ABSTRACT
PCR and quantitative PCR (qPCR) primers targeting the 16S rRNA gene were used to detect and quantify members of the genus Fibrobacter in lake water, sediment and colonized cotton taken from two freshwater lakes. Phylogenetic analysis identified two groups of sequences; those clustered with Fibrobacter succinogenes, the type species, and a defined cluster of clones loosely associated with several Fibrobacter sequences observed previously in clone libraries from freshwater environments. 16S rRNA gene sequences recovered in the same way from soil samples and ovine feces in the surrounding land were all F. succinogenes and did not include any from this group of the “freshwater” Fibrobacteres. In all cases, nested PCR was required to detect Fibrobacter 16S rRNA genes, and qPCR analysis of reverse transcribed bacterial community RNA confirmed their very low relative abundance on colonized cotton baits in the water column (at 0, 3, 7, 11, and 13 m) and on the sediment surface (<0.02% of total bacterial rRNA). However, in Esthwaite Water sediment itself, the relative abundance of fibrobacters was 2 orders of magnitude higher (ca. 1% of total bacterial rRNA). The presence of fibrobacters, including the cellulolytic rumen species F. succinogenes, on colonized cellulose samples and in lake sediment suggests that these organisms may contribute to the primary degradation of plant and algal biomass in freshwater lake ecosystems.
Cellulose is the most abundant organic polymer on Earth, and its decomposition under anoxic conditions in freshwater, marine, and estuarine sediments is an essential process in the global carbon cycle. The complete decomposition of lignocellulosic material under anaerobic conditions is mediated by a metabolically diverse microbial community comprised of several trophic groups of microorganisms (16). However, the composition of the anaerobic cellulolytic population in freshwater lakes is barely understood, and our knowledge of their identity is restricted to a small number of isolated strains belonging to the genus Clostridium (17, 25, 36).
Both molecular and culture-based studies have suggested that Clostridium spp. are likely to be the predominant cellulolytic microorganisms in anoxic environments, such as landfill sites (32) and freshwater sediments (17). Conversely, it is well established that cellulose hydrolysis in the herbivore gut is due largely to members of the bacterial genus Fibrobacter and the anaerobic fungal order Neocallimastigales (14, 19), and the distribution of both groups had been thought to be restricted entirely to the gut environment. Recently anaerobic fungi (Neocallimastigales) were detected in a cellulose-rich landfill site using molecular biological methods (20). The presence of Fibrobacter spp. was inferred from the analysis of DNA extracted from a number of municipal waste sites using genus-specific primers, with quantitative reverse transcriptase PCR (qRT-PCR) data demonstrating that they can comprise a significant proportion of the landfill microbial community (23). It is possible that Fibrobacter spp. occupy other terrestrial and aquatic habitats where cellulose is degraded under anoxic conditions, such as freshwater and marine sediments, anaerobic sludge, and waterlogged soils. Furthermore, since all cultured strains assigned to the genus Fibrobacter thus far are capable of cellulose hydrolysis (1, 24), their detection in these environments suggests a role for Fibrobacter spp. in cellulose hydrolysis beyond the herbivore gut.
The genus Fibrobacter contains two cultivated species, F. succinogenes and F. intestinalis, and these organisms have long been thought to exclusively occupy gut environments (24). Direct molecular ecological approaches have provided evidence for these and new lineages within Fibrobacter, but the taxonomic limits of the genus have been questioned (1). Recently Fibrobacter-related sequences were described for termite gut environments, and these sequences have been assigned to a subphylum of the Fibrobacteres (13). There are currently two subphyla in the Fibrobacteres lineage, and clone sequences detected in termite gut environments have been assigned to Fibrobacteres subphylum 2 (13), with representatives of the genus Fibrobacter classified as Fibrobacteres subphylum 1. The genus Fibrobacter, as currently defined, therefore encompasses only the two cultivated “gut” species F. succinogenes and F. intestinalis and those related sequences obtained from landfill sites (23).
A small number of sequences belonging to the Fibrobacteres phylum or described as Fibrobacter related have been reported in general bacterial 16S rRNA gene clone libraries generated from geographically and geochemically distinct freshwater lake samples (3, 10, 18, 22, 27). These data demonstrate that while novel Fibrobacter-related populations are present in freshwater lakes, their poor representation in clone libraries suggests a negligible contribution to the freshwater microbial community. However, Fibrobacter sequences are frequently underrepresented in general 16S rRNA gene and ribosomal intergenic spacer clone libraries generated from rumen samples, an environment where they are known to predominate (5, 15, 29-31, 34). Therefore, the importance of Fibrobacter spp. in aquatic ecosystems has yet to be properly addressed and may have been underestimated.
Here, genus-specific PCR primers targeting the 16S rRNA gene of Fibrobacter spp. were applied to DNA extracted from colonized cotton samples, membrane-filtered lake water, and sediment samples from two productive lakes in the English Lake District. The relative abundance of Fibrobacter spp. within the samples was determined by qRT-PCR, and this is important because simple PCR-based detection of fibrobacters in the freshwater lake environment could merely be indicative of runoff from land grazed by ruminant animals.
Freshwater lake samples were obtained from Esthwaite Water and Priest Pot, situated in the English Lake District, United Kingdom. Esthwaite Water is 2.4 km in length and ca. 16 m deep. It is regarded as one of the most productive lakes in the area and is eutrophic. Priest Pot is a small, hypereutrophic lake with a depth of around 3 m and 1 ha in area. Three types of sample were obtained from Esthwaite Water in November 2007 and June 2008. Dewaxed cotton string (35) in nylon mesh bags was suspended on a nylon rope at various depths within the water column (surface, 3 m, 7 m, 11 m, 13 m, and sediment of Esthwaite Water; surface, 0.5 m, 1 m, 2 m, and sediment of Priest Pot) (6). The cotton samples were retrieved after 1 month's residence within the lakes. A Freedinger apparatus was used to sample 1-liter volumes of water at depths equivalent to those of the cotton baits, and sediment samples were obtained using a Jenkin corer (26).
The oxygen concentration and temperature of the water column were measured using a YSI 85 dissolved-oxygen, conductivity, salinity, and temperature meter. Samples of waterlogged soil and ovine feces from the fields adjacent to the lakes were also obtained. Immediately after removal from the lake, colonized cotton and sediment samples were frozen on dry ice and stored at −80°C until use. Water samples were stored in an insulated bag for transport to the laboratory and immediately filtered through sterile Supor200 (0.2-μm pore diameter) membranes (PALL Life Sciences), which were subsequently stored at −80°C.
For DNA and RNA coextraction, an entire 0.2-μm (pore diameter) membrane filter (through which 1 liter of lake water was filtered) or 0.5 g of waterlogged soil, ovine feces, colonized cellulose, or lake sediment was subjected to the nucleic acid extraction method of Griffiths et al. (11). The method is based on the use of hexadecyltrimethylammonium bromide and phenol-chloroform-isoamyl alcohol extraction with mechanical bead beating. Extracted nucleic acid from each sample was resuspended in 50 μl nuclease-free water and divided into two 25-μl aliquots. One aliquot was treated with RNase A (Sigma) as described previously (11) for use as a community DNA sample. The second aliquot was DNase treated using a Turbo DNA-free kit (Ambion) for use as a community RNA sample. Extracted RNA (0.4 μg) was reverse transcribed using BioScript RNase H Low (Bioline), following the manufacturer's protocol. These processed DNA and cDNA samples were stored at −80°C until required.
Direct and nested PCRs were performed on community DNA for each environmental sample (50 ng DNA per PCR assay) using Fib1 and Fib2A as the primer set (Table 1) and SuperTaq DNA polymerase as previously described (23). Fibrobacter spp. could only be detected in lake samples via nested PCR (Fig. 1) in which an initial round of amplification with the general bacterial primer set, pA and pH′, was followed by a second round of amplification with the Fibrobacter genus-specific primer set (Fib1 and Fib2A).
Data on the nested PCR amplification of community DNA extracted from colonized cotton, membrane-filtered lake water, and sediment samples using the Fibrobacter genus-specific 16S rRNA gene primer set (Fib1F and Fib2AR). The 17-m-depth sample represents the lake sediment, and 0- to 13-m samples represent the water column. Percentage oxygen saturation and temperature readings of the water column and sediment of Esthwaite Water in the English Lake District were also measured. +, nested PCR amplification product obtained; −, no PCR amplification observed; Sed, lake sediment sample.
PCR primer sets targeting the 16S rRNA gene
Fibrobacters were detected in water, colonized cotton, and sediment samples in the lower depths of Esthwaite Water and Priest Pot but were more readily detected in colonized cotton samples than in lake water, suggesting that they are members of the biofilm community on particulate organic matter in the water column (Fig. 1). Priest Pot has a depth of 3 m, and fibrobacters were detected only at depths of 1 m and 2 m and in sediment (3 m) colonized cotton samples but not from cotton baits moored at the surface and at a 0.5-m depth. Fibrobacter spp. were detected in Esthwaite Water only between a depth of 7 m and the sediment, where the percentage of oxygen saturation varied from 67% at 7 m to 3% in the sediment (Fig. 1). Their restriction to the more anoxic regions of the water column is consistent with the obligately anaerobic physiology of all known cultured members of this genus. Amplification products were cloned into Escherichia coli JM109 cells, and a total of 46 clones were sequenced by Macrogen Inc. (Seoul, South Korea) in both orientations to obtain coverage of the entire clone sequence. Analysis by both the RDP (4) and Pintail (2) chimera detection software packages demonstrated that none of the sequences were chimeric.
Phylogenetic analysis of Fibrobacter amplification products obtained using the Fib1 and Fib2A primer set demonstrated the presence of both F. succinogenes-like and novel Fibrobacter sequences in lake samples (Fig. 2). Of 19 clones obtained from colonized cotton samples from both Esthwaite Water and Priest Pot, only 3 clones were affiliated with the cultivated rumen species F. succinogenes, while the remaining 16 formed a distinct lineage with bootstrap support of >95% (Fig. 2). Six clone sequences from this cluster (PP-C-2.0-1 to PP-C-2.0-6) were obtained from reverse-transcribed community RNA extracted from colonized cotton at a depth of 2 m in the Priest Pot water column, demonstrating that these organisms were likely to be metabolically active on the colonized cellulosic material. These clones were related to but distinct from Fibrobacteres sequences obtained from geographically distinct freshwater environments in other studies (3, 27) (Fig. 2). The branching pattern within this freshwater Fibrobacter lineage was supported by high bootstrap values (>95%), suggesting that although this cluster represents a related group of organisms, there are also new taxa within the lineage (Fig. 2).
Maximum-likelihood tree of 16S rRNA gene sequences amplified from lake water, sediment, and colonized cotton samples using the Fibrobacter genus-specific Fib1 and Fib2A PCR primer set. Contiguous sequences were aligned with their three nearest neighbor sequences from both cultured and uncultured isolates using the Greengenes NAST aligner (7, 8), in addition to a number of representatives from other bacterial phyla. The alignment was imported into the Arb software program (21) and manually optimized. A final alignment corresponding to E. coli 16S rRNA gene positions 153 to 1017 was exported from Arb, and a maximum-likelihood tree was constructed by PhyML online (12) using the HKY85 substitution model and the Shimodaira-Hasegawa (SH)-like aLRT branch support method. The tree was imported back into Arb, where non-Fibrobacteres sequences were omitted from the tree to better demonstrate the phylogeny of the Fibrobacteres phylum. Fibrobacteres subphylum 2 contains sequences drawn from previous studies of the termite gut (13, 33). Filled circles indicate those nodes at which a bootstrap value of >95% was observed, and unfilled circles denote nodes with bootstrap values between 75 and 95%. The accession number of each sequence is displayed in parentheses alongside the sequence. Numbers shown on collapsed branches indicate the number of sequences within the branch. Sequences shown in bold were amplified using the Fib1 and Fib2A primer set applied in this study. The scale bar represents 0.1 base substitution per nucleotide.
16S rRNA gene amplification products from ovine feces and soils from the adjacent fields in which sheep were grazing were included to provide evidence that the detection of fibrobacters was not due simply to runoff and/or fecal contamination from the surrounding land. All clone sequences obtained from ovine feces were restricted to a distinct lineage that contained F. succinogenes strain MC1, an isolate from the ovine rumen (1) (Fig. 2).
Nine of the 10 sequences obtained from soil samples clustered with the type strain F. succinogenes; the remaining clone was clustered with a group of sequences previously detected in landfill (Fig. 2).
Phylogenetic analysis of cloned Fibrobacter DNA amplification products from freshwater lakes has demonstrated that in addition to F. succinogenes-like sequences, novel Fibrobacter spp. are present. This confirms that fibrobacters truly inhabit lakes and are not present due to fecal runoff. Sequence analysis of Fibrobacter-specific amplification products from a range of samples has demonstrated a surprising diversity of Fibrobacter spp. The assignment of novel lake Fibrobacter clones to both new and existing lineages supports our suggestion that fibrobacters are widely distributed in terrestrial and aquatic ecosystems and are not restricted to the gut environment.
Reverse-transcribed community RNA, extracted from samples of six colonized cotton baits and one lake sediment core obtained in June 2008, was subjected to quantitative PCR in triplicate using the 1369F/Prok1492R (general bacterial) and FibroQ153F/FibroQ258R (Fibrobacter genus-specific) primer sets (Table 1), as described previously (23). Insufficient RNA could be extracted from lake water samples for qRT-PCR analysis. The relative abundance of Fibrobacter spp. in comparison to total bacterial RNA varied from 0.005 to 0.02% on colonized cotton samples throughout the water column of Esthwaite Water (at 0, 3, 7, 11, and 13 m) but ca. 1% in the sediment. The relative abundance of Fibrobacter spp. on colonized cotton samples in the water column increased with depth, but the absolute values for relative abundance were always very low (<0.02%). The Fibrobacter abundance value for sediment is comparable to quantitative-PCR-derived data for landfill sites in a previous study (23), suggesting that fibrobacters are members of the indigenous lake microbial community, particularly in the more anoxic regions of the lake ecosystem.
The failure to amplify 16S rRNA genes from fibrobacters other than F. succinogenes in ovine feces samples here, and previously in bovine rumen samples (23), suggests that this group exhibits restricted diversity in the herbivore gut and is perhaps strongly niche adapted compared to environments such as freshwater lakes and landfill sites. Consequently, fibrobacters in environments beyond the rumen may be a source of novel cellulases and hemicellulases, which are important in many industrial applications, for example, in second-generation biofuel production. Molecular biological approaches for the identification of novel cellulases have already revealed the presence of such enzymes in the termite gut (33), and metagenomic analyses of lake sediment and landfill samples could be similarly fruitful.
Nucleotide sequence accession numbers.
The sequence data obtained in this work have been submitted to the GenBank database under accession numbers FJ711708 to FJ711753.
ACKNOWLEDGMENTS
This research was funded by the Natural Environment Research Council (NERC), United Kingdom.
We thank Roger Pickup (Centre for Ecology and Hydrology, Lancaster, United Kingdom), J. Paul Loughnane and David J. Rooks (University of Liverpool, United Kingdom), and the Freshwater Biological Association (Ambleside, United Kingdom) for assistance with sampling and access to the sampling sites. We are also grateful to Ali Waheed and Harry Tomlinson for assistance with the cloning of Fibrobacter amplification products.
FOOTNOTES
- Received 25 March 2009.
- Accepted 29 May 2009.
- Copyright © 2009 American Society for Microbiology
