Influence of Elevated CO2 on the Fungal Community in a Coastal Scrub Oak Forest Soil Investigated with Terminal-Restriction Fragment Length Polymorphism Analysis

Sixteen open-top chambers (diameter, 3.66 m) were establishedin a scrub oak habitat in central Florida where vegetation wasremoved in a planned burn prior to chamber installation. Eightcontrol chambers have been continuously exposed to ambient airand eight have been continuously exposed to elevated CO₂ attwice-ambient concentration ( $~$ 700 ppm) for 5 years. Soil coreswere collected from each chamber to examine the influence ofelevated atmospheric CO₂ on the fungal community in differentsoil fractions. Each soil sample was physically fractionatedinto bulk soil, rhizosphere soil, and roots for separate analyses.Changes in relative fungal biomass were estimated by the ergosteroltechnique. In the bulk soil and root fractions, a significantlyincreased level of ergosterol was detected in the elevated CO₂treatments relative to ambient controls. Fungal community compositionwas determined by terminal-restriction fragment length polymorphism(T-RFLP) analysis of the internal transcribed spacer (ITS) region.The specificities of different ITS primer sets were evaluatedagainst plant and fungal species isolated from the experimentalsite. Changes in community composition were assessed by principalcomponent analyses of T-RFLP profiles resolved by capillaryelectrophoresis. Fungal species richness, defined by the totalnumber of terminal restriction fragments, was not significantlyaffected by either CO₂ treatment or soil fraction.

INTRODUCTION

Top

Introduction

During the last decade a number of studies have been conductedto assess the influence of elevated atmospheric CO₂ on abovegroundplant community diversity and physiology as well as on belowgroundroot dynamics, root exudates (6, 8, 22, 56, 60, 66), and nutrientavailability (27, 29). In most cases these investigations documentan impact of elevated CO₂ upon plant growth and water requirements,but the reaction of the belowground microbial community hasbeen less clear-cut. Some studies have reported no effect ofelevated CO₂ (45, 67), while others have reported increasesin microbial activity (50), in microbial biomass (9, 49), orin the number of mycorrhizal infections (45). Few studies haveattempted to directly assess changes in the composition andrichness of the microbial community as a whole in response toelevated CO₂. Among these studies, few have documented changesin either the richness or evenness of microbial community composition.For example, Griffith et al. (23) used community DNA hybridizationand percent G+C base profiling to detect changes in rhizospheremicrobial communities of wheat. They found only insignificantdifferences between treatments. Zak et al. (67), using phospholipidfatty acid analysis, did not find any significant changes inmicrobial community composition caused by elevated CO₂ in soilbeneath the bigtooth aspen (Populus grandidentata). Both Griffithet al. (23) and Zak et al. (67) used relatively broad-scaleprofiling techniques that are only able to detect changes atthe genus level or higher. Montealegre et al. (40) on the otherhand, using PCR-based fingerprinting of genomic DNA, reportedthat Rhizobium leguminosarum isolated from white clover grownat elevated CO₂ concentrations were genetically different fromisolates obtained from plants grown at ambient CO₂ concentrations.Taken in aggregate, these results suggest that microbial responsesto elevated CO₂ concentrations may be multifactorial, difficultto measure, and easily confounded by other factors. Measuresof microbial community composition may have been influencedby the experimental system used to elevate CO₂ (45), the specifictypes of plants examined (51), the sensitivity of the microbialcommunity profiling method (23, 67), or the diversity of microbialtypes present within the community (40). Molecular methods arenot immune to sampling bias and technology limitations thatmay obscure the subtle signal of microbial compositional shiftsamidst the noise of community heterogeneity. It has thereforebeen problematic to clearly separate CO₂ treatment effects fromartifacts of either experimental design or sampling.

Terminal-restriction fragment length polymorphism (T-RFLP) analysisis a direct DNA profiling method that has been used extensivelyto assess microbial community structure in habitats whose communitycompositions are both complex and diverse (11, 12, 35, 38, 46).The method has been targeted to multiple scales of phylogeneticresolution from species to domain level and is applicable tobacteria as well as archaea and eukarya (for a review, see reference31). The primary molecular target of T-RFLP community profilinghas been small-subunit rRNA (i.e., 16S and 18S ribosomal DNA[rDNA] genes), although functional gene applications have beendeveloped for the mer (4), amoA (26), nosZ (52), and nirS (3)genes among bacteria. Multiple PCR primers that target the fungalrDNA subunits and intergenic spacer regions of diverse taxonomicgroups have been reported for the strain typing of fungal isolates(58, 65). Only recently, however, have primers targeting small-and large-subunit rDNA or ribosomal intergenic spacer regionsbeen developed for the specific amplification of fungal rDNAsequences directly from environmental and clinical samples (2,28, 39, 57, 59). Here we describe the application of T-RFLPfungal community profiling of direct environmental extractswith primers targeting the internal transcribed spacer (ITS)regions of ascomycetes and basidiomycetes, or of basidiomycetesonly, in order to detect changes in fungal community compositionin response to elevated atmospheric CO₂.

A second aim was to examine the effect of elevated CO₂ on thequantity and partitioning of fungal biomass belowground as determinedfrom the ergosterol content in the soil and root fractions,respectively. Ergosterol has been extensively used to quantifyfungal biomass in soil (10, 21, 53, 64) and mycorrhizal systems(14, 44). As a specific measure of fungal biomass, ergosterolhas several advantages. It is relatively easy to extract (13,47, 55) and is in many cases the dominant fungal sterol (34,61). Ergosterol is not exclusively found in fungi, but severalworkers have shown that it is a more sensitive and reliableindicator of viable fungal biomass than other biochemical molecules(7, 36, 42, 64).

There are, however, methodological limitations to this measure.Foremost among these is the large variation in ergosterol contentamong different fungal species (34, 41, 42) and the presenceof ergosterol in a few nonfungal eukaryotes. Some green algae,for example, have been shown to produce ergosterol in smallamounts, complicating the interpretation of this measure inaquatic environments (7, 43). In addition, changes in ergosterolcontent may primarily reflect changes in species compositionrather than changes in total biomass in some environments (1,47)—a problem exacerbated by the paucity of fungal speciesand environments examined so far (see, e.g., reference 34, 62,and 63). Although these problems limit interpretation of ergosterolas a quantitative measure of total fungal biomass across environments,the ergosterol assay remains the single most reliable relativeindicator of fungal biomass in natural biological systems (7,15, 21, 42, 55, 64).

MATERIALS AND METHODS

Top

Materials and Methods

Study site.
An experiment to monitor the long-term effects of elevated atmosphericCO₂ in a nutrient-limited coastal scrub oak forest was initiatedand funded by the U.S. Department of Energy in 1995 at the JohnF. Kennedy Space Center in Florida as part of a collaborativeresearch project between the National Aeronautics and SpaceAdministration and the Smithsonian Institution (27). The sitewas cleared of scrub vegetation in a controlled burn in 1995prior to installation of 16 open-top chambers, of which 8 weresupplied with ambient air and 8 were supplied with ambient airplus 350 µl of CO₂ liter of air^-1 (29).

Sampling.
A single soil core (diameter, 5 cm; mean depth, 10 cm) was takenfrom each of the chambers and separated into three fractions:bulk soil, rhizosphere soil, and roots. The soil fractions wereobtained by passing freshly collected sample through a 2-mm-pore-sizesieve. The sieved soil was named the bulk soil. The soil attachedto roots (rhizosphere) was washed off with sterile deionizedwater and collected in a sterile bucket. The resulting soilslurry was centrifuged at 500 x g for 20 min to separate thewater from the rhizosphere soil.

From each of the three subsamples (bulk soil, rhizosphere soil,and roots), approximately 5 g (wet weight) was used to determinedry weight after drying at 80°C for 24 h and ash content(ash generated at 600°C for 6 h). The soil moisture contentwas 6.5% ± 2.7% (wt/wt) (mean ± standard deviation)and the soil pH was in the range of 4.9 to 5.5, with no differencebetween CO₂ treatments.

Extractions.
The bulk and rhizosphere soils were divided into subsamplesfor ergosterol and DNA analyses, respectively. The scarce amountof roots in the samples yielded only sufficient material fordry weight and ergosterol determinations. Bulk and rhizospheresoil samples for DNA extraction were stored at -80°C untilanalysis, while ergosterol was extracted immediately using theprocedure of Pennanen et al. (48).

DNA was extracted from 0.25-g soil fractions using the UltraCleansoil DNA isolation kit (Mo Bio Laboratories, Solana Beach, Calif.).The manufacturer's protocol was followed, with the modificationssuggested by Clement and Kitts (5). The yield was determinedusing the Pico-green assay (Molecular Probes Inc., Portland,Oreg.) according to the manufacturer's protocol in a Dynex MFXmicroplate fluorometer with serial dilutions of each extract.Bulk soil yielded 1.8 to 12.8 ng of DNA/µl, and rhizospheresoil yielded 3.3 to 45.9 ng of DNA/µl.

PCR was performed in 50-µl reaction volumes using fluorescentlylabeled forward (6-FAM) and reverse (HEX) oligonucleotide primers(synthesized by Operon Technologies, Alameda, Calif.) targetingthe ITS region of the rDNA operon. Different primer sets weredesigned to be specific for different groups of fungi as follows:ITS 1F-ITS 4, specific for higher fungi; ITS 1F-ITS 4B, specificfor basidiomycetes (18); and ITS 1F-ITS 4A, specific for ascomycetes(33). Finally we tested the combination ITS 1-ITS 4B at increasedannealing temperature (60°C) in order to increase specificityfor basidiomycetes. Final concentrations in the PCRs were: 2.0mM MgCl₂, 1x buffer (Applied Biosystems [Foster City, Calif.]buffer II), a 200 µM concentration of each deoxynucleosidetriphosphate, a 1.0 µM concentration of each primer, 0.4µg of bovine serum albumin/µl, and 1.25 U of AmpliTaqDNA polymerase (Applied Biosystems). The thermocycler reactionconditions were as follows: 5-min initial denaturation at 94°Cfollowed by 35 cycles of 0.5 min at 94°C, 2 min of annealingusing different temperatures for the different primer sets (asindicated in Table 1), and a 3-min extension at 72°C. Thefinal extension was 5 min at 72°C. Reaction yield and targetspecificity were determined by agarose gel electrophoresis.The PCR products were cleaned using a QIAquick PCR purificationkit (Qiagen, Hilden, Germany), digested with different restrictionendonucleases (HinfI, TaqI, and MseI at 5 U per reaction accordingto manufacturer's instructions), cleaned with a QIAquick nucleotideremoval kit (Qiagen), denatured at 95°C for 10 min, andseparated by capillary electrophoresis on an ABI 310 geneticanalyzer in GeneScan mode (Applied Biosystems).

View this table:
[in this window]
[in a new window]
TABLE 1. Specificities of primer sets used in this study

In addition to collecting these soil samples, we also collectedsoil for isolation of fungi by enrichment plating accordingto the procedure described by Klamer and Søchting (32).The soil particles were plated on malt extract agar and carboxymethylcellulose agar (17) and incubated at room temperature. Fungigrowing on the plates were identified morphologically at leastto the genus level. Mushrooms found on the surface in the areawere also collected during the autumn of 2000. Because of asevere drought, mushrooms were collected in a wider area thanthe study site. DNA was extracted and analyzed using the samemethods as described above with the sole exception being thatfungal tissue from fruiting bodies was extracted using bead-beatinglysis instead of heat.

The fungal DNA extracted from the pure cultures was used tooptimize PCR parameters, to test the specificity of primer sets,and to select appropriate restriction enzymes for use in thecommunity analysis.

To further test the specificity of the primer sets, DNAs fromleaves from 11 plant species from the site (Table 1) constitutingmore than 90% of the aboveground biomass (54) were extractedand amplified using the technique described above (yield, 0.7to 2.7 ng of DNA/µl). The DNAs from three oak speciesand Serenoa repens were extracted using DNeasy (Qiagen) forplant DNA extraction, since the Mo Bio kit did not yield amplifiableDNA for these species. Prior to extraction the leaves were surfacesterilized using the following procedure: 1 min in 95% ethanol,3 min in 4% sodium hypochlorite solution, 30 s in 95% ethanol,5 min of washing in sterile water. To determine if the plantDNA was suitable for amplification, PCR was conducted usinga universal eukaryotic primer set: Euk SSU_f-Euk SSU_r (37)(Table 1).

Principal component analyses (PCA) were performed using SPSS10 software. Only terminal restriction fragments (TRFs) thatwere present in at least four samples were included. Each individualTRF was scored as present or absent and analyzed as binary data.

RESULTS

Top

Results

Fungal biomass.
The amount of ergosterol calculated on the basis of organicmatter (OM) found in the three soil compartments is shown inFig. 1. In the bulk soil and root fractions, the increase inergosterol in the elevated CO₂ chambers was significant (two-tailedt test, P = 0.01 and 0.05, respectively [n = 8]) compared tothe ambient chambers. There was no difference in ergosterolcontent between the soil fractions for either ambient or elevatedchambers.

View larger version (18K):
[in this window]
[in a new window]
FIG. 1. Influence of elevated atmospheric CO₂ concentrations on ergosterol content in three different fractions of soil (n = 8). Closed bars, ambient CO₂ level; open bars, elevated CO₂ level. Error bars represent one standard error.

Fungal community analysis.
As shown in Table 1, none of the plant species amplified withthe primer set specific for higher fungi (ITS 1F-ITS 4). Theannealing temperature was lowered to 50°C with this primerset since some of the ascomycetes failed to amplify at 55°Cin pure culture. The three oak species and approximately two-thirdsof the ascomycetes did, however, amplify with the basidiomycete-specificprimer set (ITS 1F-ITS 4B), although the amplifications wereweaker than those for the basidiomycetes. This led us to testthe ITS 1-ITS 4B combination, since these primers have higherpredicted melting temperatures (64.5 and 66.3°C, respectively).This allowed us to increase the annealing temperature to 60°Cand thereby increase the specificity while maintaining amplificationof all the tested fungi extracted from pure culture. As a resultof these primer evaluations, the ITS 1F-ITS 4 primer set wasused as a specific primer set for the fungal community (ascomycetesand basidiomycetes), and the ITS 1-ITS 4B primer set was usedfor basidiomycetes. We also tested the primer set ITS 1F-ITS4A, designed to be specific for ascomycetes (33), but we werenot able to obtain satisfactory specificity with this primerset and its application for community analyses was terminated(Table 1).

In both the bulk and rhizosphere soils the increase in atmosphericCO₂ did not result in significant changes (t test) in the totalrichness of either the fungal community or the basidiomycetes,as evaluated by the total number of TRFs (Table 2). Only fewTRFs very common for both bulk and rhizosphere soils, e.g.,for TaqI TRFs 1 to 9 (average, 3.1) TRFs were common for thetwo soil fractions from the 16 samples. In the rhizosphere soilthere was no difference in species richness between the fungalcommunity and the basidiomycetes evaluated with HinfI. Whenevaluated with TaqI, on the other hand, the number of TRFs wassignificantly lower for the basidiomycete fraction comparedto the whole fungal community for both the ambient and elevatedCO₂ treatment (t test, P = 0.04 and 0.003, respectively).

View this table:
[in this window]
[in a new window]
TABLE 2. Species richness evaluated by number of terminal restriction fragments

No impact of the CO₂ treatment was observed in the fungal communityin either the bulk or the rhizosphere soil as evaluated fromthe PCA (Fig. 2). There was a large variation between samplesfor both treatments.

View larger version (12K):
[in this window]
[in a new window]
FIG. 2. PCA showing the influence of elevated atmospheric CO₂ concentrations on the fungal community in soil. Symbols: •, ambient CO₂ level; ${blacktriangledown}$ , elevated CO₂ level. (A) Bulk soil; (B) rhizosphere soil.

A PCA of the basidiomycetes in the rhizosphere soil is shownin Fig. 3. Although there is a large variation, there is a groupingof the ambient samples to the right and the elevated CO₂ tothe left, indicating that there may have been a shift in thecomposition of the community of Basidiomycetes caused by theelevated CO₂. When tested using a Mantel test (Mantel-Struct1.0) on all the samples, this change was not significant (P= 0.646), but when the two outliers in Fig. 3 were excluded,a significant treatment effect was found (P = 0.040). No treatmenteffect was observed for the basidiomycetes in the bulk soil(data not shown).

View larger version (10K):
[in this window]
[in a new window]
FIG. 3. PCA showing the influence of elevated atmospheric CO₂ concentrations on the basidiomycetes in the rhizosphere soil. Symbols: •, ambient CO₂ level; ${blacktriangledown}$ , elevated CO₂ level.

DISCUSSION

Top

Discussion

Fungal biomass increased in each of the three soil fractionsin response to the CO₂ treatment (Fig. 1), although this increasewas only significant for the bulk soil and root fractions. Theincreased ergosterol content can be interpreted as a secondaryresponse to elevated atmospheric CO₂ in this way: in the elevatedCO₂ treatments, the tree species Quercus myrtifolia, Q. geminata,and Q. chapmanii, which comprise more than 60% of the abovegroundbiomass (54), allocate more carbon to the ectomycorrhiza inthe root fraction. The carbon is transferred from the rootsto the extraradical mycelium via the Hartig net, resulting inan increase in fungal biomass in the root and the two soil fractions.

An increase in root length densities was also observed in theelevated CO₂ treatment compared to the ambient treatment, 14.2versus 8.7 mm cm^-2, respectively (8). This result cannot explainthe increased amount of ergosterol in the root fraction, sincethe ergosterol concentration was measured per unit root. Therange found in this study, 80 to 130 µg/g of OM, fallswithin the range reported by Frostegård and Bååth(16). They measured ergosterol content in a series of Spruceand Beech forests, and found concentrations in the range of35 to 218 µg/g of OM. In contrast to the findings in thisstudy, however, Wiemken et al. (66) found increased fungal biomassonly in a nutrient-rich calcareous soil and no effect in a nutrient-poorsiliceous soil. They used the amount of the phospholipid fattyacid, 18:2 ${omega}$ 6,9, to estimate the fungal biomass in a Beech-Sprucemodel system exposed to elevated atmospheric CO₂.

With the reported combinations of ITS primers, it was possibleto amplify higher fungi and basidiomycetes, respectively, withoutamplification of plant nucleic acids. The combination of ITS1-ITS 4B with an elevated annealing temperature of 60°Cyielded a higher degree of specificity for basidiomycetes inthis study than the ITS 1F-ITS 4B primer set. The ITS 1-ITS4B primer combination remains to be tested on a larger collectionof different basidiomycete species in order to verify whetherthis primer set actually can amplify all the basidiomycetesfound in this ecosystem. Glen et al. (20) tested three differentprimer sets targeting the ITS region and found a high specificityfor basidiomycetes with one of them. However, none of the primersets were able to amplify all 94 of the basidiomycetes tested.

The species richness, evaluated as the total number of TRFswas not influenced by the CO₂ treatment, and the richness seemedto be identical in the bulk and rhizosphere soil, respectively(Table 2). Relative species richness was expected to decreasein the basidiomycete community compared to the whole fungalcommunity, as the basidiomycetes only constitute a fractionof the whole fungal community. In the rhizosphere soil the speciesrichness evaluated with HinfI was about 21 for both the wholefungal community and the basidiomycete community, while a significantreduction in richness was observed with TaqI. The pattern obtainedwith TaqI was confirmed with another tetrameric restrictionenzyme, MseI (data not shown). The reason why the results fromHinfI differed from those observed with TaqI and MseI is notknown, but this difference emphasizes the importance of usingmore than one restriction enzyme when evaluating microbial communityprofiles from environmental samples.

PCA of both the bulk and rhizosphere soil fraction fungal communitiesfailed to reveal any significant CO₂ treatment effect (Fig.2), but any signal may have been obscured by the large variationbetween samples. Mycorrhizal fungi are known to be nonrandomlydistributed (i.e., spatially clustered) in soil systems. Gardesand Bruns (19) have reported that the majority of mycorrhizalfungal species occurred in less than 10% of the soil samplestaken. In a recent review by Horton and Bruns (25), it was notedthat the most abundant types of fungal biomass occurred in onlyone or two soil samples, suggesting a highly heterogeneous distribution.Furthermore, canopy cover has been shown to impact the amount,if not also the type, of ectomycorrhizal fungi associated withoak (68). Changing the sample locations only slightly wouldchange the perception of species presence and estimates of richnessconsiderably. In this study, we have eight replicate samplesfrom the two treatments, which should yield sufficient sampledensity to detect major changes in fungal community compositionresulting from the CO₂ treatments. It is obvious, however, thatadditional spatial and temporal sampling intervals will be requiredin future studies to dampen the noise inherent in patchy ecosystemsand augment any possible signal of community compositional changesarising from elevated CO₂ treatments.

Despite a large variation among replicate samples in the rhizospheresoil PCA plot, the data support a change in the basidiomycetecommunity as a consequence of the elevated atmospheric CO₂ (Fig.3). Other unaccounted factors may, however, play a role as well.For example, there is an indication of a change in the compositionof oak species, with Q. geminata being less abundant in theplots with elevated CO₂ levels. Changes in the composition ofthe basidiomycete community may therefore be a secondary effectarising from the altered oak species composition, as it is likelythat a large proportion of the basidiomycetes are ectomycorrhizalfungi associated with particular plant species. There is otherevidence, however, that at least part of the ectomycorrhizalfungi are able to colonize a variety of tree species (24, 30),which could counteract the importance of changes in the abovegroundplant species composition. Any change in either the oak speciescomposition or their physiology would result in changes in rootphysiology and the rhizosphere that would affect the associatedectomycorrhiza and fungi. Fungi may be critical reservoirs ofcarbon in systems with elevated atmospheric CO₂ since they constitutethe highest amount of microbial biomass in the soil system andare intimately involved in the flux of carbon channeled directlyfrom the aboveground biomass and indirectly from the atmosphere.Given the rather limited knowledge of fungal diversity and ecosystemfunction in elevated CO₂ environments, it is unclear what potentialimpact any changes in fungal composition would have on communitystability at either a functional diversity or richness level.

We conclude that 5 years of continuous treatment with elevatedatmospheric CO₂ concentrations has resulted in a higher fungalbiomass in the soil. Neither fungal community composition norrichness, however, was demonstrably affected by the elevatedCO₂ treatment. The change in the composition of basidiomycetesis believed to be a secondary effect reflecting changes in eitherthe plant species composition (e.g., oaks) or a shift in theectomycorrhizal community on oak roots arising from changesin carbon flux.

ACKNOWLEDGMENTS

This project was supported by a NASA Technology DevelopmentGrant from Kennedy Space Center through the University of CentralFlorida (M.K.).

We are grateful for the support by S. and G. Sheine in collectingand identifying basidiomycetes. We thank G. Hymus and P. Schmalzerfor their help with identification of the plant species.

FOOTNOTES

* Corresponding author. Mailing address: Lund University, Department of Ecology, Ecology Building, Getingegatan 60, S-223 62 Lund, Sweden. Phone: 46 46 222 37 64. Fax: 46 46 222 4716. E-mail: morten.klamer{at}zooekol.lu.se.

REFERENCES

Top