Chapter 11. Fungal Diversity in Molecular Terms: Profiling, Identification, and Quantification in the Environment

DK3133_C011.fm Page 216 Monday, April 18, 2005 12:56 PM



216



Bidartondo and Gardes

Animals



50,000

45,000



Plants



40,000

35,000



Bacteria

30,000

25,000

20,000

Fungi



15,000

10,000

5,000



02

20



01

20



00

20



99

19



98

19



97

19



96

19



95

19



19



94



0



Figure 11.1 Cumulative taxonomic nodes (families, genera, and species) in GenBank on a yearly

basis.



et al., 2002). These methods, although popular, are ultimately limited because the sporebearing structures they rely upon are inconsistently produced or detectable and the correlation between reproductive structures and vegetative structures is sometimes poor (Clapp

et al., 1995; Gardes and Bruns, 1996). Furthermore, the classical methods often overlook

asexual, cryptic, and obligately biotrophic species, all of which are exceptionally common

in fungi. Axenic isolation-dependent methods to assess diversity are highly selective; only

ca. 17% of the 70,000 described species of fungi have been successfully cultured (Hawksworth, 1991). To overcome the limitations imposed by morphology and culturability, and

to obtain a comprehensive view of fungal organismal biology and fungal communities,

ecologists have been intensively using molecular methods since the 1990s; these methods

have been applied successfully to vegetative structures, as well as reproductive structures,

in many complex natural environments. The information gained can provide the basis for

the statistical analysis of fungal communities to infer the processes that regulate the

maintenance of diversity. However, because fungi (like bacteria or insects) are too diverse

to count, exhaustive inventories of community richness remain impractical. Accordingly,

community-level studies typically address how diversity or the abundance of selected

species changes in relation to the abiotic, biotic, and temporal heterogeneity of the environment within a comparative framework.

DNA-based approaches have confirmed the extensive diversity of fungal communities and revealed previously undetected diversity (Redecker, 2000; Berch et al., 2002;

Vandenkoornhuyse et al., 2002; Schadt et al., 2003). The discovery of genomic regions

with DNA sequences conserved across entire clades and flanking variable regions led to

the design of universal primers for the polymerase chain reaction (PCR). Five major

strategies have since been developed to qualitatively and quantitatively analyze the composition of environmental samples containing multiple lineages, and they are consequently

emphasized in this chapter:



DK3133_C011.fm Page 217 Monday, April 18, 2005 12:56 PM



Fungal Diversity in Molecular Terms



1.



2.

3.



4.



5.



217



PCR products can be separated electrophoretically to obtain whole-community

genetic profiles and putative lineage assignments (e.g., via DGGE, RFLP, or TRFLP; see Table 11.1).

PCR products can be sequenced and unknown taxa assigned to a phylogenetic

clade.

Taxon-specific oligonucleotide probes can be designed and used in hybridization

experiments against target molecules to assess diversity in environmental samples (e.g., oligonucleotide fingerprinting, microarrays).

PCR can be used to quantify the concentration of DNA to estimate the biomass

of a target lineage via end-point or real-time automated techniques (i.e., competitive PCR, quantitative PCR).

Sequence-specific fluorescent probes can be used for direct absolute counts of

the organisms in a target clade via FISH, or PCR can be combined with in situ

hybridization within intact cells or tissues to attain the same goal.



This review provides an overview of the advantages as well as the technical limitations of both biochemical and DNA-based approaches used in fungal molecular ecology

and discusses recent methodological developments (i.e., microarrays) not yet applied to

fungi. For a glossary of terms and acronyms used in fungal molecular ecology and

throughout this chapter, see Table 11.1.



11.2



BIOCHEMICAL METHODS FOR ENVIRONMENTAL

MONITORING



11.2.1

Substrate Utilization Assays Using Microtitration Plates

Differences in physiological and metabolic activities between fungal species can be determined as patterns of substrate–source utilization (most often C sources) using 96-well

microtitration plates originally developed for the identification of single bacterial strains.

There are three main ways in which fungal characterizations have been carried out to

assess biochemical and community diversity using substrate utilization profiles:

1.

2.



3.



Single fungal strains have been used to inoculate a microtitration plate and each

well has been scored for fungal growth after incubation (Wildman, 1995).

Soil fungal community-wide metabolic potential has been monitored by turbidimetry, which is correlated with biomass accumulation, in the presence of

antibacterials (Buyer et al., 2001) and using commercially available fungal

plates (SFN2 and SFP2, BIOLOG Inc.) (Classen et al., 2003).

Fungal activity has been estimated in the presence of antibacterials after inoculation with decomposing litter (Dobranic and Zak, 1999) or soil organic matter

(Sobek and Zak, 2003) by adding a dye (dimethylthiazolyl–diphenyltetrazolium

bromide (MTT)) that is reduced by fungi.



The fungal communities that develop in individual wells have not been characterized, but

they probably comprise fast-growing fungi that can grow in liquid and on a single substrate

(e.g., many yeasts). Thus, it is unlikely that substrate utilization assays will provide a

comprehensive view of community diversity, either functional or phylogenetic. However,

they will remain useful to identify culturable plant pathogens, clinically important fungi,

and strains used by industry.



DK3133_C011.fm Page 218 Monday, April 18, 2005 12:56 PM



218



Bidartondo and Gardes



Table 11.1 Glossary of Terms and Acronyms Used in Fungal Molecular Ecology

Term

AFLP

ARDRA

BIOLOG

Calibrator

CGA

Chimera

Competitive PCR

DGGE

Exogenous control

FAME

FGA

GC-FAME

Heteroduplex

Hot start



Immunocapture PCR



Inhibition

Internal positive control

ITS

Metagenomics

Microarray

MPN

Multiplex PCR

Nested PCR

OFRG

Phylo-placement



PLFA

POA



Definition

Amplified fragment length polymorphism.

Amplified ribosomal DNA restriction analysis.

A commercially available microtitration plate for substrate utilization

assays.

A sample used as the basis for comparative results.

Community genome (micro)array; it is constructed using genomic

DNA isolated from axenic strains.

A sequence generated when an incompletely extended PCR product

acts as a primer on a heterologous sequence.

End-point PCR quantification using coamplification of an internal

standard.

Denaturing gradient gel electrophoresis.

DNA added to samples at a known concentration.

Fatty acid methyl ester.

Functional gene (micro)array; it contains DNA encoding enzymes

involved in ecosystem processes (e.g., nitrogen cycling).

Gas chromatography fatty acid methyl ester analysis.

A sequence formed by cross-hybridization of heterologous

sequences.

A method to avoid annealing during PCR until a threshold

temperature is reached in order to minimize nonspecific annealing,

usually by using a reversibly inactivated polymerase.

Template enrichment by binding antibody–cell complexes to

magnetic beads; recovery from solution with a magnet, DNA

isolation, and PCR.

PCR failure due to the presence of extraneous chemicals despite the

presence of sufficient template.

DNA used to distinguish true target negatives from PCR inhibition,

or to normalize for differences in sample extraction.

Internal transcribed spacer of the ribosomal repeat.

Archiving environmental DNA in genomic libraries (up to 100 kb)

without cultivation or PCR.

Orderly arrangement of large sets of DNA molecules of known

sequence for hybridization experiments.

Most probable number, a statistical method for estimating abundance.

PCR using more than one primer pair.

Using PCR products as templates for a second PCR.

Oligonucleotide fingerprinting of ribosomal genes.

Identification using phylogenetic methods of the least inclusive

monophyletic clade to which a DNA sequence of unknown origin

belongs.

Phospholipid fatty acid analysis.

Phylogenetic oligonucleotide (micro)array or “PhyloChip.” Contains

oligonucleotide probes derived primarily from ribosomal gene

segments.



DK3133_C011.fm Page 219 Monday, April 18, 2005 12:56 PM



Fungal Diversity in Molecular Terms



219



Table 11.1 Glossary of Terms and Acronyms Used in Fungal Molecular Ecology (Continued)

Term

Real time

Reporter gene

RFLP

Selectivity

Sensitivity

Specificity

SSCP

Standard

TGGE

T-RFLP

TSOP



Definition

Online quantification during logarithmic phase of PCR.

A gene whose protein product is quantified or visualized.

Restriction fragment length polymorphism.

Inclusion of all intended target templates (= universality).

Minimum number of detectable templates.

Exclusion of all unintended target templates.

Single-stranded conformational polymorphism.

A sample of known concentration used to construct a standard curve.

Temperature gradient gel electrophoresis.

Terminal restriction fragment length polymorphism.

Taxon-specific oligonucleotide probe.



11.2.2

Profiling Using Phospholipids (PLFAs or FAMEs)

Phospholipids are essential components of cell membranes, and they rapidly degrade after

cell death via enzymatic action. In the lipid analysis method, phospholipids are extracted

using various solvents and then fractionated into different categories. Phospholipid fatty

acids (PLFAs) that are liberated from the phospholipid fraction are transformed to fatty

acid methyl esters (FAMEs), which are identified and quantified using gas chromatography.

This biochemical method has been used to monitor shifts in overall microbial community

structure and in community subsets such as bacteria, actinomycetes, and fungi in soil

(Zelles, 1999; Myers et al., 2001). Measurements of stable isotope (as 13C) incorporation

into PLFAs can also be used to detect changes in metabolic activities of organisms (Arao,

1999). However, relatively little information has been compiled so far on fatty acid

composition of fungi (Graham et al., 1995; Madan et al., 2002), and it is unlikely that

phospholipids will provide detailed pictures of fungal community structure in the future.

Because many fatty acids are common to different species of fungi, they have limited

value for identification of the organisms from which they derive, especially when complex

fungal communities are analyzed (Olsson, 1999; Zelles, 1999). Another important limitation is that phospholipid concentration can change with the physiological state of the

organism in response to environmental factors (Olsson, 1999; Zelles, 1999). In summary,

this technique appears useful as a complementary approach for gross characterization of

communities (Hedlund, 2002; Olsson et al., 2003) or as a living biomass estimator for

subsets of fungi in laboratory experiments (Olsson et al., 1998; Olsson, 1999; Wallander

et al., 2001).



11.3



DNA-BASED METHODS FOR ANALYZING COMMUNITY

DIVERSITY



In the 1980s, emphasis was placed on restriction fragment length polymorphism (RFLP)

patterns of ribosomal genes to characterize fungal strains using DNA hybridization. Axenic

isolates were often used because large amounts of pure DNA were required to perform

the analysis. The discipline of molecular ecology emerged in the 1990s with the develop-



DK3133_C011.fm Page 220 Monday, April 18, 2005 12:56 PM



220



Bidartondo and Gardes



ment of PCR technology, which considerably increased the sensitivity and specificity of

all molecular tools (Mullis and Faloona, 1987). Within a few years, progress was made

in the development of fungal primers (White et al., 1990) and application of PCR to

environmental samples for community analysis without the need for cultivation (Gardes

et al., 1991; Henrion et al., 1992; Simon et al., 1992; Gardes and Bruns, 1993). In principle,

these techniques can be applied to any gene. However, ribosomal DNA molecules have

been the most popular in fungal ecology for three main reasons:

1.

2.

3.



They are present and typically in multiple identical copies within all organisms.

They comprise highly conserved sequence domains that have allowed for the

design of fungal-specific primers (White et al., 1990; Gardes and Bruns, 1993).

Their sequence variation spans various hierarchical levels of phylogenetic diversity, from species to kingdoms.



More than a decade later, numerous complementary techniques are now available to assess

the diversity of fungal communities. We will examine the principles, advantages, and

limitations of PCR-based methods and illustrate their applications in community studies

with a few selected examples. A summary is provided in Table 11.2.

11.3.1

Rapid Profiling Techniques

Fingerprinting techniques have been developed to profile mixed fungal populations in a

quick, reliable, and cost-efficient way. They all rely on extraction of environmental DNA,

amplification of targeted sequences, and electrophoresis of PCR products in an agarose

or polyacrylamide gel to generate genetic fingerprints. Differential migration of the amplified fragments depends on their length, their sequence and/or their conformation. Unless

they are automated, rapid profiling methods produce results that can vary from one

experiment to another and do not allow for the development of high-resolution databases.

11.3.1.1



Profiling Based on Size Differences of PCR Products after Restriction

Digestion: PCR-RFLP and T-RFLP



The earliest molecular attempts to analyze fungal communities relied heavily upon PCR

amplification of the internal transcribed region of the nuclear ribosomal unit (internal

transcribed spacer, ITS), followed by digestion of the amplified product using a few

tetrameric restriction enzymes, and finally electrophoretic separation of the fragments. The

patterns obtained from different species are polymorphic due to mutations in the restriction

sites and to base insertions or deletions (indels) within the amplified fragments. In fungi,

the ITS consists of two noncoding spacers, ITS1 and ITS2, separated by the highly

conserved 5.8S rRNA gene. It is typically amplified using either universal eukaryotic or

fungal-specific primers targeted to conserved regions in the 3′ end of the 18S rRNA gene

(small subunit, SSU) and the 5′ end of the 28S rRNA gene (large subunit, LSU). The ITS

region has been extensively used in fungal ecology because a low level of intraspecific

variation and a high level of interspecific variation were empirically observed. Thus, for

example, most studies of ectomycorrhizal fungal communities have involved the following

steps: (1) PCR amplification of the fungal symbiont from individual root tips, (2) digestion

of the amplified DNA product, (3) separation of the restriction fragments by electrophoresis, and (4) comparisons of patterns with restriction fragment length polymorphism databases from sporocarps (for a review, see Horton and Bruns, 2001). This relatively simple

approach is also commonly used as an intermediate screening step prior to sequencing

PCR clones obtained from soil DNA samples (Viaud et al., 2000; Chen and Cairney, 2002).



DK3133_C011.fm Page 221 Monday, April 18, 2005 12:56 PM



Fungal Diversity in Molecular Terms



221



Table 11.2 Advantages and Limitations of PCR-Based Fingerprinting Techniques

Method



Acronym



Denaturing

gradient gel

electrophoresis



DGGE



Temperature

gradient gel

electrophoresis

Single-stranded

conformational

polymorphism

Restriction

fragment length

polymorphism

Terminal RFLP



TGGE



Sequencing and

phylogenetic

affiliation

Oligonucleotide

fingerprinting



—



Microarrays for

detection



Advantages and Limitations

A rapid profiling technique for detecting community changes;

does not provide phylogenetic information directly; a lowresolution alternative or a screening tool for sequence-based

methods

Idem as DGGE



SSCP



Idem as DGGE; possibility of automation and standardization

using fluorescence-based technologies



RFLP



A rapid profiling technique extensively applied in fungal

ecology using ribosomal genes and spacers; a low-resolution

alternative or a screening tool for sequence-based methods

A modification of RFLP using fluorescence-based

technologies with higher throughput, resolution, and

reproducibility; a lower-resolution alternative or a screening

tool for sequence-based methods

The state of the art in community analysis; limited by the

pitfalls of PCR and the phylogenetic saturation of DNA

databases

Hybridization-based identification that permits taxonomic

discrimination using taxon-specific oligonucleotide probes

(TSOPs); reliable and cost-effective, but its optimization is

labor-intensive

Potentially the most powerful hybridization-based

identification system; limited by (1) the design and testing of

probes, (2) the sensitivity and specificity of the hybridization

step, and (3) the processing of results



T-RFLP



TSOP



POA

FGA

CGA



The accuracy of RFLPs is highest when they are automated and run on polyacrylamide gels using PCR products generated with fluorescently labeled nucleotides (Peter et

al., 2001; Kennedy et al., 2003). A recent modification to increase the resolution of the

RFLP technique for multitemplate sample screening involves the use of fluorescently

tagged oligonucleotide primers for PCR amplification and subsequent automated electrophoresis of the fragments generated. Only the end-labeled fragments are detected, which

reduces the complexity of the profile. This technique, referred to as T-RFLP (terminal

restriction fragment length polymorphism) is a rapid and sensitive approach that can be

used to assess the similarities between communities (Liu et al., 1997). However, the

richness and evenness of a community are only qualitatively estimated, and they depend

on the restriction enzyme used to derive the profile. In order to select optimal restriction

enzymes, optimization of the screening step is possible using computer-simulated RFLP

analysis of DNA sequence database accessions (Moyer et al., 1996; Marsh et al., 2000;

Lord et al., 2002). A basic limitation of T-RFLP is that all bands are counted equivalently,

and consequently the phylogenetic disparity among DNA sequences is unknown. In addi-



DK3133_C011.fm Page 222 Monday, April 18, 2005 12:56 PM



222



Bidartondo and Gardes



tion, its interpretation is problematic for organisms that contain different rDNA repeats

within individuals (i.e., arbuscular mycorrhizal fungi; Vandenkoornhuyse et al., 2003).

Another potential problem is the formation of pseudoterminal restriction fragments (Egert

and Friedrich, 2003).

To gain insight into the phylogenetic composition of the community, terminal restriction fragments can be recovered from gels, cloned, and sequenced (Nagashima et al.,

2003). Alternatively, the construction of single-strain profile databases directly from specimens, or indirectly from DNA sequences retrieved from databases if their amplification

primers are known, is an efficient approach. In summary, considering the high-throughput

capacity and sensitivity of T-RFLP (Moeseneder et al., 1999), it is likely that it will become

one of the most useful rapid profiling methods for the study of complex fungal communities

in the near future. It has been used already to detect and identify fungal mycelia in forest

soils and decaying plant leaves (Buchan et al., 2002; Dickie et al., 2002; Klamer et al.,

2002; Nikolcheva et al., 2003), and for the identification of ectomycorrhizal fungi from

mycorrhizae and sporocarps (Zhou and Hogetsu, 2002; Nara et al., 2003).

11.3.1.2



Profiling Based on Differences in Melting Temperatures: DGGE

and TGGE



In denaturing gradient gel electrophoresis (DGGE) and temperature gradient gel electrophoresis (TGGE), DNA fragments of the same size are separated through a linear gradient

of increasing denaturant (chemical or heat, respectively). In such environments, doublestranded DNA molecules partially unwind and undergo both a conformational and mobility

change. For two fragments with the same sequence length, the melting temperature (Tm)

of the DNA molecule is determined by the proportion and position of G + C bases. Both

techniques are considered interchangeable (Heuer et al., 1997), and they are both capable

of separating nearly identical sequences. Community patterns can subsequently be interpreted either by comparison with reference patterns or phylogenetically when individual

bands in the profile are assigned to taxonomic units by hybridization or sequencing. A

significant amount of work must be undertaken to optimize the separation of the PCR

products before screening, and relatively long GC-tailed primers must be used to stabilize

the transitional molecule. In addition, a single band in community profiles may be composed of several unrelated sequences (Schmalenberger and Tebbe, 2003). Such comigrating

fragments are relatively common when environmental samples are analyzed. Resolution

is also affected by the gel staining procedure and the electrophoretic conditions. To improve

resolution, the community can be divided into smaller phylogenetic groups using taxonspecific primers that are then individually analyzed.

DGGE and the related technique TGGE have been increasingly popular tools to

qualitatively describe bacterial communities since the first report of DGGE profiles of

bacterial mats and biofilms (Muyzer et al., 1993). They offer a relatively good compromise

between the number of samples processed and the information that can be obtained

(Muyzer and Smalla, 1998). One of the first applications of DGGE to fungal communities

was a study of pathogens in plant roots (Kowalchuk et al., 1997). In this study, a nested

PCR approach was used to specifically recover fungal 18S rDNA sequences from infected

roots. The amplified products were subjected to DGGE analysis, and the major bands

excised, reamplified, cloned, sequenced, and included in a phylogenetic analysis. The

sequence data revealed fungal diversity not detected in previous isolation-based surveys,

and the presence of several taxa not related to the existing fungal 18S rDNA sequences

in the databases. In a related study on the same plant and habitat, the DGGE strategy

proved useful in comparing communities of arbuscular mycorrhizal fungi, as well as



DK3133_C011.fm Page 223 Monday, April 18, 2005 12:56 PM



Fungal Diversity in Molecular Terms



223



comparing the arbuscular mycorrhizal fungi detected in spore banks vs. living plant roots

(Kowalchuk et al., 2002). A similar procedure was developed by Smit et al. (1999) using

TGGE to compare fungal diversity in a wheat rhizosphere vs. bulk soil in a microcosm

experiment. These rapid profiling methods have also been used to analyze fungal communities in other complex habitats, such as historical church window glass (SchabereiterGurtner et al., 2001), tree stumps (Vainio and Hantula, 2000), and grassland (Brodie et

al., 2003) and maize rhizospheres (Gomes Newton et al., 2003).

11.3.1.3



Profiling Based on Differences in Conformation: SSCP



In single-stranded conformational polymorphism (SSCP) analysis, PCR products of the

same length but with different sequences are converted into profiles composed of multiple

bands that migrate differently under nondenaturing conditions because of their conformation. SSCP is based on the principle that single-stranded DNA adopts a tertiary

structure that is sequence specific (even to single-nucleotide polymorphisms). However,

depending on the resolution of the gel, a single band may contain more than one sequence

(Schmalenberger and Tebbe, 2003). Similarly to DDGE and TGGE, it is a technique that

was originally developed to detect point mutations and allelic variants in human genes

using electrophoretic gels (Orita et al., 1989; Hayashi, 1991). However, this technique

does not require the construction of gradient gels, and its sensitivity and reproducibility

have been significantly improved using automated fluorescent electrophoresis (Lee et al.,

1996; Zumstein et al., 2000), which notably facilitates comparisons. PCR combined with

SSCP analysis has also been applied to microbial communities as an alternative fingerprinting approach to DGGE or TGGE (Lee et al., 1996; Schmalenberger and Tebbe,

2003). Despite its potential as a relatively simple and effective method for the detection

of minor sequence changes in PCR-amplified products, this method has been rarely

applied to fungal communities. Earlier attempts were made to use PCR-SSCP for the

identification of fungal species from roots, axenic isolates, or spores using the ribosomal

genes and spacers (e.g., Simon et al., 1993; Kumeda and Asao, 1996; Moricca and

Ragazzi, 1998; Kjøller and Rosendhal, 2000). To our knowledge, only one preliminary

study has applied SSCP to changes in fungal communities in response to environmental

factors (Lowell and Klein, 2001).

11.3.2

11.3.2.1



Sequencing and Phylogenetic Affiliation

Identification through Phylogenetic Placement



The PCR, clone, and sequence approach to studying microbial community structures can

be thought of as a gold standard. It is, or should be, used to validate screening techniques

(e.g., DGGE, T-RFLP, SSCP). However, bacterial and archaeal 16S community studies

have clearly highlighted the biases and artifacts introduced by this popular approach (Qiu

et al., 2001; Speksnijder et al., 2001), as well as its practical limitations (Dunbar et al.,

2002). Biases are due to over- and underrepresentation of phylotypes, depending on

amplification and cloning conditions used for multitemplate samples. Artifacts are extensive and varied, and they will be discussed in detail in Section 11.3.3.1.

The practical limitations of the PCR, clone, and sequence approach in comparative

bacterial community analysis have been highlighted by Dunbar et al. (2002), both by

reanalyzing empirical data and through modeling. While species richness is likely to be

reproducibly estimated with the typical number of clones analyzed per soil sample, species

composition is not, and it would require orders-of-magnitude-larger samples than those

typically used. A recently developed alternative to commonly used, but highly sensitive



DK3133_C011.fm Page 224 Monday, April 18, 2005 12:56 PM



224



Bidartondo and Gardes



diversity indices, lineage-per-time plots are a robust method for phylotype frequency

estimation (Martin, 2002). The use of this and other analytical tools for molecular biodiversity data has been reviewed recently (Hughes et al., 2001; Curtis et al., 2002; Bohannan

and Hughes, 2003).

The molecular analysis of ectomycorrhizas, typically screened by RFLP and followed by direct sequencing, has been relatively straightforward to apply to large numbers

of individual root samples because sufficiently pure single-strain fungal DNA can be

obtained from roots without cloning (Horton and Bruns, 2001). High-throughput sequencing using capillary arrays has greatly accelerated this trend (>60,000 roots were analyzed

by Tedersoo et al., 2003). Identification of ectomycorrhizal fungi in soil outside roots has

required cloning (Landeweert et al., 2003) or T-RFLP (Dickie et al., 2002). In contrast,

there is yet no widely used optimal molecular method to assess diversity in arbuscular

mycorrhizas, particularly in field roots or soil. This is because

1.

2.



3.

4.



The fungi colonize individual roots less extensively than ectomycorrhizal fungi,

so that target template molecules are less abundant relative to PCR inhibitors.

Fungal rDNA is highly diverse and under relaxed concerted evolution, so that

primer design is problematic (Schüßler et al., 2001), cloning is usually necessary, and sequence interpretation is troublesome.

No genes of phylogenetic utility other than nuclear rDNA have yet been widely

sequenced and made available.

The fungi contain endosymbiotic ascomycetes and bacteria.



Here we briefly describe 4 representative methods (of over 20 published to date)

selected based on their popularity or uniqueness. Helgason et al. (1998) directly amplified,

cloned, and sequenced approximately 500 base pair (bp) of the 18S gene with one primer

set. Simon et al. (1993) used nested PCR with a eukaryotic primer set, and then amplified

approximately 150 to 400 bp with three lineage-specific primer sets. Clapp et al. (1995)

used a selective enrichment PCR approach targeting approximately 150 bp of 18S rDNA

with four steps: (1) PCR with a universal primer set was done on root and leaf samples,

using a biotinylated base for leaf samples; (2) leaf and root products were combined,

denatured, and reannealed; (3) plant DNA was subtracted using streptavidin; and (4) PCR

with three lineage-specific primer sets was done. Redecker (2000) used a nested PCR

approach targeting approximately 1 kb of rDNA (3′ 18S + ITS1) with a eukaryotic primer

set, and then with five lineage-specific primer sets. In optimal conditions, fungi can be

stained and visualized in the root tissue after DNA extraction. The last approach was the

only one to target all known glomalean lineages to date. Thus, although popular, singleprimer studies of arbuscular mycorrhizal communities must be regarded with caution.

Nonetheless, all approaches can amplify nonglomalean fungi, and all lineage-specific

primer sets can cross-amplify glomalean lineages so that sequencing exemplars were

always necessary to validate putative phylo-placements.

11.3.2.2



Oligonucleotide Fingerprinting and Microarrays



Oligonucleotide fingerprinting and microarrays are hybridization tools that can be used

as genomic identification systems through the analysis of small DNA segments. In these

systems, diversity among DNA sequences is exploited to identify organisms.

In conventional oligonucleotide fingerprinting, an array is constructed by applying

spots of PCR-amplified DNA onto a nylon membrane, which is then subjected to a series

of hybridization experiments (often referred to as dot-blots), each using a single DNA



DK3133_C011.fm Page 225 Monday, April 18, 2005 12:56 PM



Fungal Diversity in Molecular Terms



225



taxon-specific oligonucleotide probe (TSOP). Fingerprinting works by sorting arrayed

DNAs into taxonomic clusters based on their hybridization patterns. When fully developed

and tested, this identification system provides a reliable and cost-effective method for

high-throughput species identification in multitemplate environmental samples. However,

this tool has been slow to be applied in molecular ecology, which is explained partly by

the time and labor required for the development of an operational identification system.

The literature offers a few examples of the successful application of TSOPs for the analysis

of fungal communities using conventional dot-blot experiments (e.g., Bruns and Gardes,

1993; Valinsky et al., 2002).

DNA microarrays (or DNA chips) were originally designed for large-scale sequencing and genetic analyses (Southern et al., 1992; Guo et al., 1994), and they represent a

significant advance in hybridization technology. They are used extensively in functional

genomics (for fungi, see Cavalieri et al., 2000; Nowrousian et al., 2003). Microarrays also

hold much promise as tools in environmental studies because of their exceptional highthroughput capacity. Thus, phylogenetic oligonucleotide microarrays (POAs) will allow

for the simultaneous application of hundreds of thousands of TSOPs in a single hybridization experiment. POA is a reversed hybridization system; matrix-immobilized oligonucleotide probes are used to capture labeled target molecules (e.g., amplified DNA segments,

rRNAs). In addition to qualitative analysis of communities (Loy et al., 2002; Wilson et

al., 2002), phylogenetic oligonucleotide microarrays (or PhyloChips) have the potential

to be sensitive enough for quantitative detection of intact rRNAs from environmental

samples (Small et al., 2001). However, to realize the full potential of POAs for the analysis

of fungal communities, significant advancements will need to be made in (1) the design

and validation of TSOPs (Wilson et al., 2002); (2) the sensitivity and specificity of the

hybridization procedure and technology, particularly for soil samples (Small et al., 2001;

Urakawa et al., 2003); and (3) the processing of hybridization results (Townsend, 2003).

In addition, two kinds of microarrays can be used to monitor the activities of organisms

in microbial communities (for a review, see Zhou and Thompson, 2002). First, functional

gene arrays (FGAs) containing genes involved in ecosystem processes (e.g., nitrogen

cycling) can be used to reveal the physiological and metabolic activities of living organisms

in environmental samples (Wu et al., 2001). Second, community genome arrays (CGAs)

can be constructed using whole genomic DNA from pure strains to examine their activities

in the environment. Both of these remain to be tested in fungal ecology.

11.3.3

Pitfalls of PCR-Based Fingerprinting Methods

Two main sources of experimental error have been recognized: (1) PCR biases and artifacts,

and (2) interpretation of the data (Wintzingerode et al., 1997; Martin, 2002; Bruns and

Shefferson, 2004). Sample processing, cell lysis, and nucleic acid extraction are significant

sources of variation for all subsequent analyses using a PCR approach (Stach et al., 2001),

but surprisingly, they appear to be very rarely optimized.

11.3.3.1



PCR Biases and Artifacts



PCR amplification requires special attention to several problems intrinsic to experiments

with multitemplate samples. With some primer–template combinations, some phylotypes

can be disproportionately amplified due to preferential priming or differences in elongation

rates among amplicons (Wintzingerode et al., 1997). A different bias can occur when a

reaction undergoes too many cycles because there is a tendency for the different amplicons

to reach equimolarity because reannealing of frequent PCR products progressively inhibits

primer–template annealing (Suzuki and Giovannoni, 1996). If quantitative inferences of



DK3133_C011.fm Page 226 Monday, April 18, 2005 12:56 PM



226



Bidartondo and Gardes



relative phylotypes are intended, adequate controls have to be included for each

primer–template (Lueders and Friedrich, 2003).

Another set of potential PCR-generated artifacts (i.e., heteroduplexes, chimeras, and

mutagenesis) have received much attention in the context of PCR, cloning, and sequencing

approaches (Qiu et al., 2001; Speksnijder et al., 2001), but artifacts are also sources of

spurious bands (or peaks) in rapid profiling methods. For instance, artifacts specific to TRFLP have been recently reported (Egert and Friedrich, 2003). PCR artifacts may lead to

the commonly observed “bushes” of closely related cloned sequences. Detecting such

aberrant microvariation is a problematic and largely unresolved concern. Heteroduplexes

are heterologous hybrids formed from different gene templates; some heteroduplexes can

separate from homoduplexes during electrophoresis (due to conformational differences),

thereby producing additional bands or peaks, which may be removed by gel purification.

The occurrence of heteroduplexes increases with increasing similarity between sequences;

this may happen even within individuals for multicopy genes. When heteroduplexes are

cloned, they can be converted into a single hybrid sequence that is a mosaic of the two

parent heterologs (because the insert is not methylated, excision-repair enzymes cannot

distinguish parent DNA strands and independently use either strand as a template for each

base). In any multitemplate PCR, heteroduplexes can be as abundant as homoduplexes

(Thompson et al., 2002) and can make up to 10% of clones (Lowell and Klein, 2000).

Heteroduplexes can be resolved before genomic library construction by digestion with a

single-strand cleaving endonuclease (Lowell and Klein, 2000), or they can be minimized

by nested PCR, increasing primer concentration, and using a low number of cycles for

reamplification (Thompson et al., 2002). Chimeras are sequences generated when an

incompletely extended PCR product acts as a primer on a heterologous sequence. For SSU

rDNA, there are programs that can detect potentially aberrant variation (e.g., Chimera

Check). In a recent study, 27 of 29 environmental fungal phylotypes recovered were thus

assessed to be potentially chimerical (Jumpponen, 2003). PCR mutagenesis can be due to

base misincorporation, or deletions when the template has secondary structure, and can

be minimized by the use of DNA polymerase mixtures that contain a proofreading enzyme.

In general, harvesting PCR products as early as possible during the exponential phase of

the reaction is widely regarded as the most efficient way of minimizing all artifacts. While

DNA sequencing is as sensitive to chimeras as rapid DNA profiling methods, the latter

are more sensitive to heteroduplexes and mutagenesis (Qiu et al., 2001). Clean sequences

cannot be obtained if a clone results from a heteroduplex, and single-base mutations may

have little or no effect on overall phylogenetic topology. However, heteroduplexes and

mutagenesis can have potentially large impacts on nonsequencing methods (i.e., RFLP, TRFLP, DGGE, TGGE, SSCP).

11.3.3.2



Data Interpretation Pitfalls



It must be noted that DNA-based ecological methods still rely heavily on single-locus

analysis of multicopy noncoding gene regions. However, multilocus genotyping provides

much more precise molecular identification of species. For a review of how DNA

sequence data alone can be used to define and recognize fungal species through the use

of multigene phylogenies (i.e., the phylogenetic species concept), see Taylor et al. (2000).

The use of single-copy genes avoids paralogous comparisons, and sequences from protein-coding genes are significantly less ambiguous to align than noncoding DNA. The

realization of both of these loci’s potential for fungal identification from environmental

samples awaits further growth of their representation in the public database. However,

the PCR amplification single-copy genes and protein-coding genes from the environment