3
Soil ecosystems include a diverse array of organisms, ranging from macroorganisms, such as plants and animals, to microorganisms, such as bacteria, archaea, and fungi. In most natural ecosystems each of these groups is composed of a range of species and thus contributes to the overall biological diversity of the system. Some notable exceptions to this principle are the plant monocultures created by modern agriculture, and the monocultures that can occur due to the invasion of habitats by some exotic plant species (e.g. Thompson et al. 1987). The interacting macro‐ and microorganisms within soils have significant impacts on the nature of the soil ecosystem and the dynamics of the biological processes occurring therein. Therefore, the diversity of all of these classes of organisms can have significant implications for ecosystem function, stability, and resilience. The value of diversity to biological communities is exemplified by the increased pathogen susceptibility of plant monocultures (Leonard and Czochor 1980). Thus there is value in exploring and understanding the diversity of all of the classes of organisms found within soil.
Within soil ecosystems the diversity present within the microbial communities far outstrips the diversity of even the most diverse communities of macroorganisms. For example, typical bacterial cell densities in soil range from 108 to 109 cells per gram of dry soil; DNA reassociation analysis has shown that a single gram of soil contains 4000–7000 unique bacterial genomes (Torsvik et al. 1990). Many questions remain about the depth and significance of this diversity, as methodological constraints have made it virtually impossible to completely assess all of the bacterial diversity within a typical soil. Due to these methodological limitations the vast majority of bacterial diversity and the metabolic potential within these genomes have yet to be discovered. Also significant is the fact that the species composition of soil bacterial communities seems especially responsive to ecosystem changes and environmental perturbations, making soil bacterial diversity a potentially useful bioindicator of soil quality (Hartmann and Widmer 2006). For these reasons there has been and continues to be a great deal of interest in exploring and understanding soil microbial diversity. The objective of this chapter is to review some of the most commonly used methods for assessing soil microbial diversity, to discuss the limitations of the various methods, and to elucidate the types of information that are and are not provided by assessments of soil microbial diversity.
3.1 Classical Culture‐Based Studies of Soil Microbial Diversity
Early studies of soil bacterial diversity involved isolating and purifying the bacterial species occurring in the soil of interest via culture‐based methods and classifying the isolates according to species based on phenotypical characteristics. This approach can be used to assemble soil bacterial community surveys and to compare the species composition of communities from different sites or different experimental treatments. Various numerical approaches can be used to quantify diversity and assess differences in composition between multiple communities based on these culture‐based surveys. For example, the diversity of a community can be quantified through the calculation of a diversity score based on one of several diversity indices such as the Shannon index (Shannon 2001). The composition of multiple communities can also be compared by calculating similarity scores for each pair of communities via a variety of similarity indices and performing a cluster analysis (see Russek‐Cohen and Colwell (1986) for a review of this topic).
3.1.1 Value of Culture‐Based Studies of Soil Microbial Diversity
Culture‐based studies are relatively inexpensive to conduct and do not require advanced analytical equipment. For these reasons standardized protocols can be used throughout the world. Another major advantage of culture‐based studies is that the organisms that have been isolated can be maintained in the laboratory and used to address additional research questions. For example, the specific biochemical composition and metabolic capabilities of organisms of interest can be analyzed for microorganisms that have been isolated in pure culture. Finally, with a culture‐based approach the investigator can be sure that the organisms that have been isolated are viable, i.e. capable of growth.
3.1.2 Limitations of Culture‐Based Studies of Soil Microbial Diversity
Culture‐based studies are extremely time consuming and labor intensive in that large numbers of bacterial strains must be isolated from soil and characterized physiologically, which limits the depth to which a given community can be characterized. Another of the main challenges associated with culture‐based approaches to assessing diversity is that less than 1% of the microbes occurring in soil can be grown on standard laboratory media. Thus, the organisms isolated for the diversity characterization are a minor subset of the organisms existing in situ. This limitation could be acceptable if the isolated organisms represented a truly random and unbiased sample of the total bacterial community in the soils, or at least of the dominant populations. Unfortunately, this is clearly not the case, as the cultivation procedure by definition selects for certain organisms. Specifically, the cultivated organisms are those that can utilize the energy source provided in the medium under the physical and chemical limitations of the growth medium and incubation conditions. Furthermore, the organisms generally selected for study are those that grow rapidly on the isolation medium. Thus, slow‐growing microbes or those that produce small colonies are commonly overlooked. Therefore, all that can be said about the strains isolated in a culture‐based survey is that propagules of these strains existed in the soil system, but they may or may not have been active in situ (in surface soils, most of the microbes are inactive or are in a resting state at any given time), and they may or may not have been among the numerically dominant populations (i.e. other species or strains may have been numerically dominant but incapable of growth on the media used to select the bacterial isolates). Thus, the only conclusion that can be made is that the organisms isolated from the soil samples are the numerically dominant strains that could be cultivated.
3.1.3 The Challenge of Defining Bacterial Species
Another inherent difficulty associated with culture‐based studies, and indeed with all studies based on species identification of soil bacteria, is the difficulty of defining a bacterial species. Bacterial species can be loosely defined as a collection of closely related bacterial strains. Classically, the strains forming a particular bacterial species have been grouped based upon possession of a series of common physiological capabilities, structural properties, and miscellaneous traits – such as motility and growth characteristics (e.g. pH or temperature tolerance ranges). A small percentage of the potential “species” in the bacterial world have been classified by this approach. Of greater concern to these definitions is that the entity that might be classed as a particular bacterial species is mutable. That is, its basic properties change. Mutations or transfers of genetic material from one species to another can readily blur the distinctions that were originally used to define the individual species. Thus, species definitions in bacteria are rather blurred and are to a large degree artificial.
Arguably, it could be stated that specific species designations are of greater significance when considering public health problems when evaluating soil populations. This statement is based on the observation that it is the overall metabolic capability of the community and degree of similarity of the individuals that define the functional soil community, rather than the presence of nonchanging individual bacterial species – because they simply do not exist in the complex soil community.
3.1.4 Alternatives to Bacterial Strain Isolation
Ideally, a method for analysis of soil microbial diversity should involve evaluation of the total population. The method should include those species that grow on laboratory media and those that don't; those species that have been described as known species and those that possibly never will be; and those species that are active in the community as it currently exists plus those that have the potential to become active when conditions change. Since the simple task of culturing all members of the total microbial community from soil is insurmountable, a reasonable compromise to achieve this ideal objective would be to use a surrogate to assess microbial diversity in soil. This surrogate should be associated with living cells only, vary in a meaningful way in relationship to the overall microbial diversity in the system, and be sufficiently variable that differences in overall microbial diversity between communities could be detected. Candidates for use as surrogates in estimating soil microbial diversity include community metabolic (physiological) capabilities and various molecular components of cells. These will be discussed in more detail below.
3.2 Surrogate Measures of Soil Microbial Diversity
Fundamentally, to be aligned with classical considerations of biological diversity, surrogate measures of soil microbial diversity should be based on a delineation that varies in parallel with actual species diversity. Furthermore, the population sampling of the system of interest should either be all inclusive or at least fully representative of the diversity of the community. Few systems exist, either above or below ground, wherein enumeration of the totality of microbial species present is possible, so a means of assuring that those species detected are a truly representative subsample of the community is necessary. However, as the foregoing discussion of bacterial isolation and species characterization has indicated, these requirements present significant challenges for analyses of microbial diversity. Therefore, an estimate of the biological diversity of a soil microbial community must be based on selection of a surrogate that meets the above specifications of an ideal measure as closely as possible.
Further characteristics of an ideal surrogate measure of biological diversity are of a more practical nature and are derived primarily from two factors. First, the medium itself, soil, is extremely complex and heterogeneous. Second, there are a large number of factors that may significantly alter the soil community composition with time. Therefore, the procedure should be quick and easy to conduct so that a large number of samples can be processed in a short period of time. This requirement is likely one of the largest impediments to the use of culture‐based methods to measure microbial diversity. Ideally the method should also be reasonably economical to conduct and require only moderately specialized equipment and skill so that it can be generally applied to the study of the world's soil ecosystems. As with the more scientifically based specifications discussed above, these requirements are only partially fulfilled by methods currently available.
The procedures most commonly used to estimate soil biological diversity are community‐level physiological profiling (CLPP), phospholipid fatty acid (PLFA) analysis, and a variety of DNA‐based procedures. For each of these techniques, the nature of the soil biological variable assessed, the potential for provision of useful information about community structural variation in soil systems, and the limitations of the data will be discussed below.
3.3 Diversity Surrogates: Physiological Profiling
Bacterial species vary in many aspects of their physiology, including their sources of carbon and energy and their tolerances for a range of environmental factors (e.g. temperature, pH, salinity). One way in which bacterial species can vary is in their ability to utilize different carbon substrates, and this ability is one characteristic that can be used to identify and type bacterial isolates. One example of this is a tool known as BIOLOG (Miller and Rhoden 1991). The BIOLOG assay consists of a 96‐well microtiter plate in which 95 of the wells each contain a different carbon substrate, and all of the wells contain a tetrazolium dye that undergoes a color change when it is reduced. During the assay a microbial suspension is added to all the wells within a plate and oxidation of individual carbon substrates triggers reduction of the tetrazolium dye. The resultant color formation in the well can be detected and quantified spectrophotometrically with a microplate reader. The BIOLOG assay was developed for the discrimination and identification of axenic bacterial and fungal cultures based on the pattern of oxidation of the substrates within the BIOLOG plates. There are several variations of the BIOLOG plates (e.g. gram positive, gram negative, fungal) each of which contains a different set of 95 substrates chosen for their ability to discriminate among species within the target microbial group.
3.3.1 Physiological Profiling of Isolates
The BIOLOG assay is effective at identifying axenic cultures (Klingler et al. 1992; Miller and Rhoden 1991), and so it can be useful for the identification of isolates in a culture‐based assessment of soil microbial diversity as described above. Specifically, individual strains can be isolated in pure culture, and the BIOLOG assay can be used to identify these isolated strains (e.g. Behrendt et al. 1997; Timonen et al. 1998). However, this approach still suffers from the limitations discussed above in regard to the use of any culture‐based approach that requires isolation of bacterial strains from soil.
3.3.2 Community‐Level Physiological Profiling
To avoid the limitations inherent in isolating individual bacterial strains from soil, some researchers have taken the approach of assessing the metabolic capabilities of whole microbial communities as a surrogate for diversity (Garland and Mills 1991). The use of the metabolic capabilities of a whole community as an indicator of microbial diversity is predicated on the assumption that significant differences in this parameter occur between soil samples that differ in microbial community composition. The truth of this assumption is not obvious. First, the prime energy source for soil ecosystems is photosynthetically fixed carbon. Second, although there are differences in the biochemical make‐up of different plant species or of the same plant species growing under varying conditions, the general array of biochemicals encountered by soil microbes is not highly different from ecosystem to ecosystem. Therefore, soil microbes are facing a reasonably common group of biochemicals of varying concentrations in most ecosystems. It could thus be anticipated that all soil microbial communities would possess a reasonably similar arsenal of metabolic capacities to recover energy and nutrients from these substances. Therefore, variations in metabolic diversity between soils could be predicted to be minimal. Fortunately, in practice metabolic diversity does provide information that varies with ecosystem type and the conditions of particular ecosystems.
The metabolic capabilities of microbial communities have frequently been assessed using a variation of the BIOLOG assay in which soil extracts are directly inoculated into the wells of the BIOLOG plates, without any isolation of individual microbial species, in order to determine which substrates can be utilized by the whole microbial community. This technique has been referred to as a CLPP (Garland and Mills 1991). It has been found to provide reproducible profiles of model bacterial communities (e.g. Haack et al. 1995) as well as reproducible metabolic diversity patterns for soil ecosystems under a variety of cropping conditions and plant types (Bossio and Scow 1995; Garland and Mills 1991; Lupwayi et al. 1998; Winding 1994; Zak et al. 1994). CLPP has also been used to assess changes in microbial community composition caused by heavy metal contamination (Kelly and Tate 1998; Knight et al. 1997), variations in agricultural management of soils (Bossio and Skow 1998), variations in soil moisture (Zak et al. 1994), plant species variation (e.g. Ellis et al. 1995; Garland 1996a,b; Grayston et al. 1998) and the presence of chemical pollutants (e.g. Fuller et al. 1997; Thompson et al. 1999).
In the studies discussed above CLPP profiles were generated using one of the BIOLOG plates designed for identification of a specific microbial group (e.g. gram‐positive bacteria, gram‐negative bacteria, or fungi). Most commonly the plates designed for gram‐negative bacteria were used. These plates are not ideal for community‐level analysis because the substrates were chosen for their ability to discriminate species in pure culture, not for their environmental relevance. In addition, there was some concern that there was no interplate replication of the substrates, which limited statistical analysis of the data. In response to the interest in using the BIOLOG technology for CLPP profiling, BIOLOG developed the EcoPlate which contains 31 substrates that were selected and optimized from past environmental applications and are arranged in triplicate on each of the plates. Many recent studies have utilized the EcoPlates for CLPP profiling of field soils (e.g. Aira et al. 2000; Fontúrbela et al. 2012; Garau et al. 2007; Rutgers et al. 2016).
3.3.3 Value of Community‐Level Physiological Profiling
CLPP has been shown to differ for bacterial communities with distinct species composition (Haack et al. 1995), and, as reviewed above, a number of studies have shown that CLPP generates reproducible profiles for microbial communities from a variety of soil types. Therefore, it is reasonable to consider CLPP as a useful tool for the determination of differences or changes in the composition of soil microbial communities. The diversity of substrates oxidized by a community in the CLPP assay can be quantified by using the CLPP data to calculate a diversity index, such as the Shannon index. The CLPP profiles of multiple communities can also be compared by multivariate techniques such as principal components analysis (PCA) (e.g. Kelly and Tate 1998; Kelly et al. 1999a) or multidimensional scaling (MDS) (e.g. Broughton and Gross 2000). CLPP analysis has some advantages over other methods for assessing diversity. For example, it is relatively inexpensive, requires a minimum of equipment (a microplate reader) and requires a minimum of training. Finally, the fact that the CLPP assay is based on measurement of metabolic processes, rather than being limited to species identification, may have some value for some research applications, as connecting species identification to functional capabilities can be a challenge for methods based strictly on species identification.
3.3.4 Limitations of Community Level Physiological Profiling
A primary question related to physiological profiling data is whether the patterns of metabolism of the individual substrates in the BIOLOG plate can be used to specifically describe how the various metabolic processes are occurring in the field. Generally they cannot. This conclusion is based mainly on the fact that the conditions within the BIOLOG plates (e.g. pH, nutrient concentrations) and the incubation conditions used (e.g. temperature, light avilability) can be significantly different than what a community would experience in the field. In addition, the length of the CLPP assay (plates can be incubated from hours to days) provides ample opportunity for enzymes that were not expressed in the field to be expressed in the plates, and some evidence suggests that population changes can occur within the communities in the wells. For example, Smalla et al. (1998) used denaturing gradient gel electrophoresis and temperature gradient electrophoresis to demonstrate that significant changes in microbial community composition can occur within the wells during the CLPP assay. Therefore, it is reasonable to infer from a positive result for a substrate within the BIOLOG plate that the organism or organisms responsible for oxidation of that susbtrate existed in the soil sample, but no inference can be gained regarding the function of these organisms in situ since the environmental conditions may not have been appropriate for them to function in the soil. Conversely, a negative result for a substrate on the BIOLOG plate does not mean that organisms capable of oxidizing this substrate were absent from the soil, as these organisms may simply have been inactive under the conditions of the BIOLOG assay. Given these limitations, it is best to view the BIOLOG results as a “fingerprint” indicative of the potential metabolic capabilities of the community in question, and not as an indicator of metabolic processes actually occurring in the field.
3.4 Diversity Surrogates: Phospholipid Fatty Acid Analysis
A second alternative indicator of variation in microbial diversity in soils involves determination of the diversity of phospholipid fatty acids (PLFAs). PLFAs are the fatty acids that are structural components of the phospholipids that comprise cell membranes in eukaryotes and bacteria. The fatty acids found within cell membrane phospholipids can vary significantly in composition (e.g. chain length and degree of unsaturation), and there are consitent differences in PLFA composition across some microbial groups.
3.4.1 PLFA Analysis of Isolates
PLFA analysis has been used to characterize axenically grown bacterial isolates from soil for assessment of microbial community diversity (e.g. Germida et al. 1998), but this approach still suffers from the limitations discussed above in regards to the use of any culture‐based approach that requires isolation of bacterial strains from soil.
3.4.2 Community PLFA Analysis
A more meaningful characterization of soil community dynamics can be attained by extracting phospholipids directly from a soil sample. For this approach it is assumed that all phospholipids contained within cells in the soil are equally accessible to the extracting agent, and that the PLFA profile of the extract is representative of the total PLFA content of the soil microbial community. For community PLFA analysis, phospholipids are extracted from a soil sample with solvent mixtures such as chloroform, methanol, and phosphate buffer (see White et al. (1979) for a detailed description of the extraction procedure). Following extraction, the chloroform fraction is concentrated and the phospholipids are separated on a silicic acid column. (see method of King et al. (1977)). The polar lipid fraction is collected, dried under nitrogen, saponified, and methylated. The resultant fatty acid methyl esters (FAMEs) are then identified via gas chromatography (e.g. with the MIDI protocol [MIDI 1995]). The MIDI system identifies and quantifies fatty acids of 9–20 carbon chain lengths.
The nomenclature for the fatty acids (e.g. 16 : 1 ω7t) refers to the total number of carbon atoms in the chain: the number of double bonds, the position of the double bond(s) from the methyl end of the chain (indicated by the number after the ω), and the conformation (cis vs. trans). Different subgroups of microorganisms (e.g. fungi, mycorrhizal fungi, gram‐positive bacteria, actinomycetes, gram‐negative bacteria) are represented in the profile by the quantities of specific fatty acids. For example 16 : 1 ω5c is an indicator for arbuscular mycorrhizal fungi (Haack et al. 1995), 16 : 0 10 for nocardioform actinomycetes, TBSA 18 : 0 10 Me and 17 : 0 10 Me for actinomycetes (Frostegård et al. 1993), 18 : 2 ω6c for fungi (Guckert et al. 1985), 15 : 0iso, 16 : 0iso, and 10Me16 : 0 for gram‐positive bacteria (Zak et al. 1996), and 16 : 1w7c, cy17 : 0, 18 : 1w7c, and cy19 : 0 for gram‐negative bacteria (Zak et al. 1996). Thus, the resultant phospholipid fatty acid profile provides an indication not only of changes in the composition of the microbial community (by changes in the overall composition of the profile) but also of the changes in the relative abundances of several broad groups of microbes. This procedure has been shown to be a sensitive indicator of shifts in soil microbial community composition resulting from human impacts, such as heavy metal contamination (Frostegård et al. 1996; Kelly et al. 1999a; Pennanen et al. 1996), contamination by organic pollutants (Wang et al. 2016), soil restoration (Kelly et al. 1999b), and increased atmospheric carbon dioxide (Janus et al. 2005). The method is sufficiently sensitive that impacts of root herbivory on rhizosphere communities (Denton et al. 1999) as well as shifts in community composition due to soil management (Bardgett et al. 1997; Bossio et al. 1998; Frostegård et al. 1993; Klamer and Bååth 1998; Kelly et al. 2007) have also been detected. A useful adjunct to the analysis of phospholipid fatty acid diversity is that the total phospholipid fatty acid content of the extract can be used to estimate total soil microbial biomass (e.g. Denton et al. 1999; Janus et al. 2005; Kelly et al. 2007; Klamer and Bååth 1998).
3.4.3 Value of PLFA Analysis
PLFA analysis is an effective tool for the comparison of microbial diversity between samples. PLFA generates quantitative data on the relative abundance of specific PLFAs, and as was described above for the CLPP assay, PLFA data can be used in the calculation of diversity indices, such as the Shannon index. PLFA data can also be used to assess the similarity of communities via multivariate techniques such as PCA (e.g. Kelly et al. 1999b) or MDS (e.g. Baniulyte et al. 2009; Broughton and Gross 2000; Kelly et al. 2007). Because PLFA is a relatively inexpensive procedure, with the main disposable cost being the solvents used in the phospholipid extraction, large numbers of samples can be run in parallel and compared. These factors can allow for higher degrees of replication and more complex experimental designs than might be possible with more expensive methods. Indicator PLFAs (discussed above) also enable discrimination of some broad microbial groups (e.g. fungi, gram‐positive bacteria, gram‐negative bacteria). The ability to assess relative abundance of both bacteria and fungi in a single assay is a unique feature and a major advantage of the PLFA assay.
3.4.4 Limitations of PFLA Analysis
The main limitation of the PLFA procedure is that it does not have a very high level of taxonomic resolution. This assay can provide insight into the relative abundances of broad microbial groups (e.g. fungi, gram‐positive bacteria, gram‐negative bacteria), but it cannot approach the level of resolution needed to identify bacterial species. Another concern is that there is not a simple, direct correlation between the abundance of a particular PLFA and cell number. Cellular fatty acid composition can vary between bacterial species, and individual bacterial species can alter the fatty acid composition of their cell membranes in response to environmental factors such as temperature (e.g. Könneke and Widdel 2003). Therefore, while increases or decreases in certain indicator fatty acids can be inferred to represent increases or decreases in the abundance of the microbial group linked to that fatty acid, care should be exercised in linking PLFA abundance to cell abundance.
3.5 Nucleic Acid‐Based Analyses of Soil Microbial Diversity
Nucleic acids represent another potential surrogate for the analysis of soil microbial diversity. The genomes of all cells are composed of DNA, and variations in DNA sequences can be used to assess microbial community composition and diversity with a level of resolution far greater than any of the other surrogates discussed above. A variety of genes or genomic regions can be used for the analysis of soil microbial diversity, but the most popular is the small subunit ribosomal RNA (SSU rRNA) gene, which was first proposed as a phylogenetic marker by Carl Woese in 1977 (Fox et al. 1977). The advantages of the SSU rRNA gene as a phylogenetic marker are that (i) it is present in all cells, (ii) its sequence is fairly well conserved across phylogenetic groups, allowing routine comparisons of sequences across these groups, and (iii) there is enough variation within the gene to infer phylogenetic relationships between taxonomic groups. Based on these characteristics, SSU rRNA gene sequences have been used to develop a comprehensive phylogenetic framework for all organisms (Woese et al. 1990), and these gene sequences are routinely used for the identification and phylognetic classification of microorganisms. Molecular techniques that analyze prokaryotic SSU rRNA genes, specifically the 16S rRNA genes, have been widely used for the analysis of soil microbial diversity (Macrae 2000).
3.5.1 Nucleic Acid Based Analysis of Isolates
As with the other diversity surrogates discussed above, nucleic acid‐based tools can be applied to prokaryotic organisms isolated from soil. Specifically, DNA can be extracted from a pure culture isolated from soil, the 16S rRNA gene of the isolate can be amplified and sequenced, and the isolate can be identified by comparing its 16S rRNA gene sequence to those in one of several large databases, such as GenBank (www.ncbi.nlm.nih.gov/genbank) or the Ribosomal Database Project (https://rdp.cme.msu.edu). Comparison with these databases will generally enable identification of an isolate to a high level of taxonomic resolution, often to the species or strain level, although these identifications are limited by the extent to which the databases cover the diversity present within soil microbial communities. Identification of soil isolates by DNA sequencing has become a standard method for strain identification and characterization, but as a tool for assessing soil microbial community diversity, DNA sequencing of strains isolated from soil has all of the limitations discussed above with respect to the inherent difficulty in culturing environmental microorganisms in the lab.
3.5.2 Community Nucleic Acid Analysis
Analogous to the approach discussed above for the PLFA assay, an alternative to a culture‐based application of DNA sequencing is a culture‐independent approach in which DNA from an entire microbial community is extracted directly from a soil sample and analyzed. There are multiple methods that have been used to extract DNA directly from soil samples, and the goal of all of these methods is to isolate DNA from the microbial community in an unbiased way, such that the DNA extracted will be representative of all of the taxa within the community, and to purify this DNA adequately to enable downstream molecular analysis.
3.5.3 DNA Extraction
Methods used to extract community DNA directly from soil are designed to accomplish the following tasks: break open cells to release DNA, protect DNA from degradation, and separate DNA from the other components of the soil sample. A variety of methods to accomplish these tasks have been published, and some are better suited to certain soil types than others (e.g. see Braid et al. 2003; Moyer et al. 1996; Tebbe and Vahjen 1993; Tsai and Olson 1992; Zhou et al. 1996). The methods also vary in terms of the equipment needed and the use of hazardous chemicals. Due to the popularity of DNA methods for analyzing microbial communities, several companies have developed commercially available DNA extraction kits, many of which are specialized for different sample types (e.g. soil, water, or fecal samples). Many comparisons of various DNA extraction methods can be found in the literature (e.g. Stach et al. 2001). Researchers should be careful in selecting a DNA extraction approach that is appropriate for the soil types they wish to analyze.
Analysis of soil DNA extracts is complicated by the heterogeneity of the chemical components found in soil and the potential for some of these chemicals to interfere with the extraction or subsequent analysis of the DNA. A primary concern is the potential for interference in DNA analytical procedures by humic acids that can co‐extract with DNA. Thus, the extraction method selected must either minimize co‐extraction of humic acids or involve a purification step to remove the humic acid fraction. Commercially available kits for DNA extraction from soil generally include purification steps to limit co‐extraction of humic acids.
Another concern associated with DNA extraction from soils is the potential for the extraction to introduce bias into the analysis if DNA is not isolated from all cell types equally, and multiple studies have suggested that DNA extraction can introduce bias into analyses of soil microbial community composition (e.g. Carrigg et al. 2007; Feinstein et al. 2009; Martin‐Laurent et al. 2001). Researchers applying these methods should be cognizant of the issue of extraction bias when interpreting results.
Finally, a key assumption of techniques based on analysis of community DNA extracted from soils is that the DNA has been extracted from living cells. Recent evidence has shown that DNA can persist in soils for significant periods of time by binding to soil particles and that this binding can protect free DNA in soils from DNase degradation. Thus, it is possible that some of the DNA extracted from soil could actually have come from cells that are no longer living in the soil being analyzed. The detection of DNA from dead cells, refered to as “relic DNA,” can be a confounding factor in studies of microbial community structure (Carini et al. 2017). Therefore, both of these limitations, extraction bias and DNA persistence in soils, should be considered when interpreting results from community DNA‐based analyses.
3.5.4 Analysis of Community DNA
DNA extracted directly from soil samples can be analyzed by a wide array of different techniques. The most widely used require amplification of target genes (e.g. 16S rRNA genes) from community DNA via the polymerase chain reaction (PCR). PCR‐based techniques include clone library sequencing, fingerprinting techniques such as T‐RFLP and denaturing gradient gel electrophoresis (DGGE), and more recently a variety of high‐throughput amplicon sequencing approaches. There are also a number of techniques for the analysis of community DNA that do not require PCR amplification of target genes, including metagenomics. Each of these approaches will be discussed below.
3.6 PCR‐Based Methods
PCR‐based methods analyze single gene targets from within community DNA by amplifying target genes via PCR prior to downstream analyses. As described above, the 16S rRNA gene is the most common target for analyses of prokaryotic community diversity, but a variety of other genes have been used as well. For example, functional genes associated with specific metabolic pathways, e.g. genes encoding enzymes involved in processes such as ammonia oxidation, denitrification, and sulfate reduction, are also popular targets for the PCR‐based methods described below.
3.6.1 Clone Library Sequencing
A phylogenetic survey can be conducted for a soil bacterial community by extracting total community DNA directly from a soil sample and amplifying the 16S rRNA genes from within the community DNA via PCR using universal primers that have been designed to target the 16S rRNA genes from all bacteria (e.g. Borneman et al. 1996). The amplicons produced are then inserted into cloning vectors (plasmids designed to carry PCR amplicons) and these cloning vectors are inserted into bacterial cells (often Escherichia coli) that have been pretreated (e.g. electroporation or chemical treatment) to enable plasmid uptake. This collection of cells containing amplified 16S rRNA genes within plasmid vectors is known as a clone library, and the 16S rRNA genes within this library can be sequenced to produce a random survey of the 16S rRNA genes found within the original community. This technique is known as clone library sequencing. The sequenced genes can be assigned to taxa by comparison to Genbank or the RDP database (as described above) or they can be binned into operational taxonomic units (OTUs) based on degree of sequence similarity. The collection of taxa or OTUs within a clone library can then be used to calculate community diversity (e.g. Dunbar et al. 1999) or to compare communities via multivariate techniques such as PCA (e.g. Hur and Chun 2004) or MDS (e.g. Borneman et al. 1996).
3.6.1.1 Value and Limitations of Clone Library Sequencing
Clone library sequencing has significantly increased our understanding of microbial diversity by revealing much greater diversity than was detected by classical culture‐based studies. However, clone library sequencing is time consuming, labor intensive, and expensive. Although the cost and time required for large scale sequencing projects are decreasing quickly, the incredible diversity of microbial communities makes it impractical at this point to attempt to sequence every amplicon produced from PCR amplification of an environmental sample. For example, soil from a beech forest showed total bacterial counts of 1.5 × 1010 cells g soil−1 and approximately 4000 unique bacterial genomes based on DNA reassociation analysis (Torsvik et al. 1990), and a more recent study used reassociation kinetics to determine that a pristine soil contained more than one million distinct genomes. Attempting to capture all of this biodiversity by sequencing a clone library is impractical. Therefore, although phylogenetic inventories of microbial communities based on cloning and sequencing 16S rRNA genes can provide useful information, these inventories will generally be incomplete, so populations comprising very small fractions of the overall community may not be detected. In addition, clone library sequencing is too slow to be useful for routine profiling and comparison of large numbers of environmental samples.
3.6.2 DNA‐Based Fingerprinting Techniques
As an alternative to cloning and sequencing, techniques such as DGGE and terminal restriction fragment length polymorphism analysis (T‐RFLP) can be used to analyze amplicons produced by PCR amplification of DNA extracted from an entire microbial community. These techniques generate profiles for microbial communities based on differences in the gene sequences of their constituents. Specifically, these techniques involve separation of amplicons on an acryamide electrophoresis gel. In DGGE the amplicons are separated based on sequence differences that lead to differences in the denaturing properties of the amplicons, and in T‐RFLP the amplicons are separated based on sequence differences that lead to differences in the sizes of fragments produced after digestion of amplicons with restriction enzymes. Both DGGE and T‐RFLP produce patterns of bands on a gel, with the bands representing distinct genotypes within a community of organisms. The presence or absence of specific bands and the intensity of the bands produced from a community (i.e. the fingerprint) can be compared between different communities and can provide insight into differences in microbial population structure between different soil habitats or changes in soil microbial population structure over time or with different experimental treatments.
3.6.2.1 Value and Limitations of DNA‐Based Fingerprinting Techniques
DNA based fingerprinting techniques, including DGGE and T‐RFLP, can be effective tools for comparisons of microbial communities across sites or treatments, and they can both be used to determine if significant differences in community composition exist. Multivariate ordination techniques such as PCA (e.g. Janus et al. 2005) or MDS (e.g. Baniulyte et al. 2009) can be used with either DGGE or T‐RFLP data to compare profiles. DGGE and T‐RFLP can be used with any gene target, including genes used as phylogeneitc markers (e.g. 16S rRNA genes) or functional genes, as long as suitable primers exist. DGGE and T‐RFLP both also have very high resolution, as they are able to discriminate amplicons differing by a single base. Finally, both DGGE and T‐RFLP are relatively inexpensive to run, although they do require specialized equipment. However, DGGE and T‐RFLP have two key limitations. The first is that it is not possible to identify the taxonomic affiliations of the bands within the profiles, or to determine their specific DNA sequences. So while these techniques can be used to determine if soil microbial communities are significantly different in composition, they cannot identify which specific taxa are differentially distributed across the communities. Second, both DGGE and T‐RFLP significantly underestimate the total diversity of a community and are biased toward the most numerically dominant organisms within a community. Therefore, researchers should be aware of these limitations before selecting these techniques and should consider these limitations when intepreting their results.
3.6.3 High‐Throughput Amplicon Sequencing
In the late 1990s, efforts to sequence the human genome led to the development of DNA sequencing technologies that were significantly higher throughput than the traditional Sanger sequencing approach. In 2005 a pyrosequencing‐based system was introduced that provided a 100‐fold increase in throughput (i.e. base reads per unit time) over Sanger sequencing (Margulies et al. 2005), and in 2006 Sogin et al. first demonstrated the use of this system for the simultaneous sequencing of a pool of 16S rRNA gene amplicons produced from an entire microbial community without the need to insert amplicons into cloning vectors (Sogin et al. 2006). In their study, Sogin et al. applied this approach to the analysis of marine microbial communties, and showed that it could produce 100–1000‐fold more sequences per sample than a traditional clone library approach (Sogin et al. 2006). These groundbreaking studies led to a revolution in the field of soil microbiology, enabling vastly more extensive analyses of soil microbial community composition and diversity than had been possible previously.
Another key innovation that was developed for high‐throughput sequencing‐based profiling of microbial communities was the use of barcoding. Barcodes are short DNA sequences (generally eight bases) that are designed in silico and attached to the 5′ ends of PCR primers, resulting in the integration of these barcodes into amplicons during the PCR reaction. If distinct barcodes are attached to amplicons from different samples, amplicons from multiple samples can be pooled and sequenced together, and the sequences from individual samples can later be separated based on these barcodes (Binladen et al. 2007). This barcoding strategy enables researchers to analyze large numbers of samples in parallel to whatever level of sequencing depth (i.e. number of sequences per sample) is desired.
Since the introduction of pyrosequencing, newer high‐throughput sequencing platforms have continued to emerge, and microbiologists have applied these new platforms to the analysis of 16S rRNA gene amplicons. Examples include Illumina (Caporaso et al. 2010; Lazarevic et al. 2009), Ion Torrent (Brown et al. 2013; Whiteley et al. 2012) and PacBio (Mosher et al. 2013; Schloss et al. 2016). The common feature of these newer sequencing platforms (sometimes refered to as “next‐generation sequencing”) is that they provide much higher throughput than Sanger sequencing (i.e. more bases sequenced per unit time) at a much lower cost per base, but each of these platforms has specific advantages and disadvantges. Researchers interested in using one of these platforms to analyze soil microbial community diversity should carefully consider factors including cost, speed, read length, and accuracy.
The incredible increase in the amount of sequence data provided by high‐throughput sequencing has revolutionized our understanding of microbial community diversity, but it has also necessitated the development of more advanced computational tools to analyze these incredibly large data sets. Specific tasks that must be accomplished for analysis of high‐throughput sequencing data sets include separation of sequences by sample based on barcodes, removal of low quality sequences, alignment, binning, and taxonomic assignment of sequences based on comparison to available databases. Many researchers have developed computational tools to accomplish these individual tasks, but two comprehensive software packages have been developed specifically for the analysis of 16S rRNA gene amplicon libraries produced via high‐throughput sequencing: mothur (Schloss et al. 2009) and QIIME (Caporaso et al. 2010). These two packages are freely available to researchers, and both are widely used in the field of soil microbiology. Both mothur and QIIME can accomplish the basic tasks listed above, but they differ in the specific methods used to accomplish these tasks. Researchers interested in the analysis of 16S rRNA gene amplicon libraries produced via high‐throughput sequencing are encouraged to explore both of these tools as possible options for data analysis.
3.6.4 Limitations of PCR‐Based Methods
It should be noted that there are some significant limitations to techniques (such as those described above) that are based on PCR amplification of DNA extracted from soil. One assumption of these techniques is that the DNA from all organisms present in the sample will be amplified with the same efficiency. This assumption may not be correct in all cases, and in some cases the process may favor the amplification of DNA from certain organisms. Biases in amplification efficiency can be caused by differences in cell lysis or DNA extraction efficiency, differences in gene copy number within the organisms or differences in the efficiency of the PCR reaction itself. PCR bias can result in misrepresentation of phylogenetic diversity in phylogenetic surveys and T‐RFLP and DGGE profiles, and thus PCR bia can be problematic in microbial ecological studies. Researchers applying PCR‐based approaches to the analysis of soil microbial community diversity need to be cognizant of these limitations.
3.7 Metagenomics
Metagenomics can be defined as the simultaneous analysis of the genomes of a population of microorganisms (Handelsman 2004). Metagenomics seeks to analyze not only the taxonomic composition of microbial communities based on marker genes such as the 16S rRNA gene (as described above), but also to document the collection of genes present within an entire microbial community. Metagenomics does not require PCR amplification, so it avoids some of the limitations of PCR‐based approaches that are described above.
The earliest metagenomic studies were based on direct isolation of community DNA from an environmental sample (as described above) followed by random insertion of fragments of this extracted DNA into cloning vectors and transformation of the clones into host bacterial cells. These metagenomic clone libraries were then either sequenced randomly in a process known as shotgun metagenomics (e.g. Tyson et al. 2004; Venter et al. 2004), or the clones were first screened for the presence of phylogenetic indicator genes (e.g. 16S rRNA genes) or for the presence of genes expressing specific functions (e.g. antibiotic resistance; Handelsman 2004) and then selected clones were sequenced. The first metagenomic studies were conducted on marine samples (Stein et al. 1996), and metagenomic studies of soil microbial communities soon followed (Henne et al. 1999; Rondon et al. 2000). Metagenomic studies of soil microbial communities have provided a wealth of insights into the functional potential present within these communities, including the discovery of novel antibiotics (Gillespie et al. 2002), antibiotic resistance genes (Riesenfeld et al. 2004; Torres‐Cortés et al. 2011), and degradative enzymes (Henne et al. 1999, 2000).
The earliest metagenomic studies (including those listed above) relied on Sanger sequencing of clones, and were thus limited by the capabilities of that sequencing platform. Advances in high‐throughput sequencing technology (as discussed above) have enabled direct, untargeted sequencing of DNA isolated from environmental samples without the need for cloning. This approach has been applied to the analysis of soil microbial communities (e.g. Luo et al. 2014; Mackelprang et al. 2011; Yergeau et al. 2010). For high‐throughput metagenomic studies, the DNA isolated from an environmental sample is first randomly fragmented, using either enzymatic or physical fragmentation, to produce a sequencing library, and this library is then sequenced using a high‐throughput sequencing platform. The Illumina HiSeq platform is the most popular for metagenomic studies, due to its high output and high accuracy, but other technologies are beginning to be adopted as well (Quince et al. 2017). After sequence data are obtained and low‐quality data removed, the first step involved in analysis is assembly of short reads into larger contigs. In some cases, including samples with low diversity and a high degree of sequencing depth, it may be possible to assemble contigs into complete bacterial genomes, but for most soil microbial communities, this will not be possible for the vast majory of the data obtained. In this case the assembled contigs can be compared to publically available databases to identify phylogenetic indicator genes (e.g. 16S rRNA genes) or genes associated with specific functions (e.g. antibiotic resistance) or metabolic pathways. Numerous approaches for computational analysis of metagenomic data are currently available, with new tools continuously being developed. The choice of appropriate tools and strategies can be challenging (see discussion of limitations below), but these choices should always be based on the goals of the study.
3.7.1 Limitations of Metagenomics
A key parameter in metagenomic analysis is the sequencing depth, i.e. the amount of base‐pairs of data produced per sample. There is no simple answer to how much sequencing is enough, but the amount of sequencing depth required for a specific study will depend on the complexity of the community being analyzed and the questions being asked. In general, deeper sequencing will provide better coverage and higher liklelihood of retrieving sequences from low‐abundance members of a community, but deeper sequencing increases costs significantly and creates challenges associated with analysis of massive data sets. Currently, microbiome studies tend to generate between 1 and 10 Gb of data (Quince et al. 2017), but these values may change rapidly based on changes in sequencing technology and cost. At this time the costs for this level of sequencing are quite high and can limit access to this technique to well financed researchers. The high costs of metagenomic sequencing cannot also make it difficult for researchers to analyze enough samples and replicates to permit rigorous statistical analysis.
Metagenomic sequencing of complex microbial communities, such as those found in soils, generates unprecedented amounts of data, resulting in tremendous computational challenges. The power needed to analyze metagenomic sequence data sets far outstrips the capacity of even the most powerful desktop computers and generally requires the use of computing clusters, with many processors running in parallel. This type of computing power is expensive to buy and to manage, and thus computational needs can limit the ability of researchers at all but the most well‐financed institutions to work with metagenomic data. Furthermore, there are no simple plug‐and‐play pipelines available for rigorous analysis of metagenomic data. Rather, significant expertise is required to navigate and apply the multitude of tools that are available for metagenomic data analysis, most of which must be run in Unix‐ or Linux‐based systems. Therefore, the use of metagenomic sequencing for the analysis of complex microbial communities relies on the researchers either having advanced training in bioinformatics or collaborating with researchers with this expertise. Therefore, lack of appropriate expertise can also limit researchers' access to metagenomic tools.
3.8 Conclusions: Utility and Limitations of Diversity Analysis Procedures
A primary conclusion from our review of each of the major types of methods used to characterize soil microbial community diversity is that no single technique provides an all‐inclusive, definitive description of soil microbial community diversity. As discussed above, each of the methods has specific advantages and disadvantages, and researchers need to be cognizant of these when choosing a method for their project and when interpreting results from a project. Moreover, the specific goals of the research project should always be the key factor in choosing a method for analyzing microbial community diversity, as no single tool is optimal for every job.
References
1. Aira, M., Monroy, F., and Domínguez, J. (2000). Changes in microbial biomass and microbial activity of pig slurry after the transit through the gut of the earthworm Eudrilus eugeniae (Kinberg, 1867). Biol. Fertil. Soils 42: 371–376.
2. Baniulyte, D., Favila, E., and Kelly, J.J. (2009). Shifts in microbial community composition following surface application of dredged river sediments. Microb. Ecol. 57: 160–169.
3. Bardgett, R.D., Leemans, D.K., Cook, R., and Hobbs, P.J. (1997). Seasonality of the soil biota of grazed and ungrazed hill grasslands. Soil Biol. Biochem. 29: 1285–1294.
4. Behrendt, U., Muller, T., and Seyfarth, W. (1997). The influence of extensification in grassland management on the populations of micro‐organisms in the phyllosphere of grasses. Microbiol. Res. 152: 75–85.
5. Binladen, J., Gilbert, M.T.P., Bollback, J.P. et al. (2007). The use of coded PCR primers enables high‐throughput sequencing of multiple homolog amplification products by 454 parallel sequencing. PLoS One 2 (2): e197.
6. Borneman, J., Skroch, P.W., O'Sullifan, K.M. et al. (1996). Molecular microbial diversity of an agricultural soil in Wisconsin. Appl. Environ. Microbiol. 62: 1935–1943.
7. Bossio, D.A. and Scow, K.M. (1995). Impact of carbon and flooding on the metabolic diversity of microbial communities in soils. Appl. Environ. Microbiol. 61: 4043–4050.
8. Bossio, D.A. and Skow, K.M. (1998). Impacts of carbon and flooding on soil microbial communities – phospholipid fatty acid profiles and substrate utilization patterns. Microb. Ecol. 35: 265–278.
9. Bossio, D.A., Scow, K.M., Gunapala, N., and Graham, K.J. (1998). Determinants of soil microbial communities – effects of agricultural management, season, and soil type of phospholipid fatty acid profiles. Microb. Ecol. 36: 1–12.
10. Braid, M.D., Daniels, L.M., and Kitts, C.L. (2003). Removal of PCR inhibitors from soil DNA by chemical flocculation. J. Microbiol. Methods 52 (3): 389–393.
11. Broughton, L.C. and Gross, K.L. (2000). Patterns of diversity in plant and soil microbial communities along a productivity gradient in a Michigan old‐field. Oecologia 125: 420–427.
12. Brown, S.P., Callaham, M.A. Jr., Oliver, A.K., and Jumpponen, A. (2013). Deep Ion Torrent sequencing identifies soil fungal community shifts after frequent prescribed fires in a southeastern US forest ecosystem. FEMS Microbiol. Ecol. 86 (3): 557–566.
13. Caporaso, J.G., Kuczynski, J., Stombaugh, J. et al. (2010). QIIME allows analysis of high‐throughput community sequencing data. Nat. Methods 7 (5): 335.
14. Carini, P., Marsden, P.J., Leff, J.W. et al. (2017). Relic DNA is abundant in soil and obscures estimates of soil microbial diversity. Nat. Microbiol. 2 (3): 16242.
15. Carrigg, C., Rice, O., Kavanagh, S. et al. (2007). DNA extraction method affects microbial community profiles from soils and sediment. Appl. Microbiol. Biotechnol. 77 (4): 955–964.
16. Denton, C.S., Bardgett, R.D., Cook, R., and Hobbs, P.J. (1999). Low amounts of root herbivory positively influence the rhizosphere microbial community in a temperate grassland soil. Soil Biol. Biochem. 31: 155–165.
17. Dunbar, J., Takala, S., Barns, S.M. et al. (1999). Levels of bacterial community diversity in four arid soils compared by cultivation and 16S rRNA gene cloning. Appl. Environ. Microbiol. 65 (4): 1662–1669.
18. Ellis, R.J., Thompson, I.P., and Bailey, M.J. (1995). Metabolic profiling as a means of characterizing plant‐associated microbial communities. FEMS Microbiol. Ecol. 16: 9–17.
19. Feinstein, L.M., Sul, W.J., and Blackwood, C.B. (2009). Assessment of bias associated with incomplete extraction of microbial DNA from soil. Appl. Environ. Microbiol. 75 (16): 5428–5433.
20. Fontúrbel, M.T., Barreiro, A., Vega, J.A. et al. (2012). Effects of an experimental fire and post‐fire stabilization treatments on soil microbial communities. Geoderma 191: 51–60.
21. Fox, G.E., Pechman, K.R., and Woese, C.R. (1977). Comparative cataloging of 16S ribosomal ribonucleic acid: molecular approach to procaryotic systematics. Int. J. Syst. Evol. Microbiol. 27 (1): 44–57.
22. Frostegård, A., Bååth, E., and Tunlid, A. (1993). Shifts in structure of soil microbial communities in limed forests as revealed by phospholipid fatty acid analysis. Soil Biol. Biochem. 25: 723–730.
23. Frostegård, A., Tunleid, A., and Bååth, E. (1996). Changes in microbial community structure during long term incubation in two soils experimentally contaminated with metals. Soil Biol. Biochem. 28: 55–63.
24. Fuller, M.E., Scow, K.M., Lau, S., and Ferris, H. (1997). Trichloroethylene (TCE) and toluene effects on the structure and function of the soil community. Soil Biol. Biochem. 29: 75–89.
25. Garau, G., Castaldi, P., Santona, L. et al. (2007). Influence of red mud, zeolite and lime on heavy metal immobilization, culturable heterotrophic microbial populations and enzyme activities in a contaminated soil. Geoderma 142: 47–57.
26. Garland, J.L. (1996a). Patterns of potential C source utilization by rhizosphere communities. Soil Biol. Biochem. 28: 223–230.
27. Garland, J.L. (1996b). Analytical approaches to the characterization of samples of microbial communities using patterns of potential C source utilization. Soil Biol. Biochem. 28: 213–221.
28. Garland, J.L. and Mills, A.L. (1994). A community‐level physiological approach for studying microbial communities. In: Beyond the Biomass (eds. K. Ritz, J. Dighton and K.E. Giller), 77–83. London: Wiley‐Sayce.
29. Germida, J.J., Siciliano, S.D., Defreitas, J.R., and Seib, A.M. (1998). Diversity of root‐associated bacteria associated with field grown canola (Brassica napus L.) and wheat (Triticum aestivum L.). FEMS Microbiol. Ecol. 26: 43–50.
30. Gillespie, D.E., Brady, S.F., Bettermann, A.D. et al. (2002). Isolation of antibiotics turbomycin A and B from a metagenomic library of soil microbial DNA. Appl. Environ. Microbiol. 68 (9): 4301–4306.
31. Grayston, S.J., Wang, S.Q., Campbell, C.D., and Edwards, A.C. (1998). Selective influence of plant species on microbial diversity in the rhizosphere. Soil Biol. Biochem. 30: 369–378.
32. Haack, S.K., Garchow, H., Klug, M.J., and Forney, L.J. (1995). Analysis of factors affecting the accuracy, reproducibility, and interpretation of microbial community carbon source utilization patterns. Appl. Environ. Microbiol. 61: 1458–1468.
33. Handelsman, J. (2004). Metagenomics: application of genomics to uncultured microorganisms. Microbiol. Mol. Biol. Rev. 68 (4): 669–685.
34. Hartmann, M. and Widmer, F. (2006). Community structure analyses are more sensitive to differences in soil bacterial communities than anonymous diversity indices. Appl. Environ. Microbiol. 72: 7804–7812.
35. Henne, A., Daniel, R., Schmitz, R.A., and Gottschalk, G. (1999). Construction of environmental DNA libraries in Escherichia coli and screening for the presence of genes conferring utilization of 4‐hydroxybutyrate. Appl. Environ. Microbiol. 65 (9): 3901–3907.
36. Henne, A., Schmitz, R.A., Bömeke, M. et al. (2000). Screening of environmental DNA libraries for the presence of genes conferring lipolytic activity on Escherichia coli. Appl. Environ. Microbiol. 66 (7): 3113–3116.
37. Hur, I. and Chun, J. (2004). A method for comparing multiple bacterial community structures from 16S rDNA clone library sequences. J. Microbiol. 42: 9–13.
38. Janus, L.R., Angeloni, N.L., McCormack, J. et al. (2005). Elevated atmospheric CO2 alters soil microbial communities associated with trembling aspen (Populus tremuloides) roots. Microb. Ecol. 50 (1): 102–109.
39. Kelly, J.J. and Tate, R.L. III (1998). Effects of heavy metal contamination and remediation on soil microbial communities in the vicinity of a zinc smelter. J. Environ. Qual. 27: 609–617.
40. Kelly, J.J., Haggblom, M., and Tate, R.L. (1999a). Changes in soil microbial communities over time resulting from one time application of zinc: a laboratory microcosm study. Soil Biol. Biochem. 31 (10): 1455–1465.
41. Kelly, J.J., Haggblom, M., and Tate, R.L. (1999b). Effects of the land application of sewage sludge on soil heavy metal concentrations and soil microbial communities. Soil Biol. Biochem. 31 (10): 1467–1470.
42. Kelly, J.J., Favila, E., Hundal, L.S., and Marlin, J.C. (2007). Assessment of soil microbial communities in surface applied mixtures of Illinois River sediments and biosolids. Appl. Soil Ecol. 36 (2): 176–183.
43. King, J.D., White, D.C., and Taylor, C.W. (1977). Use of lipid composition and metabolism to examine structure and activity of estuarine detrital microflora. Appl. Environ. Microbiol. 33: 1177–1183.
44. Klamer, M. and Bååth, E. (1998). Microbial community dynamics during composting of straw material studied using phospholipid fatty acid analysis. FEMS Microbiol. Ecol. 27: 9–20.
45. Klingler, J.M., Stowe, R.P., Obenhuber, D.C. et al. (1992). Evaluation of the Biolog automated microbial identification system. Appl. Environ. Microbiol. 58 (6): 2089–2092.
46. Knight, B.P., McGrath, S.P., and Chaudri, A.M. (1997). Biomass carbon measurements and substrate utilization patterns of microbial populations from soils amended with cadmium, copper, or zinc. Appl. Environ. Microbiol. 63: 39–43.
47. Könneke, M. and Widdel, F. (2003). Effect of growth temperature on cellular fatty acids in sulphate‐reducing bacteria. Environ. Microbiol. 11: 1064–1070.
48. Lazarevic, V., Whiteson, K., Huse, S. et al. (2009). Metagenomic study of the oral microbiota by Illumina high‐throughput sequencing. J. Microbiol. Methods 79 (3): 266–271.
49. Leonard, K.J. and Czochor, R.J. (1980). Theory of genetic interactions among populations of plants and their pathogens. Annu. Rev. Phytopathol. 18: 237–258.
50. Luo, C., Rodriguez‐R, L.M., Johnston, E.R. et al. (2014). Soil microbial community responses to a decade of warming as revealed by comparative metagenomics. Appl. Environ. Microbiol. 80 (5): 1777–1786.
51. Lupwayi, N.Z., Rice, W.A., and Clayton, G.W. (1998). Soil microbial diversity and community structure under wheat as influenced by tillage and crop rotation. Soil Biol. Biochem. 30: 1733–1741.
52. Mackelprang, R., Waldrop, M.P., DeAngelis, K.M. et al. (2011). Metagenomic analysis of a permafrost microbial community reveals a rapid response to thaw. Nature 480 (7377): 368.
53. Macrae, A. (2000). The use of 16S rDNA methods in soil microbial ecology. Braz. J. Microbiol. 31: 77–82.
54. Margulies, M., Egholm, M., Altman, W.E. et al. (2005). Genome sequencing in microfabricated high‐density picolitre reactors. Nature 437 (7057): 376.
55. Martin‐Laurent, F., Philippot, L., Hallet, S. et al. (2001). DNA extraction from soils: old bias for new microbial diversity analysis methods. Appl. Environ. Microbiol. 67 (5): 2354–2359.
56. MIDI (1995). Sherlock Microbial Identification System Operating Manual: Version 5. Newark, DE: MIDI Inc.
57. Miller, J.M. and Rhoden, D.L. (1991). Preliminary evaluation of Biolog, a carbon source utilization method for bacterial identification. J. Clin. Microbiol. 29 (6): 1143–1147.
58. Mosher, J.J., Bernberg, E.L., Shevchenko, O. et al. (2013). Efficacy of a 3rd generation high‐throughput sequencing platform for analyses of 16S rRNA genes from environmental samples. J. Microbiol. Methods 95 (2): 175–181.
59. Moyer, C.L., Tiedje, J.M., Dobbs, F.C., and Karl, D.M. (1996). A computer‐simulated restriction fragment length polymorphism analysis of bacterial small‐subunit rRNA genes: efficacy of selected tetrameric restriction enzymes for studies of microbial diversity in nature. Appl. Environ. Microbiol. 62: 2501–2507.
60. Quince, C., Walker, A.W., Simpson, J.T. et al. (2017). Shotgun metagenomics, from sampling to analysis. Nat. Biotechnol. 35 (9): 833.
61. Riesenfeld, C.S., Goodman, R.M., and Handelsman, J. (2004). Uncultured soil bacteria are a reservoir of new antibiotic resistance genes. Environ. Microbiol. 6 (9): 981–989.
62. Rondon, M.R., August, P.R., Bettermann, A.D. et al. (2000). Cloning the soil metagenome: a strategy for accessing the genetic and functional diversity of uncultured microorganisms. Appl. Environ. Microbiol. 66 (6): 2541–2547.
63. Russek‐Cohen, E. and Colwell, R.R. (1986). Applications of Numerical Taxonomy Procedures in Microbial Ecology. In: Microbial Autecology. A Method for Environmental Studies (ed. R.L. Tate III), 133–146. New York: Wiley‐Interscience.
64. Rutgers, M., Wouterse, M., Drost, S.M. et al. (2016). Monitoring soil bacteria with community‐level physiological profiles using Biolog™ ECO‐plates in the Netherlands and Europe. Appl. Soil Ecol. 97: 23–35.
65. Schloss, P.D., Westcott, S.L., Ryabin, T. et al. (2009). Introducing mothur: open‐source, platform‐independent, community‐supported software for describing and comparing microbial communities. Appl. Environ. Microbiol. 75 (23): 7537–7541.
66. Schloss, P.D., Jenior, M.L., Koumpouras, C.C. et al. (2016). Sequencing 16S rRNA gene fragments using the PacBio SMRT DNA sequencing system. PeerJ 4: e1869.
67. Shannon, C.E. (2001). A mathematical theory of communication. ACM SIGMOBILE Mob. Comput. Commun. Rev. 5: 3–55.
68. Smalla, K., Wachtendorf, U., Heuer, H. et al. (1998). Analsysi of BIOLOG GN substrate utilization patterns by microbial communities. Appl. Environ. Microbiol. 64: 1220–1225.
69. Sogin, M.L., Morrison, H.G., Huber, J.A. et al. (2006). Microbial diversity in the deep sea and the underexplored “rare biosphere”. Proc. Natl. Acad. Sci. U.S.A. 103 (32): 12115–12120.
70. Stach, J.E., Bathe, S., Clapp, J.P., and Burns, R.G. (2001). PCR‐SSCP comparison of 16S rDNA sequence diversity in soil DNA obtained using different isolation and purification methods. FEMS Microbiol. Ecol. 36 (2–3): 139–151.
71. Stein, J.L., Marsh, T.L., Wu, K.Y. et al. (1996). Characterization of uncultivated prokaryotes: isolation and analysis of a 40‐kilobase‐pair genome fragment from a planktonic marine archaeon. J. Bacteriol. 178 (3): 591–599.
72. Tebbe, C.C. and Vahjen, W. (1993). Interference of humic acids and DNA extracted directly from soil in detection and transformation of recombinant DNA from bacteria and yeast. Appl. Environ. Microbiol. 59: 2657–2665.
73. Thompson, D.W., Stuckey, R.L., and Thompson, E.B. (1987). Spread, Impact, and Control of Purple Loosestrife in North American Wetlands. Washington, D.C: US Fish and Wildlife Service.
74. Thompson, I.P., Bailey, M.J., Ellis, R.J. et al. (1999). Response of soil microbial communities to single and multiple doses of an organic pollutant. Soil Biol. Biochem. 31: 95–105.
75. Timonen, S., Jorgensen, K.S., Haahtela, K., and Sen, R. (1998). Bacterial community structure at defined locations of Pinus sylvestris‐Suillus bovinus, and Pinus sylvestris‐Paxillus involutus mycorrhizospheres in dry pine forest humus and nursery peat. Can. J. Microbiol. 44: 499–513.
76. Torres‐Cortés, G., Millán, V., Ramírez‐Saad, H.C. et al. (2011). Characterization of novel antibiotic resistance genes identified by functional metagenomics on soil samples. Environ. Microbiol. 13 (4): 1101–1114.
77. Tsai, Y.‐L. and Olson, B.H. (1992). Rapid method for separation of bacterial DNA from humic substances in sediments for polymerase chain reaction. Appl. Environ. Microbiol. 58: 2292–2295.
78. Tyson, G.W., Chapman, J., Hugenholtz, P. et al. (2004). Community structure and metabolism through reconstruction of microbial genomes from the environment. Nature 428 (6978): 37.
79. Venter, J.C., Remington, K., Heidelberg, J.F. et al. (2004). Environmental genome shotgun sequencing of the Sargasso Sea. Science 304 (5667): 66–74.
80. Wang, L., Li, F., Zhan, Y., and Zhu, L. (2016). Shifts in microbial community structure during in situ surfactant‐enhanced bioremediation of polycyclic aromatic hydrocarbon‐contaminated soil. Environ. Sci. Pollut. Res. 23 (14): 14451–14461.
81. White, D.C., Davis, W.M., Nickels, J.S. et al. (1979). Determination of the sedimentary microbial biomass by extractable lipid phosphate. Oecologia 40: 51–62.
82. Whiteley, A.S., Jenkins, S., Waite, I. et al. (2012). Microbial 16S rRNA Ion Tag and community metagenome sequencing using the Ion Torrent (PGM) platform. J. Microbiol. Methods 91 (1): 80–88.
83. Winding, A. (1994). Fingerprinting bacterial soil communities using Biolog microtitre plates. In: Beyond the Biomass (eds. K. Ritz, J. Dighton and K.E. Giller), 85–94. London: Wiley‐Sayce.
84. Woese, C.R., Kandler, O., and Wheelis, M.L. (1990). Towards a natural system of organisms: proposal for the domains Archaea, Bacteria, and Eucarya. Proc. Natl. Acad. Sci. U.S.A. 87 (12): 4576–4579.
85. Yergeau, E., Hogues, H., Whyte, L.G., and Greer, C.W. (2010). The functional potential of high Arctic permafrost revealed by metagenomic sequencing, qPCR and microarray analyses. ISME J. 4 (9): 1206.
86. Zak, J.C., Willig, M.R., Moorhead, D.L., and Wildman, H.G. (1994). Functional diversity of microbial communities – a quantitative approach. 26: 1101–1108.
87. Zak, D.R., Ringelberg, D.B., Pregitzer, K.S. et al. (1996). Soil microbial communities beneath Populus grandidentata grown under elevated atmospheric CO2. Ecol. Appl. 6: 257–262.
88. Zhou, J., Bruns, M.A., and Tiedje, J.M. (1996). DNA recovery from soils of diverse composition. Appl. Environ. Microbiol. 62: 316–322.