Pagination not final (cite DOI) / Pagination provisoire (citer le DOI) 1

ARTICLE Recovering glycoside hydrolase genes from active tundra cellulolytic bacteria

Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 07/08/14 For personal use only.

Lee J. Pinnell, Eric Dunford, Patrick Ronan, Martina Hausner, and Josh D. Neufeld

Abstract: Bacteria responsible for cellulose hydrolysis in situ are poorly understood, largely because of the relatively recent development of cultivation-independent methods for their detection and characterization. This study combined DNA stableisotope probing (DNA-SIP) and metagenomics for identifying active bacterial communities that assimilated carbon from glucose and cellulose in Arctic tundra microcosms. Following DNA-SIP, bacterial fingerprint analysis of gradient fractions confirmed isotopic enrichment. Sequenced fingerprint bands and clone library analysis of 16S rRNA genes identified active bacterial taxa associated with cellulose-associated labelled DNA, including Bacteroidetes (Sphingobacteriales), Betaproteobacteria (Burkholderiales), Alphaproteobacteria (Caulobacteraceae), and Chloroflexi (Anaerolineaceae). We also compared glycoside hydrolase metagenomic profiles from bulk soil and heavy DNA recovered from DNA-SIP incubations. Active populations consuming [13C]glucose and [13C]cellulose were distinct, based on ordinations of light and heavy DNA. Metagenomic analysis demonstrated a ⬃3-fold increase in the relative abundance of glycoside hydrolases in DNA-SIP libraries over bulk-soil libraries. The data also indicate that multiple displacement amplification introduced bias into the resulting metagenomic analysis. This research identified DNA-SIP incubation conditions for glucose and cellulose that were suitable for Arctic tundra soil and confirmed that DNA-SIP enrichment can increase target gene frequencies in metagenomic libraries. Key words: stable-isotope probing, cellulose, tundra, Arctic, soil, metagenomics. Résumé : Les bactéries responsables de l’hydrolyse de la cellulose in situ sont mal connues, principalement en raison du développement relativement récent de méthodes de détection et de caractérisation indépendantes de la culture. La présente étude a jumelé le traçage de l’ADN par isotope stable (DNA-SIP) et la métagénomique aux fins d’identification des communautés bactériennes actives qui assimileraient du carbone a` partir de glucose et de cellulose dans des microcosmes de la toundra arctique. À la suite du DNA-SIP, on a analysé les empreintes bactériennes de fractions de gradients et confirmé l’enrichissement isotopique. Le séquençage de bandes d’empreintes et l’analyse de banques clonales de gènes d’ARNr 16S ont permis d’identifier les taxons bactériens actifs qui étaient liés aux ADN indirectement marqués par la cellulose, soit Bacteroidetes (Sphingobacteriales), Betaproteobacteria (Burkholderiales), Alphaproteobacteria (Caulobacteraceae) et Chloroflexi (Anaerolineaceae). Nous avons également comparé le profil métagénomique de la glycoside hyrolase obtenu d’ADN de sol en vrac avec celui de l’ADN lourd obtenu d’incubations de DNA-SIP. Les populations actives consommant du glucose et de la cellulose marqués au 13C étaient distinctes, selon l’ordination des ADN légers et lourds. L’analyse de métagénomique a révélé un triplement de l’abondance relative de glycoside hydrolases dans les banques issues du DNA-SIP par rapport aux banques issues de sol en vrac. Les données indiquent également que l’amplification a` déplacement multiple a introduit un biais dans l’analyse de métagénomique résultante. La présente recherche a su établir des conditions propices d’incubation avec du glucose et de la cellulose pour les besoins du DNA-SIP appliqué au sol de la toundra arctique et a confirmé que l’enrichissement par DNA-SIP pouvait hausser la fréquence de gènes ciblés dans des banques destinées a` la métagénomique. [Traduit par la Rédaction] Mots-clés : traçage par isotope stable, cellulose, toundra, arctique, sol, métagénomique.

Introduction Soil contains 1500 gigatons of carbon, representing one of the largest reservoirs of carbon on Earth (Zimov et al. 2006). Within this global carbon pool, northern ecosystems possess a disproportionately large amount of carbon. The tundra represents only 6.8% of the Earth’s soil but contains 13.7% of the planet’s soil carbon pool (Post et al. 1982). The overwhelming majority of this carbon is stored within permafrost (Mack et al. 2004) and is unavailable for microbial decomposition. However, as the climate warms, this carbon sink is becoming a net carbon source (Houghton et al. 1998; Melillo et al. 2002; Schuur et al. 2009). Cellulose is the most abundant organic compound on Earth (Lynd et al. 2002; O’Sullivan et al. 2007)

and is a substantial component of this carbon reservoir. Characterizing the microorganisms responsible for decomposing cellulose in tundra ecosystems is an integral step toward better understanding a global biogeochemical response to a changing climate. Because only a small proportion of microorganisms in soil are readily cultivable, culture-independent approaches are important for microbial community characterization. Two of the most promising culture-independent approaches for linking microbial activity with function are DNA stable-isotope probing (DNA-SIP) and metagenomics. Although each of these methodologies has been used independently, the combination of the 2 has enormous and largely unexplored potential with respect to the range of substrates used and environmental samples analyzed.

Received 17 March 2014. Revision received 6 June 2014. Accepted 10 June 2014. L.J. Pinnell, E. Dunford, and J.D. Neufeld. Department of Biology, University of Waterloo, 200 University Avenue West, Waterloo, ON N2L 3G1, Canada. P. Ronan and M. Hausner. Department of Chemistry and Biology, Ryerson University, Toronto, ON M5B 2K3, Canada. Corresponding author: Josh D. Neufeld (e-mail: [email protected]). Can. J. Microbiol. 60: 1–8 (2014) dx.doi.org/10.1139/cjm-2014-0193

Published at www.nrcresearchpress.com/cjm on 11 June 2014.

Pagination not final (cite DOI) / Pagination provisoire (citer le DOI)

Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 07/08/14 For personal use only.

2

A major issue associated with traditional screening of metagenomic libraries is that a very large number of clones must be screened to identify target genes because of the low relative abundance of target clones containing the desired genes within a complex mixture of microbial community DNA (Schwarz et al. 2006; Pinnell et al. 2011). DNA-SIP is a powerful tool for linking the taxonomic identity with metabolic function of cellulolytic microorganisms by the incorporation of isotopically labelled substrate into nucleic acids, which has been used successfully to increase the occurrence of target genes in metagenomic libraries (Dumont et al. 2006; Schwarz et al. 2006; Chen et al. 2008; Kalyuzhnaya et al. 2008; Neufeld et al. 2008; Sul et al. 2009). Prior to this work, no study has combined DNA-SIP and metagenomics for accessing the metagenomes of active cellulolytic consortia in Arctic tundra soil. The goals of this study were to identify a protocol for conducting DNA-SIP on an Arctic tundra soil sample, to test the enrichment of glycoside hydrolases in metagenomic libraries derived from microorganisms that assimilated cellulose carbon, and to assess the bias of multiple displacement amplification (MDA) for generating metagenomic libraries from recovered [13C]DNA. We used [13C]cellulose and [13C]glucose as substrates for tundra incubations, followed by DNA-SIP, MDA, and next-generation sequencing. We identified specific bacteria that assimilated carbon from cellulose and glucose, which may represent cold-adapted soil community members active in cellulolysis. Importantly, this research provides foundational support for future efforts involving additional soils, substrates, and the application of functional metagenomics for glycoside hydrolase recovery.

Methods Samples Arctic tundra soil was collected from Resolute Bay, Nunavut Territory, Canada, on 3 September 2009 (68°46=04.6092==N, 109°04=17.972==W). Organic horizon surface soil (top 10 cm) was collected from a sample area vegetated by cottongrass (Eriophorum vaginatum L.) hummocks with willow (Salix arctica Pall.) shrubs, cloudberry (Rubus chamaemorus L.), and flowers such as the Arctic poppy (Papaver radicatum Rydb.). The sample was placed in a plastic container and stored at 4 °C prior to being transported on ice to the University of Waterloo. The sample was sieved (4.75 mm) and stored at 4 °C. Soil pH, total carbon content, total nitrate content, and soil texture were analyzed by the University of Guelph Laboratory Services (Ontario, Canada) for nitrate (2.07 mg·kg−1), organic carbon (43.6% d.w.), pH (6.7), and texture (16.6%:32.6%:50.8% for sand–silt– clay). Soil bulk density was 0.714 g·cm−3 (mass of soil per volume). Gravimetric moisture was approximately 82% of maximum water holding capacity, based on calculated water-filled pore space. This moisture level was considered sufficient for incubation without further adjustment. A cellulolytic enrichment was prepared by incubating household compost through a series of sequential enrichments at 60 °C in RM medium (Ronan et al. 2013). This enrichment was analyzed to ensure that we could detect glycoside hydrolases and to characterize its community and diversity in comparison with our SIPenriched DNA. Production of [13C]cellulose For use in DNA-SIP incubations, cellulose was generated using Gluconacetobacter xylinus KCCM 10100 grown on native glucose (BioBasic) or [13C6]glucose (99 atom % 13C; Sigma-Aldrich). The G. xylinus culture was precultured in liquid media containing 2% glucose (m/v). Aliquots were then cultured in liquid media according to a modified protocol (Schramm and Hestrin 1954). The medium (20 g·L−1 D-glucose, 5 g·L−1 bactopeptone, 5 g·L−1 yeast extract, 2 g·L−1 dipotassium phosphate) was adjusted to a pH of 7.0 using hydrogen chloride solution. Bacteria were cultured for 30 days at

Can. J. Microbiol. Vol. 60, 2014

30 °C under static conditions. Bacterial cellulose formed as a thick opaque pellicle at the surface of the medium. A modified version of a cellulose purification protocol (Schramm and Hestrin 1954) was used to recover pure cellulose pellicles. Pellicles were removed from media and washed thoroughly using distilled and deionized water (ddH2O). Washed pellicles were placed in a 1 L flask containing 1% sodium hydroxide and boiled for 2 h. Pellicles were then soaked in a 5% acetic acid solution for 1 h at room temperature, then washed with ddH2O. Purified cellulose pellicles were frozen and lyophilized to remove moisture. Lyophilized pellicles were ground with a mortar and pestle in liquid nitrogen. The resulting cellulose particles were lyophilized again to remove condensation introduced during grinding. The final cellulose product consisted primarily of amorphous cellulose particles (Koizumi et al. 2008). Bacterial amorphous cellulose is more bioavailable than crystalline cellulose (Hall et al. 2010), partly because carbohydratebinding modules are not required for hydrolysis (Wilson 2011), and will likely reflect only a subset of plant-derived cell-wall polymers commonly degraded by soil microorganisms. DNA-SIP incubations The DNA-SIP incubations were prepared in 100 mL crimp-top vials using the sieved and refrigerated tundra soil sample on 9 October 2009. A total of 6 incubation vials each received 10 g of soil. Two microcosms were used as incubation controls, and therefore, no substrate was added to these vials. Two other microcosms were used as DNA-SIP controls, each receiving 200 mg of [12C]glucose or [12C]cellulose. The other 2 vials received 200 mg of either [13C]glucose or [13C]cellulose. Microcosms were incubated under oxic conditions for a period of 54 days at a temperature of 15 °C. This temperature was chosen because it reduced incubation times expected for lower temperatures, and because 15 °C is within the optimum growth temperatures for psychrophilic organisms (Morita 1975). Samples of the original tundra soil were kept for molecular analysis. Soil was retrieved from the microcosms on days 14, 28, and 54. On days 14 and 28, 2 g of soil was retrieved from each sample vial. On day 54, the remaining mass (⬃6 g) was recovered from the microcosms. All soil samples were preserved in a −80 °C freezer for longterm storage. Two DNA extraction methods were used, depending on the subsequent application (i.e., rapid extraction with bead beating for initial profiling or gentle lysis prior to MDA and metagenomic library preparation). For initial community characterization, DNA extraction was performed according to a modified protocol (Griffiths et al. 2000). In brief, cells were lysed using a FastPrep instrument (MP Biomedicals) in a lysis buffer followed by phenol–chloroform purification. Precipitation used 2 volumes of a polyethylene glycol solution (30% PEG 6000, 1.6 mol·L−1 NaCl), and ethanol-washed DNA was suspended in sterile ddH2O. For metagenomic analyses, DNA was extracted using a modified procedure (Zhou et al. 1996). This extraction involved a high-salt buffer and SDS detergent to lyse cells, and precipitation involved ammonium acetate (final concentration of 2.5 mol·L−1) and isopropanol. After the initial extraction and purification, DNA was further purified using SCODA (synchronous coefficient of drag alteration; Boreal Genomics) according to a previously published protocol (Engel et al. 2012). The metagenomic DNA purification procedure was also applied to bulk-soil samples and a cellulose-degrading enrichment culture sample (Ronan et al. 2013). Following DNA extraction and purification, density gradient ultracentrifugation was used to separate “light” and “heavy” DNA (Neufeld et al. 2007b). Briefly, extracted DNA was added to a cesium chloride (CsCl) gradient and ultracentrifuged for 40 h at 44 100 r·min−1 (⬃177 000gav) using a Vti 65.2 rotor (Beckman Coulter, California, USA). Based on density, the gradient was subsequently separated into 12 fractions, retrieving both heavy [13C]DNA and light unlabelled DNA. For analysis, fractions 5–7 were considered heavy (⬃1.725 g·mL−1); fractions 10–11 were considered light (⬃1.700 g·mL−1). These pooled heavy fractions were Published by NRC Research Press

Pagination not final (cite DOI) / Pagination provisoire (citer le DOI) Pinnell et al.

Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 07/08/14 For personal use only.

used directly for metagenomic libraries or were subjected to MDA using the illustra GenomiPhi V2 DNA Amplification kit (GE Healthcare) according to the manufacturer’s protocol to increase the amount of DNA. General bacterial 16S rRNA gene profiles were generated by DGGE (denaturing gradient gel electrophoresis), according to a previously published protocol (Green et al. 2010), with primers 341f-GC and 518r, 30%–70% denaturing gradients, and a 10% polyacrylamide gel. Gels were run for 14 h at 85 V in a DGGEK-2001 system (C.B.S. Scientific Inc.), stained with SYBR Green I (Invitrogen), and visualized using the Pharos FX Plus Molecular Imager (Bio-Rad). Community characterization Bands were excised from DGGE gels and re-amplified using the same 341f and 518r primers as described previously, this time without the associated GC clamp. To further characterize light and heavy DNA in addition to the original soil community, PCR products were sequenced directly using the 518r primer and BigDye terminator version 3.1 (Applied Biosystems, California, USA). Products were run on an ABI 3730xl DNA Analyzer (Applied Biosystems) by the Toronto Centre for Applied Genomics (Hospital for Sick Children, Toronto, Ontario, Canada). Retrieved sequences were matched to their closest phylogenetic relatives using BLAST (Basic Local Alignment Search Tool; Benson et al. 2000). Clone libraries were generated for the original soil sample; the glucose-enriched heavy DNA at days 14, 28, and 54; the celluloseenriched heavy DNA at days 28 and 54; and the cellulose-enriched light DNA at 54 days. Libraries were generated from products of PCR conducted with the primers 27f (AGAGTTTGATCMTGGCTCAG; Lane 1991) and 1492r (ACCTTGTTACGACTT; Lane 1991) and amplification conditions as follows: initial denaturation at 95 °C for 5 min followed by 30 cycles of 95 °C for 1 min, 55 °C for 1 min, and 72 °C for 1 min. Final extension occurred at 72 °C for 7 min. PCR products were cloned into TOPO-TA cloning vectors (Invitrogen), and a total of 96 clones from the original tundra soil library were screened via colony PCR in 50 ␮L amplification volumes using the –21M13 (TGTAAAACGACGGCCAGT) and M13 reverse (CAGGAAACAGCTATGACC) primers as described previously (Neufeld et al. 2004). A total of 48 clones were collected from each heavy and light DNA-SIP fraction. Sanger sequencing of cloned inserts was conducted in one direction with the 27f primer at the Beckman Coulter Genomics sequencing facility in Danvers, Massachusetts, USA. Sequence identities were assigned to all sequences using the RDP Classifier (Wang et al. 2007). Sequences obtained from the DGGE bands and clone libraries were uploaded to GenBank with accession numbers as follows: native tundra soil (KF275192–KF275284), 54-day “incubation control” SIP light DNA (KF275516–KF275560), 14-day glucose SIP heavy DNA (KF275285–KF275329), 28-day glucose SIP heavy DNA (KF275330–KF275376), 54-day glucose SIP heavy DNA (KF275377–KF275423), 28-day cellulose SIP heavy DNA (KF275424– KF275470), 54-day cellulose SIP heavy DNA (KF275471–KF275516). Sequences were also analyzed using the QIIME software package, managed by AXIOME (Lynch et al. 2013), to evaluate relative similarities between the phylogenetic distributions of heavy fractions. Retrieved sequences were clustered at 97% identity using cd-hit (clustering database at high identity with tolerance; Li and Godzik 2006) and aligned using the PyNAST (Python Nearest Alignment Space Termination) program (Caporaso et al. 2010a). Phylogenetic classification was determined using the RDP Classifier (Wang et al. 2007). Category-based clustering of phylogenetic identities of sequences obtained from the original soil sample, the heavy DNA-SIP fractions from the glucose enrichments (days 14, 28, 54), the heavy DNA-SIP fractions from the cellulose enrichments (days 28, 54), and the light DNA-SIP fractions from the cellulose enrichment (day 54) was performed using PCoA (princi-

1

3

pal coordinate analysis) based on the unweighted UniFrac metric (Lozupone and Knight 2005; Lozupone et al. 2010). Ordinations were generated using QIIME (Caporaso et al. 2010b) and rendered with the KiNG (Kinemage Next Generation) molecular modeling program (Chen et al. 2009). The MRPP (Multi-Response Permutation Procedures) data were generated by PC-ORD (MJM Software Design, Gleneden Beach, Oregon). Metagenomic analysis Illumina-based sequencing was done on replicate DNA samples from the heavy cellulose DNA-SIP fractions, original soil sample, and cellulose-degrading enrichment culture samples by the DNA Services Facility (University of Illinois at Chicago). Briefly, DNA samples were prepared for sequencing with the Nextera DNA Sample Prep kit (Epicentre) prior to 100-base paired-end sequencing using the HiSeq2000 (Illumina). Unassembled forward and reverse reads from each replicate were analyzed individually, serving as annotation replicates, and uploaded separately to MG-RAST (accession Nos. 4474939.3–4474949.3, 4474980.3–4474985.3, and 4475899.3). The number of reads for each library were as follows: bulk soil 1 (5 204 312), bulk soil 2 (7 400 852), bulk soil 1 MDA (15 218 443), bulk soil 2 MDA (14 358 008), SIP (6 295 723), SIP MDA 1 (15 212 696), and SIP MDA 2 (15 591 124). Libraries were processed by the MG-RAST quality control pipeline, omitting the “demultiplexing” and “model organism” screening steps. Libraries were annotated using SwissProt with a 54% minimum percentage identity cutoff instead of an E value cutoff, and a 30-base minimum alignment length cutoff. All reads annotated as glycoside hydrolases were manually assigned to respective CAZy families (www.cazy.org). A Bray–Curtis dissimilarity matrix was constructed for all samples using the number of gene sequences in each CAZy family normalized to the proportion of representatives per one million sequences in each library. From the Bray–Curtis distance matrices, unrooted trees were constructed using the neighbor-joining method (Saitou and Nei 1987) with the APE package in R (Paradis et al. 2004). Two trees were constructed, one for the matrix relating to all glycoside hydrolase families, and the other with selected cellulose-degrading glycoside hydrolase families (GH 5–12, 26, 44, 48).

Results Following glucose and cellulose DNA-SIP incubations, 12C-labelled and 13C-labelled gradient fractions were characterized by bacterial DGGE (Fig. 1). The results demonstrated that in all cases, the heavy fractions from 13C-labelled incubated samples were distinct from corresponding fractions of 12C-labelled incubated samples, which confirms that populations of active organisms were labelled under the SIP conditions employed in this study. Although labelling of phylotypes for the [13C]glucose incubations was unequivocal, some heavy-fraction variation in community profiles was visible for [12C]cellulose incubated samples. This may be attributed to the lower amount of labelling associated with all cellulose SIP incubations, relative to glucose incubations, enabling G+C content density separation to become more apparent (Neufeld et al. 2007a). Regardless, heavy fractions of [13C]cellulose SIP incubations were clearly distinct from heavy fractions of [12C]cellulose incubations, confirming enrichment of active phylotypes (Fig. 1). Selected light and heavy SIP fractions were chosen based on distinct DGGE fingerprints (Fig. 1) for 16S rRNA gene sequencing, and 374 Sanger sequences were generated across all samples (Table S11). The UniFrac-based analysis of these libraries demonstrated that sequence profiles of [13C]glucose and [13C]cellulose heavy fractions were distinct from the original soil community and distinct from each other (Fig. 2). Microbial communities from the control fraction (fraction 11 from day 54 [13C]cellulose incubation) and those from the

Supplementary data are available with the article through the journal Web site at http://nrcresearchpress.com/doi/suppl/10.1139/cjm-2014-0193. Published by NRC Research Press

Pagination not final (cite DOI) / Pagination provisoire (citer le DOI) 4

Can. J. Microbiol. Vol. 60, 2014

Fig. 2. Principal coordinate analysis ordination of 374 bacterial clone sequences retrieved from soil stable-isotope probing incubations using an unweighted UniFrac dissimilarity matrix. The original soil library is shown in blue. The cellulose-amended libraries are shown in orange; glucose-amended libraries are shown in green. The incubation control library was constructed from the light DNA of the 54-day cellulose incubation (red). This biplot also shows grey spheres that demonstrate the top 10 bacterial classes responsible for sample differentiation. The size of the spheres reflects the abundance of that group in the sequence data. The location reflects the importance of that group in differentiating samples. Labels for these lineages are given at the taxonomic level where a classification could be made by the RDP Classifier. Numbers in the labels correspond to operational taxonomic unit ID numbers listed in Table S11.

Resolute Bay soil (T=0) Principal coordinate 2 (18%)

Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 07/08/14 For personal use only.

Fig. 1. Temporal changes in bacterial community composition for the heavy (lanes 6 and 7) and light (lanes 10 and 11) fractions from [13C]glucose and [13C]cellulose microcosms and corresponding 12C-labelled controls. Fingerprints are generated based on 16S rRNA gene amplicons separated by denaturing gradient gel electrophoresis.

Gammaproteobacteria; 36 Bacteria; 52

Glucose (14 d) Firmicutes; Sporolactobacillus; 15

Cellulose light DNA control (54 d) Acidobacteria; Gp4; 11

Glucose (28 d)

Alphaproteobacteria; Caulobacter; 3

Cellulose (28 d) Firmicutes; Clostridium; 2

Cellulose (54 d) Betaproteobacteria; Burkholderiales; Ralstonia; 23

Firmicutes; Clostridium; 9

Glucose (54 d)

Bacteroidetes; Sphingobacteriales; 92 Betaproteobacteria; Burkholderiales; Comamonadaceae; 1

Principal coordinate 1 (49%) Published by NRC Research Press

Pagination not final (cite DOI) / Pagination provisoire (citer le DOI) Pinnell et al.

Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 07/08/14 For personal use only.

Fig. 3. A 10% denaturing gradient gel electrophoresis polyacrylamide gel with a 30%–70% denaturing gradient containing the bacterial 16S rRNA gene fingerprints of the 12 DNA stableisotope probing density fractions for the 2-month [13C]cellulose incubation (decreasing density from lanes 1 through 12), flanked by 2 ladder lanes labelled “L”. The numbered arrowheads indicate community DNA bands that were excised for sequencing. Numbers correspond to sequences shown in Table 1.

5

Table 1. Taxonomic classification of [13C]cellulose-associated denaturing gradient gel electrophoresis (DGGE) band sequences with associated RDP-II confidence scores. DGGE band

Length

Classification

1

135

2

131

3

148

4

144

Chloroflexi (59%), Anaerolineae (59%), Anaerolineales (59%), Anaerolineaceae (59%) Proteobacteria (95%), Alphaproteobacteria (95%), Caulobacterales (94%), Caulobacteraceae (94%) Bacteroidetes (95%), Sphingobacteria (90%), Sphingobacteriales (90%), Cytophagaceae (71%) Bacteroidetes (89%), Sphingobacteria (87%), Sphingobacteriales (87%), Cytophagaceae (60%)

Table 2. The number of enzymes annotated as glycoside hydrolases and as cellulose-degrading glycoside hydrolases per 1 million sequences following the MG-RAST quality control pipeline.

Sample

day 14 [13C]glucose incubation grouped together and were similar to the original tundra soil community. An MRPP analysis conducted on a Bray–Curtis distance matrix confirmed that cellulose heavy DNA, glucose heavy DNA, and a group consisting of both original soil and light DNA were all distinct from one another (p = 0.02), with high within-group homogeneity (A = 0.44). Few taxa dominated collected sequences from each clone library, which raised concerns about library coverage. As a result, sequencing of DGGE bands specifically associated with 13C-associated heavy DNA fractions was essential for confirming specific community composition changes represented in the clone libraries. Four prominent DGGE bands were excised and PCR-amplified (Fig. 3); sequences affiliated with the Bacteroidetes (Sphingobacteriales) and Alphaproteobacteria (Caulobacteraceae) were identified using BLAST (Table 1) and these same taxa associated with clone data from [13C]cellulose heavy DNA fractions (Fig. 2). Although Chloroflexi (Table 1) was not displayed in the ordination taxonomic overlay (Fig. 2), nor Betaproteobacteria (Fig. 2) among the sequenced DGGE bands (Table 1), many other DGGE bands were not sequenced in the DGGE fingerprint (Fig. 3), which might account for the discrepancy. Indeed, the operational taxonomic unit (OTU) affiliated with Chloroflexi was identified in both [13C]cellulose SIP clone libraries (Table S11). Together, clone library analysis and DGGE band sequencing identified dominant celluloseassociated OTUs as Bacteroidetes (Sphingobacteriales), Betaproteobacteria (Burkholderiales), Alphaproteobacteria (Caulobacteraceae), and Chloroflexi (Anaerolineaceae). Metagenomic sequencing was conducted on 8 samples associated with this study (Table 2). We sequenced DNA from the original soil sample (in duplicate), directly from the pooled SIP heavy DNA (fractions 5–7; Fig. 3), and from a cellulose-degrading enrichment (not duplicated). We also tested amplification with MDA prior to sequencing for the original soil sample (in duplicate) and for the SIP heavy DNA (in duplicate). Prior to sequencing, we used DGGE to monitor potential representational bias introduced by amplification (Fig. 4). The results suggest that although soil samples amplified with MDA resembled those of the original soil patterns, bias was introduced by amplification of SIP DNA template based on differences in DGGE patterns.

Resolute Bay soil 1 (forward/reverse) Resolute Bay soil 2 (forward/reverse) Resolute Bay soil 1 MDA (forward/reverse) Resolute Bay soil 2 MDA (forward/reverse) DNA-SIP (forward/reverse) DNA-SIP MDA 1 (forward/reverse) DNA-SIP MDA 2 (forward/reverse) Cellulose-degrading enrichment (forward/reverse)

Glycoside hydrolases per million sequences

Cellulases* per million sequences

359/353

54/57

326/308

53/43

263/231

38/36

238/234

36/36

536/511

157/147

294/293

61/55

326/243

57/46

2649/2669

1461/1458

Note: Data are presented as values for replicate 1/replicate 2 or as forward reads/reverse reads. MDA, multiple displacement amplification; SIP, stableisotope probing. *Cellulases were determined as enzymes belonging to GH 5–12, 26, 44, 45, and 48 CAZy families, which were previously referred to as cellulase families A–H, I, J, K, and L, respectively.

For each of the 8 samples, shotgun metagenomic libraries were generated using Illumina paired-end sequencing. Following the download of all enzymes annotated as glycoside hydrolases using SwissProt, each sequence was assigned to its corresponding CAZy family (Table S2). Using the CAZy database, enzyme representation in sequence data was compared across the approximately 130 glycoside hydrolase families (www.cazy.org; January 2012). The cellulose-degrading enrichment had the highest proportion of glycoside hydrolase annotations and the highest proportion of cellulases (Table 2). Within the cellulose-degrading enrichment library, approximately 55% of all glycoside hydrolases were cellulase enzymes. The DNA-SIP library contained the next highest proportion of cellulases, with approximately 30% of all glycoside hydrolases within the sample annotated as cellulases. A total of 157 and 147 annotated cellulases per 1 million sequences in our DNA-SIP library, in contrast to 54 and 57 for bulk soil A (Resolute Bay soil 1) and 53 and 43 for bulk soil B (Resolute Bay soil 2) (Table 2). Thus, there was a ⬃3-fold increase in the abundance of cellulases following enrichment using DNA-SIP. Of the 5 most abundant annotated glycoside hydrolase sequences in the DNA-SIP library, 2 were associated with cellulosedegrading glycoside hydrolase families. However, in the bulk-soil libraries, none of the 5 most abundant glycoside hydrolase gene sequences were associated with cellulases. Similarly, none of the Published by NRC Research Press

Pagination not final (cite DOI) / Pagination provisoire (citer le DOI) 6

Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 07/08/14 For personal use only.

Fig. 4. A 10% denaturing gradient gel electrophoresis polyacrylamide gel with a 30%–70% denaturing gradient containing the bacterial 16S rRNA gene fingerprints of all samples sent for sequencing. Samples are flanked by 2 ladder lanes labelled “L”. The DNA-SIP sample was recovered from fraction 7 of the ultracentrifuge tube following DNA-SIP. MDA, multiple displacement amplification; SIP, stable-isotope probing.

5 most abundant glycoside hydrolase enzymes sequences from the SIP-MDA libraries were associated with cellulase glycoside hydrolase families, suggesting that MDA introduced a bias into the resulting libraries. The total number of glycoside hydrolases per million sequences decreased from ⬃525 in the DNA-SIP library to ⬃290 in the SIP-MDA libraries (Table 2). This MDA-associated decrease in annotated glycoside hydrolase gene abundance was greater for cellulose-degrading glycoside hydrolases and they were 3 times less abundant following MDA than they were in the DNA-SIP library. The bias introduced by MDA on the bulk-soil libraries was much less pronounced, which was expected based on the patterns observed with DGGE fingerprinting. The Bray–Curtis dissimilarity metric was used to compare the compositional representation between sample libraries. The branch lengths between the cellulose-degrading enrichment library and the closest related library (DNA-SIP) demonstrated that the low-diversity and glycoside-hydrolase-rich cellulose-degrading enrichment library was distinct from all other samples (Fig. 5). The DNA-SIP library, with the next highest proportion of glycoside hydrolases, was most closely related to the cellulose-degrading enrichment library. The similarity dendrograms confirmed that MDA bias was greater for the SIP DNA samples than for the bulk-soil samples. Note that the DNA-SIP library was more similar to the bulk-soil samples than to the SIP-MDA libraries (Fig. 5), for both total glycoside hydrolases and cellulases. Representational bias was less in the bulk-soil libraries; these MDA libraries had considerably shorter branch lengths.

Discussion Using multiple DNA-SIP incubations and time points for an Arctic tundra soil, we demonstrate that DNA-SIP is an effective method to increase the screening efficiency of metagenomic analyses when tar-

Can. J. Microbiol. Vol. 60, 2014

geting cellulose- and glucose-degrading communities. These data demonstrate that performing a DNA-SIP enrichment before the generation of metagenomic libraries increased the abundance of glycoside hydrolase enzyme annotations associated with the corresponding sequence data sets. The inclusion of a DNA-SIP preenrichment step to increase the abundance of sequences from organisms containing cellulases will allow for more efficient metagenomic research in the future and is consistent with previous studies showing that DNA-SIP increases the representation of targeted genes in metagenomic libraries (reviewed in Pinnell et al. 2011). Prominent phyla we implicated in the assimilation of cellulose carbon included members of Bacteroidetes, Proteobacteria, and Chloroflexi. The Proteobacteria are known to contain a diverse range of cellulolytic microorganisms and previous studies identified cellulolytic phylotypes, including Burkholderia (Belova et al. 2006), Sorangium (Hou et al. 2006), Novosphingobium (Landy et al. 2008), Pseudomonas (Dumova and Kruglov 2009), and Cellvibrio (Pang et al. 2009). In particular, members of the Gammaproteobacteria were implicated as active cellulolytic organisms in soil environments (Bernard et al. 2007, 2009; Lee et al. 2011). Previous DNA-SIP studies identified that Chloroflexi were associated with cellulose carbon assimilation (Schellenberger et al. 2010). At the genus level, Bacteroidetes, Burkholderia, and Ralstonia were abundant in the cellulose-amended microcosms, whereas Sporolactobacillus and Clostridium were abundant in glucose-amended microcosms (Fig. 2, Table S11). An additional excised DGGE sequence yielded a lowquality read (data not shown) that was most closely related to a sequence from an uncultured Actinobacteria species identified in another soil cellulose DNA-SIP experiment performed by el Zahar Haichar et al. (2007). Note that the current study cannot confirm whether labelled bacteria were directly involved in both cellulolysis and assimilation or simply benefitted from liberated cellulose carbon through food-web dynamics via “cross-feeding” (Neufeld et al. 2007a). We confirmed that MDA introduces bias during amplification, particularly with lower amounts of template DNA. Glycoside hydrolase gene sequences were less abundant following MDA (Table 2), suggesting amplification bias. Previous research has shown that the MDA protocol used in this research (illustra GenomiPhi kit) exhibited bias against high G+C content template DNA (Yilmaz et al. 2010). This was supported by our observation that the proportion of Actinobacteria, a bacterial phylum with high G+C content (Ventura et al. 2007), decreased from ⬃17% in the DNA-SIP library to ⬃4% in the SIP-MDA libraries, which is an important note for researchers using MDA to study soil communities. The mean G+C contents of pre- and post-MDA libraries further suggested bias against the high G+C content template DNA. To minimize or eliminate this bias, we recommend introducing maximal amounts of recovered DNA from DNA-SIP incubations to density gradient prior to ultracentrifugation. If this is not possible (e.g., minimal labelling results in low yields of labelled DNA), we recommend an initial fingerprint analysis as a reliable initial indicator of introduced bias (Fig. 4), which was also documented previously (Chen et al. 2008; Neufeld et al. 2008). Although DGGE is far less powerful than next-generation sequencing, we demonstrate it as a rapid and affordable technique for determining if MDA-amplified samples are suitable for subsequent large-insert metagenomic library analyses. The work presented here demonstrates a coupled DNA-SIP and metagenomics experiment targeting glycoside hydrolases and cellulase enzymes directly from active Arctic microorganisms. Tundra soil communities are at present poorly characterized and the data collected during this experiment provide beneficial methodological insight into ways of increasing information about active tundra cellulolytic communities. Identifying suitable SIP conditions for Arctic tundra samples will now help with subsequent experimental design involving multiple samples and replicated Published by NRC Research Press

Pagination not final (cite DOI) / Pagination provisoire (citer le DOI) Pinnell et al.

7

Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 07/08/14 For personal use only.

Fig. 5. Unrooted neighbor-joining trees constructed using Bray–Curtis similarity coefficients in R. (A) All glycoside hydrolases (GH) and (B) GH belonging to cellulose-degrading families (GH 5–12, 26, 44, 45, 48). R1, represents the forward read; R2, represents the reverse read; RB, represents Resolute Bay soil; MDA, multiple displacement amplification; SIP, stable-isotope probing.

experimental design; this will greatly increase insight into Arctic tundra biogeochemical cycling and cold-adapted microorganisms. Indeed, our ongoing research has incorporated the results of this study to analyze additional soils and substrates, combining isotope enrichment with functional metagenomics for glycoside hydrolase recovery from enriched DNA in surrogate bacterial hosts.

Acknowledgements Daniela Loock, Jennifer Joubert, and Kenneth Reimer of the Royal Military College of Canada are thanked for providing the tundra soil sample used in this study. Heidi Swanson is thanked for assistance with sample site characterization. We acknowledge funding from a Strategic Project Grant from the Natural Sciences and Engineering Research Council of Canada (NSERC).

References Belova, S., Pankratov, T., and Dedysh, S. 2006. Bacteria of the genus Burkholderia as a typical component of the microbial community of Sphagnum peat bogs. Microbiology, 75: 90–96. doi:10.1134/S0026261706010164. PMID:16579449. Benson, D.A., Karsch-Mizrachi, I., Lipman, D.J., Ostell, J., Rapp, B.A., and Wheeler, D.L. 2000. Genbank. Nucleic Acids Res. 28: 15–18. doi:10.1093/nar/ 28.1.15. PMID:10592170. Bernard, L., Mougel, C., Maron, P.A., Nowak, V., Lévêque, J., Henault, C., et al. 2007. Dynamics and identification of soil microbial populations actively assimilating carbon from 13C-labelled wheat residue as estimated by DNA- and RNA-SIP techniques. Environ. Microbiol. 9: 752–764. doi:10.1111/j.1462-2920. 2006.01197.x. PMID:17298374. Bernard, L., Maron, P.A., Mougel, C., Nowak, V., Leveque, J., Marol, C., et al. 2009. Contamination of soil by copper affects the dynamics, diversity, and activity of soil bacterial communities involved in wheat decomposition and carbon storage. Appl. Environ. Microbiol. 75: 7565–7569. doi:10.1128/AEM.00616-09. PMID:19801474. Caporaso, J.G., Bittinger, K., Bushman, F.D., DeSantis, T.Z., Andersen, G.L., and

Knight, R. 2010a. PyNAST: a flexible tool for aligning sequences to a template alignment. Bioinformatics, 26: 266–267. doi:10.1093/bioinformatics/btp636. PMID:19914921. Caporaso, J.G., Kuczynski, J., Stombaugh, J., Bittinger, K., Bushman, F.D., Costello, E.K., et al. 2010b. QIIME allows analysis of high-throughput community sequencing data. Nat. Methods, 7: 335–336. doi:10.1038/nmeth.f.303. PMID:20383131. Chen, V.B., Davis, I.W., and Richardson, D.C. 2009. KiNG (Kinemage, Next Generation): a versatile interactive molecular and scientific visualization program. Protein Sci. 18: 2403–2409. doi:10.1002/pro.250. PMID:19768809. Chen, Y., Dumont, M.G., Neufeld, J.D., Bodrossy, L., Stralis-Pavese, N., McNamara, N.P., et al. 2008. Revealing the uncultivated majority: combining DNA stable-isotope probing, multiple displacement amplification and metagenomic analyses of uncultivated Methylocystis in acidic peatlands. Environ. Microbiol. 10: 2609–2622. doi:10.1111/j.1462-2920.2008.01683.x. PMID:18631364. Dumont, M.G., Radajewski, S.M., Miguez, C.B., McDonald, I.R., and Murrell, J.C. 2006. Identification of a complete methane monooxygenase operon from soil by combining stable isotope probing and metagenomic analysis. Environ. Microbiol. 8: 1240–1250. doi:10.1111/j.1462-2920.2006.01018.x. PMID:16817932. Dumova, V., and Kruglov, Y.V. 2009. A cellulose-decomposing bacterial association. Microbiology, 78: 234–239. doi:10.1134/S0026261709020155. el Zahar Haichar, F., Achouak, W., Christen, R., Heulin, T., Marol, C., Marais, M.F., et al. 2007. Identification of cellulolytic bacteria in soil by stable isotope probing. Environ. Microbiol. 9: 625–634. doi:10.1111/j.1462-2920.2006. 01182.x. PMID:17298363. Engel, K., Pinnell, L., Cheng, J., Charles, T.C., and Neufeld, J.D. 2012. Nonlinear electrophoresis for purification of soil DNA for metagenomics. J. Microbiol. Methods, 88: 35–40. doi:10.1016/j.mimet.2011.10.007. PMID:22056233. Green, S.J., Leigh, M.B., and Neufeld, J.D. 2010. Denaturing gradient gel electrophoresis (DGGE) for microbial community analysis. In Microbiology of hydrocarbon and lipid microbiology. Edited by K.N. Timmis. Springer-Verlag Berlin Heidelberg, Berlin, Germany. pp. 4137–4158. Griffiths, R.I., Whiteley, A.S., O’Donnell, A.G., and Bailey, M.J. 2000. Rapid method for coextraction of DNA and RNA from natural environments for analysis of ribosomal DNA- and rRNA-based microbial community composition. Appl. Environ. Microbiol. 66: 5488–5491. doi:10.1128/AEM.66.12.54885491.2000. PMID:11097934. Published by NRC Research Press

Pagination not final (cite DOI) / Pagination provisoire (citer le DOI)

Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by YORK UNIV on 07/08/14 For personal use only.

8

Hall, M., Bansal, P., Lee, J.H., Realff, M.J., and Bommarius, A.S. 2010. Cellulose crystallinity — a key predictor of the enzymatic hydrolysis rate. FEBS J. 277: 1571–1582. doi:10.1111/j.1742-4658.2010.07585.x. PMID:20148968. Hou, P.-b., Li, Y.-z., Wu, B.-h., Yan, Z.-c., Yan, B.-x., and Gao, P.-j . 2006. Cellulolytic complex exists in cellulolytic myxobacterium Sorangium. Enzyme Microb. Technol. 38: 273–278. doi:10.1016/j.enzmictec.2004.08.044. Houghton, R.A., Davidson, E.A., and Woodwell, G.M. 1998. Missing sinks, feedbacks, and understanding the role of terrestrial ecosystems in the global carbon balance. Global Biogeochem. Cycles, 12: 25–34. doi:10.1029/97GB02729. Kalyuzhnaya, M.G., Lapidus, A., Ivanova, N., Copeland, A.C., McHardy, A.C., Szeto, E., et al. 2008. High-resolution metagenomics targets specific functional types in complex microbial communities. Nat. Biotechnol. 26: 1029– 1034. doi:10.1038/nbt.1488. PMID:18711340. Koizumi, S., Yue, Z., Tomita, Y., Kondo, T., Iwase, H., Yamaguchi, D., and Hashimoto, T. 2008. Bacterium organizes hierarchical amorphous structure in microbial cellulose. Eur. Phys. J. E. Soft Matter, 26: 137–142. doi:10.1140/ epje/i2007-10259-3. PMID:18311475. Landy, E.T., Mitchell, J.I., Hotchkiss, S., and Eaton, R.A. 2008. Bacterial diversity associated with archaeological waterlogged wood: ribosomal RNA clone libraries and denaturing gradient gel electrophoresis (DGGE). Int. Biodeterior. Biodegrad. 61: 106–116. doi:10.1016/j.ibiod.2007.07.007. Lane, D.J. 1991. 16S/23S rRNA sequencing. In Nucleic acid techniques in bacterial systematics. Edited by E. Stackebrandt and M. Goodfellow. John Wiley & Sons, Inc., Chichester, UK. pp. 115–175. Lee, C.G., Watanabe, T., Sato, Y., Murase, J., Asakawa, S., and Kimura, M. 2011. Bacterial populations assimilating carbon from 13C-labeled plant residue in soil: analysis by a DNA-SIP approach. Soil Biol. Biochem. 43: 814–822. doi:10. 1016/j.soilbio.2010.12.016. Li, W., and Godzik, A. 2006. Cd-hit: a fast program for clustering and comparing large sets of protein or nucleotide sequences. Bioinformatics, 22: 1658–1659. doi:10.1093/bioinformatics/btl158. PMID:16731699. Lozupone, C., and Knight, R. 2005. UniFrac: a new phylogenetic method for comparing microbial communities. Appl. Environ. Microbiol. 71: 8228–8235. doi:10.1128/AEM.71.12.8228-8235.2005. PMID:16332807. Lozupone, C., Lladser, M.E., Knights, D., Stombaugh, J., and Knight, R. 2010. UniFrac: an effective distance metric for microbial community comparison. ISME J. 5: 169–172. doi:10.1038/ismej.2010.133. PMID:20827291. Lynch, M.D.J., Masella, A.P., Hall, M.W., Bartram, A.K., and Neufeld, J.D. 2013. AXIOME: automated exploration of microbial diversity. GigaScience, 2: 3. doi:10.1186/2047-217X-2-3. PMID:23587322. Lynd, L.R., Weimer, P.J., Van Zyl, W.H., and Pretorius, I.S. 2002. Microbial cellulose utilization: fundamentals and biotechnology. Microbiol. Mol. Biol. Rev. 66: 506–577. doi:10.1128/MMBR.66.3.506-577.2002. PMID:12209002. Mack, M.C., Schuur, E.A.G., Bret-Harte, M.S., Shaver, G.R., and Chapin, F.S., III. 2004. Ecosystem carbon storage in arctic tundra reduced by long-term nutrient fertilization. Nature, 431: 440–443. doi:10.1038/nature02887. PMID:15386009. Melillo, J., Steudler, P., Aber, J., Newkirk, K., Lux, H., Bowles, F., et al. 2002. Soil warming and carbon-cycle feedbacks to the climate system. Science, 298: 2173–2176. doi:10.1126/science.1074153. PMID:12481133. Morita, R.Y. 1975. Psychrophilic bacteria. Bacteriol. Rev. 39: 144. PMID:1095004. Neufeld, J.D., Yu, Z., Lam, W., and Mohn, W.W. 2004. Serial analysis of ribosomal sequence tags (SARST): a high-throughput method for profiling complex microbial communities. Environ. Microbiol. 6: 131–144. doi:10.1046/j.1462-2920. 2003.00547.x. PMID:14756878. Neufeld, J.D., Dumont, M.G., Vohra, J., and Murrell, J.C. 2007a. Methodological considerations for the use of stable isotope probing in microbial ecology. Microb. Ecol. 53: 435–442. doi:10.1007/s00248-006-9125-x. PMID:17072677. Neufeld, J.D., Vohra, J., Dumont, M.G., Lueders, T., Manefield, M., Friedrich, M.W., and Murrell, J.C. 2007b. DNA stable-isotope probing. Nat. Protoc. 2: 860–866. doi:10.1038/nprot.2007.109. PMID:17446886. Neufeld, J.D., Chen, Y., Dumont, M.G., and Murrell, J.C. 2008. Marine methylotrophs revealed by stable-isotope probing, multiple displacement ampli-

Can. J. Microbiol. Vol. 60, 2014

fication and metagenomics. Environ. Microbiol. 10: 1526–1535. doi:10.1111/j. 1462-2920.2008.01568.x. PMID:18294205. O’Sullivan, C., Burrell, P., Clarke, W., and Blackall, L. 2007. A survey of the relative abundance of specific groups of cellulose degrading bacteria in anaerobic environments using fluorescence in situ hybridization. J. Appl. Microbiol. 103: 1332–1343. doi:10.1111/j.1365-2672.2007.03362.x. PMID:17897237. Pang, H., Zhang, P., Duan, C.-J., Mo, X.-C., Tang, J.-L., and Feng, J.-X. 2009. Identification of cellulase genes from the metagenomes of compost soils and functional characterization of one novel endoglucanase. Curr. Microbiol. 58: 404–408. doi:10.1007/s00284-008-9346-y. PMID:19159974. Paradis, E., Claude, J., and Strimmer, K. 2004. APE: analyses of phylogenetics and evolution in R language. Bioinformatics, 20: 289–290. doi:10.1093/ bioinformatics/btg412. PMID:14734327. Pinnell, L.J., Charles, T.C., and Neufeld, J.D. 2011. Stable isotope probing and metagenomics. In Stable isotope probing and related technologies. Edited by J.C. Murrell and A.S. Whitely. ASM Press, Washington, D.C. pp. 97–114. Post, W.M., Emanuel, W.R., Zinke, P.J., and Stangenberger, A.G. 1982. Soil carbon pools and world life zones. Nature, 298: 156–159. doi:10.1038/298156a0. Ronan, P., Yeung, C.W., Schellenberg, J., Sparling, R., Wolfaardt, G.M., and Hausner, M. 2013. A versatile and robust aerotolerant microbial community capable of cellulosic ethanol production. Bioresour. Technol. 129: 156–163. doi:10.1016/j.biortech.2012.10.164. PMID:23238345. Saitou, N., and Nei, M. 1987. The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol. Biol. Evol. 4(4): 406–425. PMID: 3447015. Schellenberger, S., Kolb, S., and Drake, H.L. 2010. Metabolic responses of novel cellulolytic and saccharolytic agricultural soil bacteria to oxygen. Environ. Microbiol. 12: 845–861. doi:10.1111/j.1462-2920.2009.02128.x. PMID:20050868. Schramm, M., and Hestrin, S. 1954. Factors affecting production of cellulose at the air/liquid interface of a culture of Acetobacter xylinum. J. Gen. Microbiol. 11: 123–129. doi:10.1099/00221287-11-1-123. PMID:13192310. Schuur, E.A.G., Vogel, J.G., Crummer, K.G., Lee, H., Sickman, J.O., and Osterkamp, T.E. 2009. The effect of permafrost thaw on old carbon release and net carbon exchange from tundra. Nature, 459: 556–559. doi:10.1038/ nature08031. PMID:19478781. Schwarz, S., Waschkowitz, T., and Daniel, R. 2006. Enhancement of gene detection frequencies by combining DNA-based stable-isotope probing with the construction of metagenomic DNA libraries. World J. Microbiol. Biotechnol. 22: 363–367. doi:10.1007/s11274-005-9042-z. Sul, W.J., Park, J., Quensen, J.F., Rodrigues, J.L.M., Seliger, L., Tsoi, T.V., et al. 2009. DNA-stable isotope probing integrated with metagenomics for retrieval of biphenyl dioxygenase genes from polychlorinated biphenylcontaminated river sediment. Appl. Environ. Microbiol. 75: 5501–5506. doi: 10.1128/AEM.00121-09. PMID:19648381. Ventura, M., Canchaya, C., Tauch, A., Chandra, G., Fitzgerald, G.F., Chater, K.F., and van Sinderen, D. 2007. Genomics of Actinobacteria: tracing the evolutionary history of an ancient phylum. Microbiol. Mol. Biol. Rev. 71: 495–548. doi:10.1128/MMBR.00005-07. PMID:17804669. Wang, Q., Garrity, G.M., Tiedje, J.M., and Cole, J.R. 2007. Naïve Bayesian classifier for rapid assignment of rRNA sequences into the new bacterial taxonomy. Appl. Environ. Microbiol. 73: 5261–5267. doi:10.1128/AEM.00062-07. PMID: 17586664. Wilson, D.B. 2011. Microbial diversity of cellulose hydrolysis. Curr. Opin. Microbiol. 14: 259–263. doi:10.1016/j.mib.2011.04.004. PMID:21531609. Yilmaz, S., Allgaier, M., and Hugenholtz, P. 2010. Multiple displacement amplification compromises quantitative analysis of metagenomes. Nat. Methods, 7: 943–944. doi:10.1038/nmeth1210-943. PMID:21116242. Zhou, J., Bruns, M.A., and Tiedje, J.M. 1996. DNA recovery from soils of diverse composition. Appl. Environ. Microbiol. 62: 316–322. PMID:8593035. Zimov, S.A., Schuur, E.A., and Chapin, F.S., III. 2006. Permafrost and the global carbon budget. Science, 312: 1612–1613. doi:10.1126/science.1128908. PMID: 16778046.

Published by NRC Research Press