Journal of Microbiological Methods 60 (2005) 417 – 422

www.elsevier.com/locate/jmicmeth

Note

Insight into pollutant bioavailability and toxicity using

Raman confocal microscopy
Andrew C. Singer*, Wei E. Huang, Jana Helm, Ian P. Thompson
Environmental Biotechnology Section, Centre for Ecology and Hydrology-Oxford, Mansfield Road, Oxford, OX1 3SR, UK
Received 30 July 2004; received in revised form 8 October 2004; accepted 21 October 2004
Available online 19 November 2004

Abstract

Raman confocal microscopy was used to discriminate between cultures of Burkholderia xenovorans LB400 exposed to four
different common environmental pollutants: phenanthrene, dodecane, 3-chlorobiphenyl and pentachlorophenol. Evidence is
presented for the application of Raman spectroscopy as a bioassay for pollutant bioavailability and toxicity.
D 2004 Elsevier B.V. All rights reserved.

Keywords: Raman confocal microscopy; Pollutant bioavailability; Burkholderia xenovorans

Raman spectroscopy utilizes differences in the
vibrational energy of chemical bonds, manifested as
scattered light following excitation, to produce spec-
trograms, which, upon analysis, can be easily used to
differentiate biologically associated chemicals (e.g.,
nucleic acids, protein, lipids, carbohydrates) and
biological materials (e.g., bacteria, human cells,
viruses). In previous studies, whole-organism finger-
prints were used to differentiate fungi (Edwards et al.,
1995), bacteria (Rosch et al., 2003; Zeiri et al., 2002),
bacterial strains (Jarvis and Goodacre, 2004; Schuster
et al., 2000), bacterial growth states, isotopic differ-
ences in growth substrates (Huang et al., 2004),
* Corresponding author. Tel.: +44 1865 281630; fax: +44 1865
281696.
E-mail address: acsi@ceh.ac.uk (A.C. Singer).
0167-7012/$ - see front matter D 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.mimet.2004.10.016
accumulated chemicals within bacteria (Pasteris et al.,
2001), and antibiotic susceptibility (Sockalingum et
al., 1997).
In this study, we test three null hypotheses, that (i)
there are no differences in the Raman spectra from a
bacterium exposed to different pollutants; (ii) the
adsorbed pollutant will not be distinguishable in the
bacterial Raman spectra; and (iii) the Raman spectra
will not facilitate the determination of pollutant
toxicity.
We employ a well-described polychlorinated
biphenyl (PCB) degrading bacteria, Burkholderia
xenovorans LB400 (previously B. cepacia LB400,
generously donated by Prof. William Mohn, Univer-
sity of British Columbia), as a model environmental
microorganism reasonably tolerant of persistent
organic pollutants (Bopp, 1986). B. xenovorans
LB400 was maintained on Standard Plate Count Agar
418 A.C. Singer et al. / Journal of Microbiological Methods 60 (2005) 417–422

(CM463, Oxoid Limited, Hampshire, England). A
colony was cultured in a 100-ml Erlenmeyer flask
containing 20 ml of mineral medium No. 457
(Brunner; Deutsche Sammlung von Mikroorganismen
und Zellkulturen (DSMZ) Braunschweig, Germany)
and 10 mM glucose. A 600-Al aliquot of the overnight
culture was transferred to 50 ml of mineral medium
and 10 mM glucose, shaken at 180 rpm at 30 8C to an
optical density (A 600) of 1.286. Owing to variations in
the Raman spectra from bacteria in different physio-
logical states (Huang et al., 2004), intratreatment
variability was minimized by harvesting cultures after
they had reached stationary phase. The bacteria were
concentrated by filtration through a 47-mm diameter,
0.22-Am Millipore nitrocellulose filter paper (Fisher
Scientific UK, Leicestershire, UK) in lieu of centri-
fugation to minimize damage and stress to the cells.
The filtered bacteria were then gently flushed with
0.85% saline whilst on the filter paper/filtration
apparatus to remove any excess growth media. The
washed bacteria were displaced from the filter paper
with a stream of saline and resuspended to an optical
density (A 600) of 0.997.
The pollutants, phenanthrene (PAH), pentachlor-
ophenol (PCP), and 3-chlorobiphenyl (PCB), were
aliquoted into 7-mL glass vials with a PTFE-lined cap
from an acetone stock. The solvent was allowed to
evaporate to dryness at room temperature in a fume
hood. A 1-AL aliquot of neat dodecane was directly
added to a 7-mL vial to achieve the desired concen-
tration. A final concentration of 1000 mg/L was
achieved for all pollutants with the addition of 1 mL
of the saline bacterial suspension. The pollutant–
bacterial suspensions were inverted for 1 h at 23 8C
prior to analysis. Sterile glass Pasteur pipettes were
employed to spot the pollutant–culture suspension
(~50 AL) on a quartz slide. The cultures were dfixedT
to the slides by drying at room temperature. The
resuspended saline bacterial suspension was included
as a control.
Raman microscopy was performed using a Lab-
RAM 300 confocal Raman microscope (Jobin–Yvon,

Fig. 1. Pollutant exposure differentiation. Plot of the average of five replicate spectra for cells exposed to dodecane, phenanthrene, 3-
chlorobiphenyl, and glucose. Arrows point to the frequencies of the major peaks for the pollutants in the absence of bacteria. Shadowed
frequency ranges denoted by capital letters dA–RT represent areas of known biological significance (Table 1). Bold labels denote frequencies that
share both biological (Table 1) and chemical (neat pollutant) significance in this study. Underlined labels denote frequencies responsible for
N5% of the separation detected by PCA (Table 2).
 A.C. Singer et al. / Journal of Microbiological Methods 60 (2005) 417–422 419

Fig. 2. Principal component analysis of pollutant exposure spectral differences. Plot (A) shows separation between PC1 and PC2, plot (B) shows
separation between PC2 and PC3, and plot (C) shows separation between PC1, PC2, and PC3. Treatments are denoted, (a) dodecane; (b)
control; (c) phenanthrene; (d) 3-chlorobiphenyl.

UK), as previously described (Huang et al., 2004).
The spectra of five individual cells within each of the
five treatments was acquired in the frequency range
from 1972 to 544 cm 1, totaling 1014 data points, and
an accumulation time of 60 s. Spectra were processed
for baseline correction and normalization by Labspec
software (Jobin–Yvon). The normalized data were
then imported into MVSP version 3.12d (Kovach
Computing Services, Wales, UK) and Matlab R12
(The Maths Works, MA, USA), for principal compo-
nent analysis (PCA), as previously described (Huang
et al., 2004).
We reject the null hypothesis that (i): there are
numerous differences in the Raman spectra from B.
xenovorans LB400 after exposure to different pollu-
tants (Figs. 1 and 2). The averaged and normalised
spectra of the five replicates from each pollutant–
culture combination are presented in Fig. 1. Included
in Fig. 1 are arrows which denote the frequencies of
the major peaks for the respective pollutant in the
absence of the bacteria. Shaded frequency ranges,
denoted by capital letters dA–RT, represent regions
determined to have general or specific biological
significance (Maquelin et al., 2002). The attributes of
these regions are listed in Table 1. Shaded regions
with underlined labels denote frequencies most
responsible for the separation observed in the PCA
plots, as shown in Fig. 2A–C. The frequencies which
explain N5% of the variation for each of the three
PCA axis are listed in Table 2. Shaded regions with
bolded dA–RT labels denote frequencies which share
both biological (Table 1) and chemical (neat pollutant)
significance.
Axis 1 of the PCA (Fig. 2A) explains 73% of the
variation between treatments, while an additional 9%
and 4% of the variation can be distinguished by PC2
420 A.C. Singer et al. / Journal of Microbiological Methods 60 (2005) 417–422

Table 1 carbohydrates (Schuster et al., 2000). Interestingly, all

Assignment of some biologically relevant Raman spectra (adapted three axes are influenced by the 960–1010 cm 1
from Maquelin et al., 2002 and Schuster et al., 2000)
frequency range, which is correlated with the amino
Label Frequency (cm 1) Assignment
acid phenylalanine and neat 3-chlorobiphenyl. The
A 544–553 Carbohydrate frequency ranges correlating to DNA (1482–1484,
B 640 Tyrosine
dNT) and RNA (774–834, dET) are also defining peaks
C 665 Nucleotide (guanine)
D 720 Nucleotide (adenine) in PC1. The implication is that metabolic activity
E 778–785 Nucleotides (cytosine, uracil) (e.g., increased RNA and enzyme production and
F 829 Protein (e.g., Tyrosine) expression or mitosis (Huang et al., 2004)) within the
G 897 C–O–C exposed cells is altered upon exposure to the
H 1001–1005 Phenylalanine
pollutants, potentially in response to a toxic com-
I 1061 Carbohydrate (C–N and C–C)
J 1085 C–O pound (e.g., stress people) and/or exposure to a
J 1098 Carbohydrate (C–C, C–O–C) growth substrate, as is the case with 3-chlorobiphenyl.
J 1102 NPO2 We cannot reject the null hypothesis that (ii):
K 1122–1129 C–N and C–C several of the major peaks within the Raman spectra
(unsaturated fatty acid/lipid)
L 1295 CH2 def
of the neat pollutants (arrows in Fig. 1) correlate with
L 1230–1295 Amid III (N–H, C–N) those peaks which are among the most discriminating
M 1440–1460 (–C–H2) frequencies of the PCA (Figs. 1 and 3). Most notably,
N 1484 Nucleotides (guanine, adenine) five of the nine major peaks in the neat phenanthrene
O 1575 Nucleotides (guanine, adenine) Raman spectrum correlate with frequencies respon-
P 1606 Phenylalanine
sible for separations on PC1-3. A more in-depth study
P 1614 Tyrosine
Q 1650–1680 Amid I and unsaturated lipids containing additional replicates and a wider range of
R 1735 NC=O ester pollutants should be carried out before this null
hypothesis can be rejected; however, there is evidence
to suggest differences exist in the Raman spectrum,
and PC3 (Fig. 2B, C). Specifically, frequencies that which might be attributed to the adsorbed pollutant.
denote carbohydrates (A, M), nucleotides (E, N), The data indicate that we might be able to reject the
protein (F, H, P), and lipids (J, Q, R) are responsible null hypothesis that (iii): the energy intensity of the
for the majority of the differences between treatments. laser as well as the length of scan can be used to
The most discriminating frequencies can be visualized differentially interrogate a bacterial cell to assess its
in Fig. 3 for each of the three axes. dSignificantT degree of exposure to a toxic pollutant. If the cell is
frequencies exhibit relatively high or low peaks in this stressed after exposure to a pollutant, as was the case
representation. Separation on PC1 is largely influ- with pentachlorophenol, the cell becomes increasingly
enced by the b630 cm 1 range, which is attributed to more susceptible to burning whilst being scanned. The

Table 2
Raman frequencies responsible for N5% of the graphical separation between treatments observed in axis 1–3 of the PCA (Figs. 2 and 3)
Axis 1 Axis 2 Axis 3 Pollutant(s) Bio-attributes
A 544–628 544–556 PAH Carbohydrates
E, F 774–834 PCB; dodecane; PAH RNA, protein
H 983–1004 959–1010 PCB Phenylalanine
J 1101–1106 1101–1106 NPO2 (lipid)
Unknown 1344–1350 1344–1350 PAH
M 1444–1466 1444–1466 Dodecane; PAH y(–C–H2)
N 1482–1484 1482–1484 DNA
P 1634 PAH Protein
Q 1652 Amid I and lipids
R 1724–1763 NC=O ester
 A.C. Singer et al. / Journal of Microbiological Methods 60 (2005) 417–422 421

Fig. 3. Graphical representation of PCA raw data from axes 1–3 (black, red, and blue, respectively). A greater deviation from 0 indicates greater
contribution of that frequency to the PCA separation. (For interpretation of the references to colour in this figure legend, the reader is referred to
the web version of this article.)

standard laser intensity (5 mW) and scan time (60 s) can
be adjusted to a lower setting until a full Raman profile
can be acquired. The laser intensity and scan time can,
in this way, be an indirect measure of cellular stress.
If a pollutant is bioavailable, (i) structural changes
in the cell can be discerned from the Raman spectrum;
(ii) characteristic peaks from the neat chemical can
potentially be distinguished in the Raman spectrum;
and (iii) cell integrity can be altered effecting the laser
intensity and scanning time needed to acquire a
satisfactory spectrum. Future work will investigate
the use of Raman microscopy as a rapid bioassay for
bioavailable pollutants in (contaminated) water, as
well as a tool to interrogate the biotoxicity of different
compounds.
Acknowledgment
We thank Jobin Yvon, UK, for supplying the
Raman confocal microscope and for the technical
advice and Dr. Komang Ralebitso-Senior for her
feedback of this paper. Dr. Helm has been supported
by a Marie Curie Fellowship of the European
Community programme Improving Human Potential
under contract number HPMF-CT-2002-01762.
References
Bopp, L.H., 1986. Degradation of highly chlorinated PCBs by
Pseudomonas strain LB400. Journal of Industrial Microbiology
1, 23 – 29.
Edwards, H.G.M., Russell, N.C., Weinstein, R., Wynn-Williams,
D.D., 1995. Fourier transform Raman spectroscopic study of
fungi. Journal of Raman Spectroscopy 26, 911 – 916.
Huang, W.E., Griffiths, R.I., Thompson, I.P., Bailey, M.J., Whiteley,
A.S., 2004. Raman microscopic analysis of single microbial
cells. Analytical Chemistry 76, 4452 – 4458.
Jarvis, R.M., Goodacre, R., 2004. Discrimination of bacteria using
surface-enhanced Raman spectroscopy. Analytical Chemistry
76, 40 – 47.
Maquelin, K., Kirschner, C., Choo-Smith, L.P., Braak, N.v.d.,
Endtz, H.P., Naumann, D., Puppels, G.J., 2002. Identification of
medically relevant microorganisms by vibrational spectroscopy.
Journal of Microbiological Methods 51, 255 – 271.
422 A.C. Singer et al. / Journal of Microbiological Methods 60 (2005) 417–422
Pasteris, J.D., Freeman, J.j., Goffredi, S.K., Buck, K.R., 2001.
Raman spectroscopic and laser scanning confocal microscopic
analysis of sulfur in living sulfur-precipitating marine bacteria.
Chemical Geology 180, 3 – 18.
Rosch, P., Schmitt, M., Kiefer, W., Popp, J., 2003. The identification
of microorganisms by micro-Raman spectroscopy. Journal of
Molecular Structure 661–662, 363 – 369.
Schuster, K.C., Urlaub, E., Gapes, J.R., 2000. Single-cell analysis of
bacteria by Raman microscopy: spectral information on the
chemical composition of cells and on the heterogeneity in a
culture. Journal of Microbiological Methods 42, 29 – 38.
Sockalingum, G.D., Bouhedja, W., Pina, P., Allouch, P., Mandray,
C., Labia, R., Millot, J.M., Manfait, M., 1997. ATR-FTIR
spectroscopic investigation of imipenem-susceptible and-resist-
ant Pseudomonas aeruginosa isogenic strains. Biochemical and
Biophysical Research Communications 232, 240 – 246.
Zeiri, L., Bronk, B.V., Shabtai, Y., Czege, J., Efrima, S., 2002.
Silver metal induced surface enhanced Raman of bacteria.
Colloids and Surfaces A: Physicochemical and Engineering
Aspects 208, 357 – 362.