Abstract
In this study, 12 strains of Thermoanaerobacter were isolated from a single decaying wood compost sample and subjected to genetic and phenotypic profiling. The 16S rRNA encoding gene sequences suggested that the isolates were most similar to strains of either Thermoanaerobacter pseudethanolicus or Thermoanaerobacter thermohydrosulfuricus. Examination of the lesser conserved chaperonin-60 (cpn60) universal target showed that some isolates shared the highest sequence identity with T. thermohydrosulfuricus; however, others to Thermoanaerobacter wiegelii and Thermoanaerobacter sp. Rt8.G4 (formerly Thermoanaerobacter brockii Rt8.G4). BOX-PCR fingerprinting profiles identified differences in the banding patterns not only between the isolates and the reference strains, but also among the isolates themselves. To evaluate the extent these genetic differences were manifested phenotypically, the utilization patterns of 30 carbon substrates were examined and the niche overlap indices (NOI) calculated. Despite showing a high NOI (> 0.9), significant differences existed in the substrate utilization capabilities of the isolates suggesting that either a high degree of niche specialization or mechanisms allowing for non-competitive co-existence, were present within this ecological context. Growth studies showed that the isolates were physiologically distinct in both growth rate and the fermentation product ratios. Our data indicate that phenotypic diversity exists within genetically microdiverse Thermoanaerobacter isolates from a common environment.
Introduction
Genomic and phenotypic heterogeneity within an operational taxonomic unit (OTU) is expanding our understanding of microbial diversity. Conventional diversity studies have relied upon a ribosomal RNA gene sequence (rrn) based approach, but increasing evidence suggests that using the 16S rRNA-encoding gene can significantly underestimate (Staley, ; Hong et al., ) or overestimate (Acinas et al., ,b) the true breadth of microbial diversity. Furthermore, clustering similar ribotypes into a single OTU fails to account for phenotypic diversity among ecotypes – physiologically distinct genetic lineages that occupy the same ecological niche and emerge through natural selection (Cohan, ; Rocap et al., ). Understanding the extent to which microdiversity exists in the microbial world dictates that a combined phenotypic and genetic approach be utilized.
Sub-species variation has previously been described as ‘microdiversity’ (Schloter et al., ). It has been used to describe genotypic variation (Rocap et al., ; Brown & Fuhrman, ; Cuadros-Orellana et al., ), phenotypic variation (Moore et al., ; Haverkamp et al., ), or a combination of both (Jaspers & Overmann, ; Acinas et al., ; Peña et al., ; Vermette et al., ), among isolates that conventionally fall into a single OTU. Most of these studies investigating sub-species variation have involved the examination of strains from geographically widespread, and thus heterogeneous environments with varying ecological pressures. The studies conducted by Moore et al. (), Vermette et al. (), and Peña et al. () report the co-occurrence of different ecotypes from a single environmental sample. Furthermore, it has been shown that in certain cases, ecotypes inhabiting and exploiting distinct microenvironments within a sample can be linked to distinct sequence clusters in which the average divergence within a cluster is less than the distance between clusters (Ramsing et al., ; Vermette et al., ).
The co-occurrence of highly similar strains seemingly violates the traditional ‘one niche-one strain’ philosophy in which two strains cannot occupy the same niche simultaneously (Staley, ). Competition for resources would, through time, select for a single variant, arisen from random mutation and genetic drift, best suited to occupy a specific niche. The existence of microdiverse communities therefore requires that strains exhibit either a high degree of niche specialization (Rainey & Travisano, ; Buckling et al., ; Bessen et al., ) or a mechanism of non-competitive niche diversification (Rainey et al., ). Selection based on niche differentiation would thus ultimately lead to phylogenetic divergence at the sub-species level (Cohan, ).
The mechanisms governing natural phenotypic variation have been attributed to spontaneous mutation (Schübbe et al., ), genomic rearrangement (Zverlov et al., ), and DNA acquisition (de la Cruz & Davies, ; Van Ham et al., ). Furthermore, it has been shown that under conditions of stress, these events increase as a coping mechanism (Woese, ). In coping with the stresses of the natural environment, frequent mutagenic events leading to niche diversification may act as an adaptive mechanism and lead to the establishment of a microdiverse community. Understanding the ecological significance of microdiversity has implications in terms of microbial ecology, evolutionary studies, taxonomy, bioprospecting, and biotechnology.
In this study, decaying wood compost was investigated as a potential source of novel lignocellulolytic, biofuel-producing bacteria. From a single sample, 12 highly similar, co-existing strains of Thermoanaerobacter were isolated. All strains were subjected to genetic and phenotypic characterization to identify, and evaluate differences between the strains. Small, albeit significant, differences were identified suggesting the existence of a natural microdiverse population consistent with previous studies in other environments (Acinas et al., ; Vermette et al., ). This is the first known report of microdiversity involving the genus Thermoanaerobacter.
Materials and methods
Reference strains
Thermoanaerobacter brockii ssp. brockii DSM 1457T, Thermoanaerobacter pseudethanolicus DSM 2355T, and Thermoanaerobacter thermohydrosulfuricus DSM 567T were purchased from the Deutsche Sammlung von Mikroorganismen und Zellkulturen (DSMZ), Braunschweig, Germany.
Media and substrates
All experiments were performed using ATCC 1191 medium as described (Sparling et al., ) with 2 g L−1 cellobiose unless otherwise specified. Bottles containing medium were sealed using butyl rubber stoppers and made anaerobic through gassing : degassing (1 min : 4 min) four times with a final gas cycle using 100% N2 in accordance with established protocols (Daniels et al., ). Reducing solution, 200 mM sodium sulfide, was added at 1% (v/v) prior to autoclaving. Filter sterilized, anaerobic sugar solutions were added, as necessary, to the medium using a needle and syringe post-autoclaving.
Acid hydrolyzed hemicellulose (AHH) and neutral hydrolyzed hemicellulose (NHH) were gifts from Pin-Ching Maness at the National Renewable Energy Laboratory (NREL) in Golden, CO. Hydrolysates were prepared from corn-stover according to the protocol of Datar et al. (). Prior to use as a fermentation substrate, the AHH (pH = 1.6) required over-liming to neutralize acetic and phenolic acids. To over-lime, AHH was heated to 42 °C and Ca(OH)2 was added until a pH of 10 was reached. The alkaline solution was allowed to mix for 30 min forming calcium phenate, which was removed via filtration through a 0.22 μm filter. The pH of the filtrate was adjusted to 5.6 corresponding to that of NHH, which was not over-limed.
Enrichment and isolation
Environmental isolates were obtained from a self-heating thermophilic, decaying maple, pine, and spruce wood compost in Arundel, PQ, Canada. Woodchips were collected from the center of the compost and stored in a 1 L stoppered Corning bottle (Corning Life Sciences, Lowell, MA) under an atmosphere of 100% N2 at 4 °C until use. Woodchips (~ 10 g) were transferred to 500 mL sterile 1191 medium inside an anaerobic chamber, stoppered, and incubated at 60 °C until growth was evident via gas production. Inoculations (10% v/v) were then performed into serum bottles containing 45 mL of 1191 and arabinose, cellulose, cellobiose, glycerol, xylose, AHH, or NHH were added as enrichment substrates. To avoid potential carbon limitation, all substrates were added at concentrations of 5 g L−1 except for AHH and NHH, which were added at 2 g L−1 (total sugar content) due to limited quantities. Cultures were plated in a Forma Scientific anaerobic glove chamber complete with incubator set at 60 °C on 1191 medium with agar (2% w/v) containing the respective enrichment substrate, and colonies were picked from these plates.
Microscopy
Cells were grown to an optical density (OD600) of 0.75 ± 0.03, and wet mounts were prepared for visualization using an Axio Imager.Z1 (Zeiss, Oberkochen, Germany). Images were captured using an AxioCam MRm camera (Zeiss) at a magnification of 1000×. Cell lengths were calculated using AxioVision version 4.7.2 software and averages calculated for each strain. Only individual cells, showing no signs of constriction or septum formation, and not those growing in chains, were used for measurements. A minimum of 65 measurements were used in calculating the average for each strain.
Substrate use and niche overlap
To determine the range of fermentable substrates for each isolate, 30 different carbon substrates were tested. Complex substrates tested included casamino acids, tryptone, and yeast extract. Nineteen sugar and sugar derivatives were used including arabinose, cellobiose, esculin, fructose, galactose, galacturonic acid, glucose, lactose, maltose, mannitol, mannose, melibiose, raffinose, rhamnose, ribose, salicin, sorbitol, sorbose, sucrose, and xylose. Tested carboxyl and amino acids included acetate, citrate, isoleucine, lactate, pyruvate, succinate, and tartaric acid. All substrates tested were at a final concentration of 2 g L−1 unless otherwise specified. To minimize growth on yeast extract components, a chemically defined medium, MJ medium, (Johnson et al., ), supplemented with 2 g L−1 cellobiose was initially tested. No growth was evident unless the medium was supplemented with yeast extract (data not shown). Neither vitamin nor mineral solutions (Sparling et al., ), casamino acids or tryptone were suitable replacements for yeast extract. As such, ATCC medium 1191 with reduced concentrations (0.67 g L−1) of yeast extract was used for the subsequent substrate utilization experiments. Growth was quantified via OD600. Carbon utilization was considered positive if the OD600 value of the test condition ± standard deviation was greater than that of the inoculated non-substrate control ± standard deviation. Initial experiments were conducted in 96-well polystyrene plates in a manner similar to that described by Jaspers & Overmann (). Sterile plates were placed into an anaerobic chamber 3 days prior to use to remove trace oxygen. The design of the plates was such that wells contained either: (i) no substrate + cells + medium; (ii) substrate + medium (negative control); (iii) substrate + cells + medium. The three wells together comprise a single condition. Conditions were replicated four times in each row per plate such that each row had a different carbon source. Seven test substrates and cellobiose (positive control) were tested on each plate. The total volume in the wells was 200 μL. Three plates for each set of conditions were tested and incubated at 60 °C inside the anaerobic chamber for 24, 72, or 96 h. One plate was removed after each incubation period (24, 72 or 96 h) and the growth determined by measuring the OD600. Upon removal from the chamber, plates were left for 20 min prior to reading allowing for resazurin oxidation. Wells containing substrate + medium only (negative control) typically had OD600 values ranging from 0.05 to 0.08. Conditions in which the OD600 value of the negative control was > 0.08 were discarded, as the possibility of contamination could not be discounted. Rows containing more than one negative control in which the OD600 value was > 0.08 were discarded and the experiment repeated. When contamination could not be discounted, new working stocks of cultures, derived from isolated colonies of the master stocks, were prepared and used for subsequent experiments. OD600 values were read using a PowerWave-XS single channel spectrophotometer using KCjunior software (BIO-TEK Instruments Inc., Winnoski, VT).
Substrate use was confirmed by repeating the above conditions in Balch (Bellco Glass Inc., Vineland, NJ) tubes using 10 mL cultures in triplicate. All conditions were independently verified a minimum of three times. Conditions where the OD600 reached low values (OD600 = ~ 0.10) were re-tested using culture serially transferred three times in 5 g L−1 test substrate prior to OD600 values being measured.
Test inocula were prepared by growing the cells under previously determined (not shown) carbon limiting conditions (0.5 g L−1 cellobiose). Under these conditions, the maximum OD600 of all strains ranged from 0.25 to 0.30. Thus, all inocula were grown to an OD600 of 0.20–0.25 to ensure near cellobiose depletion and to minimize the amount of residual cellobiose introduced into the test condition upon inoculation.
The niche overlap index (NOI), representing the potential for each isolate to occupy an identical carbon catabolism niche, was calculated according to the protocol of Wilson & Lindow (). In brief, the NOI between any pair of strains is the ratio of the number of substrates utilized by both strains, divided by the sum of the total substrates used by either strain. Results are scored as a binary code matrix, where substrate consumption (e.g. potential niche occupation) is scored as a ‘1’, whereas lack of substrate consumption is scored as a ‘0’. Data analysis was performed as described (Jaspers & Overmann, ).
16S rRNA and cpn60 UT amplification, sequencing and phylogenetic analysis
Isolated strains were grown overnight and DNA was extracted using the Wizard® Genomic DNA Purification kit (Promega Corp., Madison, WI) according to the manufacturer's protocols. Amplification of the 16S rRNA-encoding gene was performed from genomic DNA using previously described primers, 8F and 1541R (Löffler et al., ). PCR was performed using DreamTaq™ (Fermentas Canada Inc., Burlington, Canada). The mixtures were comprised of 25 μL 2× PCR buffer, 40 ng DNA, 0.30 μM of each primer and H2O to a final volume of 50 μL. The PCR cycling conditions were as follows: (i) initial denaturation at 96 °C for 1 min; (ii) 35 cycles of 96 °C for 45 s, 54 °C for 1 min and 72 °C for 1 min; and (iii) a final elongation step at 72 °C for 10 min.
The chaperonin-60 universal target (cpn60 UT) region was amplified using the H279/H280 primers previously described (Hill et al., ). The reaction mixture contained 0.5 μM of each primer, 2.5 mM MgSO4, 500 nM each dNTP, 2.5 U Taq polymerase (HP Taq) and 100–1200 ng DNA template. The cycling conditions involved an initial denaturation step at 95 °C for 3 min, followed by 40 cycles of 94 °C for 30 s, 50 °C for 30 s; 72 °C for 45 s. A final extension of 72 °C for 5 min was performed.
PCR product was purified prior to sequencing using Micron YM-30 ultrafiltration columns (Millipore, Billerica, MA) pre-rinsed with 200 μL water and centrifuged for 4 min at 14 000 g. PCR product (30–80 μL) was added to the column, centrifuged for 12 min at 14 000 g, and the flow through discarded. The purified product was recovered by adding 35 μL of H2O to the column followed by centrifugation at 2500 g for 5 min. The concentration of the purified product was determined by NanoDrop (Thermo Scientific, Wilmington, DE), adjusted to 10–20 ng μL−1, and sequenced. Contigs of the raw data were assembled using Pregap4 and Gap4 as part of the Staden software package (http://sourceforge.net/projects/staden).
Sequences were aligned with the default parameters of ClustalW (Thompson et al., ). Phylogenetic trees were constructed using the maximum composite likelihood method in mega4 (Tamura et al., , ) by the neighbor-joining algorithm with Jukes–Cantor correction (Jukes & Cantor, ). Bootstrap values for 1000 replicates were calculated. Phylogenetic relatedness was verified through Bayesian analysis using the Mr. Bayes v3.1.2 program (Ronquist & Huelsenbeck, ). Analysis was run for two million generations sampling every 10th tree. A majority rule consensus tree was constructed, and posterior probability analysis was performed. Sequence alignments were conducted and visualized using BioEdit v7.0.9.0 (Hall, ). Previously published sequences were accessed through GenBank or through the Joint Genome Institute (http://genome.jgi-psf.org/).
Genetic fingerprinting
BoxA1R (BOX) (Martin et al., ) and the repetitive extragenic palindromic (REP) (Stern et al., ) fingerprinting profiles were investigated for all strains. BOX-PCR was performed using the BOXA1R primer (Versalovic et al., ) as previously described (Urzi et al., ). REP-PCR involved the REP 1RI/REP 2I primers and protocol described by Versalovic et al. (). Five microliters of PCR product was used for agarose gel electrophoresis (1% w/v).
Growth and metabolic profiling
Cell growth was in 1191 medium with 2 g L−1 cellobiose as described above. Growth was monitored via OD600 using a Biochrom Novaspec II (Biochrom Ltd, Cambridge, UK) spectrophotometer. Cultures were grown to an average OD600 = 0.75 ± 0.03, at which time three biologic replicates were removed from incubation for metabolite analysis and pH determination. Three additional tubes were removed immediately post-inoculation and were used for time zero measurements. Gas production (H2 and CO2) was measured using a Multiple Gas Analyzer #1 Gas Chromatograph System Model 8610-0070 (SRI Instruments, Torrance, CA) as previously described (Rydzak et al., ). Total production was determined by correcting for solubility (Sander, ) and the bicarbonate equilibrium (Darrett & Drisham, ). One milliliter of samples were removed, centrifuged (13 000 g, 2 min), and the supernatant stored at −20 °C until analysis could be performed. Biomass calculations, organic acids, ethanol, and residual sugars were determined according to established protocols (Rydzak et al., ). Residual sugars and organic acids were analyzed by high-performance liquid chromatography (Dionex, ICS 3000) using an anion-exchange Carbo-PacPA1 (residual sugars) or an IonPac AS11-HC (organic acids) anion-exchange column. Ethanol measurements were determined spectrophotometrically at 340 nm using a R-Biopharm UV-Test kit (Darmstadt, Germany). All standards were prepared using 1191 medium to account for background signals. The supernatant from final tubes was diluted 1 in 4 to ensure that concentrations were in the detectable range. Measurements taken at time zero were subtracted from final measurements to determine total metabolite production. All growth studies were replicated three independent times.
Nucleotide accession numbers
Nucleotide accession numbers for sequences determined in this study can be accessed through GenBank. Accession numbers for 16S rRNA gene sequences are HM585213–HM585225. Accession numbers for cpn60 UT regions are HM623896–HM623910. Accession numbers for all strains can be found in Table S1 (Supporting Information).
Results and discussion
Enrichment and isolation of strains
The production of H2 or CO2 (coupled to ethanol production) on select enrichment substrates was used as an initial screening procedure to identify strains potentially suitable for biofuel production. Both gases were evident in all enrichment cultures and thus, all cultures were selected for further investigation. Gas production on arabinose, cellulose, and glycerol was much lower than for the other enrichment cultures, which was later attributed to fermentation of the yeast extract in the medium, rather than the enrichment substrate itself (Table S3). Undiluted enrichment culture was transferred onto plates, which yielded relatively uniform colonies among all of the test substrates. Colonies were smooth, uniformly round, mucoid, flat, and white. After 24 h incubation, the colonies could only be distinguished by size, ranging from 2 to 4.5 mm. Thirteen colonies (WC1–WC13) from different enrichment substrates (Table 1), and those that varied in size, were selected for subsequent genetic analysis.
Enrichment substrate and nucleotide at three divergent loci within the 16S rRNA sequence for isolates WC1–WC13
Available 16S rRNA sequence lengths (in base pairs) for Thermoanaerobacter brockii ssp. brockii DSM 1457 = 1513, Thermoanaerobacter pseudethanolicus DSM 2355 = 1601, 1527, 1527, 1769, and Thermoanaerobacter thermohydrosulfuricus DSM 567 = 1768.
NA, not applicable as WC13 is a strain of Thermoanaerobacterium thermosaccharolyticum.
Position determined in reference to WC8.
Enrichment substrate and nucleotide at three divergent loci within the 16S rRNA sequence for isolates WC1–WC13
Available 16S rRNA sequence lengths (in base pairs) for Thermoanaerobacter brockii ssp. brockii DSM 1457 = 1513, Thermoanaerobacter pseudethanolicus DSM 2355 = 1601, 1527, 1527, 1769, and Thermoanaerobacter thermohydrosulfuricus DSM 567 = 1768.
NA, not applicable as WC13 is a strain of Thermoanaerobacterium thermosaccharolyticum.
Position determined in reference to WC8.
Molecular identities and microdiversity of isolates
DNA was extracted from all 13 isolates and specific gene sequences were determined. The 16S rRNA-encoding sequence from isolate WC13 was significantly shorter than the sequences for isolates WC1–WC12 (Table 1). Isolates WC1–WC12 shared > 99% sequence identity with T. pseudethanolicus ATCC 33223 (DSM 2355) (Onyenwoke et al., ), whereas the 16S rRNA-encoding gene from isolate WC13 was most similar in sequence (> 99%) to Thermoanaerobacterium thermosaccharolyticum W16, an isolate from a hot spring in China (Ren et al., ).
Alignment of the 16S rRNA gene sequences generated from WC1–WC12 to the four 16S rRNA gene sequences in the published genome of T. pseudethanolicus (GenBank accession no. CP000924.1) showed that the shared sequence identity exists only with the gene copy found at nucleotide position 2 265 749–2 267 517 within the genome. This particular 16S rRNA gene copy contains four intervening sequences (IVS) making it significantly longer than the other gene copies found within the T. pseudethanolicus genome. These IVS were also present in the 16S rRNA gene sequences for isolates WC1–WC12 as evidenced by sequence alignments (Fig. S1). The alignment of the WC1–WC12 sequences identified three single nucleotide polymorphisms (SNPs) within the gene. The divergence at positions 1185, 1382, and 1413 is such that from the 12 sequences examined, six non-identical sequences were identified with no more than three isolates sharing an identical sequence. Furthermore, no two isolates sharing a common sequence were derived from the same enrichment treatment (Table 1). Despite only three SNPs of variation, WC1–WC12 fell into two separate clades as shown in Fig. 1. Phylogenetic analysis clustered isolates WC1–WC12 with T. thermohydrosulfuricus DSM 567, another species known to have 16S gene IVS (Rainey et al., ).
Neighbor-joining tree of 16S rRNA sequences from isolates WC1–WC13 and from selected reference strains. Clostridium thermocellum was used to root the tree. The tree was constructed as described in . Confidence levels >50% are shown after bootstrap values (outside brackets) were calculated and posterior probability analysis (inside brackets) was performed. Nucleotide accession numbers for sequences determined in this work used can be found in Table S1. Numbers in brackets after Thermoanaerobacter pseudethanolicus refer to relative position within the genome. Scale bar represents the number of changes per nucleotide. Members of the genus Caldanaerobacter were formerly described as Thermoanaerobacter (Fardeau et al., ).
Accurate strain taxonomic assignments were largely complicated by the presence of IVS within the 16S rRNA gene sequences of isolates WC1–WC12, T. pseudethanolicus and T. thermohydrosulfuricus. Such IVS have been described in numerous members of the Thermoanaerobacteriaceae family (Rainey et al., ). Alignments of the 16S rRNA genes identifies that the positioning of all four IVS sequences within the rRNA gene is conserved among isolates WC1–WC12, T. pseudethanolicus and T. thermohydrosulfuricus. Furthermore, the sequence identity of all four IVS is 100% conserved across these strains (Table S2) with the exception of IVS 1 and IVS 4 found in T. thermohydrosulfuricus. The IVS 1 sequence from T. thermohydrosulfuricus shows 97.3% homology to that of T. pseudethanolicus and the WC isolates, whereas IVS 4 shares 91.5% sequence identity across strains. However, these values may be artificially low as the published 16S rRNA gene sequence for T. thermohydrosulfuricus DSM 567 (GenBank accession no. L09161 – Rainey et al., ) contains residues identified as ‘N’ within the IVS regions. The 16S rRNA gene sequence of T. brockii ssp. brockii contains no IVS.
Additional evaluation of the published genomes of Thermoanaerobacter strains available from the Joint Genome Institute (http://genome.jgi-psf.org/) has identified that not all gene copies within a common genome necessarily contain IVS. For example, of the four T. pseudethanolicus rrn operons, one 16S rRNA gene sequence (2 265 749–2 267 517) has four IVS, whereas another copy (438 455–440 055) has a single IVS. The additional two 16S rRNA gene copies have no IVS.
As IVS are not found in all 16S rRNA gene copies within a genome (e.g. T. pseudethanolicus), yet the sequence is conserved across strains, it suggests that gene transfer may have occurred in the evolutionary history of these strains. Typically, genes with core functions, such as rrn genes, are thought to be less susceptible to horizontal gene transfer (HGT) (Jain et al., ; Daubin et al., ). However, high levels of rrn interoperonic variability among thermophiles, particularly within Caldanaerobacter subterraneus ssp. tengcongensis (formerly Thermoanaerobacter tengcongensis; Fardeau et al., ), have been attributed to HGT (Acinas et al., ). The potential for horizontal gene acquisition is further supported by the recent realization that members of the Thermoanaerobacteriaceae family exhibit varying levels of natural competence (Shaw et al., ).
High interoperonic diversity observed in the T. pseudethanolicus genome and in highly related strains (Fardeau et al., ; Pei et al., ) suggests that using 16S rRNA gene sequences for phylogenetic classification of Thermoanaerobacter strains may not accurately reflect true evolutionary relationships. In addition, the possibility that rrn genes may be subjected to HGT in this genus creates ambiguity in taxonomic assignment on the basis of 16S rRNA gene sequence alone. Therefore, as an alternative phylogenetic signature gene, the more variable cpn60 UT region was investigated. The cpn60 UT region has previously been shown to accurately discriminate between closely related strains (Marston et al., ). Furthermore, the study by Verbeke et al. () has shown that the cpn60 UT region can predict whole genome relatedness with the same accuracy as some multi-locus and even whole genome strategies and correlates well with DNA–DNA hybridization values. The phylogenetic relatedness of strain WC13 to Th. thermosaccharolyticum DSM 571 observed using 16S rRNA gene sequence analysis is further supported by the shared sequence identity of the cpn60 UT region (97.8%), thus suggesting it is a strain of this species. Isolates WC2–WC7 shared the greatest cpn60 UT sequence identity with T. thermohydrosulfuricus (> 98.7%), whereas isolates WC1 and WC8–WC12 shared the greatest sequence identity (> 98.9%) with Thermoanaerobacter sp. Rt8.G4 (Truscott & Scopes, ), a strain formerly identified as T. brockii (Truscott et al., ). In contrast, the cpn60 UT sequence identity between any of the isolates and T. pseudethanolicus was much lower ranging from 87.1 to 89.3%. Recently (June 2010), the draft genome sequence for Thermoanaerobacter wiegelii Rt8.B1 (Cook et al., ) was released by the Joint Genome Institute and isolates WC1 and WC8–WC12 shared a greater cpn60 UT sequence identity (> 99.1%) with this strain. It should be noted that the original designation of T. wiegelii Rt8.B1 as a novel species (Cook et al., ) is largely based on phenotypic and 16S rRNA data without using the previously recognized standard (DNA–DNA hybridization) (Stackebrandt et al., ) for species designation. Unfortunately, the lack of a partial or complete genome sequence for the type strain of T. thermohydrosulfuricus prevents whole genome comparisons of the type described by Richter & Rosselló-Móra (); however, mathematical models for predicting whole genome sequence identities, and thus phylogenetic relatedness, (Verbeke et al., ) suggest that T. wiegelii, T. thermohydrosulfuricus, and the isolates (WC1–WC12) described here are probably related at the species level (data presented in Verbeke et al., ).
Overall, sequence divergence was greater in the cpn60 UT region than was found using the 16S rRNA-encoding gene, identifying 12 unique genetic signatures among isolates WC1–WC12 and suggesting the presence of a genetically microdiverse community. Furthermore, different clustering patterns (Fig. 2) were observed using the cpn60 UT region than were observed using 16S rRNA gene sequences. It would be expected that if isolates WC1–WC12 represented a continuum of genetic drift, the clustering patterns would remain constant. Thus, these findings suggest that within these isolates, divergence has occurred in multiple directions before the initial laboratory enrichment process.
Neighbor-joining tree of cpn60 UT sequences from isolates WC1–WC13 and from selected reference strains. Clostridium thermocellum was used to root the tree. The tree was constructed as described in Materials and methods. Confidence levels >50% are shown after bootstrap values (outside of brackets) were calculated, and posterior probability analysis (inside brackes) was performed. Nucleotide accession numbers for sequences determined in this work used can be found in Table S1. Scale bar represents the number of changes per nucleotide. Members of the genus Caldanaerobacter were formerly described as Thermoanaerobacter (Fardeau et al., ).
Genetic fingerprinting of isolates
To help determine the extent of genetic diversity among the isolates, fingerprinting profiles based on the BOX A1R (Versalovic et al., ) and REP (Versalovic et al., ) sequences were investigated. Genetic fingerprinting based on repetitive DNA sequences is effective for identifying genetic diversity at the sub-species level (Dombek et al., ; Tacão et al., ), and has previously been effectively employed with strains of Thermoanaerobacter (Roh et al., ).
In this study, the type T. brockii strain, T. brockii ssp. brockii DSM 1457 (Lee et al., ), T. pseudethanolicus DSM 2355, T. thermohydrosulfuricus DSM 567, and Th. thermosaccharolyticum WC13 were used as reference strains. Although the banding patterns for isolates WC1–WC12 and the three Thermoanaerobacter reference strains were identical using REP-PCR (data not shown), distinct banding patterns were identified among isolates WC1–WC12 using BOX-PCR (Fig. 3). Isolate WC5 showed three distinct bands between 400 and 500 nucleotides in length, which was similar to both T. brockii ssp. brockii and T. pseudethanolicus, but distinct from T. thermohydrosulfuricus. All other isolates showed banding patterns most comparable to T. thermohydrosulfuricus, though distinct patterns among isolates were evident (e.g. WC3 and WC4). In addition, identical banding patterns existed between isolates that were not consistent with the phylogenetic clustering observed (Figs 1 and 2).
Banding patterns of BOX-PCR profiles for isolates WC1–WC13, Thermoanaerobacter brockii ssp. brockii, Thermoanaerobacter pseudethanolicus, and Thermoanaerobacter thermohydrosulfuricus. Lanes (left to right): ladder; WC1; WC2; WC3; WC4; WC5; WC6; WC7; WC8; WC9; WC10; WC11; WC12; Thermoanaerobacter thermohydrosulfuricus; Thermoanaerobacter pseudethanolicus; Thermoanaerobacter brockii ssp. brockii; Thermoanaerobacterium thermosaccharolyticum WC13; ladder; water (negative control).
These results further support the idea that these isolates form a genetically microdiverse group of T. thermohydrosulfuricus isolated from a common sampling environment. The differential banding patterns provide evidence that among isolates WC1–WC12, mechanisms allowing for genomic diversification exist. Although the mechanisms responsible (gene deletions, gene acquisitions, mobile element insertions, etc.) are unknown, they provide evidence of evolutionary progression potentially leading to strain differentiation (Cohan, ). Also, although the significance of these specific genomic variations is unknown, previous studies have shown the importance of chromosomal rearrangements in strain diversification and niche specialization (Römling et al., ; Kresse et al., ).
Cell morphology and size
Cells were grown to an OD600 = 0.75 ± 0.03 and then examined via differential interference contrast microscopy. Cultures WC1–WC12 showed indistinct morphologies as cells in all cultures were rod-shaped and arranged singly or in chains (Fig. S2a). Periodically, long filamentous cells were observed, which may occur prior to cell division. All of the above characteristics are typical of cells within the Thermoanaerobacter genus (Wiegel & Ljungdahl, ) and are morphologically indiscriminate from the descriptions for T. pseudethanolicus (Onyenwoke et al., ), T. brockii ssp. brockii (Zeikus et al., ), and T. thermohydrosulfuricus (Lee et al., ). Cell division was often unequal as suggested by differences in cell lengths found in chains. Cells removed from growth medium and immediately examined were motile for short periods of time. The lengths of the cells varied from 1.68 to 17.83 μm. Average cell length was calculated for each culture and was shown to vary from 3.66 μm for WC4 to 6.24 μm for WC6 (Fig. S2b). The average lengths for isolates WC1–WC12 were greater than those observed for T. pseudethanolicus and T. brockii ssp. brockii, but were smaller than those for T. thermohydrosulfuricus.
Niche occupation and specialization
Considerable niche overlap (NOI > 0.9) was found among strains WC1–WC12 (Fig. 4). Strain WC13, identified by DNA sequence analysis as a strain of Th. thermosaccharolyticum, also showed a NOI > 0.9 in comparison to isolates WC1–WC12 confirming high potential niche overlap independent of species and even genus designation. Previous studies investigating the potential niche overlap of genetically microdiverse strains have reported a wide range of NOIs ranging from 0.25 to 0.59 for 11 Brevundimonas alba strains (Jaspers & Overmann, ) to 1.00 in various Pseudomonas syringae strains (Wilson & Lindow, ).
NOI of isolates WC1–WC12, Thermoanaerobacter brockii ssp. brockii, Thermoanaerobacter pseudethanolicus, and Thermoanaerobacter thermohydrosulfuricus as determined by substrate utilization patterns. WC13, identified as a strain of Thermoanaerobacterium thermosaccharolyticum, is included as a phylogenetic outgroup.
Despite the high NOIs observed, seven utilization profiles (Fig. 4) for WC1–WC12 could be identified based upon the differential fermentation capabilities of only five carbon sources (Table 2). WC8 showed the least metabolic diversity as growth was supported using only 13 of the substrates, which all other strains were also capable of using. In contrast, WC1 was the most versatile, and was capable of fermenting 18 substrates.
| Lactose | Mannitol | Melibiose | Raffinose | Sorbitol | Sucrose | |
| WC1 | + | + | + | + | + | + |
| WC2 | − | − | + | + | − | − |
| WC3 | + | + | + | − | − | − |
| WC4 | + | + | + | − | − | − |
| WC5 | − | − | + | + | − | − |
| WC6 | − | − | + | + | − | − |
| WC7 | − | − | + | + | − | − |
| WC8 | − | − | + | − | − | − |
| WC9 | + | + | + | + | − | − |
| WC10 | − | − | + | + | − | − |
| WC11 | + | + | + | + | − | + |
| WC12 | + | − | + | + | + | − |
| T. brockii ssp. brocki | + | + | + | + | − | + |
| T. pseudethanolicus | + | + | + | + | − | + |
| T. thermohydrosulfuricus | − | + | − | − | − | + |
| Lactose | Mannitol | Melibiose | Raffinose | Sorbitol | Sucrose | |
| WC1 | + | + | + | + | + | + |
| WC2 | − | − | + | + | − | − |
| WC3 | + | + | + | − | − | − |
| WC4 | + | + | + | − | − | − |
| WC5 | − | − | + | + | − | − |
| WC6 | − | − | + | + | − | − |
| WC7 | − | − | + | + | − | − |
| WC8 | − | − | + | − | − | − |
| WC9 | + | + | + | + | − | − |
| WC10 | − | − | + | + | − | − |
| WC11 | + | + | + | + | − | + |
| WC12 | + | − | + | + | + | − |
| T. brockii ssp. brocki | + | + | + | + | − | + |
| T. pseudethanolicus | + | + | + | + | − | + |
| T. thermohydrosulfuricus | − | + | − | − | − | + |
Complete utilization profile can be found in Table S2.
| Lactose | Mannitol | Melibiose | Raffinose | Sorbitol | Sucrose | |
| WC1 | + | + | + | + | + | + |
| WC2 | − | − | + | + | − | − |
| WC3 | + | + | + | − | − | − |
| WC4 | + | + | + | − | − | − |
| WC5 | − | − | + | + | − | − |
| WC6 | − | − | + | + | − | − |
| WC7 | − | − | + | + | − | − |
| WC8 | − | − | + | − | − | − |
| WC9 | + | + | + | + | − | − |
| WC10 | − | − | + | + | − | − |
| WC11 | + | + | + | + | − | + |
| WC12 | + | − | + | + | + | − |
| T. brockii ssp. brocki | + | + | + | + | − | + |
| T. pseudethanolicus | + | + | + | + | − | + |
| T. thermohydrosulfuricus | − | + | − | − | − | + |
| Lactose | Mannitol | Melibiose | Raffinose | Sorbitol | Sucrose | |
| WC1 | + | + | + | + | + | + |
| WC2 | − | − | + | + | − | − |
| WC3 | + | + | + | − | − | − |
| WC4 | + | + | + | − | − | − |
| WC5 | − | − | + | + | − | − |
| WC6 | − | − | + | + | − | − |
| WC7 | − | − | + | + | − | − |
| WC8 | − | − | + | − | − | − |
| WC9 | + | + | + | + | − | − |
| WC10 | − | − | + | + | − | − |
| WC11 | + | + | + | + | − | + |
| WC12 | + | − | + | + | + | − |
| T. brockii ssp. brocki | + | + | + | + | − | + |
| T. pseudethanolicus | + | + | + | + | − | + |
| T. thermohydrosulfuricus | − | + | − | − | − | + |
Complete utilization profile can be found in Table S2.
Three distinct clusters, where all members of the cluster demonstrated identical utilization patterns, were observed. One cluster grouped isolate WC11 with T. pseudethanolicus and T. brockii ssp. brockii (Fig. 4). Isolates WC3 and WC4 comprised another cluster with both utilizing 15 test substrates, whereas WC2, WC5, WC6, WC7, and WC10 formed the final cluster capable of fermenting 14 substrates. This largest cluster was the only one observed that grouped all the isolates from a common enrichment together (WC5, WC6, WC7 – xylose enrichment). These isolates, with the exception of WC10, also formed a subgroup based on cpn60 UT analysis (Fig. 2). The sugar utilization pattern exhibited by T. pseudethanolicus DSM 2355 agrees with previously published descriptions (Onyenwoke et al., ). Thermoanaerobacter brockii ssp. brockii was shown to ferment xylose and mannose under these conditions in contrast to previous reports (Zeikus et al., ). This discrepancy may be due to differences in experimental design as this study included incubation periods of up to 96 h, whereas those of Zeikus et al. () were only 16 h, prior to measuring the OD600. Long lag periods and slow generation times were commonly observed when cultures were transferred to a new carbon substrate (data not shown) potentially accounting for the lack of growth previously observed. The substrate utilization capabilities of T. thermohydrosulfuricus agree with previous reports (Hollaus & Sleytr, ; Wiegel et al., ; Lee et al., ) except in the utilization of arabinose and raffinose. In this study, T. thermohydrosulfuricus DSM 567 did not utilize arabinose in agreement with Hollaus & Sleytr (), in contrast to the report by Wiegel et al. (). The inability of T. thermohydrosulfuricus to utilize both sugars in this study may be due to the use of lower amounts of yeast extract in these experiments, which can significantly impact the extent of microbial growth for T. thermohydrosulfuricus (Wiegel et al., ). Surprisingly, despite the observed genetic similarities (Figs 1–3), T. thermohydrosulfuricus DSM 567 showed the most distinct profile of all of the strains examined (Fig. 4).
The co-occurrence of such highly similar strains from a common environment suggests that, for strain survival, niche diversification and specialization may exist. Competition for resources in a natural environment is thought to deter similar ecotypes from sharing a common ecological niche and thus prevent their co-existence (Staley, ). The similar substrate utilization patterns (high NOIs) observed for strains WC1–WC12 thus seemingly violates this principle. These results may be explained in two ways. First, niche specialization may involve parameters not investigated in the experiments performed here. Second, wood compost is a dynamic and nutritionally heterogeneous environment conducive to the formation of micro-niches. Thus, spatial and temporal variation in the evolutionary and ecological forces exerted on these micro-environments will ultimately result in diversified localized cell populations. Previous lab studies have shown that in heterogeneous environments, adaptive radiation through a combination of mutation and natural selection allows for phenotypic and niche diversification from an initial clonal population of cells (Rainey & Travisano, ; Brockhurst et al., ). Although the environmental and evolutionary pressures isolates, WC1–WC12, were subjected to in situ are unknown, it is evident that these pressures have been selected for divergence in the potential carbon catabolism capabilities of these isolates and resulted in a genetic and phenotypic microdiverse community.
Metabolic profiling
Growth rate and fermentation end products on 2 g L−1 cellobiose were investigated to compare differences between the central metabolisms of isolates WC1–WC12, T. brockii ssp. brockii, T. pseudethanolicus, and T. thermohydrosulfuricus. Significant differences were observed (Table 3; Fig. S3) in both growth rate and ratio of end products produced. Although the O/R indices suggest accurate measurement of all end products, the carbon recovery for all strains, except WC1, WC10, and T. thermohydrosulfuricus was slightly higher than 100%. Presumably, this is due to the utilization of yeast extract in the medium, as all strains have shown to support growth using yeast extract alone (Table S3). This is further supported by HPLC analysis indicating all cultures had consumed > 91% of the cellobiose substrate with an average usage by all cultures > 98 ± 3.2% upon reaching an OD600 = 0.75 ± 0.03 (not shown). As significant lag periods were not observed, we assume that the additional substrates were consumed simultaneously with the cellobiose for direct biosynthetic use or after the exponential phase of growth when cellobiose was depleted.
| Culture | Growth rate (h gen−1) | Concentrations produced (mM) | O/R index | % C recovery | |||||
| Acetate | Lactate | Ethanol | H2 | CO2 | Biomass | ||||
| WC1 | 1.86 | 8.30 | 3.80 | 7.12 | 6.60 | 13.95 | 3.92 | 1.04 | 99.14 |
| WC2 | 2.50 | 2.05 | 17.00 | 3.03 | 1.73 | 6.59 | 3.82 | 0.97 | 102.64 |
| WC3 | 2.32 | 3.10 | 16.23 | 4.49 | 2.10 | 8.22 | 3.78 | 0.98 | 108.95 |
| WC4 | 6.74 | 7.11 | 8.22 | 7.96 | 4.82 | 13.38 | 3.96 | 1.00 | 107.14 |
| WC5 | 2.16 | 4.12 | 13.06 | 3.64 | 1.75 | 7.02 | 3.79 | 0.95 | 109.43 |
| WC6 | 2.28 | 3.62 | 15.10 | 2.82 | 1.56 | 6.09 | 3.81 | 0.94 | 110.88 |
| WC7 | 2.40 | 2.01 | 14.89 | 3.10 | 1.78 | 6.42 | 3.81 | 0.94 | 107.11 |
| WC8 | 2.65 | 6.80 | 13.53 | 4.94 | 2.88 | 9.98 | 3.77 | 1.08 | 115.39 |
| WC9 | 3.01 | 2.85 | 14.17 | 2.84 | 2.10 | 6.11 | 3.66 | 0.92 | 111.59 |
| WC10 | 2.62 | 2.84 | 12.83 | 2.78 | 2.52 | 6.83 | 3.82 | 0.99 | 97.34 |
| WC11 | 1.88 | 2.46 | 11.89 | 2.21 | 1.39 | 5.40 | 3.85 | 0.93 | 101.17 |
| WC12 | 2.45 | 0.98 | 16.52 | 3.94 | 2.23 | 7.23 | 3.84 | 0.91 | 109.49 |
| T. brockii ssp. brockii | 2.68 | 2.26 | 11.92 | 10.83 | 0.66 | 11.84 | 3.68 | 0.98 | 107.05 |
| T. pseudethanolicus | 2.14 | 6.12 | 4.74 | 12.23 | 0.98 | 15.62 | 3.77 | 1.00 | 107.95 |
| T. thermohydrosulfuricus | 5.57 | 9.33 | 4.03 | 7.63 | 6.88 | 13.09 | 3.96 | 0.93 | 96.67 |
| Culture | Growth rate (h gen−1) | Concentrations produced (mM) | O/R index | % C recovery | |||||
| Acetate | Lactate | Ethanol | H2 | CO2 | Biomass | ||||
| WC1 | 1.86 | 8.30 | 3.80 | 7.12 | 6.60 | 13.95 | 3.92 | 1.04 | 99.14 |
| WC2 | 2.50 | 2.05 | 17.00 | 3.03 | 1.73 | 6.59 | 3.82 | 0.97 | 102.64 |
| WC3 | 2.32 | 3.10 | 16.23 | 4.49 | 2.10 | 8.22 | 3.78 | 0.98 | 108.95 |
| WC4 | 6.74 | 7.11 | 8.22 | 7.96 | 4.82 | 13.38 | 3.96 | 1.00 | 107.14 |
| WC5 | 2.16 | 4.12 | 13.06 | 3.64 | 1.75 | 7.02 | 3.79 | 0.95 | 109.43 |
| WC6 | 2.28 | 3.62 | 15.10 | 2.82 | 1.56 | 6.09 | 3.81 | 0.94 | 110.88 |
| WC7 | 2.40 | 2.01 | 14.89 | 3.10 | 1.78 | 6.42 | 3.81 | 0.94 | 107.11 |
| WC8 | 2.65 | 6.80 | 13.53 | 4.94 | 2.88 | 9.98 | 3.77 | 1.08 | 115.39 |
| WC9 | 3.01 | 2.85 | 14.17 | 2.84 | 2.10 | 6.11 | 3.66 | 0.92 | 111.59 |
| WC10 | 2.62 | 2.84 | 12.83 | 2.78 | 2.52 | 6.83 | 3.82 | 0.99 | 97.34 |
| WC11 | 1.88 | 2.46 | 11.89 | 2.21 | 1.39 | 5.40 | 3.85 | 0.93 | 101.17 |
| WC12 | 2.45 | 0.98 | 16.52 | 3.94 | 2.23 | 7.23 | 3.84 | 0.91 | 109.49 |
| T. brockii ssp. brockii | 2.68 | 2.26 | 11.92 | 10.83 | 0.66 | 11.84 | 3.68 | 0.98 | 107.05 |
| T. pseudethanolicus | 2.14 | 6.12 | 4.74 | 12.23 | 0.98 | 15.62 | 3.77 | 1.00 | 107.95 |
| T. thermohydrosulfuricus | 5.57 | 9.33 | 4.03 | 7.63 | 6.88 | 13.09 | 3.96 | 0.93 | 96.67 |
End product data correspond to measurements taken at an average OD600 = 0.75 ± 0.03 and corrected for carryover from inoculation. Values represent averages from three independent experiments using three biologic replicates/experiment.
| Culture | Growth rate (h gen−1) | Concentrations produced (mM) | O/R index | % C recovery | |||||
| Acetate | Lactate | Ethanol | H2 | CO2 | Biomass | ||||
| WC1 | 1.86 | 8.30 | 3.80 | 7.12 | 6.60 | 13.95 | 3.92 | 1.04 | 99.14 |
| WC2 | 2.50 | 2.05 | 17.00 | 3.03 | 1.73 | 6.59 | 3.82 | 0.97 | 102.64 |
| WC3 | 2.32 | 3.10 | 16.23 | 4.49 | 2.10 | 8.22 | 3.78 | 0.98 | 108.95 |
| WC4 | 6.74 | 7.11 | 8.22 | 7.96 | 4.82 | 13.38 | 3.96 | 1.00 | 107.14 |
| WC5 | 2.16 | 4.12 | 13.06 | 3.64 | 1.75 | 7.02 | 3.79 | 0.95 | 109.43 |
| WC6 | 2.28 | 3.62 | 15.10 | 2.82 | 1.56 | 6.09 | 3.81 | 0.94 | 110.88 |
| WC7 | 2.40 | 2.01 | 14.89 | 3.10 | 1.78 | 6.42 | 3.81 | 0.94 | 107.11 |
| WC8 | 2.65 | 6.80 | 13.53 | 4.94 | 2.88 | 9.98 | 3.77 | 1.08 | 115.39 |
| WC9 | 3.01 | 2.85 | 14.17 | 2.84 | 2.10 | 6.11 | 3.66 | 0.92 | 111.59 |
| WC10 | 2.62 | 2.84 | 12.83 | 2.78 | 2.52 | 6.83 | 3.82 | 0.99 | 97.34 |
| WC11 | 1.88 | 2.46 | 11.89 | 2.21 | 1.39 | 5.40 | 3.85 | 0.93 | 101.17 |
| WC12 | 2.45 | 0.98 | 16.52 | 3.94 | 2.23 | 7.23 | 3.84 | 0.91 | 109.49 |
| T. brockii ssp. brockii | 2.68 | 2.26 | 11.92 | 10.83 | 0.66 | 11.84 | 3.68 | 0.98 | 107.05 |
| T. pseudethanolicus | 2.14 | 6.12 | 4.74 | 12.23 | 0.98 | 15.62 | 3.77 | 1.00 | 107.95 |
| T. thermohydrosulfuricus | 5.57 | 9.33 | 4.03 | 7.63 | 6.88 | 13.09 | 3.96 | 0.93 | 96.67 |
| Culture | Growth rate (h gen−1) | Concentrations produced (mM) | O/R index | % C recovery | |||||
| Acetate | Lactate | Ethanol | H2 | CO2 | Biomass | ||||
| WC1 | 1.86 | 8.30 | 3.80 | 7.12 | 6.60 | 13.95 | 3.92 | 1.04 | 99.14 |
| WC2 | 2.50 | 2.05 | 17.00 | 3.03 | 1.73 | 6.59 | 3.82 | 0.97 | 102.64 |
| WC3 | 2.32 | 3.10 | 16.23 | 4.49 | 2.10 | 8.22 | 3.78 | 0.98 | 108.95 |
| WC4 | 6.74 | 7.11 | 8.22 | 7.96 | 4.82 | 13.38 | 3.96 | 1.00 | 107.14 |
| WC5 | 2.16 | 4.12 | 13.06 | 3.64 | 1.75 | 7.02 | 3.79 | 0.95 | 109.43 |
| WC6 | 2.28 | 3.62 | 15.10 | 2.82 | 1.56 | 6.09 | 3.81 | 0.94 | 110.88 |
| WC7 | 2.40 | 2.01 | 14.89 | 3.10 | 1.78 | 6.42 | 3.81 | 0.94 | 107.11 |
| WC8 | 2.65 | 6.80 | 13.53 | 4.94 | 2.88 | 9.98 | 3.77 | 1.08 | 115.39 |
| WC9 | 3.01 | 2.85 | 14.17 | 2.84 | 2.10 | 6.11 | 3.66 | 0.92 | 111.59 |
| WC10 | 2.62 | 2.84 | 12.83 | 2.78 | 2.52 | 6.83 | 3.82 | 0.99 | 97.34 |
| WC11 | 1.88 | 2.46 | 11.89 | 2.21 | 1.39 | 5.40 | 3.85 | 0.93 | 101.17 |
| WC12 | 2.45 | 0.98 | 16.52 | 3.94 | 2.23 | 7.23 | 3.84 | 0.91 | 109.49 |
| T. brockii ssp. brockii | 2.68 | 2.26 | 11.92 | 10.83 | 0.66 | 11.84 | 3.68 | 0.98 | 107.05 |
| T. pseudethanolicus | 2.14 | 6.12 | 4.74 | 12.23 | 0.98 | 15.62 | 3.77 | 1.00 | 107.95 |
| T. thermohydrosulfuricus | 5.57 | 9.33 | 4.03 | 7.63 | 6.88 | 13.09 | 3.96 | 0.93 | 96.67 |
End product data correspond to measurements taken at an average OD600 = 0.75 ± 0.03 and corrected for carryover from inoculation. Values represent averages from three independent experiments using three biologic replicates/experiment.
Acetate, lactate, ethanol, H2, and CO2 were the only observed fermentation products for isolates WC1–WC12 and the three reference strains. However, significant differences existed in both the ratio of end products produced and the growth rate on 2 g L−1 cellobiose (Table 3; Fig. S3). Most isolates produced lactate as the major end product, whereas isolates WC1 and WC4 produced CO2 (coupled to acetate and ethanol production) as the major product of fermentation. There also seemed to be no obvious correlation between fermentation products and growth rate. For example, WC1 and WC4 represented the extremes in generation times (1.86 and 6.74 h gen−1 respectively), despite the fact that both strains had similar acetate and ethanol yields (Table 3). The two most similar isolates in terms of calculated generation time, WC1, and WC11, exhibited a 4.3-fold difference in lactate production.
The reference strains produced the same end products as those of the isolates. For T. brockii ssp. brockii, lactate was the major fermentation end product with only trace amounts of H2 produced in agreement with previous findings (Ben-Bassat et al., ; Lee et al., ). The fermentation profile of T. pseudethanolicus indicated that ethanol was the major reduced product similar to previous reports (Zeikus et al., ; Lovitt et al., ). The nature of the end products for T. thermohydrosulfuricus DSM 567 was consistent with previous reports (Wiegel et al., ), although the ratio of end products differed in this study. Acetate was the major organic acid produced here, whereas ethanol was predominant in Wiegel and coworkers’ experiment (). This difference may be due to variations in experimental design as the incubation temperatures varied by 10 °C between the two studies. Previous work with Thermoanaerobacter has shown that temperature can significantly impact the enzymatic activity of the core carbohydrate enzymes and thus impact end product ratios (Lamed & Zeikus, ).
The ecological significance of the metabolic differences observed between the strains is currently unclear. As the fermentation products were identical in all strains, variation in the ratios may be evidence of niche specialization, but also may simply indicate a certain level of genetic variation due to drift or mutation. Minor genetic variations may significantly impact factors including effector molecule binding affinities and rates of carbon flux. In addition, these effects may have either immediate or downstream effects. Previous studies have shown how mutations in metabolic enzymes may alter the rates of metabolic flux, while still maintaining reaction fidelity (Goupil et al., ). Alternatively, the differences may also be due to differences in the regulatory mechanisms governing protein synthesis or protein expression in the central metabolisms of these strains. Previous studies have shown how mobile element based chromosomal rearrangements can lead to phenotype variation, and ultimately differential niche occupation (Dekkers et al., ). Although spontaneous transposon-based mutagenesis has been reported in the related anaerobic thermophile, Clostridium thermocellum (Zverlov et al., ), to date, no events have been described within the Thermoanaerobacter genus. An evaluation of the Joint Genome Institute's (http://genome.jgi-psf.org/) completed genome sequences of Thermoanaerobacter sp. X514, T. pseudethanolicus DSM 2355 and the draft sequence of Thermoanaerobacter brockii ssp. finii identified 19, 46, and 25 putative annotated transposases respectively. Although the presence of mobile elements is unknown within isolates WC1–WC12, their frequent presence within the genus suggests they may exist, and may help explain the observed phenotypic variations.
Conclusions
Twelve genetically and phenotypically distinct, co-existing strains of T. thermohydrosulfuricus were isolated from a single decaying wood compost sample. The co-existence of such similar strains from a common environmental sample suggests that microbial diversification may be very tightly linked to niche specialization as has been previously reported (Rainey & Travisano, ; Ramsing et al., ). Niche diversification and specialization have previously been shown to result from fitness ‘trade-offs’ (Rainey et al., ). These ‘trade-offs’ may have evolved among strains WC1–WC12 as a means of allowing for the spatial co-existence of strains derived from a single lineage, while permitting localized populations to fill unoccupied or non-competitive niches. In brief, a ‘trade-off’ may exist in the sense that one strain may lose the ability to ferment one substrate, yet optimize its utilization of another. A second strain may evolve in an inverse manner whereby it optimizes utilization of the substrate that the first strain no longer ferments. Mechanisms through which polymorphisms leading to fitness ‘trade-offs’ and niche diversification include cross-feeding (Helling et al., ), spatial variation (Rainey & Travisano, ), and temporal variation (Turner et al., ; Horner-Devine et al., ). Furthermore, the effect that polymorphisms have on the fitness of a population is governed by the environmental context. The inability to ferment a carbon source is not a fitness disadvantage if alternative substrates are sufficient to support growth during a period of time defined by additional growth parameters (e.g. temperature, pH, oxygen concentration, etc.). Thus, differential substrate usage and phenotypic diversification may evolve as a means of non-competitive co-existence among highly similar strains resulting in a microdiverse community.
However, it is important to realize that the niche occupation potential examined here may not be reflective of the true in situ activity. Wood compost is a heterogeneous environment, and the in situ activity of a strain at a given point in time will be greatly influenced by its immediate environment and the regulatory and biochemical processes of the strain itself. As such, although it is unclear to what extent these differences in niche occupation potential are realized in the natural environment, there is significant diversity between strains in their potential to fulfill various ecological niches.
Our results indicate that within the genus Thermoanaerobacter, significant genotypic and phenotypic microdiversity can exist at the sub-species level within a common environment. Although the specific mechanisms governing these differences are unclear, it is evident that varying evolutionary pressures have acted upon these co-existing strains in the form of blended phylogenies, potential niche specialization, and variations in their central metabolism. As such, this study highlights the importance of examining sub-species diversity in future ecological, evolutionary, and bioprospecting studies.
Acknowledgements
We thank Dr Pin-Ching Maness for supplying the hemicellulose hydrolysates. We also thank Dr Bruce Ford for his assistance with the NTSYS-pc software used in evaluation of the NOI. We finally thank Marc Ranson, Florian Labat, and Andrea Wilkinson for technical contributions to this manuscript. This work was supported by a National Sciences and Engineering Research Council (NSERC) Strategic Grant (STPGP 365076), by the Manitoba Rural Adaptation Council (MRAC), Advancing Canadian Agriculture and Agri-Food (ACAAF) program (#309009), Genome Canada, and by the Cellulosic Biofuels Network (Agriculture and Agri-Food Canada).
References
Author notes
Editor: Alfons Stams
