vessel was sparged for about 5 min/h with OFN
and the off-gas bubbled though gas bottles containing 10 mM copper acetate, causing
precipitation of CuS. Concentrations of glycerol and acetate in the growth vessel were also
measured in continuous mode experiments.
2.3 Batch cultures of defined mixed SRB cultures
During the course of purifying aSRB on solid media, a novel uncolored colony
morphology was observed on several overlay plates. Physiological and phylogenetic work
confirmed that this was an Acidocella-like isolate (unpublished); bacteria of this type had
not previously been found in anaerobic bioreactor cultures. A mixed culture of this isolate,
PFBC, and Desulfosporosinus isolates from the Parys culture (isolate P1) and from the
Montserrat culture (isolate M1) was grown in batch mode in an anaerobic bioreactor. A
modified medium was used in this case, containing 5 mM glycerol, 5 mM ZnSO4 5 mM
K2SO4 and 1 mM FeSO4. Yeast extract was replaced by the vitamin mixture proposed by
Widdel and Pfennig [14] for SRB, added at 1 ml vitamin solution/l medium.
2.4 Analytical techniques
Samples were removed from the bioreactor vessels and filtered through 0.2 m
cellulose nitrate membranes (Whatman, U.K.). Concentrations of ethanol, glycerol and
acetate were measured enzymatically (R-Biopharm GmbH, Germany), and protein using
the Bradford assay [15]. Sulfate was measured turbidometrically as barium sulfate
(Hydrocheck, Cambridge, U.K.), and ferrous iron using the ferrozine reagent [16]. Sulfide
production in continuous cultures was estimated by analysis of copper concentrations in
the gas bottles by atomic absorption spectrophotometry. Proton consumption by batch and
continuous cultures was determined from the amounts of alkali that were required to
maintain the cultures at pre-set pH levels.
2.5 Calculations
Sulfate reduction rates (SRR) were calculated in mmol SO42- reduced l-1h-1.
Hydraulic residence time (HRT), i.e. the average time that the culture medium was
retained in the bioreactor, was calculated by dividing the working volume by the flow rate
of feed medium.
1181
3.
RESULTS
Mixed cultures of aSRB and other anaerobic bacteria were established readily in
enrichment cultures, as described previously [11]. Both the Parys and Montserrat cultures
contained aSRB and other bacteria (as evidenced by the formation of black and non-black
colonies, respectively, on overlay plates). Since these mixed cultures appeared to be robust
(i.e. they were maintained through many sub-cultures) it was decided to carry out initial
batch- and continuous-mode experiments with them.
Data from a batch culture run (at pH 4.0), using the Parys culture, is shown in Figure
2. Sulfate reduction in this culture corresponded with proton consumption (alkali
generation) that was due, at least in part, to production of H2S from sulfuric acid. Ferrous
sulfide was not formed in significant amounts (as would be predicted, given the pH of the
culture). Interestingly, ethanol oxidation was only about 3 mM, which was much less than
the sulfate reduced (~ 10 mM). Similar data were obtained with the Montserrat culture
(not shown). In both cultures, there were general, though sporadic, increases in protein
concentrations during culture runs, indicating that the cultures were actively growing
under these conditions (data not shown).
Figure 2. Sulfate reduction by the undefined Parys consortium, run in batch mode at
pH 4.0 Key: , sulfate; S, ethanol; , ferrous iron; , protons consumed
Sulfate reduction was observed with both cultures over the pH range 2.5-5.5. Rates of
sulfate reduction were, however, affected by culture pH, and the Parys and Montserrat
cultures showed difference responses (Figure 3). The Parys culture displayed a pH
optimum for sulfate reduction at 4.0. In contrast, rates of reduction were similar with the
Montserrat culture between pH 3.0 and 4.0 (and greater than at pH 2.5 and 4.5) and the
greatest rate of sulfate reduction was observed at pH 5.5. The shape of the graph obtained
with the Montserrat culture indicated that a "second" maximum in sulfate-reduction rate
may occur above this pH (Figure 3).
Data from the Parys culture run in continuous mode (pH 4.0; hydraulic retention time
16.5 h) is shown in Figure 4. Sulfate concentrations in the outflow liquor (which were
very similar to those in the bioreactor vessel) were 4-5 mM lower than in the inflowing
liquor, as a result of sulfate reduction. Similarly, glycerol concentrations were 3-4 mM
1182
less in the outflow than in the feed liquor, indicating that it was being utilized as the
electron donor for sulfate reduction. Acetic acid was present at about 2 mM in the
bioreactor vessel and outflow liquor, presumably as a result of incomplete oxidation of
glycerol. In cultures maintained at lower pH, glycerol oxidation again appeared to be
coupled to sulfate reduction, though concentrations of inflow/outflow glycerol were more
similar at pH 3.0 and 2.5 (indicating less active sulfate reduction) than at pH 3.5 and 4.0.
Data from the Parys culture run in continuous mode, showing the effects of culture pH and
hydraulic retention times, are shown in Table 1.
Figure 3. Sulfate reduction rates (SRR) for batch cultures of ( ) the Parys and ( )
the Montserrat undefined mixed cultures
Figure 4. Sulfidogenesis by the undefined Parys consortium run in continuous mode

(pH 4.0; HRT 16.5 h). Key: ( ) sulfate (in); ( ) sulfate (out); (S) glycerol (in); (U)
glycerol (out); (V) acetate (out)
1183
Table 1. Sulfate reduction rates (SRR) in continuous culture experiments with the
Parys undefined mixed culture, as affected by culture pH and hydraulic retention
times (HRT)
pH
2.5
3.0
3.5
4.0
4.0
4.0
HRT
(h)
66
66
66
66
33
16.5
SRR
(mol SO42- reduced/h)
40
82
166
134
340
460
Data from batch growth of the defined SRB-Acidocella culture are shown in Figure 5.
The inclusion of zinc sulfate in this instance led to the formation of zinc sulfide, thereby
reducing the amount of acid consumption (to maintain a low pH) by the culture, due to the
reaction:
Zn2+ + H2S ZnS + 2H+
Again, glycerol oxidation was coupled to sulfate reduction, which in turn resulted in
the removal from solution of stoichiometric amounts of zinc. It was also interesting that,
although some acetic acid was produced during days 1-5, this was rapidly consumed
during day 6, with concentrations falling from about 2 to <0.1 mM (Figure 5).
Figure 5. Sulfidogenesis coupled to precipitation of zinc in a defined culture of

Desulfosporosinus and Acidocella PFBC, grown in batch mode. Key: ( ) sulfate
reduced; ( ) soluble zinc; (S) glycerol; (V) acetate
4.
DISCUSSION
Dissimilatory reduction of sulfate to sulfide is of significance in two major respects,
both of which are important for the bioremediation of acidic, metal-rich waters, such as
AMD. First is the production of net alkalinity, resulting from the conversion of a strong
acid (sulfuric) to a weak acid (hydrogen sulfide), and from the generation of bicarbonate.
Secondly, sulfate reduction leads to the precipitation of a number of potentially toxic
1184
heavy metals (cadmium, copper, zinc etc.) as highly insoluble sulfides. Both of these key
features were illustrated in the present study, where mixed cultures of aSRB and other
acidophiles required addition of protons to bioreactor cultures in order to maintain the preset pH levels, and where copper and zinc were both removed from solution (copper offline by sparging the sulfide from the culture vessel with nitrogen, and zinc in situ).
Sulfidogenesis is used commercially to remediate acidic waters and to recover metals,
such as at the Budelco zinc refinery in the Netherlands [17]. However, these systems use
neutrophilic strains of SRB, and it is necessary to preclude contact between the SRB
cultures and the acidic wastewaters, which is usually achieved by stripping H2S/HS- from
the SRB reactors and allowing the sulfide to contact the waste stream in separate tanks.
Considerable savings in costs and engineering complexity could be achieved by using
single SRB/waste stream tanks, maintained at low pH to allow selective precipitation of
metals (e.g. zinc and copper, but not iron). This work has shown that this can be achieved
using acid-tolerant/acidophilic strains of SRB, grown in mixed cultures with other
acidophiles. Even though ferrous iron was present in all of the culture media used, FeS
was not formed in the pH 4.0 reactors, since its solubility product (4.0 x 10-19 M2) is
considerably greater than that of other metal sulfides such as zinc sulfide (1.0 x 10-23 M2),
which does precipitate at pH 4.0.
Other studies have indicted that, although it is possible to purify aSRB from the
consortia, the mixed cultures are inherently more robust than pure cultures (unpublished).
One reason for this is that the Desulfosporosinus isolates studied here are incomplete
oxidizers, i.e. they produce (and accumulate) acetate from ethanol and glycerol. This is a
common trait amongst SRB, but acidophiles have the added problem that organic acids
(which occur primarily as undissociated acids in low pH liquors) are toxic at relatively
low concentrations [18]. The presence of other bacteria in the anaerobic cultures might
therefore serve to lower concentrations of acetic acid to the point at which they are nontoxic. Support for this hypothesis came from the work with the defined population of
Desulfosporosinus sp. M1 and Acidocella-like acidophiles. The latter is most closely
related to Acidocella sp. WJB3, which is known to be unusually adept (for an acidophile)
at metabolizing acetic acid [19]. However, Acidocella spp. are all supposedly obligate
aerobes. Attempts to grow pure cultures of strain PFBC anaerobically on acetic acid (or
any other substrate) have all failed, yet data from molecular analysis of mixed
aSRB/PFBC cultures have shown that the anaerobic aSRB and aerobic PFBC grow
concurrently in anoxic media (data not shown). Work is currently underway to resolve this
apparent conundrum. What was apparent is that acetate concentrations in mixed cultures
of Desulfosporosinus M1/Acidocella PFBC were considerably lower than in the
"undefined" aSRB consortia (where Bacillus-, Cellulomonas- and Luteococcus-like
isolates were detected [13] and Acidocella-like bacteria were not detected).
This work has confirmed that SRB may be metabolically active in acidic liquors, and
that their potential for remediation of AMD and related wastewaters is considerable.
ACKNOWLEDGEMENTS
A.M.S. is grateful to the BBSRC (U.K.) and British Nuclear Fuels Limited (U.K.) for
provision of a research studentship. K.B.H. and D.B.J. thank the BBSRC (ref. 5/T06496)
and the DTI (U.K.; ref. BTL/20/71) for partial support of this work.
REFERENCES
1. D. B. Johnson, FEMS Microbiol. Ecol. 27 (1998) 307.
1185
2. T. A. M. Bridge and D. B. Johnson, Appl. Environ. Microbiol. 64 (1998) 2181.

3. T. A. M. Bridge and D. B. Johnson, Geomicrobiol. J. 17 (2000) 193.
4. K. Ksel, T. Dorsch, G. Acker and E. Stackebrandt, Appl. Environ. Microbiol. 65
(1999) 3633.
5. J. T. Pronk, J. C. de Bruyn, P. Bos and J. G. Kuenen, Appl. Environ. Microbiol. 58
(1992) 2227.
6. N. Ohmura, K. Sasaki, N. Matsumoto and H. Saiki, J. Bacteriol. 184 (2002) 2081.
7. J. H. Tuttle, P. R. Dugan, C. B. Macmillan and C. I. Randles, J. Bacteriol. 97 (1969)
594.
8. R. A. Gyure, A. Konopka, A. Brooks and W. Doemel, FEMS Microbiol. Ecol. 73
(1990) 193.
9. D. Fortin, M. Roy, J. P. Rioux and P. J. Thibault, FEMS Microbiol. Ecol. 33 (2000)
197.
10. K. A. Ksel, U. Roth, T. Trinkwalter and S. Peiffer, Environ. Exp. Bot. 46 (2001) 213.
11. A. M. Sen and D. B. Johnson, Biohydrometallurgy and the Environment Toward the
Mining of the 21st Century, R. Amils and A. Ballester (eds.), Elsevier, Amsterdam,
(1999) 709.
12. . Kolmert and D. B. Johnson, J. Chem. Technol. Biotechnol. 76 (2001) 836.
13. A. M. Sen, Ph.D. Thesis, University of Wales, Bangor, (2001).
14. F. Widdel and N. Pfennig, Arch. Microbiol. 129 (1981) 395.
15. M. M. Bradford, Anal. Biochem. 72 (1976) 248.
16. D. R. Lovley and E. J. P. Phillips, Appl. Environ. Microbiol. 53 (1987) 1536.
17. L. J. Barnes, P. J. M. Scheeren and C. J. N. Buisman, Emerging Technology for
Bioremediation of Metals, J. L. Means and R. E. Hinchee (eds.), Lewis Publishers,
Boca Raton, Fl., (1994) 38.
18. B. Alexander, S. Leach and W. J. Ingledew, J. Gen. Microbiol. 133 (1987) 1171.
19. K. B. Hallberg, . K. Kolmert, D. B. Johnson and P. A. Williams, Biohydrometallurgy
and the Environment Toward the Mining of the 21st Century, R. Amils and A.
Ballester (eds.), Elsevier, Amsterdam, (1999) 719.
1186

Sulfur assimilation in Acidithiobacillus ferrooxidans

J. Valds1, E. Jedlicki2 and D.S. Holmes1,3
1
Laboratory of Bioinformatics & Genome Biology, University of Santiago, Santiago, Chile,

2
Program of Cellular and Molecular Biology, University of Chile, Santiago, Chile
3
Millenium Institute of Fundamental and Applied Biology, Santiago, Chile
Abstract
A bioinformatic analysis of the nearly complete genome sequence of Acidithiobacillus
ferrooxidans, available from The Institute for Genome Research (TIGR) and Integrated
Genomics (IG), reveals the presence of a number of genes potentially involved in sulfur
uptake and metabolism. The coding potential of these genes is compared to similar genes
for known sulfur metabolic pathways in other microorganisms. In addition, the proposed
sulfur assimilation genes of A. ferrooxidans are shown to be present in operons in an
organization similar to that found in some other organisms. Genes potentially involved in
the regulation of sulfur uptake and assimilation have also been detected. Based upon these
observations, we present a metabolic model of sulfur uptake and assimilation in A.
ferrooxidans that is substantially similar to the genetic organization and metabolic
pathway structure that has been experimentally demonstrated to function in other
organisms. An interesting exception to this similarity is the observation that A.
ferrooxidans appears to have the genetic potential to encode for two different pathways of
sulfur activation whereas many organism only have one such pathway. In addition, an
interesting gene fusion event seems to have taken place in A. ferrooxidans involving the
joining of a cysN gene, encoding a sulfurylase enzyme subunit B, with a cysC-like gene
encoding an APS activity.
Keywords: sulfur assimilation, sulfur regulation, cysteine biosynthesis, genome analysis,

Acidithiobacillus ferrooxidans
1.
INTRODUCTION
The assimilation of sulfur is an essential process in all cells being required for the
formation of the amino acids, methionine and cysteine. Sulfur is also used for the
construction of FeS centers in proteins involved in electron transfer processes (1) and is
found in certain membrane components such as the nod factor of the plant symbiont
Rhizobium sp. (2) and in various sulfolipids of Mycobacterium sp. that are thought to
bevirulence factors (3).
* Corresponding author: David Holmes: dsholmes2000@yahoo.com. Work supported by Fondecyt No.
1010623 and the Millenium Institute for Fundamental and Applied Biology, Santiago, Chile. We thank the
Institute of Genome Research (TIGR) and Integrated Genomics, Inc. (IG) for the use of their partial
sequence of the Acidithiobacillus ferrooxidans genome.
1187
The process of sulfur assimilation, in all organisms begins with the uptake of sulfate
ions from the environment through the cellular membrane via an ABC-type membrane
pump mechanism. Sulfate ions are generally available in most environments and are
abundantly available in mining environments rich in sulfuric acid.
After entry into the cell, the sulfate is made biologically available by activation, a
process that involves the reaction between sulfate and ATP, catalyzed by the enzyme ATP
sulfurylase to yield adenosine 5-phosphosulfate (APS) and PPi. A second enzyme, APS
kinase, can further phosphorylate APS with another ATP to give 3-phosphoadenosine 5phosphosulfate (PAPS)(1). In plants and phototrophic bacteria the intermediate of
assimilative reduction is APS, but in fungi and some chemotrophic bacteria it is PAPS.
Depending on the organism, these intermediates are then further reduced and fed into
pathways that are shared in all organisms leading to the formation of the amino acids
cysteine and methionine. The reason for the evolution of two mechanisms for sulfur
assimilation by different organisms is unclear. The enzymes involved in the formation of
APS or PAPS show significant sequence similarity and may evolve from a common
ancestor.
We have carried out a bioinformatic analysis of the almost complete genome
sequence of A. ferrooxidans in an attempt to identify candidate genes potentially involved
in sulfur assimilation. Using this information, we suggest possible biochemical pathways
for sulfur assimilation and for the regulation of sulfur uptake. It is beyond the scope of this
paper to connect the proposed sulfur assimilation pathways with proposed sulfur oxidation
genes and pathways in A. ferrooxidans. However, the present paper will aid future studies
designed not only to make these connections but ultimately to deduce the interconnected
regulatory mechanism that balance sulfur assimilation with sulfur oxidation in a variety of
environmental situations.
2.
METHODS
Known metabolic pathways of sulfur assimilation were obtained from BIOCYC
(http://biocyc.org:1555/META/server.html), KEGG (http://genome.ad.jp/kegg/kegg4.
html) and ERGO (http://wit.integratedgenomincs.com/WIT2/CGI). Amino acid sequences
derived from genes identified as being involved in sulfate acquisition were used as query
sequences to search the partial genome sequence of A. ferrooxidans ATCC 23270 in the
TIGR (http:// www.tigr.org/) and ERGO data bases using TBlastN and BlastP
respectively. When a prospective candidate gene was identified in TIGR or ERGO its
predicted amino acid sequence was then used to formulate a BlastP
(http://www.ncbi.nlm.nhi.gov/BlastP/) search of the non-redundant database at NCBI.
Only bidirectional best hits were accepted as evidence for putative homologs. Candidate
genes and their translated proteins were further characterized employing the following
bioinformatic tools available in the web: primary structure similarity relations
(http://www.ebi.c.uk/ClustalW/), secondary structure predictions JPred http://www.
compbio.dundee.ac.uk/Software/JPred/jpred.html),
transmembrane
predictions
(http://www.ch.embnet.org/software/TMPRED_form.html), motif predictions (http://
www.blocks.fhcrc.org/, http://www.ebi.ac.uk/interpro/, and domain predictions (http://
www.tolouse.inra.fr/prodom.htlm), and prediction of protein localization sites
(http://psort.nibb.ac.jp/).
1188
3.

Sulfur assimilation, in most organisms, typically starts with sulfate uptake from the
environment by a membrane complex that includes an ABC-type membrane pump (Figure
1). A bioinformatic analysis of the predicted genes of A. ferrooxidans reveals the presence
of candidate genes that could encode the proteins for such a complex. One of these genes
encodes a possible sulfate-permease protein and lies adjacent to a gene potentially
encoding an ABC-type ATP binding membrane pump and a protein of unknown function.
The three genes appear to be organized as an operon (data not shown) lending support to
the idea that they encode proteins involved in sulfate uptake. On the other hand a positive
identification of genes and proteins involved in sulfate uptake is difficult because ABCtype membrane pump proteins belong to a large family of paralogous genes that encode
proteins with similar sequences and structures but that differ in the substrate that they
transport. Thus bioinformatic identification of an ABC transporter for sulfate uptake by
sequence similarity is insufficient to pinpoint its function and experimental validation will
be essential.
On the other hand, genes involved in the subsequent activation and utilization of
sulfur do not appear to belong to extensive paralogous families. In addition, the genes and
pathways of sulfur assimilation are quite conserved among organisms. These
considerations make it possible not only to identify candidate orthologous genes in A.
ferrooxidans by bioinformatic analysis but also to have a reasonable degree of confidence
in such assignations. It also allows us to propose a biochemical model of sulfur
assimilation for A. ferrooxidans. It should be stressed that such a model requires
experimental validation before it can be accepted. Such experiments are now underway
and will be reported in the future. However, it is useful to note that, in the absence of well
developed methods of gene transfer in A. ferrooxidans and the concomitant paucity of
genetic analyses, metabolic models derived by bioinformatics provide an important way to
investigate the potential biochemistry and genetics of the organism. Such studies have
already revealed valuable information regarding the amino acid biosynthesis (4), nitrogen
metabolism, metal fluxes and other characteristics of A. ferrooxidans. (5).
Table 1. Identification and some properties of the genes predicted to be involved in
sulfur assimilation in A. ferrooxidans. (A) Proposed gene name. (B) Putative function
assigned to each gene. (C) The best BlastP hit of each candidate gene to orthologs in
GenBank and the organism in which the ortholog is found and (D) the percent
similarity of the amino acid sequence of the candidate gene with its ortholog from
GenBank
A. Gene
CysD1
CysD2
CysN1
CysN2
CysN3
CysH
CysI
CysJ
CysB1
CysB2
B. Putative function assigned

ATP sulfurylase, catalitic subunit
ATP sulfurylase, catalitic subunit
ATP sulfurylase, GTP-binding subunit
ATP sulfurylase/Adenylylsulfate kinase
ATP sulfurylase/Adenylylsulfate kinase
APS reductase
Sulfite reductase hemoprotein
Sulfite reductase flavoprotein
Transcriptional regulator
Transcriptional regulator
C. Best Blastp hit and

organism
Neisseria meningitidis
Mesorhizobium loti
Neisseria meningitidis
Mesorhizobium loti
Coxiella burnetii
Allochromatium vinosum
Bacillus subtilis
Bacillus subtilis
Pseudomonas aeruginosa
Pseudomonas aeruginosa
D. %
Similarity
71
78
69
72
61
74
67
59
69
52
1189
A. Gene
CysQ
CysA
CysW
CysM
B. Putative function assigned

PAPS catabolism
Sulfate /thiosulfate transport
Sulfate/ thiosulfate transport
O-acetyl serine (thiol)-lyase-B
C. Best Blastp hit and

organism
Aquifex aeolicus
Mesorhizobium loti
Halobacterium sp.
Pseudomonas putida
D. %
Similarity
72
53
54
70
In contrast to many organisms, where sulfur activation and subsequent reduction

occurs either via APS or PAPS, candidate genes have been identified A. ferrooxidans that
are predicted to encode enzymes involved in both APS and PAPS formation (Figure 1 and
Table 1). In addition, genes have been identified for proteins involved in the subsequent
utilization of APS and PAPS and also in the regulation of the uptake and incorporation of
sulfur (Figure 1 and Table 1).
Figure 1. Proposed biochemical pathway and genes for the uptake of sulfate and the
assimilation of sulfur in A. ferrooxidans based upon bioinformatics analysis of its
genome. Potential genes are shown in parentheses. The reason for the dotted line that
connects PAPS and sulfite is discussed in the text. Also shown is the feedback
inhibition of serine acetyl transferase by cysteine and the regulation of the sulfur
uptake and assimilation genes by N-acetyl-L-serine which is discussed in the text
The levels of similarity (Table 1) of the predicted A ferrooxidans amino acid
sequences to known proteins involved in sulfur activation are sufficient to allow a
reasonable degree of confidence in the assignments made. In addition, many of the
candidate sulfur assimilation genes appear to have an operon-like organization similar to
those found in other organisms (Figures 2 and 3). Operon-like organization is a strong
predictor of similar function for the genes within the operon. Thus, the finding that many
1190
of the candidate sulfur assimilation genes of A. ferrooxidans are display such organization
strengthens the likelihood of the assignments of gene function which we have made using
bioinformatics predictions based on comparative protein sequence similarity.
The formation of APS in A. ferrooxidans could be accomplished by the protein
products of cysD1 and cysN1, which, in E. coli, encode the heterodimeric enzyme ATP
sulfurylase (6). In both A. ferrooxidans (Figure 2) and E. coli cysD1 and cysN1 appear to
be organized in an operon. However, the E. coli operon also includes the gene cysC that
catalyzes the phosphorylation of APS to give PAPS. This gene appears to be absent in the
equivalent A. ferrooxidans operon. However, the A. ferrooxidans genome has an
additional two copies of cysN1, which we term cysN2 and cysN3 (Figures 1,2 and Table
1). Interestingly, a detailed examination of the predicted protein sequences encoded by
cysN2 and cysN3 reveals the presence of an additional domain in their C terminal regions
that exhibit significant sequence similarity to the E. coli cysC (Figure 3). This raises the
possibility that the A. ferrooxidans cysN2 and cysN3 encode bi-functional proteins formed
by a gene fusion event between cysN1 and cysC and that the resulting fused protein can
assume the function of the missing cysC catalyzing the formation of PAPS. A similar
bifunctional protein has also been observed in Rhizobium sp. (7). Additional support for
this conjecture comes from the observation that a cloned segment of A. ferrooxidans DNA
was able to complement an E. coli cysC mutant, demonstrating that the organism has the
capacity to encode a cysC-like function (8).
It appears that a transpose-encoding element (Tpa, Figure 2) has inserted between
cysN2 and cysQ that may interrupt the coordinated regulation of the cysDNQ operon. The
significance of this observation is not known.
Upstream of cysD1 and cysN1 in the A. ferrooxidans genome there are candidate
genes for cysH, cysI and cysJ organized in an operon-like structure (Figure 2). Many
microorganisms, such as E. coli and S. typhimurium, have a similar gene organization (1).
cysI and cysJ encode a heterodimeric sulfite reductase with the form 48. This catalyzes
the reduction of sulfite to sulfide which s the final reduction state of sulfur for its
assimilation and the putative cysI and cysJ in A. ferrooxidans can be assigned with
reasonable confidence.
The role of cysH in A. ferrooxidans is more enigmatic. In E. coli, cysH encodes a
PAPS reductase without APS reductase activity, but in the plant A. thaliana the cysH
product is exclusively an APS reductase. In the bacteria Pseudomonas putida and
Mycobacterium spp., cysH exhibits both APS and PAPS reductase activity, but with
differing levels of activity (9). The A. ferrooxidans cysH ortholog exhibits a motif
characteristic of APS reductase (data not shown) (10), suggesting that it is has at least an
APS reductase activity but leaves open, for the present, the possibility that it also has a
PAPS reductase activity. For this reason we have drawn in Figure 1 a solid line connecting
APS reduction with cysH, signifying a reasonable level of confidence in the assignment,
and we have placed a dotted line connecting PAPS reduction with cysH to emphasize that
some ambiguity remains regarding cysH function.
The final step in sulfur assimilation is the synthesis of L-cysteine from O-acetyl-Lserine and sulfide. This reaction is catalyzed by one or both of two O-acetyl-L-serine
(thiol)-lyase isozymes, designated A and B and encoded by cysK and cysM genes in S.
typhimurium and E. coli (1). A. ferrooxidans has a potential cysM ortholog that is
predicted to accomplish this final step in sulfur assimilation (Table 1, Figures 1 and 2).
1191
Figure 2. Examples of two operon-like organizations of the candidate genes of A.

ferrooxidans postulated to be involved in sulfate uptake and sulfur assimilation. The
proposed function of the various genes designated with prefix "cys" can be found in
Figure 1. tpa = probable transposase. hyp = hypothetical gene of unknown function
Figure 3. Comparison between the organization and function of the genes found in
the cysDNC operon of E. coli with similar genes of A. ferrooxidans. In two instances
in A. ferrooxidans a cysN gene (N2 and N3), potentially encoding a sulfurylase
subunit, appears to be fused with a gene with significant similarity to the cysC gene
of E. coli encoding an APS activity
In addition to exhibiting potential orthologs of genes involved in sulfur assimilation,
A. ferrooxidans also appears to have characteristic mechanisms for the regulation of the
relevant genes and enzymes. In plants and enteric bacteria the activity of the sulfur uptake
and assimilation genes is regulated by a combination of feedback inhibition of serine
transacetylase by cysteine and a gene regulatory system called the cysteine regulon (1).
Excess cysteine can bind to acetylserine transferase reducing the activity of this enzyme,
which, in turn, limits the production of O-acetyl-L serine. Since this latter compound is a
precursor of cysteine its reduction causes a decrease of the synthesis of cysteine. Since A.
ferrooxidans appears to have the gene encoding acetylserine transferase it can be assumed
that it is subject to a similar feedback inhibition mechanism.
On the other hand, a reduction in the availability of sulfide leads to the accumulation
of O-acetylserine, which spontaneously isomerizes to N-acetylserine. This latter
compound binds to cysB and the resulting complex serves as an activator of several genes
involved in sulfate uptake and sulfur assimilation (11, 12). cysB is a transcriptional
regulator that belongs to the lysR family of transcriptional activators. A ferrooxidans has
1192
two copies of this gene both of which display the characteristic helix-turn-helix motif. of
this family. Binding of N-acetylserine to cysB promotes the activation of genes involved
in sulfate uptake and sulfur assimilation (1).
A third mechanism of regulation involves the control of levels of PAPS carried out by
the cysQ gene product. Reports indicates that accumulation of PAPS is toxic for the cell
(13). In A. ferrooxidans we found an ortholog of cysQ associated in an operon-like
organization with cysD2 and cysN2 genes (Figure 2). A similar organization can be
detected in Brucella abortus by inspection of the Integrated Genomics Inc. database of
genome sequences (http://wit.integratedgenomics.com/ WIT2/CGI).
3.
CONCLUSIONS
Genes and enzymes predicted to be responsible for sulfur uptake and assimilation in
A. ferrooxidans exhibit significant similarity, including conservation of key protein motifs,
to those found in a number of other microorganisms. This permits a metabolic model to be
constructed that suggests that sulfur assimilation in A. ferrooxidans exhibits substantial
similarity to other organisms. The operon-like organization of these genes in A.
ferrooxidans is also similar to that observed in other microorganisms. In addition, A.
ferrooxidans has a repertoire of genes predicted to be involved in the regulation of sulfur
assimilation and this characteristic, together with the operonic organization of the relevant
genes, implies that key aspects of the genetic and biochemical regulation of sulfur
assimilation are shared between A. ferrooxidans and other organisms.
However, two important differences between A. ferrooxidans and other organisms
stand out. First, it appears that A. ferrooxidans may have encoded enzymes that permit
sulfur to be activated by both APS and PAPS intermediates, a characteristic that is shared
by some but not all organisms. Secondly, compared to most organisms, A. ferrooxidans
exhibits a probable gene fusion event in which an ATP sulfurylase-encoding gene has
been joined to a gene encoding an APS kinase activity. The recognition and analysis of
gene fusion events are becoming increasingly important for building models of gene
evolution. In addition, gene fusions, known as Rosetta stone sequences, serve as powerful
predictors of gene and protein function in bionformatic analyses (14).
The model of sulfur assimilation shown in Figure 1 was inferred by a range of
bioinformatic inferences, including gene similarity comparisons, detection of possible
operons and identification of putative genes involved in the regulation of sulfur
assimilation. However, the model requires experimental validation, some of which is now
underway in our laboratory.
In addition to sulfur assimilation, A. ferrooxidans uses sulfur as an energy and
electron source. Several laboratories are actively engaged in deducing the transduction of
sulfur to energy in A. ferrooxidans and a comprehensive model of sulfur utilization this
organism will require an understanding of the metabolic and regulatory connections
between the two pathways.
REFERENCES
1. Kredich, N. M. Biosynthesis of cysteine, Escherichia coli and Salmonella
typhimurium: cellular and molecular biology, In F. C. Neidhardt, J. L. Ingraham, K.B.
Low, B. Magasanic, C. M. Schaechter, and H. E. Umbarger (ed.). vol. 1, pgs. 419-428.
(1987).
2. Dnari, J., Debell, F, and Rosemberg, C. Annu. Rev. Microbiol. 46, pgs. 497-531.
(1992).
1193
3. Williams, S.J., Senaratne, R. H., Mougous, J. D., Riley, L. W., Bertozzi, C. R. J.

Bacteriol. 277, pgs.32606-32615.(2002)
4. Selkov, E., Overbeek, R., Kogan, Y., Chu, L., Vonstein, V., Holmes, D., Silver, S.,
Haselkorn, R., Fonstein, M. PNAS. 7, pgs. 3509-3514. (2000).
5. Barreto, M., R. Quatrini, S. Bueno, C. Arriagada, J. Valdes, S. Silver, E. Jedlicki and
D. S. Holmes. (In press, Hydrometallurgy, 2003).
6. Leyh, T. S., Taylor, J. C., Markman, G. D. J. Biol. Chem. 263, pgs. 2409-2416. (1988)
7. Schwedock, J. S., Liu, C., Leyh, T. S., Long, S. R. J. Bacteriol. 176, pgs. 7055-7064.
(1994).
8. Fry,I. J., and E. Garcia. In J. Salley, R. G. L. McCready, and P. L. Wichlacz (ed.),
Biohydrometallurgy. pgs. 171-185. (1989).
9. Bick, J. A., Dennis, J. J., Zylstra, G. J., Nowack J, Leustek T. J. Bacteriol. 182,
pgs.135-142. (2000).
10. Kopriva, S., Buchert, T., Fritz, G., Suter, M., Benda, R., Schunemann, V., Koprivova,
A., Schurmann, P., Trautwein, AX., Kroneck, PM., Brunold, C. J. Biol. Chem. 277.
pgs. 21786-21791. (2002).
11. Hryniewicz, M. M., and Kredich, N. M. J. Bacteriol. 173, 5876-5886. (1991).
12. Monroe, R.S., Ostrwski, J., Hryniewicz, M. M., Kredich, N. M. 172, pgs.6919-6929.
(1990).
13. Neuwald, A. F., Krishnan, B. R., Brikum, I., Kulakauskas, S., Suziedelis, K.,
Tomcsanyi, T., Leyh, T. S., Berg, D.E. J. Bacteriol. 174, pgs. 415-425. (1992).
14. Marcotte, E. M., Pellegrini, M., Ng HL., Rice, D. W., Yeates, T. O., Eisenberg, D.
Science.285, pgs.751-3. (1999).
1194

Survival of acidophilic bacteria under conditions of substrate

depletion that occur during culture storage
D. Barrie Johnsona, Debby F. Bruhnb and Francisco F. Robertob
a

Lockheed Idaho Technologies Co., Idaho National Engineering and Environmental
Laboratory, P.O.Box 1625, Idaho Falls, Idaho, 83415-22032
Abstract
Survival of four mesophilic acidophilic bacteria (two chemolithotrophs,
Acidithiobacillus ferrooxidans and Leptospirillum ferrooxidans, and two heterotrophs,
"Ferrimicrobium acidiphilum" and Acidiphilium SJH) was investigated during 500 days of
starvation stress. Counts of plateable At. ferrooxidans remained fairly stable for 25 days
following completion of ferrous iron oxidation, after which they declined; however, 2 x
104 viable cells/ml stored culture were still present after 500 days of starvation stress. In
contrast, numbers of viable L. ferrooxidans went into immediate decline following iron
oxidation, and no viable planktonic cells were detected 45 days after iron oxidation. Pure
cultures of the two heterotrophic acidophiles displayed similar trends to At. ferrooxidans.
When mixed populations of acidophilic mesophiles were subjected to long-term starvation
stress, mortality rates of At. ferrooxidans and Acidiphilium SJH were greater, and those of
L. ferrooxidans and "F. acidiphilum" less than in corresponding pure cultures. The
significance of metabolic diversity, sensitivities to organic compounds and scavenging
potential on the survival of acidophilic bacteria is discussed.
Keywords: acidophilic bacteria; mixed cultures; pH stress; starvation stress

1.
INTRODUCTION
Bacteria that inhabit acidic, metal-rich environments are physiologically diverse, in
their carbon assimilation, temperature-response characteristics and abilities to catalyze the
dissimilatory transformations of iron and sulfur [1]. Interest in this group stems primarily
from their role in the genesis of acid mine drainage [2] and in their exploitation in mineral
processing operations [3]. Besides being characterized by low pH (generally 2-4),
environments populated by acidophilic microorganisms generally contain large
concentrations of sulfate, iron, other (heavy) metals and metalloids (depending on the
geochemistry of the local environment), and small concentrations of dissolved organic
carbon. Bacteria such as the iron-oxidizing chemolithotrophs Acidithiobacillus
ferrooxidans and Leptospirillum spp. are obligate acidophiles, and are inactive in neutral
and mildly acidic environments. A variety of interactions between acidophilic
microorganisms have been described [4]. For example, mutualistic associations occur
between autotrophic and heterotrophic acidophiles, whereby utilization of organic
materials (e.g. leakage products from autotrophic cells) by the latter "detoxify" the
1195
environment for acidophilic autotrophs, which are often highly sensitive to small
molecular weight organic compounds.
The effects of nutrient stress and substrate depletion on the survival of bacteria are of
fundamental importance in microbial ecology. Long-term survival in environments that
are depleted in nutrients, and which also may be stressful in other ways (e.g. pH,
temperature), may result from bacteria having specific survival strategies. Although
acidophilic bacteria are important in ore leaching and genesis of acid mine drainage
pollution, relatively little has been published on their abilities to survive starvation stress.
Zychlinsky and Matin [5] found that survival of the acidophilic mixotroph Acidiphilium
acidophilum during "long-term starvation" (up to 22 days) was similar to that reported for
neutrophilic bacteria. In this paper, we describe the responses of four mesophilic
acidophiles to starvation stress for periods of up to 500 days.
2.
2.1 Bacteria
Survival of four mesophilic acidophiles was investigated: (i) At. ferrooxidans (ATCC
23270; the type strain); L. ferrooxidans (strain CF12, isolated from the Noranda
Blackbird-cobalt mine, Cobalt, Idaho); (iii) "Ferrimicrobium acidiphilum" (strain T-23
[1]); (iv) Acidiphilium SJH [1]. Two of these bacteria are autotrophs (At. ferrooxidans and
L. ferrooxidans) and two are heterotrophs ("F. acidiphilum" and Acidiphilium SJH). Three
species (At. ferrooxidans, L. ferrooxidans and "F. acidiphilum") oxidize ferrous iron to
ferric, and three (Acidiphilium SJH, F. acidophilum and At. ferrooxidans) can reduce
ferric iron to ferrous in oxygen-limited environments.
2.2 Experimental protocols
Pure cultures of each acidophile, and a mixed culture containing all four, were grown
in 20 mM ferrous sulfate liquid medium, poised initially at pH 1.8 (to avoid precipitation
of ferric compounds in oxidized cultures). Cultures (250 ml in 500 ml conical flasks) were
incubated (shaken) for three days at 30C, and subsequently stored at room temperature
(23 2C) in the dark. Flasks were loosely covered with aluminum foil to reduce
evaporation losses.
A second experiment was set up to monitor the effect of pH on the survival of
mesophilic acidophiles under conditions of nutrient depletion. Individual cultures of the
four bacteria were grown in appropriate liquid media (20 mM ferrous sulfate for At.
ferrooxidans and L. ferrooxidans; 20 mM ferrous sulfate/0.02% (wt/vol) yeast extract for
"F. acidiphilum"; 0.002% (wt/vol) yeast extract for Acidiphilium SJH). When cell
numbers had reached approximately 107/ml, bacteria were harvested aseptically by
centrifugation, washed, and re-suspended in acidified (pH 2.5) basal salts. Aliquots (0.5
ml) of cell suspensions were added, separately, to basal salts solutions (200 ml in 500 ml
flasks) adjusted at either pH 2.0, 5.0 or 8.0, using dilute H2SO4 or NaOH. Cultures were
incubated, unshaken, at room temperature in the dark. All experiments used duplicate
cultures.
2.3 Analytical techniques
Total numbers of bacteria were assessed by acridine orange direct counts (AODCs)
using a modification of the method of Hobbie et al. [6]. This involved filtering culture
aliquots through black polycarbonate membranes (0.2 m pore size; Nuclepore, U.S.A.),
1196
rinsing first with acidified (pH 1.0; H2SO4) distilled water and then with alkaline (pH
11.0; NaOH) water, fixing cells with 4% (v/v) glutaraldehyde (10 mins.) followed by a
second pH 11.0 water rinse, staining with 0.01% (w/v) acridine orange solution (pH 11.0;
6 mins.) and a final rinse in distilled water. Counts of viable bacteria were made by plating
diluted culture aliquots onto selective solid media [7]. These were ferrous iron overlay
medium (Feo) for enumerating At. ferrooxidans, L. ferrooxidans and "F. acidiphilum", and
yeast extract medium for Acidiphilium SJH. Identification of bacteria was made on the
basis of differences in colony morphologies, combined with microscopic examination, as
described previously [7]. To prepare samples for scanning electron microscopy (SEM),
culture aliquots (containing 106-107 cells/ml) were filtered through Nuclepore membranes
(0.2 m pore size) and rinsed, sequentially, with 0.1 M H2SO4, acidified (pH 2.0) basal
salts and "nanopure" water. After fixing with 5% (v/v) glutaraldehyde, cells were
dehydrated by flushing with solutions of ethanol (30%, 60%, 90%, 100% (x2); all vol/vol)
and critical point dried. Bacteria were viewed in an Amray Model 1830 SEM; cell
dimensions were recorded and biovolumes calculated.
Other analyses included determination of ferrous iron concentrations (by titration with
1 mM KMnO4 in dilute H2SO4), dissolved organic carbon (samples pre-filtered through
0.2 m membranes and DOC estimated using an OI Corporation Model 700 DOC
analyzer), and pH using a glass electrode.
3.
RESULTS
3.1 Survival trends in pure cultures of acidophiles

The acidophilic bacteria used in these experiments varied widely in their abilities to
survive under conditions of substrate depletion. Their survival was affected by the
presence of other bacteria (i.e. mixed versus pure cultures) and by environmental factors.
Figure 1 shows how total counts of all four acidophiles remained fairly constant during the
first 100 days of substrate depletion while, in contrast, counts of cultivatable bacteria
differed markedly between isolates. Counts of plateable At. ferrooxidans remained fairly
static for ca. 25 days following completion of iron oxidation (which took 3 days) after
which time they declined, though significant numbers of cultivatable cells (ca. 2 x 104/ml)
were still present some 500 days after the start of the experiment. In contrast, plate counts
of the other iron-oxidizing chemolithotroph, L. ferrooxidans, declined dramatically
following iron oxidation, and no cultivatable bacteria were recovered on day 45. However,
cultures plated after 87 and 100 days did produce colonies (confirmed as L. ferrooxidans),
though none was detected beyond these times. Viable counts of pure cultures of the two
acidophilic heterotrophic isolates also displayed interesting differences. Plate counts of "F.
acidiphilum" showed an initial decline in contrast to Acidiphilium SJH whose numbers
increased during this time, though these stabilized from day 28 and, by the end of the
experiment, viable counts of "F. acidiphilum" were similar to those of both At.
ferrooxidans and Acidiphilium SJH. One difference between "F. acidiphilum" and the
other iron-oxidizing acidophiles is the former's more protracted ferrous iron oxidation
when cultured in "inorganic" media. Whereas iron oxidation was complete in cultures of
At. ferrooxidans and L. ferrooxidans by the time that cultures were put into storage (3 days
after inoculation), residual ferrous iron concentrations in cultures of "F. acidophilum"
were 7 mM (after 9 days) and 3 mM (after 17 days). Slow ferrous iron oxidation is a
feature of iron-oxidizing heterotrophs when grown in the absence of organic carbon (D.B.
Johnson, unpublished data).
1197
Figure 1. Effect of starvation-stress on, (a) total bacterial numbers (AODCs) and (b)
culturable bacteria. Symbols: (S) At. ferrooxidans; () L. ferrooxidans; (z) F.
acidophilum; (T) Acidiphilium SJH. Error bars represent standard deviations
Table 1. Changes in cell dimensions* of acidophilic bacteria stored under conditions
of substrate limitation
Three of the acidophiles displayed net decreases in cell biovolumes after 37 days
compared with those recorded at day 3 (Table 1). The single exception was Acidiphilium
SJH, whose cells were longer and thinner at day 37 but which showed no net change in
biovolume. Decrease in mean cell biovolume was smallest for At. ferrooxidans among the
iron-oxidizing acidophiles. Net biovolume reductions were similar for L. ferrooxidans and
"F. acidophilum" during this time; however, while L. ferrooxidans cells had shrunk by
similar proportions in both length and width, those of "F. acidophilum" were much shorter
than at day 3, though of similar width.
1198
3.2 Survival trends in mixed cultures of acidophiles

Patterns of bacterial survival were subtly different when the acidophiles were grown
in mixed culture (Fig. 2). In particular, numbers of plateable L. ferrooxidans declined at a
slower rate than was observed in pure cultures, and plate counts of "F.acidophilum"
increased (rather than decreased) during the first 25-30 days of culture storage. In contrast,
numbers of viable At. ferrooxidans and Acidiphilium SJH went into more immediate
decline than in pure cultures of these acidophiles. As with pure cultures, after 500 days
incubation viable cells of all acidophiles (except L. ferrooxidans) were detected in mixed
cultures, with "F. acidiphilum" being the most abundant viable acidophile.
Figure 2. Comparison of plate counts of acidophiles, grown in pure and mixed

cultures. Data shown are for changes in populations that occurred during the first 45
days of starvation-stress. Error bars represent standard deviations. (a) At.
ferrooxidans; (b) L. ferrooxidans; (c) `F. acidiphilum; (d) Acidiphilium SJH
3.3 Survival trends of acidophiles at different pH
The pH of the substrate-depleted media in which cultures of bacteria were maintained
had a major influence on the survival rates of the mesophilic acidophiles. Because of the
poor buffering capacity of the basal salts solutions, the addition of the bacterial
suspensions caused a shift in pH in all cases, resulting in pH values of 2.15, 4.10 and 6.20
for media adjusted initially to pH 2.0, 5.0 and 8.0, respectively. Subsequent changes in pH
in these cultures were minor (<0.1 pH unit). Concentrations of dissolved organic carbon
(ca. 4-5 mg/L at the start of the experiment) were about 50% of levels recorded in the first
experiment since, in the latter, the culture liquor was spent bacterial medium (which
would have accumulated cell leakage products etc.) rather than basal salts solution.
Changes in viable bacterial counts recorded in this experiment are shown in Fig. 3. No
colonies of L. ferrooxidans were recovered from any culture in this experiment, indicating
either that the methodology used in harvesting cells of this acidophile proved lethal, or
1199
that L. ferrooxidans lost viability very rapidly in the basal salts solutions. At pH 2.15, the
overall trends (Fig. 3a) were similar to those found in experiment 1. Viable counts of "F.
acidiphilum" exceeded those of other bacteria from day 14 onwards, and plate counts of
this heterotrophic iron-oxidizer remained fairly stable throughout the 100 days of culture
storage. Viability of At. ferrooxidans fell sharply during the first 40 days of culture
storage, but numbers stabilized at about 104 cells/ml by day 100. Counts of viable
Acidiphilium SJH, in contrast, were fairly stable for the first 40 days of storage, but then
went into severe decline, and no viable cells were detected at day 100. At pH 4.1, patterns
of survival of the three acidophiles were quite different, with both At. ferrooxidans and
Acidiphilium SJH displaying more prolonged cell viability than at the lower pH; in
contrast, plate counts of "F. acidiphilum" declined rapidly and from day 72 onwards no
viable cells of this acidophile were detected (Fig. 3b). At the highest pH (6.2), the
mortality rate of "F. acidiphilum" was even more pronounced. At. ferrooxidans also died
out more rapidly at this pH than at either pH 2.15 or 4.1, but Acidiphilium SJH survived in
reasonable numbers (ca. 103-104/ml) throughout the period of culture storage (Fig. 3c).
4.
DISCUSSION
Microorganisms have developed a variety of strategies for surviving times of
depletion of substrate or of a particular essential nutrient [8]. This may involve the
possession and expression of "survival genes", cell proliferation, and reduction in cell size
(producing "ultramicrocells" of <0.3 m diameter). Various methods have been used to
differentiate between viable and non-viable cells in a population [9]. The use of plate
counts to enumerate viable cells has been criticized, since bacteria may be able to take up
and metabolize nutrients without resuming growth and division under the imposed
conditions of plating. Oliver [10] described "viable but not culturable" cells as those that
can be demonstrated to be metabolically active but which are incapable of sustaining
cellular division in media that normally support growth. The solid media and experimental
techniques used in the present study have been developed over several years specifically
to produce high plating efficiencies of acidophilic bacteria from environmental materials
(as well as to facilitate identification of isolates). Numerical estimations of iron-oxidizing
acidophiles in laboratory cultures and environmental samples (where bacteria are
frequently in a state of substrate limitation) as determined using "overlay" plates have
been shown to be similar or greater than the "Most Probable Number" approach [7]; the
latter uses liquid media that might be considered to be more conducive to the growth of
stressed cells. It is worth noting that the stored cultures that produced zero plate counts of
L. ferrooxidans also proved negative (i.e. no iron oxidation) when samples were
subcultured in liquid media. Oliver [10] found that "starved" cells responded quickly to
the reversal of nutrient limitation whereas the response in viable non-culturable cells was
slow; zero response, as described for L. ferrooxidans, indicates that cells were indeed nonviable.
Trends of starvation-survival displayed by the four species of mesophilic acidophiles
used in the present work were often very different. The chemolithotrophic iron-oxidizer,
L. ferrooxidans displayed much more limited longevity than At. ferrooxidans. This may be
due either to an inherently greater capacity of At. ferrooxidans to survive in the absence of
extraneous substrates (using storage reserves etc.) or to its greater metabolic versatility.
Ferrous iron is the only substrate known to be utilized by L. ferrooxidans; whilst it is
possible that some ferrous iron might have been regenerated during culture storage, this is
highly unlikely. Even though both Acidiphilium SJH and "F. acidiphilum" can couple the
oxidation of organic substrates to the reduction of ferric iron [4] this is generally only
1200
Figure 3. Effect of pH on long-term survival of mesophilic acidophiles. Symbols: (U)

At. ferrooxidans; ({) "F. acidiphilum"; (V) Acidiphilium SJH. Error bars represent
standard deviations
observed under conditions of limited dissolved oxygen concentrations, for example, in
cultures containing extraneous organic materials. Since no reduced sulfur compounds
1201
were present in the spent media or the basal salts suspension, the only other alternative
potential substrates would have been organic. These would have arisen from leakage from
active cells, from lysis from dead and dying cells, and as atmospheric inputs. While At.
ferrooxidans is often considered to be an obligate autotroph, its ability to incorporate
methionine [11] indicates that it has at least some capacity to utilize organic carbon
compounds. Further evidence that the survival of At. ferrooxidans was at least in part due
to heterotrophic metabolism came from comparing its longevity in pure and mixed
cultures. More limited survival in mixed cultures may have been due to reduced
availability of organic substrates, resulting from competition with the two acidophilic
heterotrophs present. In contrast, the greater sensitivity of L. ferrooxidans to dissolved
carbon compounds [12] combined with its inability to utilize organic materials would
imply that heterotrophic activity would have a net beneficial effect on survival of this
iron-oxidizer, as was observed.
The reason for the "re-appearance" of colonies of L. ferrooxidans at days 87 and 100
is unclear, but may relate to the break-up of cell aggregates. L. ferrooxidans is known to
form flocs consisting of cells enmeshed in extracellular polymers [13] in liquid media.
These would have not been enumerated by direct or plate counts, under the protocol used.
It is possible that these entrapped cells have greater longevity that free-swimming cells;
disintegration of flocs during protracted starvation stress may have resulted in the release
of viable planktonic cells, which again tended to die out relatively rapidly.
The relative survival of the two heterotrophs, Acidiphilium SJH and "F. acidiphilum",
was also affected by competition for the limited organic resources available. Acidiphilium
SJH is a particularly adept scavenger in pure cultures, but was out-competed by the ironoxidizing heterotroph "F. acidiphilum" in highly acidic mixed cultures. Indeed, the greater
capacity of "F. acidiphilum" to survive starvation stress than the other mesophilic
acidophiles tested suggests that these bacteria may be better adapted to situations where
substrate depletion is a common phenomenon, and may also be a reason for the dominance
of such bacteria in some environments. However, mortality rates of "F. acidiphilum" in
moderately acidic and near-neutral solutions were much greater than in media poised at
about pH 2. The resulting reduction in competition for available organic materials was
considered to be a major reason for the enhanced survival of both At. ferrooxidans and
Acidiphilium SJH in mixed cultures at pH 4.10 than in those poised at pH 2.15. However,
in solutions of pH 6.10, the combined stresses of starvation and inhospitable pH resulted
in more limited survival of all acidophilic bacteria than in more acidic cultures. It is
surprising that near-neutral external pH should have lessened the survival of acidophiles,
as the internal pH of these bacteria is close to neutral in active cells, and 2-3 pH units
above external pH in starved cells [14]. It is known that the maintenance of pH values in
resting acidophiles is passive rather than active.
The trends identified in this study have highlighted the ability of acidophilic bacteria
to survive long-term starvation stress, though surviving populations may be very different
in composition to those present at the onset of nutrient depletion. Environmental factors
appear to have an important influence on survival of these bacteria. Metabolic versatility
appears to be one reason for the superior survival of some species of acidophiles.
Evidence for survival strategy based on the formation of viable ultramicrocells was
limited to the heterotrophic iron-oxidizer "F. acidiphilum". The results show that At.
ferrooxidans, Acidiphilium SJH and "F. acidiphilum" can survive in significant numbers
for up to at least 500 days in substrate-depleted situations, though L. ferrooxidans has a far
lesser capacity to survive under the conditions imposed in our experiments. A limitation of
this work was that only one strain of each acidophile was used in experimental work. It is
1202
conceivable that considerable variation also exists between strains of a particular

acidophilic species, and between species in the case of Acidiphilium.
ACKNOWLEDGEMENTS
This work was supported by funding provided by the U.S. Department of Energy,
Office of Science, to the Idaho National Engineering and Environmental Laboratory,
operated by Bechtel BWXT Idaho, LLC, under contract DE-AC07099-ID13727. In
addition, DBJ was supported during the course of this work by a Faculty Fellowship
program sponsored by the U.S. Department of Energy, Office of University and Science
Education Programs.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
K.B. Hallberg and D.B. Johnson. Adv. Appl. Microbiol. 49 (2001) 37.
D.B. Johnson. Water, Air & Soil Pollut. Foc. 3 (2003) 47.
D.E. Rawlings. Annu. Rev. Microbiol. 56 (2002) 65.
D. B. Johnson. FEMS Microbiol. Ecol. 27 (1998) 307.
E. Zychlinsky and A. Matin. J Bacteriol. 153 (1983) 371.
A.T. Hobbie, J.R. Daley and S. Jaspar. Appl. Environ. Microbiol. 33 (1977) 1225.
D.B. Johnson. J. Microbiol. Meth. 23 (1995) 205.
S. Kjelleberg. (ed.) Starvation in bacteria. Plenum Press, New York, 1993.
D.A. Siegele, M. Almiron and R. Kolter. p.151. In: S. Kjelleberg (ed.), Starvationin
bacteria. Plenum Press, New York, 1993.
10. D.J. Oliver. p.239. In: S. Kjelleberg (ed.), Starvation in bacteria. Plenum Press,
NewYork, 1993.
11. D.J. Oliver and J.K. VanSlyke. Arch. Biochem. Biophys. 263 (1988) 369.
12. D.B. Johnson, M.F. Said, M.A. Ghauri and S. McGinness. p. 403. In: J. Salley, R.G.L.
McCready and P.L. Wichlacz (eds.), Biohydrometallurgy 1989. Canmet, Ottawa,
1990.
13. R. Hallmann, A. Friedrich, H-P. Koops, A. Pommerening-Roser, K. Rohde, C.
Zenneck and W. Sand. Geomicrobiol. J. 10 (1992) 193.
14. P.R. Norris and W.J. Ingledew. p. 115. In R.A. Herbert and R.J. Sharp (eds.),
Molecular biology and biotechnology of extremophiles. Blackie, Glasgow, 1992.
1203

Synthesis of nanophase hydroxyapatite by Serratia sp. N14

P. Yonga, R.L. Sammonsb, P.M. Marquisb, H. Luggb and L.E. Macaskiea*
a
Schools of Biosciences, The University of Birmingham, Birmingham B15 2TT, UK

b
School of Dentistry, The University of Birmingham, Birmingham B15 2TT, UK
Abstract
A species of Serratia isolated from heavy metal contaminated soil and previously
identified as Citrobacter sp. N14 expresses an atypically high level of acid phosphatase. In
the presence of CaCl2 and glycerol-2-phosphate (G2P), biosynthesis of nano-crystalline
hydroxyapatite (HA) was achieved by use of immobilised Serratia cells. Calcium ions
accumulated as cell-bound calcium phosphates utilizing phosphate released by the
enzymatic cleavage of the substrate (G2P). Addition of citrate (to restrict the free Ca2+
concentration) and/or reducing the substrate concentration reduced the size of the HA
crystals and promote formation of the nano-phase. Hydroxyapatite crystals were located
using Scanning Electron Microscopy (SEM) and identified by X-ray diffraction (XRD).
The size of the crystals was calculated to be between 15 and 25 nm, a size more suitable
for biomedical applications than that obtained by chemical synthesis.
Keywords: biomineralization, calcium phosphate, hydroxyapatite, Serratia

1.
INTRODUCTION
Hydroxyapatite (HA) is a well-established bone substitute biomaterial, which has
been successfully used for replacement and augmentation of diseased or traumatised bone
[1]. The biocompatibility of this material is related to its close similarity to the
composition of the mineral component of natural bone. Hydroxyapatite can also be used to
coat metal prostheses, and for orthopaedic and dental applications, using techniques such
as plasma spraying. Unfortunately the coating methods currently employed lead to
coatings containing a variety of phosphate phases, both amorphous and crystalline, which
can be susceptible to dissolution or delamination in clinical use [2,3]. Equivalent problems
are also encountered in the fabrication of dense hydroxyapatite components, where
difficulties in densification inevitably lead to implants with poor mechanical properties,
notably strength and fracture toughness, which cannot be used in load bearing
applications. It is well known that as the grain size of materials reduces towards the
nanoscale (<200 nm, ideally below 50 nm), properties such as strength are dramatically
improved [4]. To achieve this goal requires precursor materials of this nanoscale
dimension, which can be subsequently manipulated and fabricated without loss of the fine
grain size. Conventional chemical synthesis approaches do not produce particles of the
* To whom correspondence should be addressed

(Fax: +44(0)121-414-5925; e-mail: l.e.macaskie@bham.ac.uk)
1205
required dimensions and also invariably lead to clusters, or agglomerates of particles,

which lead to inherent weakness in the final implant.
Microbially mediated biomineralization has a high potential for the production of
metallic nano-particles. Klaus [5] found that Pseudomonas stutzeri extracted silver from
ores and deposited pure metallic silver particles of up to 200 nm diameter inside the
bacteria. Palladium (II) can be reduced enzymatically to Pd(0) at the surface of
Desulfovibrio desulfuricans [6,7]; the mean crystal size of biologically-manufactured
Pd(0) was only half of that obtained by reduction of soluble Pd(II) under H2 without cells,
but it is not clear which biological features contributed to this phenomenon [7,8]. In
biomineralization, the components of microbial cell walls and their associated structural
polymers may be used as organic "scaffolds" for the patterning of extracellular vesicles
and associated inorganic minerals [9]. The compartmentalisation of the biological space
on and around cells could be utilised to promote spatial localisation for crystal growth, and
nucleation could be mediated by organic polymeric substrates in or on the spatial
boundary. The inorganic mineral crystals could be in membrane-bound vesicles, in the
mucilaginous layers of cell walls or impregnated in biopolymers in the extra-cellular space
[9].
A biomineralization method was developed for the accumulation of metal ions from
aqueous solution utilizing cells of a Serratia sp., which was originally classified as
Citrobacter [10,11]. It was recently reassigned as Serratia on the basis of biochemical and
physiological criteria and 16S rRNA homology [12]. In this bioaccumulation process,
metal uptake is mediated by the activity of phosphatase enzyme that is localised both
periplasmically or within extracellular polymeric materials [13]. This enzyme continues to
function in resting cells to liberate phosphate ions via the hydrolytic cleavage of an
organic phosphate molecule (e.g. glycerol 2-phosphate (G2P)) and promotes
stoichiometric deposition of cell-bound metal phosphate [13]. Early studies showed that
under appropriate conditions (high pH) strontium could be desolubilized using this method
[14]. The biomineralization of calcium phosphate should occur in similar manner and the
purpose of this study was to evaluate the potential of this approach for the microbially
mediated synthesis of calcium phosphate. The investigation focused on the biosynthesis of
nanocrystalline hydroxyapatite (HA) with the objective of producing a better HA
biomineral for biomedical application. The effect of the challenge solution on metal
phosphate deposition, the chemical structure and the size of crystals, and their
arrangement on the biomass, was investigated.
2.
EXPERIMENTAL
2.1 Preparation of biomass

Biomass was grown under carbon (lactose)-limited continuous culture at 30C in an
airlift fermenter (Fig. 1) as described previously [15]. Minimal medium (2.5 l) [15] was
added to the fermenter containing up to 200 polyurethane reticulated foam cubes (TM30
supplied by Recticel, Belgium, washed with distilled water three times) threaded on cotton
strings, and the culture was started by the addition of 50 ml inoculum which had been pregrown in similar medium. The culture was allowed to grow batch-wise and switched to
continuous mode after 24 h (OD600 ~ 0.6), maintained at steady-state for six days [15], and
the foam cubes (heavily loaded with biofilm) were withdrawn and kept in air in a sealed
vessel over isotonic saline (8.5 g/l) at 4C until use. The culture was monitored by
measurement of phosphatase activities and cell densities of free cells in the culture
outflow [15].
1206
2.2 Phosphatase activity assay

Phosphatase activity was assayed by the release of p-nitrophenol (PNP) from pnitrophenyl phosphate (PNPP) [14,15]. Phosphatase activity (units) is defined as nmol of
PNP liberated per min per mg bacterial protein, with OD600 (cell optical density at 600
nm) of the cells converted to bacterial protein using a conversion factor obtained by the
assay method of Lowry (0.278 mg protein/ml at OD600 =1; path length = 1 cm) [16].
2.3 Bioaccumulation of calcium phosphate by Serratia sp. N14
Ten biofilm-loaded foam cubes (~4.5 mg dry cell mass per cube, washed by dipping
into isotonic saline) were challenged in 50 ml of 50 mM TAPSO/NaOH buffer pH 9.2
with CaCl2 (1 mM or 5 mM) and sodium glycerol 2-phosphate (G2P: 1 mM or 5 mM)
with or without citrate (2 mM). The flask was shaken at room temperature. Additional
CaCl2, G2P and citrate were dosed into the flask daily (additional final concentrations of 1
mM, 5 mM and 2 mM as appropriate, after allowing for volume reduction after sampling).
The dosing regime is summarised in Table 1. Samples were taken daily before dosing for
monitoring of residual Ca(II) and phosphate concentrations. Calcium phosphate-loaded
cells harvested by centrifugation after eight days of dosing (or as otherwise specified)
were treated and analysed by X-ray powder diffraction and electron microscopy as
described. Representative data are shown throughout.
2.4 Assay of phosphate and calcium
Samples (up to 5 mM Pi) were diluted 30-fold. Sodium molybdate (0.6 ml; 2.5%
(w/v) in 1.67 M H2SO4 was added into 1 ml assay solution and phosphate was visualized
by adding 0.4 ml SnCl2. (Fresh solution of SnCl2 was made from 0.25 ml stock solution
diluted into 100 ml of 1 M HCl, and the stock was made monthly by dissolving 1.5 g
SnCl2 in 2.5 ml concentrated HCl.) The quantity of resultant blue complex was measured
at 720 nm (Perkin-Elmer spectrophotometer) and the phosphate concentration estimated
by reference to a standard curve, similarly prepared. For assay of Ca(II) 0.1 ml of sample
(up to 1 mM Ca(II)) was added to 1.9 ml 20 mM MOPS/NaOH buffer pH 7, and mixed,
and Ca(II) was visualized by the addition of 0.1 ml of 0.15% (w/v; aq.) Arsenazo III, with
estimation of the pink complex at A631 nm (Perkin-Elmer spectrophotometer) versus a
standard solution of CaCl2.
2.5 Localization of cell-bound deposits using scanning electron microscopy
Calcium phosphate loaded cells (squeezed from one foam cube) were centrifuged,
fixed with 2.5% glutaradehyde fixative (in 0.1 M Na cacodylate/HCl buffer pH 5.2) at 4C
for one hour, dehydrated at room temperature, treated and examined under a scanning
electron microscope as described previously [6].
2.6 Identification of cell bond material and crystal size determination by X-ray
powder diffraction (XRD) analysis
The Ca-loaded cells (squeezed from 10 foam cubes) were centrifuged after eight days
of challenge with the same daily addition, washed with distilled water and acetone, and
then dried at room temperature. The samples were laid on a silicon crystal plate and
analysed using a high precision X-ray powder diffractometer (School of Physics, The
University of Birmingham) for data collection as described previously [7,11]. The powder
diffraction patterns were recorded from 5 to 70 (2) with a step length of 0.05 (2). The
crystal size of cell-bound calcium phosphate/HA was calculated for samples obtained
1207
from loaded dried biomass or from commercial reference material HA (Capital 60

hydroxylapatite: Plasma Biotal Ltd., Tideswell, Derbyshire, UK) by computing the diffraction
pattern using the Langford programme [17].
3.

When the cells are challenged with calcium and a phosphatase substrate i.e. G2P,
there is a biocatalytic (enzymically-mediated) and an inorganic precipitation combined
process, in which the component reactions are:
Glycerol-2-phosphate Phosphatase> Glycerol + Phosphate
(1)
____
Phosphate + Calcium
> Calcium phosphate
(2)
The synthesis of calcium phosphate would occur where the Ca(II) ions interact with
the liberated precipitant, phosphate ions. Calcium phosphate can be synthesised
chemically, and many authors have studied the factors of affecting this synthesis [1,2,18].
The type of calcium phosphate and the crystal size are largely dependent on the conditions
of synthesis, i.e. the setting time, solution pH, ions concentrations, temperature, etc.
Indeed, the bioprecipitation of calcium phosphate by Serratia cells would be affected
similarly by these physicochemical factors. The form of phosphate ions released by
reaction (1) is pH-dependent. Phosphate is a tri-basic ion (K1 = 7.5 x 10-3, K2 = 6.2 x 10-8,
K3 = 5.0 x 10-13) [19]. The major forms of which in aqueous solution would be H3PO4 and
H2PO4- below pH 4.6, H2PO4- and HPO42- from pH 4.6 to 9.7 and HPO42- and PO43- for pH
values above 9.7. In addition, the rate of phosphate release in reaction (1) is dependent on
the solution pH since phosphatase activity is pH dependent. Phosphatase activity of the
cells was reduced by 70% when the solution pH was changed to 9.0 from 5.5 (Fig. 1a).
However, after 24 h the inorganic phosphate concentrations were not significantly
different in solutions at pH 7 to 9.2 although the phosphate released was lower in the
solution at pH 10.0 (Fig. 1b) attributable to the lower reaction rate. The observation that
phosphate release was >95% complete at 24 h (Fig.1b) justified the choice of daily
intervals for the dosing regime. The requirement for alkaline pH was shown previously in
the case of Sr(II) removal [14] where a compromise was required between the requirement
for high pH for precipitation of the alkaline earth metal phosphate and retention of high
cellular phosphatase activity.
Many authors have studied the solubility of calcium phosphate salts [1,20,21]. The
type of calcium phosphate precipitated would be dependent on the pH value of the
challenge solution [20-22]. It has been reported that hydroxyapatite (Ca5(PO4)3(OH)) is
the most stable calcium phosphate salt at pH>4, but it precipitates directly only at above
pH 8 [20,21]. The current investigation was carried out to define the conditions for the
biosynthesis of HA for potential medical applications; the quality of the calcium
phosphate crystals is much more important than the overall yield. Calcium
bioprecipitation by Serratia was tested in a solution of 50 mM TAPSO/NaOH buffer (pH
9.2) and 1 mM CaCl2, with phosphatase substrate G2P at saturating [23] (5 mM) or
limiting (1 mM) concentrations using biofilm-immobilized cells.
The biomineralization process (controlled by enzyme-mediated inorganic phosphate
(Pi) concentration), relies on the cellular microenvironment becoming supersaturated with
inorganic phosphate released from G2P by the cells. However, nucleation represents an
activation energy barrier to the spontaneous formation of a solid phase from a
supersaturated solution. Increasing the degree of super-saturation and lowering the
interfacial energy can reduce the activation energy for nucleation [24]. Indeed, for 1 mM
free Ca2+ in the bulk solution without complexing reagent, it was calculated that
1208
precipitation should occur for CaHPO4 at 0.1 mM HPO4- (Ksp=1x10-7 for CaHPO4) [25],
and for Ca3(PO4)2 at 1.4x10-7 mM PO4- (Ksp= 2.0x10-29 for Ca3(PO4)2) [25]. However,
biosorption of calcium was observed only after 24 h even in the solution with 5 mM G2P
and a released phosphate concentration of higher than 4 mM (Fig.2a.). In the solution with
1 mM substrate the initiation of precipitation was retarded even further (Fig.2b). When the
cells were challenged in the presence of citrate, the concentration of free Ca(II) ions was
reduced significantly because citrate is a calcium complexing reagent (K = 4.79x104) [25].
By calculation, there would be only 0.02 mM free ions for Ca(II) at a input concentration
of 1mM in the solution with 2 mM citrate, which, in turn, would need a phosphate
concentration above 5 mM to form calcium phosphate precipitate. It was shown that the
initial precipitation of calcium phosphate from the solution with 2 mM citrate was the
slowest (Fig. 2c) compared to that from the other two solutions. However once the
precipitation started, the rate of calcium uptake was the same in 1mM as in 5 mM G2P
solution with or without citrate, although there was a much lower phosphate or free
calcium ion concentration in the latter two solutions (Fig.2b, 2c). Therefore, for the
formation of calcium phosphate, nucleation was the controlling stage initially. Once the
barrier of nucleation was overcome, continued precipitation of calcium phosphate became
feasible. Even in a solution with 1 mM CaCl2, 2 mM citrate and 1 mM G2P (added daily),
the precipitation was completed within eight days (not shown).
140
Pi concentration (mM)
Relative activity (%)
160
120
100
80
60
40
20
5
4
3
2
1
0
0
4
10
11
pH
Figure 1a. The effect of pH on

phosphatase activity of Serratia cells.
Cells were squeezed from the foam
cubes and resuspended in 200 mM
buffer (pH 5-6: acetate buffer, pH 6-7:
MOPS/NaOH buffer, pH 8-9: TAPSO
buffer, and pH 10: AMPSO buffer), and
phosphatase activities were assayed as
described in Experimental. The activity
of the cells at pH 7 (expressed as 100%)
was 2384 units
8.6
9.2
10
pH
Figure 1b. The effect of pH on inorganic

phosphate (Pi) release by immobilized
Serratia cells. Five foam cubes were
challenged in 25 ml 50 mM buffer (pH
7-9.2: TAPSO, and pH 10 AMPSO) 5
mM G2P and 2 mM citrate. Samples
were taken after 24 hours and
centrifuged and the phosphate content
of supernatants assayed as described in
Experimental
1209
5
0
0 1 2 3 4 5 6
Time (Days)
20
15
10
5
0
25
8
7
6
5
4
3
2
1
0
20
15
10
5
0
0 1 2 3 4 5 6
0 1 2 3 4 5 6
Time (Days)
Time (days)
Ca precipitated (mM)
10
8
7
6
5
4
3
2
1
0
15
25
20
8
7
6
5
4
3
2
1
0
25
Pi concentration(mM)
a.
b.
c.
Figure 2. Phosphate release and calcium accumulation by immobilized Serratia cells
Ten biofilm-foam cubes (~4.5 mg dry cell mass per cube; phosphatase activity was
2834 units at harvest) were challenged in 50 ml of 50 mM TAPSO/NaOH buffer (pH
9.2), and: a) 1 mM CaCl2 and 5 mM G2P with same addition daily; b) 1 mM CaCl2
and 1 mM G2P with same addition daily; c) 1 mM CaCl2, 5 mM G2P and 2 mM
citrate with same addition daily. , inorganic phosphate concentration in the
challenge solution (mM), , calcium removed by Serratia cells (mM)
A scanning electron microscope study (SEM) was performed for samples harvested
after eight days using cells challenged with CaCl2 (1 mM), and citrate (0 mM or 2 mM)
without (Fig. 3a) or in the presence of 5 mM G2P (Fig. 3b) added daily. Calcium
phosphate crystals were visible heavily coating the cell surface but with no apparent
differences between the precipitate made from solutions with different concentration of
calcium chloride, G2P and citrate. HA crystals were identified by XRD by comparing the
pattern with that of the commercial HA and also the XRD standard data file (Fig. 4). The
size of HA crystals was calculated by computing the diffraction pattern using the Langford
programme [17]; the HA crystals on the cells were only 15-25 nm, substantially below a
typical commercial product (Table 1). The largest size (25 nm) of HA crystal was from
solution with addition of substrate (G2P) at its saturating concentration (5 mM) in the
absence of citrate. The crystal size was considerably reduced when either the
concentration of G2P was lowed to 1 mM or citrate (2 mM) was introduced into the
solution. When the cells were challenged with 1 mM G2P and 2 mM citrate, Pi release
was at the slowest rate and the concentration of free calcium ions was also very low due to
citrate complexing. Hence Pi had to remove Ca from the complex to form calcium
phosphate precipitate. Indeed, the smallest crystals were from solution with daily addition
of 1 mM G2P and 2 mM citrate (15-16 nm). Therefore, the crystals formed at a size
governed by the availability of free calcium ions and the rate of phosphate release using
the bacterial cell surface as the nucleation template.
4.
CONCLUSION
The synthesis of nano-particulate calcium phosphate was achieved by challenging
Serratia cells in the solution of 50 mM TAPSO buffer (pH 9.2), 1 mM calcium chloride, 2
mM (or 0 mM) citrate and 5 mM (or 1mM) G2P with the same daily doses. The
biosynthesised crystals were located by scanning electron microscopy, and identified by
X-ray powder diffraction. The sizes of crystals were 15-25 nm, substantially below a
typical commercial product. However, a more detailed study is needed for this
biomineralization system to produce the best HA biomineral for commercial exploitation.
The mechanical strength and potential biomedical applications of the biosynthesised HA
will also need further investigation.
1210
a.
b.
Figure 3. Calcium accumulation by Serratia cells viewed under SEM. Five biofilmfoam cubes were challenged in 25 ml of 50 mM TAPSO/NaOH buffer (pH 9.2)
solution with 1 mM CaCl2 only (a.) or with 1 mM CaCl2 and 5 mM G2P with or
without 2 mM citrate (b.) with the same daily addition (for eight days). Cells were
harvested and treated as described in Experimental.
Figure 4. The X-ray powder diffraction pattern of the cell-bound material and
hydroxyapatite. Upper: pattern for the sample from 50 mM TAPSO buffer (pH 9.2)
solution with 1 mM calcium chloride, and 5 mM G2P with the same addition daily
for eight days. Lower: pattern for the commercial HA crystals. Vertical lines:
superimposition of the obtained pattern upon the standard database files for HA [26]
1211
Table.1. The size of crystals for HA precipitated from different solutions with
immobilized cells and the commercialised HA. Samples were centrifuged after eight
days, treated and the calcium phosphate crystals were identified by XRD as HA as
described in Experimental. The crystal size was calculated by computing the
diffraction pattern using the Langford programme [17]
Samples
First day
addition: 50
mM TAPSO
buffer, pH 9.2
Daily addition
to final
concentration
for seven days
HA size (nm)
CaCl2
1
1 mM
2
1 mM
3
1 mM
4
1 mM
5*
5 mM
6*
5 mM
HA**
N/A
G2P
5 mM
5 mM
1 mM
1 mM
1 mM
1 mM
N/A
Citrate
CaCl2
2 mM
1 mM
0 mM
1 mM
2 mM
1 mM
0 mM
1 mM
2 mM
1 mM
0 mM
1 mM
N/A
N/A
G2P
5 mM
5 mM
1 mM
1 mM
1 mM
1 mM
N/A
Citrate
2 mM
21
0 mM
25
2 mM
15
0 mM
20
2 mM
16
0 mM
20
N/A
167
*Note that following the initial dose of 5 mM Ca2+ 1 mM Ca2+ was dosed subsequently in samples
5 and 6.
**Commercial HA: Capital 60 hydroxylapatite, Plasma Biotal Ltd., Tideswell, Derbyshire, UK.
ACKNOWLEDGEMENTS
This project was supported by the BBSRC (Grant No. 6/E11940). The authors wish to
thank Dr J. I. Langford (School of Physics & Space Research, University of Birmingham)
for assistance with the XRD technique and Mrs L. Tomkins and Mr P. Whittle for help
with electron microscopy. The authors also thank Recticel (Belgium) for the gift of
reticulated foam for the growth of biofilm.
REFERENCES
1. M. Jarcho, Clin. Orthopaed. Rel. Res., 207(1981)290.
2. S. D. Cook, J. F. Kay, K. A. Thomas and M. Jarcho, Int. J. Oral Maxillofacial Impl.,
2(1987)1.
3. A. N. Cranin, M. Mehrali and M. Baraoidan, J. Implantol., 20(1994)235.
4. R. W. Siegel, Mechanical properties of nanophase materials. In Synthesis and
Properties of Mechanically Alloyed and Nanocrystalline Materials, Parts 1 & 2,
Materials Science Forum, 235 (1997) 851.
5. T. Klaus, R. Joerger, and C. Granqvist. Proc. Nat. Acad. Sci., 96 (1999) 13611.
6. P. Yong, N. A. Rowson, J. P. G. Farr, I. R. Harris and L.E. Macaskie. J. Chem. Tech.
Biotechnol. 77 (2002) 593.
7. P. Yong, N. A. Rowson, J. P. G. Farr, I. R. Harris and L. E. Macaskie. Biotechnol.
Bioeng., 80(2002)369.
8. P. Yong, N. A. Rowson, J. P. G. Farr, I. R. Harris and L. E. Macaskie. Environ.
Technol., In press (2003).
9. S. Mann, J. Chem. Soc Dalton Trans., 21 (1997) 3953.
10. L. E. Macaskie, R. M. Empson, A. K. Cheetham, C. P. Grey, and A. J. Skarnulis,
Science, 257 (1992) 782.
11. P. Yong and L.E. Macaskie J. Chem. Tech. Biotechnol., 63(1995)101.
12. P. Pattanapipitpaisal, A. N. Mabbett, J. A. Finlay, A. J. Beswick, M. Paterson-Beedle,
A. Essa, J. Wright, M. R. Tolley, U. Badar, N.A. Ahmed, J. L. Hobman, N. L. Brown
and L. E. Macaskie, Environ. Technol., 23 (2002) 731.
1212
13. L. E. Macaskie, K. M. Bonthrone, P. Yong, and D. T. Goddard, Microbiology, 146

(2000) 1855.
14. L.E. Macaskie and A.C.R. Dean. Biotechnol. Lett., 7 (1985) 627.
15. J. A. Finlay, V. J. M. Allan, A. Conner, M. E. Callow, G. Basnakova and L. E.
Macaskie, Biotechnol. Bioeng., 63(1999)8.
16. B. C. Jeong and L. E. Macaskie, Enz. Mic. Technol., 24 (1999) 218.
17. J. I. Langford. National Institute of Standard and Technology Special Publication 846.
Proceedings of the International Conference. Accuracy in Powder Diffraction II, NIST,
Gaitherburg, MD, May 26-29, (1992) 110.
18. C. Durucan and P. Brown. J. Mat. Sci.: Materials in Medicine, 11 (2000) 365.
19. G. H. Jeffery, J. Bassett, J. Mendham, R. C. Denney (eds). Vogels Textbook of
Quantitative Chemical Analysis, Fifth Edition. Longman Scientific Technical, John
Wiley And Sons Inc., New York, USA (1989).
20. P. Boudeville, S. Serraj, J.M. Leloup, J. Margerit, B. Pauvert, and A. Terol. J. Mat.
Sci.: Materials in Medicine, 10 (1999) 99.
21. S. Matsuya, S. Takagi, and L.C. Chow. J. Mat. Sci.: Materials in Medicine, 11 (2000)
305.
22. E. Fernandez, F.J. Gil, M.P. Ginebra, F.C.M. Driessens, J.A. Planell and S.M. Best. J.
Mat. Sci.: Materials in Medicine, 10 (1999) 223.
23. P. Yong and L.E. Macaskie. Biotechnol. Bioeng., 55(1997)821.
24. S. Mann. Nature, 332 (1988) 119.
25. J.A. Dean (ed). Lange's Handbook of Chemistry, McGraw-Hill, Inc. New York, 1992.
26. Anon. Powder Diffrection File. Card No. 09-0432, JCPDS, Swarthmore, PA, USA
(1992).
1213

The effect of maintenance on the ferrous-iron oxidation kinetics

of Leptospirillum ferrooxidans
C.J.N. Dempers1, A.W. Breed2 and G.S. Hansford
Gold Fields Minerals Bioprocessing Laboratory, Department of Chemical Engineering,
University of Cape Town, Rondebosch, 7701, South Africa
Abstract
The results of batch experiments carried out using a predominantly Leptospirillum
ferrooxidans culture indicated that the maintenance requirement of energy sufficient
cultures is a combination of the variable and constant maintenance energy requirements.
For this reason, modelling the bioenergetics of energy sufficient cultures assuming a
constant maintenance energy requirement results in a significant overestimation of the
maximum bacterial yield. However, the variable maintenance equation proposed by Pirt
(Pirt, 1982) may be used to determine the maximum bacterial yields and the maintenance
coefficients from both substrate lean and substrate rich experiments. This is because this
model is able to account for variations in the growth rate as a result of changes in the
maintenance requirement due to the microorganisms being limited by more than one
factor.
Keywords: batch processing, bioreactors, Chemoautotrophs, kinetic parameters,

maintenance requirements, modelling
1.
INTRODUCTION
Recent work has provided strong evidence that the bioleaching of sulfide minerals
occurs via a multiple sub-process mechanism [1-5]. According to Sand et al. [2], the
disulfides pyrite (FeS2), molybdenite (MoS2), and tungstenite (WS2) are chemically
oxidised by the ferric-iron present in the bioleaching medium via the main intermediate
thiosulfate:
FeS2 + 6 Fe3+ + 3 H 2O S2O32 + 7 Fe2 + + 6 H +
(1)
The thiosulfate is subsequently, degraded in a cyclic process to sulfate, with elemental

sulfur being a side product [2]:
S2O32 + 8 Fe3+ + 5 H 2O 2SO 24 + 8 Fe2 + + 10 H +
(2)
The above thus explains why ferrous-iron oxidizing bacteria alone, e.g.
Leptospirillum ferrooxidans, are able to oxidise these metal sulfides.3
1
MRT Africa, Park Place East, 104 North Rand Road, Hughes, Boksburg, 1460, South Africa
Department of Chemical Engineering, University of Sydney, NSW 2006, Australia
3
The genus Leptospirillum has recently been shown to have two species, viz. Leptospirillum ferrooxidans
and Leptospirillum ferriphilum [6]. This differentiation is not however considered in this work.
2
1215
In contrast to the above, metal sulfides such as galena (PbS), sphalerite (ZnS),
chalcopyrite (CuFeS2), hauerite (MnS2), orpiment (As2S3), and realgar (As4S4) are
oxidised by both the ferric-iron and protons present in the bioleaching solution; the main
intermediates of these reactions are polysulfides and elemental sulfur [2]:
MS + Fe3+ + H + M 2 + + 0.5 H 2Sn + Fe2 + (n > 2)
(3)
0.5 H 2Sn + Fe3+ 0.125 S8 + Fe2 + + H +
(4)
0.125 S8 + 1.5 O 2 + H 2O + SO 24 + 2 H +
(5)
These metal sulfides are thus degradable by bacteria that are able to oxidize sulfur
compounds, e.g. Acidithiobacillus thiooxidans.4
If ferrous-iron oxidising bacteria are present within the system, then the ferrous-iron
produced by Reactions 1 to 4 is subsequently oxidised to the ferric form by these bacteria:
4 Fe 2+ + O 2 + 4 H + 4 Fe 3+ + 2 H 2 O
(6)
A multiple sub-process mechanism, such as the one described above, thus suggests
that the overall process can be expressed as a number of interconnected chemical and
bacterial sub-processes, the kinetics of which may be studied separately, and the results
obtained used to predict the performance of bioleach reactors for a variety of different
minerals, micro-organisms and operating conditions.
To date a number of kinetic models for bacterial ferrous-iron oxidation have been
proposed [8]. These models can be broadly classified as either empirical or MichaelisMenten/Monod based. In the proposed models, the bacterial specific growth rate, , is
usually related to the bacterial specific ferrous-iron utilisation rate, q Fe2 + , via the
, and a constant maintenance coefficient
maximum bacterial yield on ferrous-iron, YFemax
2+
X
on ferrous-iron, m Fe2 + ; viz. using the Pirt Equation [9]:
q Fe 2 + =
+ m Fe2 +
YFemax
2+
X
(7)
The bacterial specific ferrous-iron utilisation rate, q Fe2 + , is defined as follows:

q Fe2 + =
rFe2 +
cX
(8)
where:
rFe2 + is the ferrous-iron production rate, and
c X is the bacterial concentration expressed as mol C.l-1.
From Equation 7 it is thus apparent that the Pirt Equation [9] is based on the
assumption that the biomass specific maintenance requirement is constant and
independent of the growth rate; the maintenance energy requirement is the minimum
amount of substrate, per unit of biomass, which is required to maintain the vital functions
of the microorganisms.
Although the Pirt Equation [9] has been widely used to model the ferrous-iron
oxidation kinetics of the bacteria encountered in bioleaching, a dependence of the
experimentally determined yield and maintenance coefficients on the growth conditions
4
Acidithiobacillus thiooxidans was previously named Thiobacillus thiooxidans [7].
1216
has been reported [5]. Boon [5] reported that it was necessary to fit the Pirt Equation [9] to
the initial and late batch data separately, i.e. the calculated values of YFemax
and m Fe2 +
2+
X
depended on whether the growth was energy sufficient or energy limited. Similar
dependencies on the growth conditions have been observed during experiments performed
using other microorganisms [10-13].
Neijssel and Tempest [14] suggested that these deviations could be attributed to the
growth of the microorganisms being limited by factors other than the energy source. This
led these workers to propose that the maintenance requirement was not constant, but
varied with changes in the bacterial growth rate. Pirt [15] modified the relationship
developed by Neijssel and Tempest [14] by assuming that the energy required for
maintenance included a term that decreased with an increase in the specific growth rate.
Although the variable maintenance approach has been successfully used to describe the
kinetics of non-bioleaching microorganisms, it has yet to be used in the modelling of the
ferrous-iron oxidation kinetics of the bacteria used in bioleaching operations.
The primary objective of the work presented below was to determine whether or not
the variable maintenance equation proposed by Pirt [15] could be used to describe the
batch growth data of a predominantly L. ferrooxidans culture during both energy sufficient
and energy-limited growth. A further objective of the work was to determine whether or
not the kinetic parameters determined in this way were similar to those determined
previously during continuous ferrous-iron oxidation experiments performed using the
same bacterial culture and similar growth conditions.5
2.
THEORETICAL ASPECTS
If bacteria are grown in continuous culture and under conditions in which the carbon
source is limited to carbon dioxide, then the carbon dioxide production rate, rCO 2 , can be
used to estimate the biomass production rate growth rate, rX , [5]:

rX = rCO 2
(9)
However, if the bacteria are grown in batch culture and under conditions in which the
carbon source is limited to carbon dioxide, then the bacterial concentration at time, t, can
be estimated from the carbon dioxide production rate, rCO 2 , and the initial bacterial
concentration, c X 0 :
t =t
c X = c X0 +
CO 2
dt
(10)
t =0
Furthermore if the stoichiometric formula of bacteria is assumed to be CH1.8O0.5N0.2

[18, 19] and energy for bacterial growth and maintenance is obtained from the oxidation
of ferrous-iron, performing mass and charge balances and solving in terms of the carbon
dioxide and oxygen production rates, rCO 2 and rO 2 , respectively, yields the degree-ofreduction balance:
rFe2 + = 4rO 2 4.2 rCO 2
(11)
The continuous experiments are reported elsewhere [16, 17] whereas the batch experiments were
performed during the course of the current study.
1217
If the maintenance energy requirement is assumed to be constant and independent of

the growth rate and the theoretical maximum growth yield, YFemax
, occurs when the
2+
X
maintenance energy, m Fe2 + , is zero, then the relationship between the amount of substrate
consumed by the biomass for bacterial growth and maintenance can be described using
Equation 7, i.e. the Pirt Equation [9]. However, if the maintenance requirement is assumed
to vary with changes in the bacterial growth rate and if the maintenance energy term is
assumed to include a portion that decreases with an increase in the specific rate of growth,
then the relationship between the bacterial specific substrate utilisation rate, the maximum
substrate specific yield and the and the growth rate must be described using [15]:
q Fe 2 + =
max
v
+ m Fe 2 + + m Fe
2 + (1 k )
(12)
Fe 2 + X
In Equation 12, m Fe2 + is the constant maintenance coefficient on ferrous iron, whereas
v
m Fe
2 + and k are growth dependent maintenance coefficients. Pirt [15] also postulated that
the growth rate dependent maintenance energy requirement decreases to zero as the
specific growth rate, , approaches the maximum bacterial specific growth rate, max .
This in turn implies that:
k=
(13)
max
Combining Equations 12 and 13 and rearranging yields:

q Fe 2 +
v
1
m Fe
2+
=
max
max
Y 2+
Fe X
+ m 2+ + m v 2+
Fe
Fe
(14)
A graphical representation of Equation 14 is shown in Figure 1; the two lines, (a) and
(b), represent experimental data obtained under different energetic conditions.
From Figure 1 and the definition of the constant maintenance requirement, it is
apparent that, under conditions in which the growth rate is limited by the concentration of
the energy source, i.e. during energy limited growth, the maintenance energy requirement
of the microorganisms is constant and lower than the maintenance energy requirement
when the growth is limited by another factor, e.g. the concentration of a specific trace
metal. Under these conditions the constant maintenance coefficient can thus be obtained
from the intercept of the solid line and the y-axis, i.e. from line (a). Furthermore, because
the variable maintenance requirement decreases to zero under energy-limited conditions,
the maximum growth yield can also be determined from the slope of line (a).
In contrast to the above, the maintenance energy requirement of energy sufficient
cultures is a combination of the variable and constant maintenance energy requirements
(see line (b) in Figure 1). The variable maintenance energy requirement can therefore be
determined from the difference between the values of the y-intercepts determined during
energy sufficient and energy limited growth experiments; the value of the variable
maintenance coefficient will thus depend on the limiting factor. These factors include the
concentrations of phosphate, sulfate and ammonium ions, the concentration of heavy
metals and the bacterial growth rate.
1218
2+
-1
-1
qFe2+ (mol Fe .(mol C) .h )
a
qFe2+max
mFe2+ + mFe2+v
1/YFe2+max
mFe2+
0
0
max
-1
(h )
Figure 1. Relationship between the bacterial specific growth rate, , and the bacterial
specific ferrous-iron utilisation rate, q Fe2 + , assuming a) constant maintenance and b)
variable maintenance
In addition to the above, it can be seen that the maximum bacterial specific ferrousiron utilisation rate, q max
, and the maximum bacterial specific growth rate are defined by
Fe 2 +
the intersection of the solid and the broken lines.
Equation 14 can also be written in terms of the bacterial specific oxygen utilisation
rate, q O 2 , hence the maximum bacterial yield on oxygen, YOmax
, the constant maintenance
2X
coefficient on oxygen, m O 2 and the growth dependent maintenance coefficient on oxygen,
m Ov 2 , may be calculated in the same way as the ferrous-iron based parameters. The
validity of the yield and maintenance coefficients calculated determined in this manner
can in turn be can be checked using the degree of reduction balance, Equation 10; i.e.
using:
YOmax
=
2X
m O2 =
4 YFemax
2+
X
1 4.2 YFemax
2+
X
(15)
m Fe2 +
(16)
4
It is thus also apparent from Figure 1 and the preceding discussion that the values of
the kinetic parameters determined from experimental data can be highly dependent on the
conditions used.
3.

The batch ferrous-iron oxidation experiments performed during the course of this
study were carried out at pH 1.70 and 30, 35 and 40C and at 40C and pH 1.10, 1.30,
1.50 and 1.70, using a predominantly L. ferrooxidans culture. The bacterial culture used
was initially obtained from a vat-type two-stage (220 l) continuous bioleaching miniplant treating an arsenopyrite/pyrite concentrate from Fairview Gold Mine in Barberton,
South Africa [20]. The inoculum for each of the batch experiments was obtained from
steady state continuous cultures of microorganisms from the bioleaching mini-plant grown
1219
on ferrous-iron medium at a dilution rate of 0.04 h-1 and the temperature and pH to be
used in the batch experiment.
The batch experiments themselves were carried out in 2l air-sparged, agitated
bioreactors. The bioreactors had a H/D of 1.32 and a working volume of 1l. Circulating
water from constant temperature baths through the bioreactor jackets controlled the
temperature in the bioreactors. The pH of the solution in the bioreactors was monitored
continuously and maintained at the required pH by the addition of concentrated sulphuric
acid (98%). The redox potential of the bioleaching solution in the bioreactors was
measured continuously using Metrohm redox electrodes (Pt-Ag/AgCl) and logged by
computer.
The ferrous iron media used consisted of 12 g.l-1 FeSO4.7H2O, 1.11 g.l-1 K2SO4,
0.53 g.l-1 (NH4)2HPO4, 1.83 g.l-1 (NH4)2SO4 and 10 ml.l-1 trace element solution [21]
adjusted to between pH 0.95 and pH 1.30 using H2SO4conc.
Air, at a flow rate of 1 l.min-1, was supplied to the bioreactors using Brooks mass
flow controllers. The off-gas from the bioreactors was dried prior to passing through CO2
and O2 gas analysers. This enabled the oxygen utilisation rate, -rO2, carbon dioxide
utilisation rate, -rCO2, and biomass concentration, cX, to be determined [4]. The
experimental equipment used is described in greater detail elsewhere [3].
The total iron concentration in solution was determined by both atomic adsorption
spectroscopy (AAS) and by titration with potassium dichromate [22]. This enabled the
ferric/ferrous-iron ratio and the ferrous- and ferric-iron concentrations to be determined
using a calibration curve for the specific electrode and the Nernst equation [5]. Changes in
the ferrous- and ferric-iron and total iron concentrations with time allowed the ferrous-iron
production rate, rFe2 + , to be calculated.
4.

In the first instance an attempt was made to determine the kinetic parameters, YFemax2 + X ,
and m Fe 2 + using the Pirt Equation [9] i.e. by plotting vs. q Fe 2 + . A typical plot of vs.
q Fe 2 + for the experiment performed at 40C and pH 1.10 is shown in Figure 2.
From Figure 2 it is evident that there are three distinct linear regions, hence it was not
possible to determine the values of YFemax2 + X and m Fe 2 + in this way. However, the similarity
between Figs. 1 and 2 suggests that the micro-organisms used were in fact limited by
different factors during the period over which the experiment was performed and should
therefore be modelled using the variable maintenance equation, viz. Equation 14.
The values of the maximum bacterial yield on ferrous-iron, YFemax2 + X , and the constant
maintenance coefficient on ferrous-iron, m Fe 2 + , were therefore calculated by linear
regression of the data obtained during the latter stages of the batch experiments (region
cd); it was assumed that there was no variable maintenance requirement in this region.
v
This allowed the variable maintenance coefficient on ferrous-iron, m Fe
2 + , to be estimated
from a regression line drawn through the data obtained during the initial stages of the
batch experiment (region ab). Finally, the maximum bacterial specific growth rate, max ,
, were determined
and the maximum bacterial specific ferrous-iron utilisation rate, q max
Fe 2 +
1220
from the intercept of the regression lines drawn through the data obtained during the initial
and latter stages of the batch experiments, as demonstrated in Figure 1.
The average ferrous-iron and oxygen-based parameters calculated in this manner are
listed in Table 1 and Table 2. The validity of the yield and maintenance coefficients listed
in Table 1 and Table 2 was checked by comparing the experimental (calculated) values
with the values predicted by the "degree of reduction" balance i.e. Equations 15 and 16.
These comparisons are displayed in Figure 3, from which it is apparent that the correlation
is good (average correlation coefficient R2 = 0.9596).
-1
12
2+
-1
qFe2+ (mol Fe .(mol C) .h )
16
b
c
0
0.00
d
0.02
0.04
0.06
0.08
-1
(h )
Figure 2. Variation in the experimentally observed bacterial specific ferrous-iron

utilisation rate, q Fe2 + , with changes in the bacterial specific growth rate,
Table 1. Average ferrous-iron based bioenergetic parameters
Experimental conditions
30C
pH 1.70
35C
40C
14.65
22.31
(mol Fe2+.(mol C)-1.h-1)
0.0056
m Fe2 + (mol C.(mol Fe2+)-1.h-1)

v
2+
-1 -1
m Fe
2 + (mol Fe .(mol C) .h )
max
Fe 2 +
-1
-1
(mol Fe .(mol C) .h )
max
Fe 2 + X
2+
max (h-1)
pH 1.50
40C
pH 1.30
pH 1.10
10.65
14.32
18.30
16.36
0.0055
0.0036
0.0058
0.0054
0.0058
0.540
0.855
0.002
0.95
2.26
1.00
7.85
6.50
11.32
6.06
10.08
5.95
0.079
0.119
0.038
0.077
0.087
0.089
1221
Table 2. Average oxygen based bioenergetic parameters

Experimental conditions
30C
pH 1.70
35C
40C
q Omax
2
3.37
5.63
YOmax
2X
0.0228
m O 2 (mol C.(mol Fe ) .h )
pH 1.50
40C
pH 1.30
pH 1.10
2.61
3.50
4.49
4.00
0.0227
0.0158
0.0236
0.0221
0.0239
0.131
0.203
0.0917
0.2312
0.5638
0.2423
m Ov 2 (mol Fe2+.(mol C)-1.h-1)
2.32
1.59
2.74
1.52
2.51
1.49
max (h-1)
0.074
0.123
0.040
0.077
0.087
0.090
2+ -1
-1
0.032
0.60
Figure 3(b)
0.024
-1
0.016
2+
4YFe2+/(1-4.2YFe2+)
-1
(mFe2+)/4 (mol Fe .(mol C) .h )
Figure 3(a)
0.008
0.000
0.000
0.008
0.016
0.024
0.45
0.30
0.15
0.00
0.00
0.032
0.15
-1
YO (mol C.(mol O2) )
0.30
0.45
-1
0.60
-1
mO (mol O2.(mol C) .h )
2
Figure 3(c)
2+
-1
-1
(mFe2+ )/4 (mol Fe .(mol C) .h )
3.2
2.4
1.6
0.8
0.0
0.0
0.8
v
1.6
2.4
-1
3.2
-1
mO (mol O2.(mol C) .h )
2
Figure 3. Comparison of the predicted and experimental relationships for (a) the
maximum bacterial yield, (b) the constant maintenance coefficient and (c) the
variable maintenance coefficient
1222
In addition to the above, it is also apparent from the results listed in Table 1 and Table
2 that the maximum bacterial yields on ferrous-iron and oxygen and their respective
constant and variable maintenance coefficients, did not vary significantly with either
temperature or pH. For this reason, average maximum bacterial yield and constant
maintenance coefficients on ferrous-iron and oxygen were calculated by linear regression
using the data from the latter stages of all the batch experiments performed during this
investigation. These values are listed in Table 3, together with previously reported values
of these parameters. From the results presented in Table 3 it is apparent that the average
maximum bacterial yields and constant maintenance coefficients determined during the
course of this investigation are similar to previously published values.
Table 3. Average values of the maximum bacterial yield and constant maintenance
coefficient on ferrous-iron and the maximum bacterial yield and maintenance
coefficient on oxygen together with the values reported previously
Breed [3]
Van Scherpenzeel et al.

[23]
Current work
YFemax
(mol C.(mol Fe2+)-1)
2+
X
0.0075
0.011
0.0060
m Fe 2 + (mol Fe2+.(mol C)-1.h-1)
1.196
0.444
1.261
YOmax
(mol C.(mol O2)-1)
2X
0.0279
0.046
0.0247
m O 2 (mol O2.(mol C)-1.h-1)
0.2376
0.0425
0.3096
The maximum bacterial specific ferrous-iron utilisation rates, q max

, and the maximum
Fe
2+
bacterial specific growth rates, max , calculated using the values listed in Table 3 and the
variable maintenance equation, Equation 14, are listed in Table 4, together with previously
published values of these parameters.
Table 4. Comparison of the maximum bacterial specific growth rates and ferrousiron utilisation rates calculated using the variable maintenance equation and those
published by Breed [3]
Experimental
conditions
30C; pH 1.70
35C; pH 1.70
40C; pH 1.70
40C; pH 1.50
40C; pH 1.30
40C; pH 1.10
Breed [3]
max
(h-1)
0.0397
0.0638
0.0862
0.1238
0.1077
0.1027
Current work
max
Fe 2 +

8.65
11.01
13.62
19.02
15.57
15.26
max
(h-1)
0.074
0.123
0.040
0.077
0.087
0.090
q max
Fe 2 +
(mol O2.(mol C)-1.h-1)

14.65
22.31
10.65
14.32
18.30
16.36
From these results it is apparent that the values of q max

and max calculated using the
Fe
variable maintenance model are also similar to previously published values of these
parameters. The results presented thus suggest that during the initial stages of a batch
experiment (region ab in Figure 2) the microorganisms are not substrate-(energy-)limited,
whereas in the latter stages of the experiment (region cd in Figure 2) they are substrate
(energy) limited. The points, (b) and (c) in Figure 2 represent points at which a change in
the limiting substrate occurs. In fact, within the transition region (region bc in Figure 2)
the limiting substrate is continually changing, hence a line with a negative slope is
2+
1223
obtained. In other words, during batch experiments the growth of the microorganisms are
limited by more than one factor separately, hence the data cannot be modelled using the
constant maintenance equation. However, it can be modelled using the variable
maintenance equation.
5.
CONCLUSIONS
The results of the batch experiments performed during the course of this study
indicate that the maintenance requirement of microorganisms is actually a combination of
the variable and constant maintenance energy requirements. The variable maintenance
requirement depends on the limiting substrate, whereas the constant maintenance
requirement is dependent on the energy source. Where the bacterial growth rate is limited
solely by the energy source, the variable maintenance requirement decreases to zero.
Modelling the bioenergetics of energy sufficient cultures assuming a constant maintenance
energy requirement thus results in a significant overestimation of the maximum bacterial
yield. However, the variable maintenance equation proposed by Pirt [15] may be used to
determine the maximum bacterial yields and the maintenance coefficients from both
substrate lean and substrate rich experiments. This is because this model is able to account
for variations in the growth rate as a result of changes in the maintenance requirement due
to the micro-organisms being limited by more than one factor, e.g. ferrous-iron, ammonia,
phosphate, sulfate, etc.
It is thus suggested that the variable maintenance equation be used instead of the
constant maintenance equation when quantifying the bioenergetics of microorganisms,
including those encountered during the bioleaching, unless the energy source is known to
be the limiting substrate. The above is especially important for the case of bioleaching
using ferrous-iron oxidising microorganisms at low redox potentials (high ferrous-iron
concentrations), e.g. during mesophilic chalcopyrite bioleaching and during thermophilic
bioleaching.
REFERENCES
1. G.S. Hansford and T. Vargas, Hydrometallurgy, 59 (2001) 135.
2. W. Sand, T. Gehrke, P.-G. Jozsa and A. Schippers, Hydrometallurgy, 59 (2001) 159.
3. A.W. Breed, Studies on the mechanism and kinetics of bioleaching with special
reference to the bioleaching of refractory gold-bearing arsenopyrite/pyrite
concentrates, PhD Thesis, University of Cape Town, Cape Town, South Africa, 2000.
4. M. Boon, G.S. Hansford and J.J. Heijnen, Biohydrometallurgical Processing I, T.
Vargas, C.A. Jerez, J.V. Wiertz and H. Toledo (eds.), University of Chile, Santiago,
Chile, 1995.
5. M. Boon, Theoretical and experimental methods in the modelling of biooxidation
Kinetics of sulphide minerals, PhD Thesis, Technische Universiteit Delft, The
Netherlands, 1996.
6. N.J. Coram and D.E. Rawlings, Appl. Environ. Microbiology, 68 (2002) 838.
7. D.P. Kelly and A.P. Wood, Int. Jour. Systematic and Evolutionary Microbiol., 50
(2000) 511.
8. M. Nemati, S.T.L Harrison, C. Webb and G.S. Hansford, Biochem. Eng. J., 1 (1998)
171.
9. S.J. Pirt, Proceedings of the Royal Society of London. Series B: Biological Sciences,
163 (1965) 224.
1224
10. J.G. Kuenen, Growth yields and maintenance energy requirement in Thiobacillus
species under energy limitation, Arch. Microbiol., 122 (1979) 183.
11. A.J. Downs and C.W. Jones, Arch. Microbiol., 105 (1975) 159.
12. O.M. Neijssel and D.W. Tempest, Arch. Microbiol., 106 (1975) 251.
13. A.H. Stouthamer and C.W. Bettenhausen, Arch. Microbiol., 102 (1975) 187.
14. O.M. Neijssel and D.W. Tempest, Arch. Microbiol., 107 (1976) 215.
15. S.J. Pirt, Arch. Microbiol., 133 (1982) 300.
16. A.W. Breed, C.J.N. Dempers, G.E. Searby, M.N. Gardner, D.E. Rawlings and G.S.
Hansford, Biotechnol. Bioeng., 65 (1999) 44.
17. A.W. Breed and G.S. Hansford, Biochem. Eng. J., 3 (1999) 193.
18. C.A. Jones and D.P. Kelly, J. Chem. Technol. Biotechnol., 33B (1983) 241.
19. J.A. Roels, Energetics and kinetics in biotechnology, Elsevier Biomedical Press,
Amsterdam, 1983.
20. A.W. Breed, S.T.L. Harrison and G.S. Hansford, IBS-BIOMINE 97, Australian
Mineral Foundation, Glenside, Australia, 1997.
21. W. Vishniac and M. Santer, Bacteriol. Revs., 21 (1957) 195.
22. G.H. Jeffrey, J. Bassett, J. Mendham and R.C. Denney (eds.), Vogels Textbook of
Quantitative Chemical Analysis, 5th edition, Longman Scientific & Technical, New
York, 1989.
23. D.A. van Scherpenzeel, M. Boon, C. Ras, G.S. Hansford and J.J. Heijnen, Biotechnol.
Prog., 14 (1998) 425.
1225

The kinetics of thermophilic ferrous-iron oxidation

G.E. Searby and G.S. Hansford
Gold Fields Mineral Bioprocessing Laboratory, Department of Chemical Engineering,
University of Cape Town, Private Bag, Rondebosch, 7701, South Africa
Fax: +27-21-689 7579. E-mail: giles@chemeng.uct.ac.za
Abstract
The kinetics of ferrous iron oxidation by a thermophilic archaeal culture was
investigated in continuous culture, at temperatures ranging from 65 to 75C, and a pH of
1.5.
The reaction kinetics were followed by on-line monitoring of the solution redox
potential and of the concentrations of oxygen and carbon dioxide in the off gas stream.
The specific rate of ferrous iron utilisation was calculated from the rates of oxygen and
carbon dioxide utilisation via a degree of reduction balance, and examined as a function of
the ferric/ferrous iron ratio, determined from the redox potential and checked by
measuring dissolved iron concentrations.
The results were modelled using a kinetic model developed for mesophilic oxidation
based on Michaelis-Menten enzyme kinetics and proportional to the ferric/ferrous iron
ratio. The effect of temperature was accounted for by the incorporation of an Arrhenius
term.
The thermophiles achieved specific iron utilisation rates similar to those reported for
L. ferrooxidans, but was active at significantly lower redox potentials, and less active at
high redox, more comparable to At. ferrooxidans.
Keywords: thermophiles, bioleaching, kinetics, iron oxidation, continuous culture

1.
INTRODUCTION
The recalcitrance of chalcopyrite to bioleaching by mesophilic microorganisms has
led to interest in high temperature leaching. Thermophilic archaea have been shown to be
capable of effective leaching. (Le Roux, 1988, and Dew et al, 1999).
Like mesophilic systems, thermophilic mixed cultures often contain species that are
predominantly iron oxidisers and others that are predominantly sulfur oxidisers. This
indicates the possibility that the leaching mechanism is similar to that generally accepted
for conventional bioleaching microbes, i.e. a chemical ferric leach followed by microbially
mediated ferrous iron oxidation reforming the primary ferric leach reagent, and microbial
oxidation of sulphur compounds released.
An understanding of the ferrous iron oxidation kinetics of these archaea at their
operating temperature is therefore desirable so that their bioleaching capabilities may be
fully exploited.
1227
Mesophilic ferrous iron oxidation (mostly focussed on Acidithiobacillus ferrooxidans)

has been studied extensively. A number of kinetic models have been proposed, starting
with empirical systems such as the logistic equation and simple Monod models, followed
by more developed models. These were often based on Michaelis-Menten enzyme kinetics
and adapted to show substrate and/or product inhibition and effects of various parameters
such as dissolved oxygen concentration, pH, and temperature.
Ingledew (1982) proposed a chemiosmotic theory for ferrous iron oxidation in At.
ferrooxidans, involving the generation of a transmembrane potential and a proton-motive
force by splitting the two half reactions in the overall chemical reaction of ferrous iron
oxidation across the cell membrane. This model highlighted the influence of the solution
redox potential on the microbial growth and oxidation kinetics and has led to the
generation of models based on redox potential (Huberts, 1994; Boon, 1995; and Meruane
et al, 2002).
It is proposed that the oxidation of ferrous iron by thermophiles may be controlled by
a similar mechanism, and hence that the kinetics can be described using a similar model.
2.

Experiments were performed in three 1.7L, air-sparged glass vessels, each with an
H/D ratio of 1.32, using a working volume of 1L. Temperature in the reactors was
maintained at 65, 70 and 75C by a heated water jacket. pH was maintained at 1.5 by
manipulation of the feed pH and by addition of concentrated sulphuric acid.
The reactors were kept airtight, allowing the measurement of the change in oxygen
and carbon concentration in the entering and exiting gas streams. Off-gas analysis and the
measurement of the change in the solution potential have been shown to be an effective
system for the investigation of ferrous-iron oxidation (Boon, 1996).
Compressed air was passed through a gas-chiller and a series of filters to obtain bonedry medical quality air and fed into the reactor at a constant rate via a Brooks 5850S mass
flow control valve, and sparged below the impeller. Agitation and gas mixing was
achieved by a Lightnin 315 impeller running at 400 rpm. Air leaving the reactor passed
through a reflux condenser, which served the dual purpose of drying the gas before it
enters the gas analysis system, and avoiding evaporative water loss in the reactor. The
offgas then passed through a cloth filter and a Hartmann and Braun CGEK sample gas
conditioner before entering the analysers. Oxygen concentration was measured by a
Hartmann and Braun Magnos 6G paramagnetic oxygen analyser and carbon dioxide using
a Hartmann and Braun Uras 4 infrared photometer.
The culture used was a mixed culture of extreme Sulfolobus-like thermophilic archaea
grown at pH 1.5 and at a temperature of 70C. The culture was taken initially from a semicontinuous system grown on a chalcopyrite concentrate and has been maintained in
continuous culture on ferrous iron for more than 2 years.
The culture was grown in a basal salt medium containing 0.5 g/L MgSO4.7H2O, 0.4
g/L (NH4)2SO4 0.02 g/L K2HPO4.3H2O and 0.1 g/L KCl (Clark and Norris, 1996). The
energy source was 12g/L FeSO4. Reduced sulphur was added as 0.225g/L K2S4O6 as the
thermophiles are unable to assimilate sulphate.
Experiments were performed in continuous culture at dilution rates ranging from
0.015 to 0.9 hr-1. Continuous flow was obtained by feeding the fresh nutrient medium via
a variable speed peristaltic pump, and by removing the resultant reactor liquor at a fixed
liquid level by a fixed speed peristaltic pump, maintaining a constant volume. The reactors
1228
were maintained at each dilution rate for 5 residence times before steady state was
assumed. Steady state was verified by the generation of steady data for a further residence
time. Wall growth was eliminated by daily scrubbing of all reactor surfaces.
The reaction kinetics was followed by off-gas analysis. In an autotrophic system, the
only carbon source is atmospheric carbon dioxide, and thus the carbon dioxide utilisation
rate can be used to follow the growth kinetics.
rX = rCO 2
(1)
This provides a non-invasive on-line measurement of the cell concentration, in terms

of moles of carbon fixed.
cx =
rCO 2
D
Autotrophic microbes oxidise iron as their energy source,
4Fe2 + + 4H + + O 2
4Fe3+ + 2H 2O
microbes
(2)
(3)
and utilise the energy produced for cell synthesis (typical cell stoichiometry CH1.8O0.5N0.2)
CO2, H2O, NH +4
CH1.8O0.5 N 0.2
(4)
Simultaneous solution of the species and charge balances for the species in Reactions
3 and 4, or the degree of reduction balance over the same set of species (Roels, 1983)
(5)
i ri = 0
i
yields the rate of microbial ferrous iron oxidation in terms of the measured off-gas
parameters.
rFe2 + = 4.2rCO 2 4rO 2
(6)
The solution redox potential was measured using a Pt Ag/AgCl2 redox electrode. This
was related to the ferric/ferrous-iron ratio in the reactor via the Nernst equation,
approximating activities as concentrations.
E
E '0
RT [Fe3+ ]
+
ln
zF [Fe2 + ]
(7)
The ferrous-iron and total iron concentrations were determined by titration with
K2Cr2O7 (Jeffrey et al, 1989). Ferric iron concentrations can then be calculated by the
difference between the ferrous and the total iron concentrations. This provided a check for
the ferric/ferrous-iron ratio as determined by redox measurements.
The specific iron utilisation rate was modelled using a Michaelis-Menten-form model
dependent on the ferric/ferrous ratio, previously used to model L. ferrooxidans (van
Scherpenzeel, 1998, Breed et al, 1999) and At. ferrooxidans (Boon, 1996)
q Fe 2 + =
q max
Fe 2 +
1 + K Fe 2 +
[ Fe 3+ ]
[ Fe 2 + ]
(8)
Breed et al (1999) described the effect of temperature on Leptospirillum ferrooxidans

as an Arrhenius function of the qmax constant
1229
max
Fe 2 +
= k 0e
Ea
RT
(9)
and a linear function of the KFe2+

K Fe 2 + = 0.0002T 0.0453
(10)
Nemati and Webb described the effect of temperature on At. ferrooxidans simply as
an Arrhenius function of qFe2+max.

Continuous reactors were run at temperatures of 65, 70, and 75C, pH 1.5, 12gFe2+.L-1
and steady state data was obtained at dilution rates of 0.02, 0.033, 0.04, 0.05, 0.059, 0.071
and 0.08 h-1. Washout occurred at dilution rates between 0.09 and 0.11 hr-1.
All three systems are characterised by a maximum cell concentration at intermediate
dilution rates decreasing at both high and low dilution rates (Figure 1). At low dilution
rates the system is energy-limited increasing the cell maintenance and reducing the cell
concentration, whilst at high dilution rates, the high growth rates stress the cells, which
also leads to increased maintenance requirements.
The rates of ferrous iron oxidation determined from the rates of oxygen and carbon
dioxide utilisation via Equation 6 were compared in Figure 2 to those determined from the
difference between feed and reactor contents iron concentrations. The results confirm the
applicability of the degree of reduction balance in thermophilic systems.
15
-1
Cell concentration (mmolC.l)
3.
1.5
-rFe2+
10
5
0.5
0
0
0
0.05
0.1
-1
Dilution Rate (hr )
0.15
Figure 1. Steady state cell

concentrations for each dilution
rate at 65(), 70 () and 75C ()
10
15
-4rO2 - 4.2 rCO2

Figure 2. Confirmation of degree
of reduction balance at 65(), 70
() and 75C ()
The dependence of the system on the redox potential via the ferric/ferrous iron ratio,
as predicted by Equation 8, was demonstrated by plotting the specific iron utilisation rate
obtained at each steady state against the ferric/ferrous iron ratio measured (Figure 3).
1230
(molFe .molC .hr )
Specific iron utilisation rate
20
15
10
0
0.1
10
3+
100
1000
2+
[Fe ]/[Fe ]
Figure 3. Specific iron utilisation rates obtained at 65(), 70 () and 75C () as a

function of the ferric/ferrous iron ratio and modelled by Equation 8
The specific oxidation rates increase with an increase in temperature, this would
indicate that the experiments were run within the microbes' preferred operating range.
Above this range, thermal deactivation is expected, and a consequent drop in rate.
The results show a similar dependence on the ferric/ferrous iron ratio as was observed
in mesophilic systems. The specific utilisation rate reaches a maximum at the lowest
ferric/ferrous iron ratio corresponding to the fastest dilution rate before washout. The rate
decreases with increasing redox potential until the ferrous iron concentration reaches
levels below the minimum required for further growth. This has been described as the
threshold substrate concentration (Braddock et al, 1984).
The data was modelled using Equation 8, obtaining the kinetic constants q max
and
Fe 2+
K Fe2 + by linear regression. The results show good agreement between data and prediction
over the range of potentials measured, suggesting that thermophilic ferrous iron oxidation
is proportional to the ferric/ferrous iron ratio and in turn the redox potential.
Table 2. Kinetic constants qmax
and K Fe2+ derived from the data and compared to
Fe 2+
literature values
Present work
L. ferrooxidans
(Breed et al, 1999)
At. ferrooxidans
(Boon, 1995)
Temperature
q max
Fe 2+
K Fe2 +
65
70
75
11.08
12.32
19.36
0.155
0.098
0.227
40
13.58
0.0033
30
8.8
0.05
The K Fe2 + values obtained show more commonality with those of Acidithiobacillus
ferrooxidans, than with those of Leptospirillum ferrooxidans. This has been described as
the affinity of the microbe for ferrous iron, with L. ferrooxidans having the capacity to
oxidise much lower concentrations of iron and hence operate at much higher redox
potentials. Higher K Fe2 + values indicate that the ferric/ferrous iron ratio or redox potential
1231
at which the oxidation and growth rates begin to drop away is lower. This may be of
interest in the investigation of why thermophiles are more able to bioleach chalcopyrite,
where better leaching may be achieved at lower solution redox potentials.
(molFe .molC .h )
25
20
15
10
5
0
0.1
10
100
3+
1000
10000
2+
[Fe ]/[Fe ]
Figure 4. A comparison of the relationship between q Fe2+ and the ferric/ferrous iron
ratio for thermophiles (70C (), 75C ()), and mesophiles At. ferrooxidans (----),
and L. ferrooxidans(----)
The specific growth rate; this was compared to the specific iron utilisation rate to
determine the yield and maintenance coefficients from the Pirt equation (Pirt, 1965)
q Fe 2 + =
Y max
2+
Fe
+ m Fe 2 +
(11)
Straight lines fitting the constant maintenance Pirt equation (Pirt, 1965) are only
expected for substrate limited growth. Deviations from linearity at high redox potential
and high growth stress may be attributed to increased maintenance requirements and may
be better described using the variable maintenance equation (Pirt, 1982).
max
Table 3. Bioenergetic parameters Yield ( YFe
) and the maintenance coefficient
2+
X
(mFe2+)
Temperature
(C)
65
70
75
YFemax
2+
X
.
m Fe2 +
2+ -1
(mol C (mol Fe ) )
0.0082
0.0082
0.0071
2+.
(mol Fe mol C-1.h-1)

0.394
1.265
1.308
The maximum yield is a very weak function of temperature, showing no significant

change within the temperature range measured, and not differing greatly from values
measured at mesophilic temperatures, lying between two values obtained for
Leptospirillum-like systems 0.0059 and 0.011 molC.(molFe2+)-1 obtained by Breed et al
(1999) and van Scherpenzeel (1998) respectively. The maintenance coefficient is a much
stronger function of temperature, indicating increased stress with increased temperature.
(Figure 5) can be modelled by the Arrhenius
The effect of temperature on q max
Fe 2+
equation. The data can thus be modelled as
1232
54 .437
RT
2 .70 10 e
[ Fe 3 + ]
1 + K Fe 2 +
[ Fe 2 + ]
15
10
ln(qmax)
-1
qFe2+ (molFe .molC .h )
20
-1
(12)
2+
q Fe 2 + =
0
0.00285
0.05
-1
0.1
y = -6547.6x + 21.718
R2 = 0.8814
0.0029
0.00295
-1
0.003
1/T (degC )
Specific Growth Rate (h )
Figure 5. Determination of the

max
)
energetic parameters, yield ( YFe
2+
X
Figure 6. Effect of temperature on

qmax
, showing fit of Arrhenius
Fe 2+
and the maintenance coefficient

( m Fe2+ ) at 65(), 70 () and 75C ()
equation
(molFe .molC .h )
-1
15.00
-1
2+
20.00
10.00
5.00
0.00
0.1
10
3+
100
1000
2+
[Fe ]/[Fe ]
Figure 7. Specific iron utilisation rates obtained at 65(), 70 () and 75C () were
compared to values predicted using Equation 8 () and Equation 12 (---)
4.
CONCLUSIONS
The kinetics of ferrous iron oxidation using thermophilic archaea was investigated in
continuous culture. The rate of iron consumption calculated from measured iron
concentrations correlated well with that calculated from off gas analysis via the degree of
reduction balance. This indicates that the overall reaction stoichiometry for microbial
ferrous iron oxidation remains the same as at mesophilic temperatures.
1233
The results were successfully modelled using a Michaelis-Menten form kinetic model,
developed by Boon (1996) for mesophilic bioleaching, in which the specific rate of iron
oxidation is proportional to the ferric/ferrous iron ratio. The effect of temperature was
accounted for by the incorporation of an Arrhenius term.
The maximum specific iron utilisation rates determined for the thermophilic systems
were not greatly different to those achievable in mesophilic systems, so simple rate of iron
oxidation cannot explain the advantage of bioleaching at elevated temperatures. However,
the lower tolerance of high redox potential can lead to a bioleach system that self-regulates
its redox to a lower level where the rate of ferric leaching is enhanced.
NOMENCLATURE
cX
concentration of bacteria
mmol C.L-1
dilution rate
h-1
redox potential of the solution (Pt-Ag/AgCl)
mV
E '0
equilibrium redox potential for Ag/AgCl electrode
mV
Ea
activation energy
kJ.mol-1
Faraday constant
coulombs.mol-1
[Fe2+]
concentration of ferrous-iron
mmol Fe2+.L-1
[Fe3+]
concentration of ferric-iron
mmol Fe3+.L-1
Ko
Arrhenius constant
mol Fe2+.(mol C)-1.h-1
KFe2+
ferrous-iron based kinetic constant in bacterial

ferrous-iron oxidation
dimensionless
mFe2+
maintenance coefficient on ferrous-iron
qFe2+
bacterial specific ferrous-iron utilisation rate
q max
Fe 2 +
maximum bacterial specific ferrous-iron utilisation

rate
Universal gas constant
kJ.K-1.mol-1
-rFe2+
ferrous-iron utilisation rate
mmol Fe2+.L-1.h-1
-rCO2
carbon dioxide utilisation rate
mmol CO2.L-1.h-1
-rO2
oxygen utilisation rate
mmol O2.L-1.h-1
rX
biomass production rate
mmol C.L-1.h-1
max
YFe
2+
X
maximum bacterial yield on ferrous-iron
mol C.( mol Fe2+)-1
number of electrons involved in a reaction
dimensionless
Degree of reduction of species i
dimensionless
bacterial specific growth rate
h-1
REFERENCES
1. Boon M., 1996, "Theoretical and Experimental Methods in the Modelling of
Biooxidation kinetics of Sulphide Minerals", PhD Thesis, Technische Universiteit
Delft, The Netherlands.
2. Boon M., Hansford G.S. and Heijnen J.J., 1995, "The Role of Bacterial Ferrous Iron
Oxidation in the Bio-Oxidation of Pyrite", Vargas T., Jerez C.A., Wiertz J.V. and
1234
Toledo H. (eds.), Biohydrometallurgical Processing, 1, Santiago: University of Chile,

153-163.
3. Braddock J.F., Luong H.V. and Brown E.J., 1984, "Growth kinetics of Thiobacillus
ferrooxidans isolated from arsenic mine drainage", Applied and Environmental
Microbiology 48, (1), 48-55.
4. Breed A.W., Dempers C.J.N., Searby G.E., Gardner M.N., Rawlings D.E. and
Hansford G.S., 1999, "The Effect of Temperature on the Continuous Ferrous-iron
Oxidation Kinetics of a Predominantly Leptospirillum ferrooxidans Culture",
Biotechnology and Bioengineering, 65, 44-53.
5. Clark, D.A., Norris, P.R., 1996, Oxidation of Mineral Sulphides by Thermophilic
Micro-organisms Minerals Engineering vol. 9 no. 11, pp 1119-1125
6. Crundwell, F. K., 1997, The kinetics of the chemiosmotic proton circuit of the ironoxidising bacterium Thiobacillus ferrooxidans, Bioelectrochemistry and
Bioenergetics, 43, 115-122
7. Dew, D.W., van Buuren, C., McEwan, K., Bowker, C., 1999, Bioleaching of Base
Metal Sulphide Concentrates: A Comparison of Mesophile and Thermophile Bacterial
Cultures, In: Proceedings of International Biohydrometallurgy Symposium IBS 99,
Ed: by R. Amils, A. Ballester, Madrid, Elsevier, pp. 229-238
8. Huberts R., 1994, "Modelling of ferrous sulfate oxidation by iron oxidising bacteria - a
chemiosmotic and electrochemical approach", PhD Thesis, University of the
Witwatersrand, Johannesburg, South Africa.
9. Ingledew W.J., 1982, Thiobacillus ferrooxidans: The bioenergetics of an acidophilic
chemolithotroph, Biochimica et Biophyisca Acta, 683, 89-117
10. Jeffery, G.H., Basset, J., Mendham, J. and Denney, R.C., 1989, Vogels textbook of
quantitative chemical analysis, 5th edition, Longman Scientific and Technical, New
York.
11. Le Roux, N.W., Wakerley, D.S., 1988, Leaching of chalcopyrite (CuFeS2) at 70 C
using Sulfolobus, In: Biohydrometallurgy: Proceedings of the International
Symposium Warwick 1987, Ed. by P.R. Norris and D.P Kelly, University of Warwick,
305-319.
12. Meruane, G., Salhe, C., Wiertz, J., Vargas, T., 2002, A novel electro-chemicalenzymatic model which quantifies the effect of the solution Eh on the kinetics of
ferrous iron oxidation of Acidithiobacillus ferrooxidans., 80, (3), 280-288
13. Nemati M. and Webb C., 1997, "A kinetic model for biological oxidation of ferrousiron by Thiobacillus ferrooxidans", Biotechnology and Bioengineering, 53, (5). 478486.
14. Pirt S.J., 1965, "The maintenance energy of bacteria in growing cultures", Proceedings
of the Royal Society B, 163, 224-231.
15. Pirt S.J., 1982, "Maintenance energy: a general model for energy-limited and energysufficient growth", Archives of Microbiology, 133, 300-302.
16. Roels J.A., 1983, Energetics and kinetics in biotechnology, Elsevier Biomedical
Press, Amsterdam, 20-31
17. van Scherpenzeel D.A., Boon M., Ras C., Hansford G.S. and Heijnen J.J., 1998,
"Kinetics of ferrous-iron oxidation by Leptospirillum bacteria in continuous cultures",
Biotechnology Progress, 14, 425-433.
1235

The role of microorganisms in dispersion of thallium

compounds in the environment
A. Sklodowska, M. Golan and R. Matlakowska
Warsaw University, Faculty of Biology, Laboratory of Environmental Pollution Analysis,
CEMERA Centre of Excellence, Miecznikowa 1, 02-096 Warsaw, Poland
E-mail: asklodowska@biol.uw.edu.pl
Abstract
Thallium is a highly toxic element and very rarely studied in the context of
environmental hazards connected with zinc and non-ferrous metal industry.
Microorganisms naturally existing in post-flotation and smelt wastes can participate in
thallium release from waste deposits and can contribute to its dispersion in the
environment. Twenty-one isolates were obtained from wastes of a non-ferrous smelter in
Southern Poland characterised by high heavy metal contamination. Ten isolates showed
high activity in thallium leaching from wastes (post-flotation and smelt wastes) as well as
from pure thallous sulphide. Additionally, cadmium and lead were bioleached from
wastes. The isolated bacteria indicated thallium resistance at a concentration up to 100
mg/l and some of them were able to survive in good condition at a concentration of up to 4
g/l. The same bacteria were isolated from rivers and wastewater in this region. A
preliminary characterisation of isolates was performed. It was shown that some petroleum
products i.e. asphalt-base crude occasionally used for waste immobilisation at the edge of
pond or flotation surfactants partially stopped the activity of sulphide oxidising bacteria.
1.
INTRODUCTION
Microorganisms naturally existing in post-flotation and smelt wastes can participate in
thallium release from waste deposits and can contribute to its dispersion in the
environment.
The role of microbes in the dispersion of inorganic metal salts (especially sulfides)
has been known for years. Oxidation of these compounds is the way of gaining energy,
needed in many biochemical processes such as CO2 fixation etc. Metal ions are unused
products of reactions, which can penetrate into the environment. Since the 70s it has been
known, that thiobacilli (today genus Thiobacillus is divided into a few genera, belonging
to another subclasses of Proteobacteria [6]) can divide thallous sulfide, to obtain sulfide
ions the energy source [5]. Free thallous ions can penetrate into the environment, and
take part in many biogeochemical cycles.
Microbes can methylate thallium ions, producing dimethylthallium Me2Tl+. A
method of estimating the concentration of this compound in environmental samples was
worked out by Schedlbauer and Heumann [8]. The mechanism of this process is still
unknown. Probably methylated cobalamine is needed, as in the mercury methylation [4].
1237
The role of this process in the environment is unknown. Me2Tl+ is very stable [8].
This can suggest, that it is the way of detoxification of the microbes environment,
similarly to the methyllation of mercury [4].
For the majority of microbes thallium is highly toxic. By disturbing metabolic
processes, it stops cell division. For Anacystis nidulans 15 ppm of thallium in medium
completely inhibit all metabolic processes, for Chlamydomanas reinhardtii only 5 ppm
[7].
Thallium is a highly toxic element and very rarely studied in the context of
environmental hazards connected with zinc and non-ferrous metal industry. The research
carried out in Southern Poland enabled the identification of several regions, which are
seriously threatened by thallium as well as to indicate direct sources of pollution. Polluted
regions included mainly the surroundings of the zinc smelter and post-flotation waste
ponds.
2.1 Collecting and storing of samples

Samples of wastes of a non-ferrous smelter in Southern Poland were collected in
October 2001. The material was stored in single-use, sterile plastic tubes. Microbiological
analysis was carried out within 24 hours. The remaining material was stored at -20C.
2.2 Bacterial strains
Ten strains were isolated from wastes. Non-modified strains of Halothiobacillus
neapolitanus and Paracoccus versutus received from The Department of Bacterial
Genetics, Warsaw University were used as a reference due to their potential similarity to
isolates from wastes.
2.3 Isolation of strains
5 g of fresh wastes were added to 50 ml of sterile Beijerincks medium. The mixture
was incubated on rotary shaker, at room temperature. After 24 hours, the culture was
transferred to solid medium and cultivated at 28C for 4 days. Single colonies were
inoculated on fresh solid medium every 5 days.
2.4 Media
1. Beijerincks medium [2] with 10 g of Na2S2O3.5H2O, as the only energy source
and 2 ml of Tuovinens salts [10]. Solid medium contained 25 ml of 3% phenol
reds solution, pH = 7,5.
2. Modified LB medium [9] with 20 g of NaCl, pH = 7,5
3. Modified LB medium with thiosulphate [9] with addition of 20 g Na2S2O3.5H2O,
pH = 7,5.
4. Davis medium [3] with glucose, as carbon source.
5. Modified Beijerincks medium I with 10g of Tl2S instead of thiosulphate.
6. Modified Beijerincks medium II: 60 g of wastes, dried to dry weight in 105C
instead thiosulphate, and without Tuovinens salts.
1238
2.5 Estimation of bacterial cell number in wastes

5 g of fresh wastes were added to 50 ml of sterile NaCl solution (0.9%) and stored on
laboratory rotary shaker for hour. Solutions of this culture were then inoculated on solid
media: Beijerinck, modified LB with NaCl and Davis. Plates were stored at 28C for 7
days. After the incubation the colonies on every plate were counted.
2.6 Characterisation of isolates
Isolated strains were inoculated on solid media to test their ability to grow in different
conditions. Beijerincks, modified LB, modified LB with Na2S2O3 and Davis media were
used. All isolates were stained using Gram method.
The resistance of strains to thallous ions was tested. Bacterial strains were inoculated
on solid modified LB medium with Na2S2O3, containing Tl+ (as TlNO3) in concentrations:
25, 50, 75, 100 ppm, or solid Beijerincks medium with analogous ratios of thallim.
Paracoccus versutus and Halothiobacillus neapolitanus were inoculated on the same
media.
2.7 Preparation of a mixture of strains
Fresh cultures of all isolated strains on Beijerincks medium were prepared. After 5
days of incubation at room temperature on a laboratory shaker 1 ml of each culture was
added to sterile medium without thiosulphate. This mixture was used as an inoculum in
next experiments.
2.8 Experiments
2.8.1 Thallium bioleaching from thallous sulphide
100 ml of Beijerincks medium with 10 g of Tl2S was inoculated with mixture of
bacterial strains. Experiment was conducted in Erlenmayers flasks (300 ml) on a
laboratory shaker at room temperature. Non-inoculated medium, cultivated under the same
conditions, as cultures was control for this experiment. Two series of culture (designated
1st culture and 2nd culture) were prepared. Every day pH, Tl+ concentration and the number
of bacterial cell were measured.
2.8.2 Thallium bioleaching from wastes
200 ml of Beijerincks medium with 60 g of wastes (dried at 105C) was inoculated
with mixture of strains. Incubation was conducted in Erlenmayer flasks (500 ml), on a
shaker at room temperature. Non-inoculated medium served as a control for this
experiment. Two series of culture (designated 1st culture and 2nd culture) were prepared.
The concentration of thallium, cadmium and lead was estimated. Additionally, the pH and
the number of bacterial cell were assessed.
The number of bacterial cells was estimated after staining with DAPI and counting on
filters under the epifluorescence microscope.
2.9 Chemical analysis

The concentration of metals in acidified supernatant and mineralised wastes was
measured using Flame Atomic Absorption Spectrometer SOLAAR M6.
Before the analysis wastes were dried at 105C and mineralised in Millestone
Laboratory Microwave System with 65% HNO3 and 36% H2O2 (9:1).
1239
The analysis of thallium in soils was carried out with diisopropylether extraction,
according to Asami et al. [1]. Mineralised wastes were filtered and distilled water was
added, giving the total volume of 100 ml in the flask 50 ml were moved to Erlenmayer
flask, 6 ml of HBr (36%) and 1 ml of 0,1% solution of CeSO4.4H2O was added. After 15
minutes the total volume was placed into a separator (200 ml), 5 ml of ether was added
and then shaken for 5 minutes. Organic phase was collected and evaporated. The sample
was then dissolved in 5 ml of 3% HNO3
3.

Concentration of heavy metals and pH of wastes of a non-ferrous smelter were
measured and presented in Table 1.
Table 1. Heavy metals concentration and pH of wastes

pH
7.00 7.10
Tl
40 50
Concentration of heavy metals [mg/kg d.w.]

Cd
Pb
Zn
120 130
18000 21000
4500 5000
The bacterial cell number isolated from wastes able to grow on different media was
estimated. For all media (Beijerinks, modified LB and Davis) similar results were
observed: 104-105 cells per mg of wastes (wet weight).
Twenty-one bacterial strains were isolated from wastes on Beijerincks medium.
From all isolates, 10 were chosen on the basis of their growth ability. All isolates were
Gram-negative, or Gram-variable (young cells were negative, and after 10 days of
incubation, positive). Eight of strains were rods, and two of them were too small to
identify their morphology under a light microscope.
Table 2. Characteristics of isolates

Growth on medium:
Strain
Gram
1.
negative
2.
negative
3.
negative
4.
variable
5.
variable
6.
variable
7.
variable
8.
negative
9.
negative
10.
negative
n/e not estimated
Morphology
rod
rod
rod
rod
rod
rod
n/e
rod
rod
n/e
Beijerinck Modified LB
+
+
+
+
+
+
+
+
+
+
+
+
Davis
+
+
+
Modified LB with
Na2S2O3
+
+
+
+
+
+
+
The isolated bacteria indicated thallium resistance at a concentration up to 100 mg/l

and some of them were able to survive in good condition at a concentration of up to 4 g/l
(data not shown).
1240
Table 3. Resistance of freshly isolated bacterial strains to different concentration of

thallium ions
Strain
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
H.neapolitanus
P. versutus
Growth on medium with thallium in concentration:

25 ppm
50 ppm
75 ppm
100 ppm
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
-
The ability of isolates to thallium leaching from pure Tl2S was tested in the first
experiment. Sulphide ion was the only energy source for microorganisms. The
concentration of Tl+ in a supernatant indicated the rate of the leaching process. The
highest concentration of Tl+ (1624 ppm) on the 9th day of cultivation was obtained (Fig.
1). In the 2nd culture the highest concentration was observed at the beginning of the
experiment 250 ppm. From the 5th day it droped to 110 ppm on the last day of
cultivation. Throughout the experiment the biggest concentration was noticed in the 1st
culture. For the first five days the concentration was stable 210-320 ppm, then it started to
increase, reaching 1624 ppm on the 9th day in the 1st culture. This phenomenon may be
explained by the irregular structure of the crystal, or some impurities in crystal net.
In both cultures brown, stable sediment, localised on flask walls, under the liquid line
was observed.
Figure 1. Bioleaching of thallium from Tl2S concentration of Tl ions in leaching

solution
Throughout the experiment bacterial cell number systematically increased in both
cultures. For both series of cultures similar results were obtained from about 1 x 104/ml
to about 7 x 105/ml (Fig. 2).
1241
cell number [104/ml]
80
60
40
20
0
1
10
time (days of experiment)

first culture
second culture
Figure 2. Bacterial cell number per ml of culture during bioleaching of Tl2S
Figure 3 Bacterial cells stained with DAPI adhered to particle of wastes during
bioleaching process
Figure 3 presents mixture of isolates cultivated in mineral medium containing wastes.
The bacterial cells stained with DAPI attached to particle of wastes are visible.
In both cultures pH decrease of about 1 unit was observed. In the first culture the
decrease from 7.40 to 6.50 and in the 2nd one from 7.00 to 6.20 was observed while in the
control from 7.29 to 7.00. In both cultures pH was stable for the first 5 days, then it
decreased (Fig.4.).
To check the ability of isolates to leach heavy metals (esp. thallium) from wastes, the
experiment using modified Beijerincks medium II was carried out. Thallium
concentration in wastes (mg/kg d.w.) was measured and its decrease showed the process
efficiency. Additional parameters of process were: bacterial cell number, pH changes. The
rate of cadmium and lead bioleaching were also measured.
Apart from the fouling of culture medium, no other differences between cultures and
control images were obsewved during the experiment.
In the first culture 40% of thallium was leached. This was more than in the second
culture, where it was only 25%. In the control less than 10% of thallium was leached. At
1242
the beginning of the experiment the concentration of thallium was different for both
cultures and the control. Probably because the wastes are an unhomogenic blend with
unspecified ratio of different compounds (Fig. 5).
Cadmium was bioleached in 100% in the 1st culture and more than in 70% in the 2nd
one. In the control flask 17% decrease of the concentration was noticed (Fig. 6).
Tl [ppm]
Figure 4. Changes of pH of supernatant during bioleaching of Tl2S

50
40
30
20
10
0
1
10
13

first culture
second culture
control
Figure 5. Thalium concetration in wastes during bioleaching experiment
Cd [ppm]
150
100
50
0
1
10
13

first culture
second culture
control
Figure 6. Cadmium concentration in wastes during bioleaching process

For lead 70% decrease of the concentration for the 1st culture and more than 50% for
the 2nd one was obtained at the end of the experiment, while in the control about 25%. For
1243
lead, differences in the concentration at the beginning of the experiment, were smaller
than for thallium and cadmium (Fig. 7).
25000
Pb [ppm]
20000
15000
10000
5000
0
1
10
13

first culture
second culture
control
Figure 7. Lead concentration in wastes during bioleaching process
cell number [106 /ml]
Bacterial cell number during the experiment was similar for both cultures: from
3.20x106 to 8.05x107 per ml in the 1st one and 5.40x106 5.00x107 for the 2nd one. Lagphase was longer in the 2nd culture, than in the 1st one (Fig. 8).
100
80
60
40
20
0
1
9 10 13

first culture
second culture
Figure 8. Bacteria cell number during bioleaching of wastes

Throughout the experiment the pH was stable due to probable buffer capacity of
wastes that includes a significant amount of carbonate. The maximal ratio of changes was
0,8 unit in the 1st culture, 1 unit in the 2nd one and only 0,1 unit in the control (Fig. 9).
Isolated strains are able to leach thallium from both: pure salts and ores. Isolates are
very interesting due to their ability to grow in high concentrations of thallium and other
heavy metals, which are thought to be highly toxic. It was shown that microbes' activity
might be one of the reasons of thallium contamination near ores treatment plants and near
postflotation wastes deposits. This problem raises many new questions concerning the
physiology of their resistance to metal ions. Additionally, isolated microbes might be used
in the biohydrometallurgical methods of this metal extraction.
This research will be continued to understand the problem of the taxonomy and
physiology of these isolates, and find possibilities of using them into biotechnological
processes.
1244
pH
Another issue which arises is how to inhibit microbes activity in the deposits, or how
to retain dissolved metal ions within the deposit.
7,6
7,2
6,8
6,4
6
1
9 10

first culture
control
second culture
Figure 9. Changes of pH of supernatant during bioleaching of wastes

REFERENCES
1. T. Asami, C. Mizui, T.Shimada, M. Kubota, Fresenius J. Anal. Chem. 356 (1996) 348.
2. M.W. Beijerinck, Zentralbl. Bakteriol. Parasitenkd. Infenktionskr. Hyg. Abt. II , 11
(1904) 593.
3. D.H. Davis, M. Doudoroff, R.Y. Stanier, M. Mandel, Int. J. Syst.Bacteriol., 19 (1969)
375.
4. G.M. Gadd, Encyclopedia of Microbiology, Vol. 2, Academic Press, Inc., 1992.
5. M. Galizzi, E. Ferrari, Appl. Environ. Microbiol., 32 (1976) 433.
6. D.P. Kelly, A.P. Wood, Int. J. Syst. Evolution. Microbiol., 50 (2000) 511.
7. B. Lustigman, L.H. Lee, J. Morate, F. Khan, Bull. Environ. Contam. Toxicol., 64
(2000) 565.
8. O.F. Schedlbauer, K.G. Heumann, Appl. Organometal. Chem., 14 (2000) 330.
9. A. Sklodowska, R. Matlakowska, W. Ludwig, Acta Microbiol. Polon., Vol. 45 No 2
(1996) 131.
10. O.H. Tuovinen, D.P. Kelly, Arch. Microbiol., 111 (1973) 257.
1245

Microbiology Fundamentals2 PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Microbiology Fundamentals2 PDF

Uploaded by

Copyright:

Available Formats

C HAPTER 4

15th International Biohydrometallurgy Symposium (IBS 2003)

A model for iron uptake in Acidithiobacillus ferrooxidans based

Laboratory of Bioinformatics & Genome Biology, University of Santiago, Santiago,

Corresponding author: David Holmes: dsholmes2000@yahoo.com.

predictions (HMM-based Protein Sequence Analysis http://www.cse.ucsc.edu/research/

RESULTS AND DISCUSSION

family of genes known as truncated hemoglobins, identified in prokaryotes, protozoa,

Figure 2. Gene organization of the proposed FeIII-siderophore uptake region of the

15th International Biohydrometallurgy Symposium (IBS 2003)

Activity and occurrence of leaching bacteria in mine waste at

MATERIALS AND METHODS

2.1 Sample sites and sample analyses

RESULTS AND DISCUSSION

3.1 Microbial activity

Figure 2. Temporal development of the microcalorimetrically measured activity of

Figure 3. Dependency of microbial activity of samples from the La Union mining

H. L. Ehrich, Geomicrobiology, Marcel Dekker, New York, 2002.

15th International Biohydrometallurgy Symposium (IBS 2003)

An AFM-study on the adhesion of Acidithiobacillus ferrooxidans

Institut fr Allgemeine Botanik, Mikrobiologie, Universitt Hamburg,

MATERIALS AND METHODS

RESULTS AND DISCUSSION

Figure. 1. AFM-image of a sterile pyrite surface with a typical layered surface

Figure 2. AFM-image of a pyrite

Figure. 3. AFM-image of a pyrite

Figure 4. AFM-image of a cleaned

Figure 5. AFM-image of a single cell of

Figure 6A-D. Typical patterns of time dependent surface colonization by

Figure 7. Schematical graph of a cell of Acidithiobacillus ferrooxidans attached to

2.2 Attachment of Leptospirillum ferrooxidans strain R3 to pyrite

Figure 8. AFM-image of cells of Leptospirillum ferrooxidans attached to pyrite after

Figure 9A-C. AFM-image of cells of Leptospirillum ferrooxidans attached to pyrite

15th International Biohydrometallurgy Symposium (IBS 2003)

An X-ray photoelectron spectroscopy study of the mechanism

A.J. Parker Cooperative Research Centre for Hydrometallurgy, School of Applied

* Corresponding author. PO Box 90, Bentley, WA 6982, Australia, e-mail: craig.klauber@csiro.au

2.1 Microbial culture

solution were measured using Inductively Coupled Plasma - Atomic Emission

RESULTS AND DISCUSSION

3.1 Solution conditions

Figure 1. Solution concentration of copper and iron and Eh variation during

Figure 2. Sulphur 2p spectra from concentrate leached in the presence of S.

Figure 3. Sulphur 2p spectra from concentrate leached in the presence of S.

Figure 4. Copper 2p spectra from concentrate leached in the presence of S. metallicus

Figure 5. Iron 2p spectra from concentrate leached in the presence of S. metallicus at

Comparison of the S 2p at 60 hours with the S. thermosulfidooxidans to that for

higher BE state as the ammonium component in ammonium jarosite (the ammonium

Figure 8. Decrease in surface dimer concentration is mirrored by a commensurate

15th International Biohydrometallurgy Symposium (IBS 2003)

Analysis of chalcopyrite (CuFeS2) electrodes utilizing galvanic

Department of Biochemistry and Chemical Technology, Institute of Chemistry, So

Corresponding author (oswaldo@iq.unesp.br)

MATERIAL AND METHODS

2.1 Bacterial strain and growth conditions

RESULTS AND DISCUSSION

Figure 3. Scanning electron microscopy of the A. ferrooxidans-LR on the surface of

15th International Biohydrometallurgy Symposium (IBS 2003)

Application of the bacterial weathering of silicate minerals in

Department of Biotechnology, Institute of Geotechnics of the Slovak Academy of

The identity of members of communities of microorganisms in natural environments,

MATERIALS AND METHODS

2.1 Granitic eluvium