Professional Documents
Culture Documents
1 s2.0 S0960852420305952 Main
1 s2.0 S0960852420305952 Main
Bioresource Technology
journal homepage: www.elsevier.com/locate/biortech
A R T I C LE I N FO A B S T R A C T
Keywords: Hemicellulose hydrolysates (HH), which could be an interesting carbon source to feed mixed microbial cultures
Levulinic acid biological production (MMC) able to accumulate high value-added compounds. This research focused on the evaluation of a culture
Polyhydroxyalkanoates and levulinic acid co- strategy to achieve the simultaneous biological production of Levulinic Acid (LA) and Polyhydroxyalcanoates
production (PHA) by MMC fed with a synthetic HH (SHH). The culture strategy involves the use of sequential batch reactors
Mixed microbial cultures
(SBR) to select microorganisms capable of producing LA and PHA. This work proved that the cultivation strategy
Hemicellulose hydrolysate
used allowed the biological production of LA, reaching 37%w/w when the SHH was composed of 85% pentoses.
In addition, the simultaneous biological production of LA and PHB was possible when the SHH was enriched
with acetate (45% pentoses − 50% acetate). Finally, this study showed that the composition of the SHH impacts
directly on the selected microorganism genus and the type and quantity of the value-added compounds obtained.
1. Introduction high production cost. One way to reduce the cost is to look for alter-
native raw materials, such as hemicellulose extracts (Brodin et al.,
One of the greatest challenges today is the transition from a fossil 2017).
fuel economy to one based on renewable resources (Jeong et al., 2018; Wheat straw and sugarcane bagasse, among others, are non-wood
Zhang et al., 2015). In this sense, levulinic acid (LA) and poly- lignocellulosic biomass derived from the agricultural sector, which has
hydroxyalkanoates (PHAs) are interesting compounds that may poten- been cited as a source of value-added products such as xylitol, ethanol,
tially be obtained from lignocellulosic biomass through chemical and cellulose nanofibers and recently, LA and PHA. In every case, fractio-
biochemical synthesis (Dietrich et al., 2019; Tiong et al., 2019; nation of the biomass is required to obtain cellulose, hemicelluloses and
Rodriguez-Perez et al., 2018). LA is the building block for at least fif- lignin (Stoffel et al., 2014). In non-woods, hemicelluloses are rich in
teen highly valued chemical compounds. It is considered the eightieth pentoses, and the best way to separate this fraction is through a hy-
major value-added compound in the USA due to its broad range of drothermal treatment (Vallejos et al., 2017). Hot water HH are then
applications such as fuel additives, flavors, fragrances and polymers rich in xylose (Queirós et al., 2014; Vallejos et al., 2016), which can be
through chemical catalysis (Lappalainen and Dong, 2019). On the other used to produce LA and PHA. LA production is commonly based on
hand, PHAs are bioplastics that have physicochemical and mechanical thermochemical reactions, which use different acid catalysts to induce
properties similar to petroleum-based plastics and they can also be used the conversion of carbohydrates to LA (Jeong et al., 2018). Further-
in medical, agroindustrial, pharmaceutical and food applications more, the biological production of LA by direct fermentation of pen-
(Cassuriaga et al., 2018; Rodrigues et al., 2019). Applications of PHAs toses using engineered Saccharomyces Cerevisiae and Pichia stipis has
as bioplastics include packaging, textiles and prostheses. However, been reported (Zhangelinni, 2017). PHA production from pentoses as a
PHAs are not used significantly in these products, possibly due to their carbon source has been widely studied using pure cultures from
⁎
Corresponding author at: Departamento de Ingeniería Química, Universidad de La Frontera, Casilla 54-D, Temuco, Chile.
E-mail address: gustavo.ciudad@ufrontera.cl (G. Ciudad).
https://doi.org/10.1016/j.biortech.2020.123323
Received 5 February 2020; Received in revised form 2 April 2020; Accepted 3 April 2020
Available online 06 April 2020
0960-8524/ © 2020 Elsevier Ltd. All rights reserved.
F. Pinto-Ibieta, et al. Bioresource Technology 309 (2020) 123323
2
F. Pinto-Ibieta, et al. Bioresource Technology 309 (2020) 123323
Fig. 1. LA, PHB, VS and biomass evolution against operating time for SBR1 (A) and SBR2 (B).
for the reactor samples. The substrate consumption was measured by extraction using the PowerSoil®DNA Isolation Kit (Mo Bio Laboratories)
determining the concentration of reducing sugars and acetate in filtered according to the manufacturer's instructions. After quantification in a
samples (0.22 µm pore size PVDF membrane, Merck). To quantify the NanoDrop spectrophotometer (Thermo Scientific), DNA samples were
total reducing sugars, the dinitrosalicylic acid (DNS) reagent method prepared using barcoded library adaptors (400 nM) containing the
was used (Prasertsung et al., 2017; Zhang et al., 2012). The DNS re- bacterial and fungi primers. PCR reactions run in triplicates were
agent was prepared by mixing dinitrosalicylic acid, potassium sodium pooled, and the amplicon libraries were purified using Agencourt®
tartrate and NaOH in hot deionized water (50 °C). 0.5 ml of the DNS AMpure XP bead protocol (Beckmann Coulter, USA). The concentration
reagent and 0.5 ml of filtered HH sample were mixed and heated for was measured by Quant-iTTM HS DNA assay (Thermo Fisher Scientific,
5 min, then the mixture was cooled and 5 ml of distilled water was USA) and quality was evaluated with a Tapestation 2200, using D1K
added. The absorbance intensity of the final solution was measured at ScreenTapes (Agilent, USA). Afterward, sequencing libraries were
540 nm using a spectrophotometer (Thermo Scientific – GENESYS 10 s). pooled in equimolar concentrations and sequenced on an Illumina
The calibration curve was performed using glucose (Merck) as a stan- MiSeq platform (v.3) with 15% PhiX spike-in and at a read-length of
dard reagent. Acetate was determined by gas-chromatography in a 300 bp paired-end by the Nucleus of Scientific and Technological
flame ionization detector (Clarus 400, PerkinElmer), using a Nukol™ Development in Bioresources of UFRO (BIOREN-UFRO). Sequenced
capillary column (Sigma-Aldrich, Darmstadt, Germany). Volatile solids sample libraries were processed following the DADA2 bioinformatics
(VS) and soluble chemical oxygen demand (SCOD) were quantified pipeline reported by Callahan et al. (2016) using version 1.3.3 of the
using a standard technique (AWWA, 2005) DADA2 R-package. DADA2 allows the inference of exact amplicon se-
quence variants (ASVs), providing several benefits over traditional OTU
2.4. DNA extraction, amplicon sequencing and reads processing clustering methods (Callahan et al., 2017). Illumina sequencing adap-
tors and PCR primers were trimmed before quality filtering. Sequences
To quantify the abundance of the different bacteria at different were truncated after 280 and 250 nucleotides (16S rDNA) and 190 and
times, biomass samples were collected in both SBRs at the initial phase 190 (ITS) for forward and reverse reads respectively, the length at
(day 0) and in the stationary state (day 70), respectively, for the culture which the average quality dropped below a Phred score of 20. After
selection. The biomass was centrifuged at 12,000g for 10 min. After truncation, reads with expected error rates higher than 3 and 5 for
removing the supernatant, a 0.25 g pellet was collected for DNA forward and reverse reads were removed. After filtering, error rate
3
F. Pinto-Ibieta, et al. Bioresource Technology 309 (2020) 123323
learning, ASV inference and denoising, reads were merged with a acid is transformed into LA via two potential pathways: (1) the trans-
minimum overlap of 20 bp. Chimeric sequences were identified and formation of pyruvate by redox reactions to 4-oxo-2 hydroxy-pentanoic
removed. For a total of 12 samples, 77,703 out of 144,741 reads from acid, which is dehydrated into 4-oxo-2 pentanoic acid that finally is
16S rDNA were kept on average per sample after processing, re- reduced to LA (Zanghellini, 2017), and (2) the oxidation of pyruvic
presenting 53.7% of the average input reads. Similarly, for ITS analysis, acid, producing acetyl-CoA which goes into the Krebs cycle and the
from the same 12 samples, 154,406 out of 180,384 reads were kept on synthesis of LA from succinyl CoA (Lee et al., 2019).
average per sample after processing, representing 85.6% of the average Fig. 1.B. showed that during the first 30 days of operation, the
input reads. Taxonomy based on 16S and ITS sequences was assigned microorganisms present in SBR2 were able to synthesize both LA and
using the SILVA v.132 (Glöckner et al., 2017) and UNITE v.8.0 (Nilsson PHB, reaching a maximum of 14% and 17% w/w respectively, unlike
et al., 2019) databases, respectively. Adequacy of sequencing depth the SBR1 where the selection of microorganisms was able to synthesize
after reads processing was corroborated with rarefaction curves at the LA almost exclusively. It has been widely reported that simple carbon
ASV level. chain molecules like acetic acid can lead to the accumulation of storage
polymers such as PHB, which could explain the greater amount of PHB
3. Results and discussion synthesized by microbial communities in the SBR2 reactor. The results
obtained showed that LA and PHB can be co-produced when the SHH is
3.1. Enrichment reactor evolution over the operation time using formulated fed with acetate as the co-substrate. The eventual benefits of this co-
SHH production should be evaluated considering the productivity of each
compound, their final properties and the difficulties and costs of their
The SBR1 was operated for 80 days without any interruptions, separation, among others. Some experiences about the recovery of
finding two value-added intracellular compounds identified as levulinic different value-added compounds coming from co-cultures have been
acid (LA) and polyhydroxybutyrate (PHB). The evolution of LA, PHB, reported (Rodrigues et al., 2019; Wang et al., 2018).
VS and biomass during the operation of the SBR1 is shown in Fig. 1.A. The initial pentose and acetate concentrations (in the SBR inputs of
An acclimatization stage during the first 40 days of operation followed the cycle) are positively correlated with the final concentrations of PHB
by a stationary stage was observed, reflected by the reach of a constant and LA (at the outputs of the cycle), since LA and PHA production in-
concentration of LA and PHB in successive cycle-ends. At the same time, creased proportionally to pentose and acetate concentrations, respec-
a cell wash-out was observed since there was a gradual reduction in VS tively. SBR1 was fed with twice the amount of pentoses as SBR2 (86%
and biomass content across the acclimatization stage until 2.4 g L−1 of vs 46%) and produced twice the amount of LA (31% versus 14%).
biomass was reached during the stationary stage. This behavior de- Moreover, the amount of acetate was 5.5 times lower in SBR1 (9%
monstrates that a selection of microorganisms became able to use SHH versus 50%) and produced 5.7 times less PHB (3% versus 18%).
as a substrate. This selection of microorganisms has been previously Pentoses, acetate and soluble COD consumption in both reactors were
observed in many studies (Guo et al., 2016; Lee et al., 2015; Oliveira monitored at the beginning and the end of some cycles during the
et al., 2016). For example, Huang et al. (2018) used different VFA to stationary stage (Fig. 2). The pentose and acetate concentration inside
feed an SBR, provoking a selection of PHB-producing microorganisms, SBR1 at the beginning of each cycle were 41 and 4.3 Cmmol L−−1,
reaching the stationary stage after 40 days. The required time for respectively (Fig. 2.A), whereas they were 19 and 25 Cmmol L−1, re-
reaching the stationary stage has been reported in a range from 30 to spectively in SBR2 (Fig. 2.B). At the end of each cycle, neither pentoses
167 days, depending on the conditions, such as the cycle length, sub- nor acetate were detected in either reactor. Furthermore, the soluble
strate complexity and origin and composition of the microbial biomass, COD concentrations at the beginning of the cycles remained constant
among others (Albuquerque et al., 2007; Morgan-Sagastume et al., around 4 g L−1, proving that the added carbon source was not accu-
2015). mulated by degrading all pentoses, VFA and some inhibitors like HMF
According to the operating conditions in SBR1, it was possible to and furfural, contained in the feed of the two reactors.
select microorganisms able to synthesize intracellularly LA and, to a
lesser extent, PHB. LA increased during the first forty days until 3.2. Evolution of the production of value-added compounds, substrate
reaching a maximum of 32%w/w at the stationary state, while only a consumption and dissolved oxygen in the different SBR operation cycles.
maximum of 3%w/w of PHB was reached at the same stage. Although
both compounds are synthesized by the microbial biomass, a higher The average value of the studied variables in the three cycles is
concentration of LA (30% w/w) than that of PHB (3% w/w) was not shown in Fig. 3. A feast/famine (F/F) dynamics can be identified from
expected. These results are similar to those reported by Zanghellini the variations observed in DO in both reactors during one operating
(2017), where pentoses were also used as a carbon source for LA cycle (Dionisi et al., 2005). According to Korkakaki et al. (2017), DO
synthesis. On the other hand, it is possible that PHB was synthesized concentrations may decrease below 2 mg L−1 during the feast phase,
only from the 9%w/w of the acetate in the SHH since MMC prefer VFA while in the famine phase it increased above 8 mg L−1, indicating no
over pentoses to synthesize PHA (Queirós et al., 2014). Consequently, it significant biological activity. It has been strongly demonstrated that F/
is likely that under the operating conditions used in this study, the F dynamic adapts biomass to stressful conditions and drives the accu-
selected microbial biomass preferred to metabolize pentoses to LA, mulation of valuable compounds (Rodriguez-Perez et al., 2018). The F/
contrary to what has been reported by other studies, where pentoses are F dynamic was observed in both conditions which, regardless of the
converted to PHA by MMC (Huang et al., 2016; Queirós et al., 2014). carbon source imposed on the cultures, stimulated the storage of dif-
The evolution of LA, PHB, VS and biomass during the 80 days op- ferent compounds inside the microorganisms.
erated in the SBR2 is shown in Fig. 1.B. An acclimatization stage during The end of the feast phase was reached at 4.8 h of operation in SBR1
the first 30 days is observed, followed by a stationary stage from day 15 (Fig. 3.A), consuming 87% of the SHH. This behavior suggests that the
onward. Less time was needed to reach a stationary stage in SBR2 SHH was used for cell growth and LA production simultaneously. The
(10 days), probably due to its higher concentration of acetate in the intracellular LA increased from 30 to 37% in the feast phase, whereas it
substrate compared to SBR1, since the metabolic pathway to synthesize decreased from 37 to 30% in the famine phase. It seems that during the
PHB from acetate seems to have fewer steps than the two metabolic famine phase the MMC can consume the LA accumulated during the
pathways to synthesize LA from pentoses reported in the literature. The feast phase, probably due to external carbon source limitation. In the
reported possible pathways for intracellular LA production start with case of SBR2, 82% pentoses and 96% acetate were consumed during the
the transformation of the pentoses to pyruvic acid via the pentose feast phase (Fig. 3.B). The simultaneous consumption of both substrates
phosphate pathway (Lee et al., 2019; Miscevic et al., 2019)then pyruvic is indicative of the presence of two groups of microorganisms
4
F. Pinto-Ibieta, et al. Bioresource Technology 309 (2020) 123323
Fig. 2. Pentose and acetate concentration at the beginning and end of the cycles as a function of the operating time during the stationary stage in SBR1 (A) and SBR2
(B).
synthesizing LA and PHB separately. LA reached 20% w/w at the end of Fungus communities are shown in Fig. 5. The predominant fungi
the feast phase, whereas PHB reached a maximum of 25% w/w at identified in the inoculum of both reactors were those of the genera
180 min, prior to the end of the feast phase. The maximum amounts of Trichosporon (30%), Cutaneotrichosporon (23%), Apiotrichum (14%) and
PHB and LA were reached when acetate and pentoses were depleted, Candida (10%). Haglerozyma (87%) predominated in the stationary
respectively. This behavior confirms the results showed in point 3.1, stage in SBR1 and Gueomyces (80%) in SBR2. These results suggest that
i.e., the preference of different groups in the selected MMC to meta- a highly selective pressure was applied to the microorganisms in gen-
bolize pentoses to LA and acetate to PHB under the operating conditions eral that favored the prevalence of the minor fraction of the bacteria
used in this study. able to resist overstressing conditions.
This result is noteworthy considering that Paracoccus, Enterococuss
3.3. Structure of mixed microbial cultures related to the biological and lactococuss of the genus Azospirillum are organisms known to ac-
production of value-added compounds. cumulate carbon as PHA (Huang et al., 2018; Kumar et al., 2018; Mota
et al., 2019; Müller-Santos et al., 2015; Sundaram et al., 2016). Azos-
The structure of the bacterial community identified for the in- pirillum is one of the best studied polymer-accumulating rhizospheric
oculum is presented in Fig. 4. The bacterial microbial community was bacteria, which base their survival strategy on PHB generation (Müller-
conformed mainly by Rhodoferax, Flavobacterium and Arcobacter, Santos et al., 2015). By contrast, the fungus haglerozyma belongs to the
showing a relative abundance of 50%, 25% and 25%, respectively. family Trichosporoneceae, which has phylogenetic relationships with
Moreover, the genera Paracoccus, Lactococcus and Enterococcus were apiotrichum and Zooglea, which have been previously associated with
found to a lesser extent. As a result of the MMC selection in the sta- the accumulation of different storage polymers such as PHAs (Queirós
tionary stage, the predominant communities were mainly Paracoccus, et al., 2017). If the two reactors are compared, the overall micro-
Lactococcus and Enterococcus in the SBR1 (relative abundances of 26%, organisms reveal a high variability in their respective community
28% and 15% respectively). On the other hand, the genus Azospirillum structures. These changes were principally guided by the feed compo-
was the main component in SBR2 with a relative abundance of 90%. sition, indicating a strong feed dependence on the MMC selection
5
F. Pinto-Ibieta, et al. Bioresource Technology 309 (2020) 123323
Fig. 3. Accumulated compounds (LA and PHB), acetate, reducing sugars and dissolved oxygen as a function of time in one cycle of SBR1 (A) and SBR2 (B).
process. Oliveira et al. (2016) have reported that the selected MMC will metabolic pathways and the way in which LA is stored by the cell.
be different as a consequence of changes in the feed strategy. In their
work, when nitrogen and carbon substrates were added simultaneously CRediT authorship contribution statement
at the beginning of the cycle, the communities were dominated by
Corynebacterium. However, when the nitrogen addition was delayed to F. Pinto-Ibieta: Writing - original draft, Conceptualization,
the end of the feast phase, Xantomonadaceae were prevailed. Investigation. M. Cea: Writing - review & editing, Resources,
Finally, this study provides a first approach to the biological Supervision, Conceptualization. F. Cabrera: Writing - original draft,
synthesis of LA using MMC, which is highly desirable given the need to Conceptualization. M. Abanto: Software, Formal analysis. F.E.
generate clean alternatives to increase LA production. Moreover, LA Felissia: Writing - review & editing. M.C. Area: Writing - review &
synthesis cannot be attributed to a specific microorganism, fungus, editing. G. Ciudad: Writing - review & editing, Funding acquisition,
bacterium or a synergic relationship between them under the evaluated Supervision, Conceptualization.
systems.
MMC in SBR, feed with HH and submitted to feast/famine regimen, The authors declare that they have no known competing financial
proved to be a good strategy for biological LA and PHB production. The interests or personal relationships that could have appeared to influ-
composition of the hydrolysate impacted directly on the type of mi- ence the work reported in this paper.
croorganism selected and the type and quantity of the final product.
When the HH was enriched with acetate, LA and PHB were produced Acknowledgments
simultaneously. Although the main selected LA-producing micro-
organisms are reported, possible synergies between them are unknown, The authors wish to thank the financial support provided by the
so further studies are required to understand the possible synergies, ELAC2015/T03-0715-ValBio-3D. The authors would also like to thank
6
F. Pinto-Ibieta, et al. Bioresource Technology 309 (2020) 123323
Fig. 4. Profile of bacterial consortium in the original sample (day 0) and in the selected MMC (steady-state) for bacteria.
Fig. 5. Profile of fungus consortium in the original sample (day 0) and in the selected MMC (steady-state) for fungus.
7
F. Pinto-Ibieta, et al. Bioresource Technology 309 (2020) 123323
8
F. Pinto-Ibieta, et al. Bioresource Technology 309 (2020) 123323
Zhao, L., Han, D., Yin, Z., Bao, M., Lu, J., 2019. Biohydrogen and polyhydroxyalkanoate
organic electrolyte solutions. Polym. Degrad. Stab. 97, 573–577. https://doi.org/10. production from original hydrolyzed polyacrylamide-containing wastewater. Bioresour.
1016/j.polymdegradstab.2012.01.010. Technol. 287, 121404. https://doi.org/10.1016/j.biortech.2019.121404.
Zhangelinni, A., 2017. Fermentation route for the production of levulinic acid, levulinate
esters and valerolantone and derivatives thereof. US 2017/0275657 A1.