An Evaluation of 18S rDNA Approaches for the Study of Fungal Diversity in Grassland Soils

An Evaluation of 18S rDNA Approaches for the Study of FungalDiversity in Grassland Soils

J. Hunt, L. Boddy, P.F. Randerson and H.J. Rogers

Cardiff School of Biosciences, Cardiff University, P.O. Box 915, Cardiff CF10 3TL, Wales, United Kingdom

Received: 6 June 2003 / Accepted: 29 July 2003 / Online publication: 4 March 2004

Abstract

Fungal community structure and diversity in two types ofagricultural grassland soil were investigated by amplified18S ribosomal DNA restriction analysis (ARDRA) and18S ribosomal DNA sequence analysis. These two grass-land sites represent a species-rich old hay meadow and anagriculturally improved site with low floristic diversity.Two primer sets were used in combination to amplifyapproximately 550 bp of rDNA from three major fungalgroups, the zygomycetes, basidiomycetes, and ascomyc-etes, and clone libraries were created for each site. 18SARDRA was used to analyze 170 rDNA clones, and threediversity indices were calculated. A small-scale culturinganalysis was also carried out and the most commonisolates analyzed using ARDRA and sequence analysis.The soil fungal community revealed by the rDNA ap-proaches was significantly different from that producedby this limited culture-based analysis. Twenty-eight soil-derived clones were sequenced, and many representedfungal taxa rarely reported in culture-based studies. ThePCR-based techniques detected differences in diversitybetween the two fungal communities and changes inpatterns of dominance that paralleled higher plant di-versity. The results suggest that 18S rDNA-based ap-proaches are a useful tool for initial screening of fungalcommunities, and that they represent a more compre-hensive picture of the community than plate culturing.

Introduction

Soil is the most diverse of terrestrial habitats, and the soilmicrobial community plays a fundamental role in de-composition, nutrient cycling, and energy flow [46]. Soilswith high biodiversity appear to be more resistant to

stress [15], and low diversity may be associated withimpaired ecosystem function [42].

Fungi are important as decomposers of organicmatter and comprise a major proportion of soil microbialbiomass. In grasslands 78–90% of the total decomposerbiomass may be fungal [21]. Fungi also function as plantand insect pathogens, predators, and mutualists. My-corrhizal plant symbioses are ubiquitous and are oftenessential for plant survival in dry or nutrient-poor con-ditions [1]. Changes in fungal diversity associated withagricultural management may, therefore, have importantimplications for soil fertility, stability, plant establish-ment, and yield. However, although it is known thatfungal communities are affected by agricultural practices[3], the effects of agricultural intensification on fungalbiodiversity remain unclear.

There are a number of difficulties involved in study-ing soil fungal biodiversity. Conventional microbiologicalculturing techniques are thought to detect <1% of bac-teria present in the soil because of the selectivity of growthmedia and conditions [4], and it is likely that fungalpresence is similarly underestimated [16]. Several fungalgroups, such as arbuscular mycorrhizas [14] and somebasidiomycetes [12], are difficult or impossible to culture,and isolation onto agar media favors fast-growing, heavilysporulating species [12]. Thus isolation techniques do notprovide an accurate picture of the in situ diversity of theactive members of fungal communities.

For basidiomycetes, field surveys of fruiting bodiescan give some idea of what species are present. However,lack of fruit bodies does not necessarily imply that aspecies is absent, since fruit body production is veryvariable and, for some species, infrequent [12, 36]. Thishas been shown very clearly for ectomycorrhizas, wheresporocarps do not reflect diversity and abundance offungi revealed in root tips by genotyping [19]. Repeatedsampling, both within a season and over many years, isnecessary to obtain a reasonable picture of basidiomyc-etes present [36].Correspondence to: L. Boddy; E-mail: [email protected]

DOI: 10.1007/s00248-003-2018-3 d Volume 47, 385–395 (2004) d � Springer-Verlag New York, LLC 2004 385

Molecular techniques such as amplified rDNAsequencing, amplified 18S ribosomal DNA restrictionanalysis (ARDRA), and temperature and denaturinggradient gel electrophoresis can be used to provide aculture-independent picture of soil microbial commu-nities. Such techniques are now widely used in the studyof soil bacterial diversity [4, 31], but there have beenrelatively few attempts to assess their suitability to studyfungal diversity in the field [2, 22, 34, 43], and apart fromone study [2], these have only provided qualitative in-formation.

Molecular identification of fungi to species level hasgenerally utilized the variable internally transcribedspacer (ITS) regions of the rDNA. However, althoughPCR primers exist with enhanced specificity for the ITSregions of basidiomycetes [13] and ascomycetes [24],they are generally unsuitable for direct amplification offungal sequences from environmental samples for tech-niques such as ARDRA because they also amplify the ITSregion from some plant material [13]. In addition thevariability of the ITS region makes it unsuitable forbroad-scale comparisons of fungal groups [23], sincesequence alignment becomes difficult [2]. Although thetaxonomic resolution of 18S rDNA may not always besufficient to identify fungal species and strains, this geneis conserved enough to allow comparison across a widerange of fungal taxa and is therefore potentially useful foran initial assessment of the total fungal community.

The aims of this work were to (i) test the suitabilityof 18S rDNA–based techniques for studying soil fungalcommunities; (ii) apply quantitative measures of diver-sity; and (iii) investigate the effects of agricultural man-agement and floristic diversity on below ground fungalcommunities.

Materials and Methods

Sample Collection. Soil samples were collected fromtwo adjacent sites—Pentwyn Farm and Bush Meadow in

Monmouthshire, UK (Nat Grid ref. SO 524 093/4; Table1). The former is characteristic of an unimproved low-land grassland, and the latter represents a semi-improvedgrassland. Although the vegetation on these two sitesreflects historic differences in management practices(Table 1), especially the addition of mineral fertilizer,they are currently managed identically by Gwent WildlifeTrust. Both sites are grazed in the spring by sheep duringlambing, then closed off for the growth of the hay crop inMay. Hay is cut at the beginning of July (depending onweather conditions). Cattle graze the sward regrowth,before they are moved inside for the winter and sheep arereturned to the sites. No organic fertilizers are used, al-though there is an input of material from animal fecesand urine.

Ten random soil cores (2 cm diameter, 10 cm depth)were taken from each site on 3 April 2000. During themonth leading up to this date the minimum and maxi-mum temperatures were 4.5 and 11.1�C, respectively.Total monthly rainfall was 30.9 mm.

Each separate core was mixed by hand and sieved(mesh size <2 mm) to remove plant material beforesubsamples from each core were stored at )70�C formolecular analysis. The remainder was used immediatelyin culturing experiments.

DNA Extraction. Total community DNA was ex-tracted from each core from each site. The extractionmethod used was a modification of that described byDempster et al. [8] for plant material high in poly-phenolics. Approximately 250 mg of soil and 250 mg ofglass beads (4 lm diam) were added to a 1.5-mLEppendorf tube. To this was added 500 lL extractionbuffer [100 mM Tris-HCl (pH 8.0), 1.4 M NaCl, 20 mMEDTA, 2% CTAB, 0.4% b-mercaptoethanol, 1% PVP],and the sample was incubated at 50�C for 15 min. Anequal volume of chloroform–isoamyl alcohol (24:1) wasadded, and the samples were shaken for 20 min on aflat-bed shaker at 200 rpm. After precipitation of the

Table 1. Summary of the soil, plant and management characteristics of the Bush Meadow and Pentwyn Farm field sites

Pentwyn Farm Bush Meadow

UK Grid reference SO 524 094 SO 524 093UK National Vegetation

Classification [30]Unimproved lowland MG5 grasslanda Semi-improved lowland MG6 grasslandb

Dominant grass species Centaurea nigra, Agrostis capillaris,Anthoxanthum odoratum

Lolium perenne, Agrostis capillaris,Anthoxanthum odoratum

Plant species recorded 48 41Nationally uncommon species Dactylorhiza fuchsii, Orchis morio,

Plantanthera chloranthaNone

Past management Never received fertilizer;traditional hay meadow

Regular fertilizer and lime;sheep grazing

Area (km2) 1.3 0.9Soil type Clay loam Clay loamSoil pH 4.95 4.95aDefined as Centaurea nigra–Cynosurus cristatus grassland [30].bDefined as Lolium perenne–Cynosurus cristatus grassland [30].

386 J. HUNT ET AL.: 18S rDNA AND FUNGAL DIVERSITY IN GRASSLAND SOILS

soil debris by centrifugation at 13,000g, the DNA waspurified by ethanol precipitation and resuspended in100 lL TE (10 mM Tris, 1 mM EDTA, pH 8.0). Thequality of extracted DNA and approximate yields weredetermined by agarose gel electrophoresis. DNA wasgenerally of a high molecular weight and yields werebetween 2 and 5 lg DNA g)1 soil. Samples that showedevidence of shearing were not used for PCR amplifica-tion.

PCR Amplification. PCR amplification of fungalsmall-subunit rDNA from soil samples was carried outusing two primer sets, EF4/fung5 and EF4/EF3 [34]. TheEF4/fung5 primer pair amplifies a fragment of �550 basepairs, and the EF4/EF3 primers amplify a 1.5-kb sectionof the 18S rDNA. Amplification was performed in a 25-lL reaction containing 1 U of Hotstar DNA Taq po-lymerase (Qiagen Ltd., Crawley, UK), a 10· dilution ofthe manufacturer’s buffer, 200 lM concentrations of eachdeoxynucleoside triphosphate, and 100 nM concentra-tion of each primer. Diluted template DNA, 1 lL of 1:10or 1:100, was used. The following thermocycling patternwas used: 94�C for 15 min (1 cycle): 94�C for 1 min, 48�Cfor 1 min, 72�C for 3 min (40 cycles), and 72�C for 10min (1 cycle). All PCR amplifications were carried outusing a PTC 200 DNA Engine (M J Research, Cambridge,UK).

Cloning. PCR products were visualized on ag-arose gel stained with ethidium bromide. Bands wereexcised and DNA purified using a Qiaquick gel extrac-tion kit (Qiagen). Products from three separate am-plifications were pooled prior to cloning to minimizePCR drift. Purified amplification products were clonedinto the pGEM T-easy vector system (Promega UK Ltd.,Southampton, UK), and ligations were transformed intoEscherichia coli XL1-Blue sub-cloning grade competentcells in accordance with the manufacturer’s directions(Stratagene Ltd., Cambridge, UK). EF4/EF3 and EF4/fung5 libraries were created from each soil sample.White colonies were screened directly for inserts byperforming colony PCR with either EF4/fung5 or EF4/EF3.

Amplified Ribosomal DNA Restriction Analysis (AR-

DRA). For each soil sample 18S rDNA was amplifiedfrom positive clones from both EF4/EF3 and EF4/fung5libraries using primers EF4/fung5. Six lL of each PCRproduct was separately digested with HaeIII and HinfIrestriction enzymes (Promega UK) at 37�C for 4 h. Re-stricted DNAs were analyzed using ICN low gel tem-perature agarose at 2.5% wt vol)1 (ICN Biomedicals Ltd.,Thame, UK). For each enzyme each unique restrictionpattern was given a single letter code. The clones were

then classified using the restriction pattern codes fromboth enzymes, in the order HinfI HaeIII. ARDRA patternab, for instance, represents HinfI pattern a and HaeIIIpattern b.

Culturing and 18S rDNA Analysis of Isolates. Tensoil plates were made from each core [45]. Briefly, 0.2 mgof sieved soil was placed in the bottom of a sterile Petridish. The soil was dispersed by adding a drop of steriledistilled water and the particles overlaid with 8–10 mL ofmolten Czapek Dox medium (pH 5) containing 0.01%kanamycin (Sigma-Aldrich Ltd., Dorset, UK). The plateswere incubated at 15�C and checked daily for �5 days,until the plates were filled with mycelia. Emerging fungalcolonies were transferred to fresh Czapek Dox mediumand incubated at 15�C, and then identified to genuswhere possible.

DNA was extracted from fungal cultures using amodification of the Cenis method, adapted for materialgrown on solid media [6]. Surface hyphae (0.1 g) werescraped from each culture and ground in 300 lL buffer(200 mM Tris-HCl pH 5.8, 250 mM NaCl, 25 mMEDTA, 0.5% SDS) using a conical grinder. Cell debris waspelleted by centrifugation at 13,000g and the DNA pu-rified by alcohol precipitation. DNA was resuspended in100 lL TE (pH 8.0).

PCR amplification was carried out as described forthe soil samples, using only primer set EF4/fung5. ThesePCR products were cloned into the pGEM T-Easy vectorsystem and clones were analyzed by ARDRA as describedpreviously.

Sequencing. Plasmids containing the 18S rDNAinserts were purified using the QIAprep spin miniprep kit(Qiagen). Twenty-eight clones derived from soil samplesand seven from cultures were sequenced on an ABI Prism3100 capillary sequencer (Applied Biosystems, FosterCity, CA, USA). Sequencing reactions were performedusing the BigDye Terminator Cycle Sequencing Kit(Applied Biosystems).

Sequence Analysis. Sequences were checked forchimeric artefacts using the CHIMERA-CHECK programof the Ribosomal Database Project, RDP-II [25], andcompared with similar rDNA sequences retrieved fromthe RDP-II and GenBank databases. Sequences werealigned with representative fungal sequences from theRDP-II and GenBank using Clustal X [40], and thealignment was corrected manually. A Neighbour- Joiningtree was constructed from a distance matrix calculatedusing Kimura’s two-parameter model [20] in the PHY-LIP program [11]. Data sets were bootstrapped usingSEQBOOT (PHYLIP ver. 3.5).

J. HUNT ET AL.: 18S rDNA AND FUNGAL DIVERSITY IN GRASSLAND SOILS 387

Statistical Analysis. The ARDRA data were usedto estimate three diversity indices:

(i) The Shannon diversity index, H¢ [33], a generaldiversity index, which considers both species richness andevenness:

H0 ¼ �X

pi ln pi

where pi = ni/N, n is the total number of clones of the ithARDRA type, and N the total number of clones in thesample.

(ii) Coverage [29], the portion of the actual diversitythat has been sampled.

C ¼ 1 � ðn1=NÞ

where n1 is the number of clones that occurred once andN is the total number of clones examined. This providesan estimate of how well the sample size reflects the ap-parent diversity within the clone library.

(iii) ARDRA group richness. The freeware programAnalytic Rarefaction [18] was used to calculate therichness of each sample by rarefaction, where the

expected number of ARDRA groups in a sample con-taining 71 clones or a data set containing 156 clones wasdetermined using the complete data sets.

Nucleotide Sequence Accession Numbers. Thepartial clone sequences determined in this study havebeen deposited in the GenBank database under accessionnumbers AY163413–AY163447.

Results

One hundred and seventy EF4/fung5 clones from thesoil-derived libraries were analyzed using ARDRA and 28were sequenced. Cultivated isolates were also included forcomparison in both the ARDRA and sequence analysis.

ARDRA of Soil-Derived Clones. A total of 84 clonesfrom Bush Meadow and 86 clones from Pentwyn Farmwere analyzed by digestion with HaeIII and HinfI.

The agriculturally unimproved Pentwyn Farm siteshowed higher clone diversity than the semi-improvedBush Meadow site (Fig. 1 and Table 2). At Pentwyn Farm

Figure 1. ARDRA groups derived from soillibraries from the improved Bush Meadowsite (Bush) and the unimproved PentwynFarm site (Pentwyn). Bars represent 95%confidence intervals.

Table 2. Estimates of richness, percent coverage, and Shannon diversity calculated for the Bush Meadow and Pentwyn Farm sites usingthe ARDRA data

No. of clones No. of ARDRA groups Richness % coverage Shannon index (H¢) ±95% CIa Range of H¢b

Bush soil library 84 20 18 87 1.95 ± 0.002 0–4.44Pentwyn soil library 86 24 22 85 2.60 ± 0.005 0–4.45Pooled soil libraries 170 29 29 92 2.54 ± 0.002 0–5.10Bush isolates 71 6 6 99 1.05 ± 0.004 0–4.26Pentwyn isolates 85 7 7 99 1.06 ± 0.003 0–4.44Pooled isolates 156 8 8 99 1.35 ± 0.005 0–4.68

Data from both soil clone libraries and cultured isolates are shown. See text for the methods used to calculate these parameters.aThe Jackknife method, described by Zahl [48], was used to estimate confidence intervals.bThe possible range of values for H¢ are the minimum and maximum values which could be obtained for libraries containing n ARDRA groups at minimumdiversity and maximum diversity.

388 J. HUNT ET AL.: 18S rDNA AND FUNGAL DIVERSITY IN GRASSLAND SOILS

the most abundant clone (ac) constituted 24% of allPentwyn clones (Fig. 1). In contrast the most abundantclone at the Bush Meadow site (aa) constituted 50% ofthe total Bush clone library. ARDRA groups containingonly a single clone were more common at Pentwyn(58%) than at Bush Meadow (55%), but the differencewas not statistically significant (P > 0.05).

Differences in diversity between the unimproved andimproved sites were evident (Table 2). The ARDRAgroup richness, coverage, and Shannon H¢ values all in-dicated that the Pentwyn site had a higher clone diversitythan the Bush site.

Culturing. The 191 isolates that were recoveredfrom the soil plates were divided into 58 OTUs (opera-tional taxonomic units) on the basis of growth rates,density of mycelial growth, spore color, and morphology.

These OTUs were designated cult 1 through to cult 58.Identification to species level was deemed beyond thescope of this study and would not have been comparableto the discriminatory power of the molecular techniquesemployed. Isolates that could be identified were fromgenera commonly isolated from soil, including Penicil-lium (cult 7 and cult 3), Trichoderma (cult 2), Absidia(cult 14 and cult 46), and Fusarium (cult 4). No isolateswere identifiable as basidiomycetes by the presence ofeither fruiting structures or clamp connections.

In both the semi-improved Bush Meadow and un-improved Pentwyn Farm sites the most commonly isolatedgroups were from the genera Trichoderma and Penicillium(Fig. 2). Comparisons between the isolates from the twosites did not show any major differences in diversity, al-though cult 2, cult 3, and cult 4 were more common in theBush Meadow site than at Pentwyn Farm (Fig. 2).

Figure 2. Cultured isolates derived fromPentwyn Farm and Bush Meadow soils. Atotal of 58 distinct OTUs were identified onthe basis of morphology. For both sites, eachOTU is expressed as a percentage of the totalnumber of isolates recovered from that site.Bars represent 95% confidence intervals.

Figure 3. A comparison of ARDRA groupsfrom the combined Bush Meadow andPentwyn Farm libraries (Clones) and fromisolates cultured from soil derived from boththese sites (Cultures). Bars represent 95%confidence intervals.

J. HUNT ET AL.: 18S rDNA AND FUNGAL DIVERSITY IN GRASSLAND SOILS 389

ARDRA of Cultures. EF4/fung5 18S rDNAsequences from 156 fungal isolates (71 from BushMeadow and 85 from Pentwyn Farm) were analyzedusing HaeIII and HinfI restriction enzymes. It was notpossible to amplify a PCR product from DNA extractedfrom the remaining 35 isolates. Only eight ARDRAgroups were found in the cultures (Table 2, Fig. 3). Twoof these ARDRA groups (dc and gf) were unique toPentwyn Farm, and group ae was unique to BushMeadow. The three diversity measurements showed verysmall differences between the two sites (Table 2).

When the ARDRA patterns from the cultures werecompared to those derived from soil library clones, cleardifferences could be seen (Fig. 3, Table 2). All of the eightARDRA groups found in the cultures were also found inthe soil libraries. The dominant ARDRA group in thecultures was db, and this group was significantly moreabundant than in clones from the soil libraries

(P = 0.05). The coverage value for both the two sitescombined and individually was 99%, suggesting thatnearly all of the available ARDRA types in the isolates hadbeen sampled.

Sequence Data. Sixteen EF4/fung5 clones fromthe Bush Meadow soil samples and 12 clones from thePentwyn Farm soil samples were chosen at random andsequenced. EF4/fung5 fragments from the most fre-quently occurring cultured isolates (cults 1, 2, 3, 4, 5, 7,and 15) were also sequenced. Three main subdivisions ofkingdom Fungi, the ascomycetes, basidiomycetes, andzygomycetes, were represented in the Bush and Pentclone libraries, but no mitochondrial, bacterial, or othereukaryotic sequences were identified.

The most similar database sequence to each clone isindicated in Table 3. Homologies ranged between 92%and 100% identity. Two pairs of the sequences (Bush 1.5

Table 3. Clones from the Bush Meadow (Bush) and Pentwyn Farm (Pent) libraries and cultured isolates (Cult) showing the nearestmatch from GenBank and the ARDRA group of each clone

Clone Fungal taxa with most similar sequence % identity ARDRA group

ZygomycetesBush 2.4e Echinosporangium transversale AF113424a 96 af

2.21e Fungal clone AF096367 99 af2.20e Glomus intraradices AJ3018591 98 da

Pent 5.3ex Fungal clone AF096367 98 ab3.2ex Mortierella chlamydospora AF157143 93 ce3.6e Mortierella chlamydospora AF157143 99 af4.10 Mortierella wolfii AF113425 95 ac

AscomycetesBush 1.1 Endomyces scopularum AF267227 96 bb

4.41e Sporothrix schenckii M85053 95 dfPent 2.8 Gibberella pulicaris AF081467 99 dc

3.2 Mycosphaerella mycopappi U43449 99 ae3.9 Paracoccidiodes brasiliensis AF238302 97 ac3.6 Phialocephala fortinii L76626 97 db3.5 Rhizina undulata U42664 96 ef

Cult 3 Eupenicillium javanicum U21298 99 ab7 Eupenicillium javanicum U21298 99 ab4 Gibberella pulicaris AF081467 99 db1 Hypocrea lutea D144071 99 db

15 Monilinia fructicola AF01505 100 gf5 Talaromyces luteus AB006716 99 hc2 Trichoderma viride AF525230 99 db

BasidiomycetesBush 2.6 Calvatia gigantea AF026622b 98 ab

3.4 Calvatia gigantea AF026622 97 aa4.3 Calvatia gigantea AF026622 97 aa4.9 Calvatia gigantea AF026622 98 ae5.4 Calvatia gigantea AF026622 97 aa5.9 Entoloma strictius AF287832b 95 cb3.7e Inocybe sp. AF287836 95 ab1.5 Lycoperdon sp. AF026619 97 aa5.0 Lycoperdon sp. AF026619 97 aa1.4e Panellus stypticus AF026589 93 ab3.2e Panellus stypticus AF026589 92 ab

Pent 5.7 Calvatia gigantea AF026622a 97 aa4.6 Entoloma strictius AF287832b 95 aa5.8 Inocybe sp. AF287836 95 ab

aGenbank accession number; bsporocarps of C. gigantia and Entoloma sp. have been seen at the sites.

390 J. HUNT ET AL.: 18S rDNA AND FUNGAL DIVERSITY IN GRASSLAND SOILS

and 5.0 and Bush 5.9 and 4.6) from the soil clones wereidentical to each other, and cult 15 was identical to asequence in the GenBank database.

A neighbor-joining tree showed the three fungalgroups, the ascomycetes, basidiomycetes, and zygomyc-etes, separated from each other with high levels ofbootstrap support (Fig. 4). Groups of clones formeddistinct clusters within the basidiomycetes and zyg-omycetes, which did not include any of the referencetaxa, probably reflecting the paucity of reference taxa inthe databases (Table 3, Fig. 4). The basidiomycetes madeup the largest fraction of soil-derived clones (54%), withhymenomycetes predominating. The ascomycetes com-prised only 21% of the clones, but all of the culture se-quences. Ascomycete sequences were more common in

the unimproved Pentwyn Farm soil than in the semi-improved Bush Meadow site. A quarter of the sequencesfrom the clone libraries clustered within the zygomycetes.The majority of these sequences formed a distinct group.One clone (Bush 2.20e) had 98% sequence homology toGlomus intraradices and these two sequences formed aseparate cluster (Fig. 4). The Zygomycota are polyphy-letic, and there is evidence that the arbuscular mycorrhiza(AM) fungi can be separated into a monophyletic cladeunrelated to any other zygomycete group [32].

Identification of ARDRA Groups. The corre-sponding ARDRA pattern for each sequence is shown inTable 3. ARDRA group ab occurred most frequently. Thesequence data indicates that the clones in this group

Figure 4. Neighbor-joining tree showingthe relationship of Pentwyn Farm (Pent),Bush Meadow (Bush), and culture-derived (Cult) clones to reference fungaltaxa, based on analysis of 550 bases ofaligned 18S rDNA sequences. Bootstrapvalues are shown for nodes that had >50%support in a bootstrap analysis of 100replicates.

J. HUNT ET AL.: 18S rDNA AND FUNGAL DIVERSITY IN GRASSLAND SOILS 391

could have come from any of the three fungal phyla.Groups ac and ae are also shared across taxa. However,other large ARDRA groups do appear to correspond tofungal taxonomic groups. Only basidiomycete sequenceswere found to have the ARDRA pattern aa, and thispattern represented 12% of all the soil clones. Patternsdb, dc, df, ef, gf, and hc were only found in the asc-omycete clones, with db being the most common (4/13).Group db was the dominant ARDRA group found in thecultured isolates (Fig. 3). ARDRA patterns af, da, and cewere exclusively found in the zygomycete clones. How-ever, in some cases, clones which matched the same da-tabase taxa produced different ARDRA groups.

Discussion

We were able to detect variations in both the levels ofdiversity and patterns of dominance in the two soil fungalcommunities, using a combination of 18S ARDRA and18S rDNA sequence analysis. By comparison with se-quence analysis, the ARDRA technique allowed rapidassessment of large numbers of clones from the soil li-braries, and it was possible to suggest phylogenetic affil-iations for some of the major ARDRA groups bycomparison with sequenced clones. Sequence analysisshowed the EF4/fung5 and EF4/EF3 PCR primers [34] tobe highly specific for fungal 18S rDNA sequences, and nononfungal sequence was retrieved from the soil libraries.This is important for fingerprinting techniques such asARDRA, where it is essential to know that all the groupsdetected are of fungal origin.

The diversity present was probably underestimatedusing ARDRA. A number of clones with the same AR-DRA profile matched different database sequences, sug-gesting that some sequence differences may be outsidethe restriction sites. The tentative phylogenetic affiliationsof ARDRA groups made in this study could be refined bya systematic analysis of sequences from fungal referencestrains. Other ways of increasing the discrimination ofARDRA would be to use more restriction enzymes [28]or to analyze a longer DNA sequence. Increasing thenumber of enzymes used, however, would also dramati-cally increase both the time and cost of analysis. A pre-liminary ARDRA study, using the longer (1.5 kb) rDNAsequence amplified by the EF4/EF3 primer set, producedpatterns that were too complicated to effectively analyzeby agarose gel electrophoresis (results not shown). An-other set of fungal specific PCR primers for the 18S rRNAgene, which cover several variable regions, have beendeveloped recently [2, 5], and these may be useful as analternative to those tested in this work.

The ARDRA patterns from the culturable fungi andthe soil-derived clones provided very different pictures ofthe fungal community. ARDRA group richness, coverage,and Shannon diversity were all lower in the isolates,

suggesting that the culturing strategy was more selectivethan direct DNA extraction and PCR. The sequencesfrom the soil libraries covered a much wider taxonomicrange than that generally reported in isolation-basedstudies of soil fungi. Borneman and Hartin [5] reportedsimilar differences in the dominant genera identified inavocado grove soil when an rDNA analysis was used,compared to a traditional culture-based approach. Fast-growing, prolifically sporulating, ruderal taxa [12], suchas species of Penicillium, Trichoderma, Fusarium, Mucor,and Mortierella [9], are commonly isolated from soil, andsimilar dominant groups were isolated onto agar mediumin the present study. Cult 7 and cult 3 were identified asPenicillium species and cult 2 as a Trichoderma species onthe basis of morphology, and this identification wassupported by a high level of sequence homology to da-tabase species in these genera. None of the sequencesderived from direct soil DNA extraction was identical toany culture-derived sequence.

Our analysis of culturable soil fungi was small-scaleand intended simply to detect the dominant groups forcomparison with the 18S rDNA techniques. Only onetype of broad-spectrum medium and a single set ofgrowth conditions was used, and it is well known thatisolation media can dramatically affect the overall di-versity of species recovered [12, 14, 38]. However, thereare a number of ecologically important fungal groupswhich are difficult to isolate, even with a more extensiveculturing strategy. Basidiomycetes are notoriously diffi-cult to isolate [12] from soil, but specialized culturingtechniques, such as the modified soil washing proceduredescribed by Thorn et al. [41], can be successful, thoughsuch approaches are time consuming, require specialistidentification skills, and still may not recover all taxapresent [41]. One of the potential advantages of PCR-based sequence analysis for ecological studies is its abilityto detect such groups. In the present study 54% of thesoil-derived sequences were from basidiomycetes, andthey appear to constitute a major ARDRA group (aa), butno basidiomycetes were isolated. Further, some mycor-rhizas and endophytes [14] are difficult or impossibleto culture away from their hosts, even with a more ex-tensive culturing strategy, nor can they be easily identi-fied, except by molecular methods [17]. Such fungiform associations with >90% of land plant families [35]and dominate grassland ecosystems, so any commu-nity analysis which does not detect their presence isseriously flawed. In the present study a clone was alsofound which contained a sequence with a high level ofhomology to an arbuscular mycorrhiza species, Glomusintraradices.

Although a more representative range of taxa appearsto be detected by PCR-based techniques, they are not freefrom problems, which can bias community analysis [re-viewed by 44]. The original DNA extraction must result

392 J. HUNT ET AL.: 18S rDNA AND FUNGAL DIVERSITY IN GRASSLAND SOILS

in complete and representative lysis of fungal specieswithout shearing of the DNA or the coextraction of hu-mic substances inhibitory to PCR [39]. The bead-beatingmethod used in the present study resulted in minimalshearing, and the amounts of DNA released were com-parable to other similar studies [39, 49]. PCR can resultin differential amplification of templates in a complexmixture [37]. Here we tried to minimize PCR drift bypooling amplified samples prior to cloning. Althoughsome amplification bias has been reported for each of thetwo primer sets chosen for this study, their use in com-bination is designed to overcome any problems of dif-ferential amplification [34]. However, the biases reportedby Smit et al. [34], on the basis of database studies andtheir own work, do not seem to be borne out in thepresent study. Notably, the EF4/fung5 primer set wasoriginally reported to have a relatively low coverage ofbasidiomycetes, and Smit et al. [34] found no basidi-omycete sequences in rhizosphere soil samples; however,67% of the EF4/fung5 sequences from the soil libraries inthe present study were basidiomycetes. This may reflectdifferences in the species present in the bulk soil and therhizosphere rather than primer bias. Such differenceshave been reported for bacteria, with reduced 16S di-versity in the rhizosphere compared with the bulk soil[26]. No nonfungal sequences were retrieved using eitherprimer set, although nonspecific amplification was re-ported as a problem with EF4/fung5 in a study of avo-cado grove soil [5]. Our findings are reinforced by arecent assessment of a series of fungal 18S rDNA primers[2], which also showed the EF4 and EF3 primers to befungal specific and suggested that any possible primerbias may be less significant than predicted from databasecomparisons.

A number of factors made it difficult to preciselyidentify fungal environmental clones by comparison withdatabase sequences. The RDP-II fungal 18S rDNA data-base contains 10 times fewer sequences than the prok-aryotic 16S rRNA database [25], and there are no 18SrDNA sequences in either the GenBank or RDP-II dat-abases for a number of basidiomycete grassland generasuch as Mycena and Hygrocybe. Studies using 18S rDNAsequences to identify fungal clones are scarce, and there islittle consensus as to what level of sequence similarityconstitutes species level identification. Smit et al. [34], ina study of fungal clones generated using the EF4/fung5primer set, suggested that similarities >95% could beconsidered to indicate the identity of the cloned se-quence. However, Kowalchuk et al. [22] considered thathomologies of >98% were required to identify fungalclones to genus level. Levels of marker variability are notuniform across fungal taxa [23], and so 18S rDNA se-quence information may provide specific definition inone taxon, while in other taxa it may provide definitionat another phylogenetic level.

Some of the newly sequenced basidiomycete andzygomycete clones formed separate well-supportedsubclades when analyzed phylogenetically. In general,these clones showed low levels of identification (between92% and 96%) with database species. Similar findingswere reported by Kowalchuk et al. [22], who foundclones with 90–96% identities to known fungal 18SrDNA sequences. Although many studies of bacterialdiversity have produced sequences unrelated to cultura-ble organisms in the databases [4, 27], it is not clear inthe present case whether these clones are from undis-covered species or from described fungi where an 18SrDNA sequence is not available in the database.

Although we found higher clone diversity on theunimproved, floristically diverse Pentwyn Farm site,some caution must be exercised in drawing conclusionsabout the effects of agricultural management on fungaldiversity. Fungal communities vary seasonally [7], andthis study only compared the two sites at a single timepoint. Previous studies have shown that fungi dominatedecomposition pathways in seminatural unfertilized soilsand not in fertilized sites [3], but it is not clear whetherbacterial or fungal species diversity varies. In contrast toour findings, a recent molecular-based survey of bacterialdiversity in agricultural grassland detected no differencesthat could be related to higher plant diversity [27].However, Donnison et al. [10] found that agriculturalmanagement had a significant effect on the isolationfrequency of culturable soil fungi. Further investigationsover time, combining 18S rDNA techniques with exper-imental manipulation of grazing and fertilizer additionon these sites, are needed to fully understand the effect ofagricultural management on the fungal community.Species diversity is only one aspect of microbial com-munities, and assessments of biodiversity must also ad-dress functional diversity and redundancy [47]. Althoughthis was beyond the scope of the present study, functionalanalysis of the microbial communities on these sites isongoing.

This work has shown that 18S rDNA based tech-niques can be used effectively as an initial screen to detectdifferences between soil fungal communities. However,18S rDNA is not variable enough to distinguish betweenclosely related species in all cases, and more variable genesor more variable parts of the rDNA are needed to com-pletely describe diversity. Although the relatively smallnumbers of reference sequences currently available makesdetection and identification of unknown clones prob-lematic at present, this situation will improve as greaternumbers of 18S fungal sequences become available.

Acknowledgments

This work was carried out as part of a Ph.D. studentshipjointly funded by Gwent Wildlife Trust and Cardiff

J. HUNT ET AL.: 18S rDNA AND FUNGAL DIVERSITY IN GRASSLAND SOILS 393

University. We thank Gareth Lewis for carrying out theDNA sequencing, and John Hedger for checking somecultures.

References

1. Allen, MF, Allen, EB (1992) Mycorrhizae and plant communitydevelopment: mechanisms and patterns. In: Carroll, GC, Wicklow,DT (Eds.) The Fungal Community. Marcel Dekker, New York, pp455–481

2. Anderson, IC, Campbell, CD, Prosser, JI (2003) Potential bias offungal 18S rDNA and internal transcribed spacer polymerase chainreaction primers for estimating fungal biodiversity in soil. EnvironMicrobiol 5: 36

3. Bardgett, RD, Frankland, JC, Whittaker, JB (1993) The effects ofagricultural management on the soil biota of some upland grass-lands. Agric Ecosyst Environ 45: 25–45

4. Borneman, J, Skroch, PW, O’Sullivan, KM, Palus, JA, Rumjanek,NG, Jansen, JL, Nienhuis, J, Triplett, EW (1996) Molecular mi-crobial diversity of an agricultural soil in Wisconsin. Appl EnvironMicrobiol 62: 1935–1943

5. Borneman, J, Hartin, RJ (2000) PCR primers that amplify fungalrRNA genes from environmental samples. Appl Environ Microbiol66: 4356–4360

6. Cenis, JL (1992) Rapid extraction of fungal DNA for PCR ampli-fication. Nucleic Acids Res 20: 2380

7. Daniell, TJ, Husband, R, Fitter, AH, Young, JPW (2001) Moleculardiversity of arbuscular mycorrhizal fungi colonising arable crops.FEMS Microbiol Ecol 36: 203–209

8. Dempster, EL, Pryor, KV, Francis, D, Young, JE, Rogers, HJ (1999)Rapid DNA extraction from ferns for PCR-based analyses. Bio-techniques 27: 66

9. Domsch, KH, Gams, W, Anderson, T (1980) Compendium of SoilFungi. Academic Press, London

10. Donnison, LM, Griffith, GS, Hedger, J, Hobbs, PJ, Bardgett, RD(2000) Management influences on soil microbial communities andtheir function in botanically diverse haymeadows of northernEngland and Wales. Soil Biol Biochem 32: 253–263

11. Felsenstein, J (1995) PHYLIP (Phylogeny Inference Package) ver-sion 3.57c. Department of Genetics, University of Washington,Seattle

12. Frankland, JC, Dighton, J, Boddy, L (1990) Methods for studyingfungi in soil and forest litter. Methods Microbiol 22: 343–404

13. Gardes, M, Bruns, TD (1993) ITS primers with enhanced specif-icity for basidiomycetes—application to the identification of my-corrhizae and rusts. Mol Ecol 2: 113–118

14. Gams, W (1992) The analysis of communities of saprophyticmicrofungi with special reference to soil fungi. In: Winterhoff, W(Ed.) Fungi in Vegetation Science. Kluwer Academic, Boston

15. Griffiths, BS, Ritz, K, Bardgett, RD, Cook, R, Christensen, S,Ekelund, F, Sorensen, SJ, Baath, E, Bloem, J, de Ruiter, PC,Dolfing, J, Nicolardot, B (2000) Ecosystem response of pasture soilcommunities to fumigation-induced microbial diversity reduc-tions: an examination of the biodiversity ecosystem function re-lationship. Oikos 90: 279–294

16. Hawksworth, DL (2001) The magnitude of fungal diversity: the 1.5million species estimate revisited. Mycol Res 105: 1422–1432

17. Helgason, T, Daniell, TJ, Husband, R, Fitter, AH, Young, JPW(1998) Ploughing up the wood-wide web? Nature (Lond) 394:431

18. Holland, S (August 2001) aRarefactWin program. University ofGeorgia, Athens. http://www.uga.edu/�strata/software/AnRareReadme.html

19. Horton, TR, Bruns, TD (2001) The molecular revolution in ec-tomycorrhizal ecology: peeking into the black box. Mol Ecol 10:1855–1871

20. Kimura, M (1980) A simple method for estimating evolutionaryrates of base substitutions through comparative studies of nucle-otide sequences. J Mol Evol 16: 111–120

21. Kjoller, A, Struwe, S (1982) Microfungi in ecosystems. Fungaloccurrence and activity in litter and soil. Oikos 39: 389–422

22. Kowalchuk, GA, Gerards, S, Woldendorp, JW (1997) Detectionand characterisation of fungal infections of Ammophila arenaria(Marram Grass) roots by denaturing gradient gel electrophoresis ofspecifically amplified 18S rDNA. Appl Environ Microbiol 63:3858–3865

23. Kuninaga, S, Natsuaki, T, Takeuchi, T, Yokosawa, R (1997)Sequence variation of the rDNA ITS regions within and betweenanastomosis groups in Rhizoctonia solani. Curr Genet 32: 237–243

24. Larena, I, Salazar, O, Gonzalez, V, Julian, MC, Rubio, V (1999)Design of a primer for ribosomal DNA internal transcribed spacerwith enhanced specificity for ascomycetes. J Biotechnol 75: 187–194

25. Maidak, BL, Cole, JR, Lilburn, TG, Parker, CTJ, Saxman, PR,Farris, RJ, Garrity, GM, Olsen, GJ, Schmidt, TM, Tiedje, JM (2001)The RDP-II (Ribosomal Database Project). Nucleic Acids Res 29:173–174

26. Marilley, L, Hartwig, UA, Aragno, M (1999) Influence of an ele-vated atmospheric CO2 content on soil and rhizosphere bacterialcommunities beneath Lolium perenne and Trifolium repens underfield conditions. Microb Ecol 38: 39–49

27. McCaig, AE, Glover, AL, Prosser, JI (1999) Molecular analysis ofbacterial community structure and diversity in unimproved andimproved grass pastures. Appl Environ Microbiol 65: 1721–1730

28. Moyer, CL, Tiedje, JM, Dobbs, FC, Karl, DM (1996) A computer-simulated restriction fragment length polymorphism analysis ofbacterial small-subunit rRNA genes: efficacy of selected tetramericrestriction enzymes for studies of microbial diversity in nature.Appl Environ Microbiol 62: 2501–2507

29. Mullins, TD, Britschgl, TB, Krest, RL, Giovannoni, SJ (1995) Ge-netic comparisons reveal the same unknown bacterial lineages inAtlantic and Pacific bacterioplankton communities. Limnol Oce-anogr 40: 148–158

30. Rodwell, JS (1992) Grassland and Montane Communities. BritishPlant Communities, vol 3. Cambridge University Press, Cam-bridge, UK

31. Scheid, D, Stubner, S (2001) Structure and diversity of Gram-negative sulfate-reducing bacteria on rice roots. FEMS MicrobiolEcol 36: 175–183

32. Schubler, D, Schwarzott, D, Walker, C (2001) A new fungal phy-lum, Glomeromycota: phylogeny and evolution. Mycol Res 105:1413–1421

33. Shannon, CE, Weaver, W (1948) The Mathematical Theory ofCommunication. University of Illinois Press, Urbana, IL

34. Smit, E, Leeflang, P, Glandorf, B, van Elsas, JD, Wernars, K (1999)Analysis of fungal diversity in the wheat rhizosphere by sequencingof cloned PCR-amplified genes encoding 18S rRNA and temper-ature gradient gel electrophoresis. Appl Environ Microbiol 65:2614–2621

35. Smith, SE, Read, DJ (1997) Mycorrhizal Symbiosis, 2nd ed., Aca-demic Press, New York

36. Straatsma, G, Ayer, F, Egli, S (2001) Species richness, abundanceand phenology of fungal fruit bodies over 21 years in a Swiss forestplot. Mycol Res 105: 515–523

37. Suzuki, MT, Giovannoni, SJ (1996) Bias caused by template an-nealing in the amplification of mixtures of 16S rRNA genes byPCR. Appl Environ Microbiol 62: 625–630

394 J. HUNT ET AL.: 18S rDNA AND FUNGAL DIVERSITY IN GRASSLAND SOILS

38. Tabacchioni, S, Chiarini, L, Bevivino, A, Cantale, C, Dalmastri, C(2000) Bias caused by using different isolation media for assessingthe genetic diversity of a natural microbial population. MicrobEcol 40: 169–176

39. Tebbe, CC, Vahjen, W (1993) Interference of humic acids andDNA extracted directly from soil in detection and transformationof recombinant DNA from bacteria and a yeast. Appl EnvironMicrobiol 59: 2657–2665

40. Thompson, JD, Gibson, TJ, Plewniak, F, Jeanmougin, F, Higgins,DG (1997) The CLUSTAL X Windows interface: flexible strategiesfor multiple sequence alignment aided by quality analysis tools.Nucleic Acids Res 25: 4876–4882

41. Thorn, RG, Reddy, CA, Harris, D, Paul, EA (1996) Isolation ofsaprophytic Basidiomycetes from soil. Appl Environ Microbiol 62:4288–4292

42. Tilman, DA, Knops, J, Wedin, D (1997) The influence of func-tional diversity and composition on ecosystem processes. Science277: 1300–1302

43. Van Elsas, JD, Duarte, GF, Keijzer-Wolters, A, Smit, E (2000)Analysis of the dynamics of fungal communities in soil via fungal-specific PCR of soil DNA followed by denaturing gradient gelelectrophoresis. J Microbiol Methods 43: 133–151

44. von Wintzingerode, F, Gobel, UB, Stackebrandt, E (1997) Deter-mination of microbial diversity in environmental samples: pitfallsof PCR-based rRNA analysis. FEMS Microbiol Rev 21: 213–229

45. Warcup, JH (1950) The soil plate method for isolation of fungifrom soil. Nature (Lond) 166: 117–118

46. Wardle, DA, Giller, KE (1997) The quest for a contemporary eco-logical dimension to soil biology. Soil Biol Biochem 28: 1549–1554

47. Yin, B, Crowley, D, Sparovek, G, DeMelo, WJ, Borneman, J (2000)Bacterial functional redundancy along a soil reclamation gradient.Appl Environ Microbiol 66: 4361–4365

48. Zahl, S (1977) Jackknifing an index of diversity. Ecology 58: 907–913

49. Zhou, J, Bruns, MA, Tiedje, JM (1996) DNA recovery from soils ofdiverse composition. Appl Environ Microbiol 62: 2316–2322

J. HUNT ET AL.: 18S rDNA AND FUNGAL DIVERSITY IN GRASSLAND SOILS 395