The ISME Journal (2013) 7, 2044–2053 & 2013 International Society for Microbial Ecology All rights reserved 1751-7362/13 www.nature.com/ismej ORIGINAL ARTICLE Linking microbial community structure to b-glucosidic function in soil aggregates Vanessa L Bailey1, Sarah J Fansler1, James C Stegen1 and Lee Ann McCue2 1Microbiology, Pacific Northwest National Laboratory, Richland, WA, USA and 2Computational Biology and Bioinformatics, Pacific Northwest National Laboratory, Richland, WA, USA To link microbial community 16S structure to a measured function in a natural soil, we have scaled both DNA and b-glucosidase assays down to a volume of soil that may approach a unique microbial community. b-Glucosidase activity was assayed in 450 individual aggregates, which were then sorted into classes of high or low activities, from which groups of 10 or 11 aggregates were identified and grouped for DNA extraction and pyrosequencing. Tandem assays of ATP were conducted for each aggregate in order to normalize these small groups of aggregates for biomass size. In spite of there being no significant differences in the richness or diversity of the microbial communities associated with high b-glucosidase activities compared with the communities associated with low b-glucosidase communities, several analyses of variance clearly show that the communities of these two groups differ. The separation of these groups is partially driven by the differential abundances of members of the Chitinophagaceae family. It may be observed that functional differences in otherwise similar soil aggregates can be largely attributed to differences in resource availability, rather than to the presence or absence of particular taxonomic groups. The ISME Journal (2013) 7, 2044–2053; doi:10.1038/ismej.2013.87; published online 30 May 2013 Subject Category: Microbial ecology and functional diversity of natural habitats Keywords: microbial community; diversity; 16S pyrosequencing; beta-glucosidase; soil aggregates; ATP Introduction been done on quantities of soil too large to consider as a defined habitat for a deconvolvable microbial The composition of soil microbial biomass is 7 community. overwhelmingly diverse, up to 10 different species Unlike more well-mixed or constrained habitats, in a gram of soil (Gans et al., 2005), and these such as termite guts or microbial mats, the remarkable assemblages mediate a vast number of 3-dimensional nature of soil, presenting chemical ecosystem processes associated with nutrient and physical heterogeneities and barriers, makes it cycling, C dynamics and soil health. Contemporary difficult to assign the microbes, reported in a high-throughput sequencing techniques are rapidly metagenome, to particular ecological roles. Addi- increasing our knowledge of what organisms live in tionally, microbial census data, such as 16S soil, and their phylogenetic diversity. However, sequences, are analyzed in the context of existing directly linking this information, essentially sequence databases, and soils are notoriously ‘census’ data, with actual ecosystem processes is home to a wealth of uncultured or unclassified needed in order to understand how interacting microorganisms (Dokic et al., 2010). Shotgun meta- microbial communities evolve and function genomes seek to provide greater genetic information (Morales and Holben, 2011). For natural microbial to yield functional inferences, but their usefulness is communities, we do not yet know the relationship often hampered by the challenges associated with between diversity and function, because it has not (1) de novo assembly of short read sequence data been analyzed at an appropriate spatial scale. from a complex assemblage, (2) the availability of Until recently, identifying such structure–function too few appropriate reference genomes to which linkages in soil has been hindered by the amount metagenome fragments can be aligned and (3) poor of biomass needed for experiments; the high- accuracy of binning of the sequenced fragments throughput sequencing of soil metagenomes has into appropriate groups or potential operational taxonomic units (OTUs) that reflect species within Correspondence: VL Bailey, Microbiology, Pacific Northwest the assemblage (Wooley et al., 2010). National Laboratory, 902 Battelle Boulevard, MSIN J4-18, Further confusing our ability to understand the Richland, WA 99352, USA. microbial communities in natural systems, such as E-mail: [email protected] Received 9 January 2013; revised 7 March 2013; accepted 23 April soil, is the fact that what the community is doing is 2013; published online 30 May 2013 likely far more complex than which species are Community structure and function in aggregates VL Bailey et al 2045 present, based on the likely presence of functional (841 mm and 1000 mm), and 450 field-moist aggre- redundancies within the community and metabolic gates (13% w/w) were collected using sterile, versatility within single species. Unfortunately, fine-point tweezers. Mass was determined for each classical soil enzyme assays, though simple to macroaggregate on a AX105DR Delta Range micro- conduct, reveal the potential function of a sample, balance (Mettler-Toledo, Greifensee, Switzerland), not the actual expressed function in situ. Efforts to and the macroaggregate was then transferred to a study the in situ activities in the soil by directly well in a 96-well plate. Aggregate-free wells were analyzing mRNAs have proven difficult, given the randomly distributed on each plate as controls. ephemeral nature of mRNA and the abundance of nucleases in soil (Cai et al., 2006; Pietramellara et al., 2009). Additionally, both enzyme assays and Biochemical assays the mRNA analyses are limited; most enzyme assays b-Glucosidase (EC: 3.2.1.21) and ATP were analyzed will probably be reflective of multiple genes and sequentially on each aggregate (Bailey et al., 2012). organisms, and mRNA analyses are dependent on a Briefly, the b-glucosidase assay (modified from priori sequence knowledge (Bailey et al., 2010; (Saiya-Cork et al., 2002)) was scaled down to McGrath et al., 2010; Wang et al., 2011). Finally, 125 ml: 100 ml of acetate buffer (50 mM, pH 5.0) was both analyses have traditionally been conducted added to each macroaggregate and the plate agitated. on relatively large sample sizes, compared with Twenty-five microliters of 4-methylumbelliferyl the scale at which microorganisms interact as a b-D-glucopyranoside (200 mM) was added to each well, community, the latter of which would better able and the plates incubated at 22 1C in the dark for 1 h. researchers to draw correlations (that is, (Kumaresan The plates were then centrifuged (10 000 Â g)for2min, et al., 2011)) between the community structure and the supernatant transferred into a fresh 96-well plate, function. and 5 mlNaOH(1.0M) added to stop the reaction. The main objective of our research was to link Methylumbelliferone fluorescence was measured microbial community structure to a measured func- immediately, using a Wallac Victor 2 1420 Multilabel tion in a natural soil, at a scale approaching a Counter (PerkinElmer, Waltham, MA, USA). The defined habitat. We used b-glucosidase activity as a pellets from this procedure were kept on ice for the target function; it is a likely rate-limiting step in ATP assay. cellulose decomposition (van Zyl et al., 2010), ATP was assayed with the BacTiter kit (Promega and we have successfully analyzed this enzyme Inc, Madison, WI, USA). Soil pellets were resus- sequentially with measurements of total soil ATP as pended in 100 ml EDTA (20 mM, pH 7.5) and a surrogate for microbial biomass (Bailey et al., sonicated for 15 min (Branson Ultrasonics Corp, 2012). These biochemical assays were performed on Danbury, CT, USA). The plate was centrifuged individual soil aggregates, naturally occurring (10 000 Â g) for 2 min, and 100 ml of supernatant ‘clumps’ of soil less than a millimeter in diameter, transferred into a fresh 96-well plate. In a darkened to constrain the microbial community to a scale room, 5 ml of MgCl (0.4 M) and 100 ml ATP assay more representative of a community capable of 2 reagent were added, the plates shaken gently interacting. These aggregates were then grouped for 5 min, and luminescence wasmeasured on according to enzyme activity level and the DNA the Wallac Victor 2 1420 Multilabel Counter. In extracted; combining similarly active aggregates was preparation for DNA extraction, the plates were then necessary to recover sufficient DNA for pyrosequen- centrifuged, the supernatant removed and the cing. In these grouped samples, we observed that pellets frozen at À 80 1C. although the diversities of the microbial commu- nities were similar, there were statistically signifi- cant differences in community structure between the high and low b-glucosidase activity aggregate Grouping of aggregates and DNA extraction groups. In order to collect enough DNA for 16S pyrosequen- cing, individual aggregates had to be pooled. The aggregates were sorted by b-glucosidase activity, Materials and methods and aggregates with b-glucosidase activities within 1 s.d. of the mean b-glucosidase activity for the whole Soils set were discarded. The remaining aggregates were Soil was collected from a grassland field at the randomly sorted into five groups of ‘high activity’ United States Department of Agriculture Conserva- aggregates (HAP) and seven groups of ‘low activity’ tion Field, near Pullman, Washington (fine-silty, aggregates (LAP). The total ATP measured in each mixed, superactive,