Comparative Analysis of Carbon Monoxide Tolerance among Thermoanaerobacter Species

An anaerobic thermophilic strain (strain PCO) was isolated from a syngas-converting enrichment culture. Syngas components cannot be used by strain PCO, but the new strain is very tolerant to carbon monoxide (pCO = 1.7 × 105 Pa, 100% CO). 16S rRNA gene analysis and DNA-DNA hybridization revealed that strain PCO is a strain of Thermoanaerobacter thermohydrosulfuricus. The physiology of strain PCO and other Thermoanaerobacter species was compared, focusing on their tolerance to carbon monoxide. T. thermohydrosulfuricus, T. brockii subsp. finnii, T. pseudethanolicus, and T. wiegelii were exposed to increased CO concentrations in the headspace, while growth, glucose consumption and product formation were monitored. Remarkably, glucose conversion rates by Thermoanaerobacter species were not affected by CO. All the tested strains fermented glucose to mainly lactate, ethanol, acetate, and hydrogen, but final product concentrations differed. In the presence of CO, ethanol production was generally less affected, but H2 production decreased with increasing CO partial pressure. This study highlights the CO resistance of Thermoanaerobacter species.


INTRODUCTION
Thermophiles thrive at austere and unusual conditions and their evolutionary significance and biotechnological potential have triggered microbiological research over the last decades (Turner et al., 2007;Wagner and Wiegel, 2008;Yoneda et al., 2015). There is a continuous biotechnological interest in highly thermostable enzymes, which make this kind of organisms very attractive. Thermophilic bacteria of the class Clostridia, such as members of the genera Clostridium, Thermoanaerobacter, Thermoanaerobacterium, and Caldicellulosiruptor, are currently used as biocatalysts for the production of biofuels or other chemicals of interest (Hemme et al., 2010;Carere et al., 2012). Specifically members of the Thermoanaerobacter genus are utilized to produce ethanol and hydrogen (H 2 ) from a variety of saccharides (Jessen and Orlygsson, 2012). Within thermophiles, an organism from Thermoanaerobacter genus-T. ethanolicus-is one of the most well-studied ethanol-producing bacteria (Wiegel and Ljungdahl, 1981;Lacis and Lawford, 1991). A less common substrate, carbon monoxide (CO), is used by T. thermohydrosulfuricus subsp. carboxydovorans and T. kivui. T. thermohydrosulfuricus subsp. carboxydovorans can grow with CO as sole electron donor (25% in the headspace), producing H 2 and CO 2 (Balk et al., 2009). T. thermohydrosulfuricus shares 99% similarity of the 16S rRNA gene sequence and over 70% DNA-DNA hybridization with T. thermohydrosulfuricus subsp. carboxydovorans, but only the latter one can use CO. Growth of the homoacetogenic T. kivui with CO diluted with CO 2 /N 2 or CO 2 /H 2 was described by Kevbrina et al. (1996). Recently, Weghoff and Müller (2016) reported the ability of T. kivui to grow on only CO (100% in the headspace), producing acetate and hydrogen. Carboxydotrophic metabolism in Thermoanaerobacter species is normally not assessed, and it is not known if they can endure CO or even adapt to grow on CO, as recently reported for T. kivui (Weghoff and Müller, 2016). In this work we isolated Thermoanaerobacter thermohydrosulfuricus strain PCO from a thermophilic syngas-converting enrichment, but this strain appears unable to oxidize CO. The main objectives of this work were (1) to characterize and determine the CO tolerance of Thermoanaerobacter thermohydrosulfuricus strain PCO, and (2) to compare the effect of CO on growth, glucose consumption and product formation of strain PCO and of four close relative species from the Thermoanaerobacter genus.

Enrichments and Isolation
Suspended sludge from a thermophilic anaerobic municipal solid waste digester (Barcelona, Spain) was used as inoculum for starting up syngas-converting enrichments. Microbial cultures were enriched with synthetic syngas (mixture of 60% CO, 10% CO 2 , and 30% H 2 , total pressure 1.7 × 10 5 Pa) as sole carbon and energy source (Alves et al., 2013). Isolation of strain PCO was done using soft agar (1.5%, w/v) incubations and liquid medium serial dilutions, with 20 mM pyruvate as sole substrate. Sodium pyruvate was added to the medium from a 1M filter-sterilized stock solution. A phosphate-buffered mineral medium was used, containing (per liter): Na 2 HPO 4 , 1.63 g; NaH 2 PO 4 , 1.02 g; resazurin, 0.5 g; NH 4 Cl, 0.3 g; CaCl 2 ·2H 2 O, 0.11 g; MgCl 2 ·6H 2 O, 0.10 g; NaCl, 0.3 g; 1 mL of acid and alkaline trace element stock each, and 0.2 ml of vitamin stock. Trace elements and vitamins were prepared as described previously (Stams et al., 1993). Before inoculation, medium was reduced with sodium sulfide (0.8 mM final concentration). Bottles were incubated in the dark at 55 • C while shaken at 100 rpm (liquid cultures) or standing (soft-agar cultures). Colonies were picked from softagar incubations, inoculated in fresh liquid medium containing pyruvate (20 mM). Cultures were further purified by subsequent serial dilutions alternating with soft-agar colony picking. Purity of the culture was checked by microscopic examination after growth with different substrates (Olympus CX41, Tokyo, Japan). Direct sequencing of the 16S rRNA gene and denaturing gradient gel electrophoresis (DGGE) were also applied to check the genetic purity of the culture.

Sequencing and Phylogenetic Analysis
PCR products obtained from 16S rRNA gene amplification were purified using the PCR Clean Up kit NucleoSpin Extract II (Macherey-Nagel, Düren, Germany) and sequenced directly at Eurofins MWG Operon (Ebersberg, Germany). Partial sequences were assembled using the alignment editor BioEdit v7.0.9 software package (Hall, 1999). Similarity searches for the 16S rRNA gene sequence derived from strain PCO were performed using the NCBI BLAST search program within the GenBank database (Altschul et al., 1990). The 16S rRNA gene sequence of Thermoanaerobacter strain PCO is available in the DDBL/EMBL/GenBank databases under the accession number HF586422.

Carbon Monoxide Tolerance Tests
Strain PCO, Thermoanaerobacter thermohydrosulfuricus, T. brockii subsp. finnii, T. pseudethanolicus, and T. wiegelii were tested for CO tolerance. All cultures were incubated with 0, 50, and 100% CO (pCO/P, where "pCO" is the CO partial pressure and "P" the total gas pressure). Additionally, strain PCO was incubated with 25 and 75% CO. Initial total pressure was 1.7 × 10 5 Pa in all the assays; N 2 was used to pressurize the headspace for CO percentages lower than 100%. The tests were performed using an anaerobic phosphate-buffered mineral salt medium and glucose (20 mM) was used as carbon and energy source. Bottles were incubated in the dark, at 55 • C and shaken at 100 rpm. All the assays were done in duplicate. Growth of the strains was determined by measuring optical density increase at 600 nm with a spectrophotometer (U-1500 Hitachi, Tokyo, Japan). The statistical significance of the differences detected in glucose conversion rates and end products production was evaluated using single factor analysis of variances (ANOVA).

Analytical Methods
Soluble substrates and intermediates (sugars, organic acids, and alcohols) were measured using a HPLC Thermo Electron equipment with a Shodex SH1821 column and equipped with a RI detector. The mobile phase used was sulfuric acid (0.01 N) at a flow rate of 0.6 mL min −1 . Column temperature was set at 60 • C. Inorganic anions were analyzed by chromatography using a HPLC Dionex system, equipped with an Ionpac AS22 column, and ED40 electrochemical detector. Column temperature and pressure varied between 35-40 • C and 130 × 10 5 -160 × 10 5 Pa. Gaseous compounds (CO, CO 2 , H 2 ) were analyzed by gas chromatography on a GC-2014 Shimadzu with a thermal conductivity detector. CO 2 was analyzed with a CP Poraplot Q column (25 m length, 0.53 mm internal diameter; film thickness, 20 µm). Helium was used as carrier gas at a flow rate of 15 mL min −1 , and the temperatures in the injector, column, and detector were 60, 33, and 130 • C. CO and H 2 were analyzed with a Molsieve 13X column (2 m length, 3 mm internal diameter). Argon was used as carrier gas at a flow rate of 50 mL min −1 , and temperatures in the injector, column, and detector were 80, 100, and 130 • C.

Physiological Characterization of Strain PCO and Comparison with Closely Related Thermoanaerobacter Species
Strain PCO was isolated from the thermophilic syngasconverting enrichment described by Alves et al. (2013). Isolation was performed using pyruvate as sole carbon and energy source. Strain PCO has a G+C content of the DNA of 34.5 mol % and shares 98% identity with the 16S rRNA gene of Thermoanaerobacter thermohydrosulfuricus (the 16S rRNA gene sequence of strain PCO is available in the DDBL/EMBL/GenBank databases under the accession number HF586422). DNA-DNA hybridization between the two strains was 100%, confirming that strain PCO is a strain of Thermoanaerobacter Growth pH (optimum) 6.5-7.5 6.9-7.5 6.5-6.8 nr 6.8    Cayol et al. (1995), Lee et al. (1993), Onyenwoke et al. (2007), Cook et al. (1996) and this study. nr, not reported. Symbols: +, utilized; ±, poorly utilized; −, not utilized; *, data from this study.  (Alves et al., 2013). However, its presence indicates that it can grow on metabolic byproducts and/or dead cells at high CO concentrations. Strain PCO formed terminal round endospores, which is a characteristic of Thermoanaerobacter species (Wiegel and Ljungdahl, 1981;Lee et al., 1993Lee et al., , 2007Kim et al., 2001;Balk et al., 2009;Shaw et al., 2010). Cells of strain PCO are straight rods and normally occur singly (Figure 1). Strain PCO had an optimum growth temperature of 70 • C; no growth was detected below 37 • C or above 75 • C. The optimum pH for growth was between 6.5 and 7.5. Strain PCO is a very versatile organism that can utilize a range of different substrates, such as: arabinose, cellobiose, cellulose, fructose, galactose, glucose, lactose, maltose, mannitol, mannose, pectin, pyruvate, raffinose, ribose, sorbitol, starch, sucrose, trehalose, xylan, xylose, and yeast extract. No growth occurred with acetate, ethanol, formate, glycerol, glycine, lactate, methanol, propionate, CO (from 20 to 100% CO, total pressure 1.7 × 10 5 Pa), and H 2 /CO 2 (80/20%, total pressure 1.7 × 10 5 Pa). The main products detected and quantified from glucose fermentation by strain PCO were lactate, ethanol, acetate, and H 2 (Figure 2) which are typically formed from glucose by most of the Thermoanaerobacter species (Wiegel and Ljungdahl, 1981;Lee et al., 1993;Kim et al., 2001;Lee et al., 2007;Balk et al., 2009;Shaw et al., 2010). Strain PCO is able to reduce elemental sulfur and AQDS, but sulfate, sulfite, thiosulfate, nitrate, and nitrite could not serve as electron acceptors. The comparison between the morphological, biochemical and physiological characteristics of strain PCO and its close relatives is presented in Table 1. All of them can use thiosulfate as electron acceptor, but strain PCO cannot. Even though strain PCO and T. thermohydrosulfuricus are the same species, strain PCO can be differentiated because of its ability to grow and ferment cellulose and reduce AQDS.  a Product yield = (C P,tf − C P,t0 )/(C S,t0 − C S,tf ); where, C P is product concentration and C S is glucose concentration measure at time zero (t 0 ) and at the end of the assay (t f ). Note: final product concentration was calculated as an average of the plateau of production curves. b Estimated CO 2 production considering that 1 mol of CO 2 is produced for each mol of ethanol or acetate formed [a maximum deviation of 20% was obtained when estimating total CO 2 concentrations from measured CO 2 concentration in the headspace summed with the correspondent calculated dissolved CO 2 (using the Henry law)]. c Carbon recovery (CR) = (n P,tf − n P,t0 )/(n S,tf − n S,t0 ) + n biomass ; where n is the number of carbon moles in products (P − acetate, lactate, ethanol, and CO 2 ) and in glucose (S) and n biomass is the estimated mol of carbon used for biomass growth. Carbon Monoxide Tolerance of Strain PCO and Related Thermoanaerobacter sp.
When strain PCO was cultured with 20 mM of glucose and subjected to different CO concentrations in the headspace (0, 25, 50, 75, and 100%, total pressure 1.7 × 10 5 Pa) no significant differences were found in glucose consumption rates ( Table 2). Carbon monoxide concentrations of 75% or lower did not affect substantially product formation from glucose (Figures 3A, 4A), except for H 2 . Hydrogen production by strain PCO decreased significantly (P = 0.0001) from 3.94 ± 0.55 to 1.52 ± 0.13 mM when cultures were incubated with 0 and 25% CO in the headspace, respectively ( Figure 3A). When testing different CO partial pressures, from 25 to 100%, a decrease in H 2 production is observed, with H 2 production of <0.02 mmol L −1 at 100% CO. The other end products from glucose conversion were only affected when strain PCO was exposed to 100% CO. There was a significant decrease on the final production of lactate (P = 0.01) and ethanol (P = 0.004) only when comparing cultures grown with 75 and 100% CO ( Figure 4A). These results could be explained due to the fact that this strain had been isolated from a syngas-converting culture, that was in contact with high CO concentrations for over 1 year (Alves et al., 2013), which might have increased the tolerance of strain PCO to CO. There are only two Thermoanaerobacter species able to use CO: T. thermohydrosulfuricus subsp. carboxydovorans and T. kivui (Balk et al., 2009;Weghoff and Müller, 2016); for other Thermoanaerobacter species the ability to convert CO was never reported. The ability to utilize CO as carbon and energy source and/or the ability to tolerate CO by T. thermohydrosulfuricus, T. brockii subsp. finnii, T. pseudethanolicus, and T. wiegelii were tested in this study. None of the tested Thermoanaerobacter species could utilize CO ( Table 1), but all species could grow and completely convert glucose in the presence of 0, 50, or 100% of CO in the headspace (Table 3, Figures 3, 4). This suggests that CO tolerance is a characteristic present among Thermoanaerobacter genus, and not only a property of strain PCO. Carbon recovery at the end of the incubations was nearly 100% for all the growth tests ( Table 3). All the tested strains produced the same identified and quantified end products from glucose fermentation, i.e., hydrogen, lactate, acetate, and ethanol, as expected from previous reports (Kim et al., 2001;Lee et al., 2007;Balk et al., 2009;Shaw et al., 2010). Nevertheless, final product concentrations (Figures 3,  4) varied for the different tested strains. Ethanol production was in general less affected by the presence of different CO concentrations. A significant decrease in ethanol production was only observed in the presence of 100% CO and just for two of the five microorganisms tested: strain PCO (P = 0.004) and T. wiegelii (P = 0.002). H 2 production by all the tested strains decreased significantly, even in the presence of low CO concentrations. In the presence of 100% CO, H 2 production by the Thermoanaerobacter species tested was decreased by 75-95%. Acetate concentration decreased with increasing CO percentage (reduction between 25 and 50%); the exception was strain PCO for which the final acetate concentration was significantly higher in the presence of high CO percentage [acetate final concentration in cultures with 0 and 100% CO were 5.4 ± 0.6 and 8.3 ± 0.5 mM, respectively (P = 0.0003)]. These results show the metabolic changes in versatile Thermoanaerobacter species upon addition of CO, resulting in a general decrease in H 2 production. Although none of the tested Thermoanaerobacter strains could convert CO, all were able to withstand CO.
The lower H 2 production suggests that CO is inhibiting the hydrogenases of Thermoanaerobacter species. Hydrogenases catalyze the oxidation of hydrogen or the reduction of protons. From recent genomic studies, it was confirmed that [Fe-Fe]hydrogenases, responsible for hydrogen production, are present and well-conserved in all of the Thermoanaerobacter species. However, another class of hydrogenases, [Ni-Fe]-hydrogenases, is present only in T. thermohydrosulfuricus and T. wiegelii (Verbeke et al., 2013;Bhattacharya et al., 2015). [Fe-Fe] or iron-only hydrogenases are known to be more sensitive to CO than [Ni-Fe]-hydrogenases (Diender et al., 2015), which corroborate the results obtained regarding to the effect of CO on hydrogen production from glucose conversion by Thermoanaerobacter species. Hydrogenases were shown to be specifically inhibited by carbon monoxide, since CO binds at the active site of the enzyme (Purec et al., 1962;Gutiérrez-Sánchez et al., 2010;Baffert et al., 2011;Matsumoto et al., 2011;Bertsch and Müller, 2015). Genomic analysis of T. kivui revealed the presence of genes encoding for carbon monoxide dehydrogenases (CODH) and Ech-hydrogenases complexes (which are responsible for hydrogen production by carboxydotrophic organisms; Hess et al., 2014), although its ability to convert CO was only reported very recently after adaption to increasing concentrations of CO (Weghoff and Müller, 2016). Proper adaption to CO may be the key for achieving CO conversion by Thermoanaerobacter species which contain the necessary genomic machinery. Therefore, the high tolerance to CO and the potential of some microorganisms for CO utilization, make the members of Thermoanaerobacter genus important for the biotechnological use of syngas/industrial COrich gases. Thermophilic microorganisms including members of Thermoanaerobacter genus are interesting catalysts for production of biofuels (Carere et al., 2012;Verbeke et al., 2013;Hess et al., 2014;Bhattacharya et al., 2015;Sant'Anna et al., 2015). From this perspective the present study is important as CO is a way to steer the formation of fermentation products. Further research is needed to get a better insight into how at a molecular level carbon monoxide affects product formation in Thermoanaerobacter species.

AUTHOR CONTRIBUTIONS
JA planned and performed the experiments, data interpretation and wrote the manuscript. MA assisted in the design of the study, participated in data interpretation as well as revisions of the final manuscript. CP assisted in the design of the study, participated in data interpretation as well as revisions of the final manuscript. AS assisted in the design of the study, participated in data interpretation as well as revisions of the final manuscript. DS conceived the study, participated in the planning and coordination of the study, and revised the manuscript. All authors read and gave approval for publication of the manuscript.