Effects of Artificial Defoliation of Pines on the Structure and Physiology of the Soil Fungal Community of a Mixed Pine-Spruce Forest

Loss of photosynthetic area can affect soil microbial communitiesby altering the availability of fixed carbon. We used denaturinggradient gel electrophoresis (DGGE) and Biolog filamentous-fungusplates to determine the effects of artificial defoliation ofpines in a mixed pine-spruce forest on the composition of thefungal community in a forest soil. As measured by DGGE, twofungal species were affected significantly by the defoliationof pines (P < 0.001); the frequency of members of the ectomycorrhizalfungus genus Cenococcum decreased significantly, while the frequencyof organisms of an unidentified soil fungus increased. The decreasein the amount of Cenococcum organisms may have occurred becauseof the formation of extensive hyphal networks by species ofthis genus, which require more of the carbon fixed by theirhost, or because this fungus is dependent upon quantitativedifferences in spruce root exudates. The defoliation of pinesdid not affect the overall composition of the soil fungal communityor fungal-species richness (number of species per core). Biologfilamentous-fungus plate assays indicated a significant increase(P < 0.001) in the number of carbon substrates utilized bythe soil fungi and the rate at which these substrates were used,which could indicate an increase in fungal-species richness.Thus, either small changes in the soil fungal community giverise to significant increases in physiological capabilitiesor PCR bias limits the reliability of the DGGE results. Thesedata indicate that combined genetic and physiological assessmentsof the soil fungal community are needed to accurately assessthe effect of disturbance on indigenous microbial systems.

INTRODUCTION

Top

Introduction

The rhizosphere supports a wide array of organisms, includingbacteria, fungi, and a host of invertebrates. Many of theseorganisms rely in large part on carbon from root exudates forgrowth (21, 29, 31, 37). Thus, any above-ground disturbancethat alters the availability of fixed carbon (e.g., elevatedCO₂, defoliation, and herbivory) could affect the structureand/or function of these soil communities (3, 18, 19, 35, 37,39, 40, 41, 45, 46, 48).

In this study, we focused on the effects of artificial defoliationon the soil fungal community. The defoliation of plants caneither increase carbon allocation to roots, thereby providingincreased resources for microbial growth (25, 29), or decreasecarbon allocation to roots (22, 42). This relationship dependson whether the plant is grazing tolerant (defoliation resultsin increased carbon to shoots) or grazing intolerant (defoliationresults in increased carbon to roots) (22) and on the age ofthe plant (12). These relationships can be complicated, however,by root death in response to defoliation, which results in therelease of nutrients that can increase the carbon and nitrogenresources available to soil communities and can alter soil pHthrough the release of organic acids and amino acids (37).

Fungi comprise one of the most important functional groups ofsoil microbes and are critical to nutrient cycling in terrestrialecosystems (28, 53). Fungi vary in their carbon demands, andthe loss of photosynthetic potential through defoliation candecrease their ability to support and maintain fungal hyphalnetworks (2). Most studies of defoliation effects on fungalcommunities have focused on grasses, and the effects can beeither positive, e.g., increased root colonization by arbuscularmycorrhizae (12), or negative, e.g., reduced ability to maintainhyphal networks (2). These differences in effects depend onthe severity of defoliation, the soil fertility, and the lifehistory strategy of both the fungal species (2, 12, 37) andthe defoliated plant.

Few studies have addressed defoliation or carbon sink effectsin the context of conifers, although soil fungal biomass isproportionate to photosynthesizing biomass (36). Defoliationdecreases fruiting body formation in Betula stands (33), decreasesectomycorrhizal (EM) infection of Douglas fir (32) and Pinyonpine (10, 17), and can alter the EM morphotype diversity ofScots pine (49). Artificial defoliation of only pines in themixed pine-spruce system studied here can significantly alterEM fungal-community structure (9) and significantly affect theEM mututalisms of neighboring, undefoliated spruce (9). Thus,the defoliation of only one tree species in a mixed-speciessetting can have relatively far-reaching effects (9).

Even if the defoliation effects on fungi are limited to EM interactions,EM fungi often form extensive hyphal networks in the soil thatfunction as nutrient-gathering instruments and effectively increasethe surface area across which trees can absorb nutrients (27,47). In this study, we used a combination of molecular geneticand culture-based methods to determine the effects of defoliationof only pines in a mixed pine-spruce forest in Yellowstone NationalPark on both genetic diversity and functional diversity in thesoil fungal community.

We tested the null hypothesis that despite the presence of asecond undefoliated tree species supplying root exudates, thedefoliation of pines in mixed pine-spruce forest would not significantlyaffect the soil fungal community. Our rationale was that thepresence of undefoliated spruce would buffer the community againstthe disturbance and that the experiment would provide insightinto the role of system complexity in maintaining diversityin the face of disturbance. We used denaturing gradient gelelectrophoresis (DGGE) to evaluate changes in soil fungal communitiesin response to defoliation in terms of the presence or absenceand richness of soil fungal species that could be amplifiedusing PCR. We also used filamentous-fungus plates (Fungilog;Biolog, Hayward, Calif.) (11, 16, 23) to determine effects ofdefoliation on soil fungal physiological biodiversity. Thesemethods provide different measures of fungal diversity, a measureof the fungal DNA resident in the soil, and a measure of theculturable members of the fungal community. DGGE measures diversityof the fungal DNA resident in the soil, while Fungilog growthassays measure the potential physiological capabilities of thefungi that are culturable. This study is the first to combinethese methods to study effects of a conifer system on the soilfungal community and to take the possible moderating effectsdue to above-ground species diversity into account.

MATERIALS AND METHODS

Top

Materials and Methods

In the summer of 1998, individuals of Pinus contorta (lodgepolepine) in a mixed P. contorta-Picea engelmannii (Engelmann spruce)stand in Yellowstone National Park, Wyo., were 50% defoliatedby removing every other pair of needles at the base of eachwhorl. The study site was in a forest approximately 100 m fromthe shores of Yellowstone Lake, and the soils were composedof obsidian sand. Coring sites were along the margin of a clear-cutarea in which the litter layer and understory were identicalto those of undisturbed forest (a 3- to 5-cm litter layer andunderstory of Vaccinium scoparium Cov [grouse whortleberry])that has been the subject of previous studies (8, 9). The clear-cutwas created by the National Park Service approximately 50 yearsago, when logs were required to build structures in the park(Bob Lindstrom, Yellowstone Center for Resources, personal communication).Treatment was performed in three paired treatment-control blocks.

Each paired treatment and control block was 5 by 3 m, and eachpair was separated from the surrounding forest by at least 5m to eliminate edge effect. Treatment and control blocks weresituated directly adjacent to each other along a 30-m transect,and the three replicate paired blocks were all within 5 m ofeach other. This spatial arrangement was arrived at empirically(8, 9), was designed to decrease the likelihood that spatialvariation in the fungal community would overshadow treatmenteffects, and has been used by us in the past to assess effectsof artificial defoliation on the EM community of this stand(9).

Trees of both P. contorta and Picea engelmannii species hadreseeded naturally and were 1 to 2 m in height, while P. contortastems significantly outnumbered Picea engelmannii stems (P <0.001). Control blocks contained 60 ± 17 (standard error)P. contorta individuals and 7 ± 2 Picea engelmannii individuals,while treatment blocks contained 68 ± 16 and 10 ±2 individuals, respectively. All individuals of P. contortain treatment plots were 50% defoliated by twisting at the baseof every other needle whorl over the entire canopy.

For DGGE analysis, seven 15-ml soil samples (approximately 2cm in diameter and 12 cm deep; number of cores, 21) were takenfrom within each block, preserved on dry ice for 3 days (untilthey were returned to the lab), and then freeze-dried for long-termstorage. Roots were concentrated in the top few centimetersof each core, and soils for analysis were taken from the rootzone. DNA was extracted from 10-mg soil samples (containingno root material) by using sonication and a detergent-basedmini-prep (7). Detergent-based preps of small amounts of soilcan provide resolution down to 1 culturable cell per 10 mg ofsoil by PCR-based assays (6). To detect fungal DNA in the soilextract, the universal fungal primer set ITS1F-ITS4 was used(13). All reactions were successful. The parameters for PCRwere as follows: initial denaturation at 95°C for 1 min35 s; 13 cycles of denaturation at 94°C, primer annealingfor 55 s at 55°C, and polymerization for 45 s at 72°C;9 additional cycles with the polymerization time extended to2 min; 9 cycles with a 3-min extension; and a final 10-min polymerizationstep at 72°C.

Amplified fungal DNA from soils were subjected to DGGE to separatefungal taxa and make identifications of individual fungal speciespossible. PCR amplification products were gel purified on agarosegels (2% low-melting temperature gels and 1% unmodified gels),the bands were excised with a sterile razor blade, the gel fragmentswere transferred to individual 1.5-ml Eppendorf tubes with 200µl of Tris-EDTA buffer, and the tubes were heated to 65°Cfor 10 min to melt the agarose gel matrix and free the DNA.To ensure consistency within samples, the analysis of sampleextracts (n = 14) from the first set of paired blocks was repeatedthree times with each sample. There was no variation among samples;thus, each of the remaining samples was assayed once.

For each DGGE assay, 12.5 µl of a 1:10 dilution of amplifiedDNA in Tris-EDTA (10 ml of 1 M Tris [pH 8.0] plus 4 ml of 0.25 M EDTA [pH 8.0] in 986 ml of distilled H₂O) was used as aPCR template in a reaction to add GC clamp tails. The PCR protocolwas the same as before except that one primer (ITS1F) had anadditional 40-bp GC tail.

A DCode DGGE universal detection system apparatus (Bio-Rad,Hercules, Calif.) was used to perform the DGGE analysis accordingto the manufacturer's specifications and as previously described(13), except that a 20 to 60% denaturant gradient was used.Individual bands were cut out and sequenced on an automatedsequencer (ABI Prism 377; Perkin-Elmer, Boston, Mass.) at CaliforniaState University, Hayward. Sequence data were run through theBLAST search programs of the National Center for BiotechnologyInformation (www4.ncbi.nlm.nih.gov/) to determine the taxonomicaffiliation of each DGGE band by sequence homology.

To determine the functional diversity of the soil fungal community,we used Fungilog microtiter plates. For each assay, a 10-g soilsample (three from each of the treatment and control blocks;a total of nine samples for each experiment) was shaken in 100ml of Ringer's solution (20) (three times for 1 min each). Thissolution was allowed to settle for 15 min, and 150-µlaliquots were pipetted onto the Fungilog microtiter plates.Each plate contained 95 wells with different carbon sourcesand a single control well with no carbon source. Plates wereincubated at room temperature with antibiotics (50 µgeach of streptomycin and ampicillin/ml) to ensure that resultswere due to fungal and not bacterial growth. Every 24 h, for5 days, plates were scanned with an optical density plate reader(at 490 nm). This method produces 96 data points (95 substratesplus one water control) per replicate, per day, with each correspondingto a specific carbon substrate. Growth is measured colorimetricallywhen the tetrazolium dye in the medium accepts electrons fromactive electron transport. Plates were read before growth becametoo massive, that is, before overrunning the plates and becomingnonlinear.

The significance of the effects of defoliation on the presenceor absence and abundance of soil fungal species was determinedwith a contingency table and ${chi}$ ² analysis for the whole communityand Fisher's exact test for the species that exhibited differencesfollowing treatment. Changes in species richness, defined asthe number of unique DGGE bands (or fungal species) per unitarea (in this case per soil sample), were analyzed by usingStudent's t test. Principal-components analysis (PCA) was usedto project the data onto a small number of statistically independentaxes that account for the majority of the variability in thedata. This projection results from identification of the eigenvectorsof the covariance matrix of the data. Then the data were projectedonto the eigendirections with the largest eigenvalues. The significanceof the effects on physiological diversity, as indicated by totalnumber of substrates utilized and rates of substrate utilization,was determined with Student's t test, with each day being analyzedindividually.

RESULTS

Top

Results

The defoliation of only pines in a mixed pine-spruce forestresulted in no significant effect on species richness of thepredominant soil fungal species (number of species per sample).Both control samples and treatment samples contained five speciesper sample (±0.4 for controls, and ±0.5 for treatments)as indicated by DGGE analysis. Likewise, ${chi}$ ² analysis of the totalcommunity indicated no significant effect on overall soil fungal-communitycomposition; the majority of fungal species were present atthe same frequency in both treatment blocks and control blocks(Table 1). Most BLAST analyses indicated that the DGGE bandswere $≥$ 95% similar (excluding unknown bases) to those of EM fungiof conifers. The two exceptions were taxon 10, which had highsequence similarity ( $≥$ 95%) with a species of the genus Eupenicillium,a free-living ascomycete, and taxon 8, which, though clearlya basidiomycetous fungus, was not closely related to any fungusrepresented by a sequence in GenBank (82% match to any fungusin GenBank and to only 300 bases of more than 600 sequenced).This sample was sequenced from multiple amplifications and inboth directions in order to rule out the presence of chimeras.

View this table:
[in this window]
[in a new window]
TABLE 1. Presence or absence and abundances of fungal taxa

Although species richness was not affected, there was a significanteffect (P < 0.001) when the frequency of individual fungiwas analyzed with Fisher's exact test. Specifically, Cenococcumsp. (taxon 9, an ascomycete fungus that is an important EM associateof pines) was common in control plots but nearly eliminatedfollowing treatment, while the unidentified taxon 8 was relativelyrare in controls but common in treatment blocks.

The defoliation of only pines in a mixed pine-spruce forestresulted in significant changes in physiological capacity asmeasured by Fungilog tests (P < 0.001), both in terms ofcarbon source utilization (with growth being indicated by tetrazoliumdye) (Fig. 1A) and in the number of carbon substrates utilized(Fig. 1B). Both measures were significantly higher in treatmentsamples from 72 h until the end of the experiment. Repeatedmeasures with a further sampling of equal size produced similar,statistically significant results.

View larger version (20K):
[in this window]
[in a new window]
FIG. 1. Results of Biolog filamentous-fungus assays. (A) Total activity; (B) number of substrates utilized. ${diamondsuit}$ , treatments; ${blacksquare}$ , controls. Only results from days 2 to 5 are depicted, as activities on day 1 were zero.

PCA placed samples into one cluster (Fig. 2) that contained8 of 9 control cores and three treatment cores. The remainingcores were in a relatively loose assemblage that did not forma cohesive cluster and were distinct and separate from the controlcluster. Separation of the control cluster was based primarilyon greater utilization of sugars. The first two principal componentsaccounted for 72 and 10% of the variation in the data. The firstcomponent was correlated with 22 sugars (n-acetyl-D-glucosamine,D-arabitol, dextrin, D-fructose, galactose, D-galacturonic acid,gentiobiose, D-gluconic acid, D-glucose, 2-keto-gluconic acid,maltose, D-mannitol, D-mannose, palatinose, D-melezitose, sucrose,D-trehalose, D-xylose, D-cellobiose, glycogen, D-glucuronicacid, and D-ribose), seven fatty acids (Tween 80, ketoglutaricacid, L-lactic acid, L-maleic acid, quinic acid, D-saccharicacid, and succinic acid), and seven amino acids (L-alanine,L-alanylglycine, L-asparagine, L-phenylalanine, L-glutamic acid,glycyl-L-glutamic acid, and L-serine), while the second componentwas correlated with 15 sugars (D-arabitol, dextrin, D-fructose,gentiobiose, D-gluconic acid, 2-ketogluconic acid, maltose,D-mannitol, D-mannose, D-melezitose, palatinose, sucrose, D-trehalose,and D-ribose), seven fatty acids (Tween 80, ketoglutaric acid,L-lactic acid, L-maleic acid, quinic acid, D-saccharic acid,and succinic acid), and five amino acids (L-alanine, L-glutamicacid, glycyl-L-glutamic acid, L-serine, and uridine [a nucleicacid]).

View larger version (10K):
[in this window]
[in a new window]
FIG. 2. PCA of Fungilog data. T corresponds to treatments, and C corresponds to controls.

DISCUSSION

Top

Discussion

Changes in below-ground allocation of carbon resulting fromabove-ground disturbances (such as defoliation and herbivory)can alter both the structure and the physiological diversityof microbial communities (e.g., references 3, 12, 37, 39, 40,41, 45, and 46).

DGGE analysis indicated that defoliation of only pines in experimentalblocks of a mixed stand of pine and spruce can affect the abundanceof individual soil fungal species but that species richness(number of species per core) and overall soil fungal-communitycomposition are not significantly affected. Most of the fungidetected by DGGE have high sequence similarity to EM fungi,with a few being similar to free-living fungal taxa. None ofthe fungi with high sequence similarity to free-living fungiwere affected by defoliation.

One fungus that was affected by defoliation, Cenococcum, wasmore abundant in controls, and a second, taxon 8, was more abundantin treatments and thus appeared to respond positively to thechanges induced by defoliation. Cenococcum is a well-studiedEM fungus that is very common in pine forests, although littleis known about how this fungus reacts to defoliation. SinceCenococcum can form a large biomass and extensive rhizomorphsin forest soil (14, 30), a reduction in photosynthetic potentialand concomitant reduction in the fixed carbon allocated to theroots of the EM plants could reduce the amount of this funguspresent.

We also hypothesize that the presence of nondefoliated sprucehelped maintain fungal-species richness in these nutrient-poorsoils and may also have contributed to the measured increasein Cenococcum organisms. Our data suggest that roots of thenondefoliated spruce tree can support the EM fungi of the defoliatedpines (9) and hence help to maintain EM fungal-community compositionand infection levels. In addition, growth of fungal endophytescan be influenced by the presence of specific host plant carbohydrates(24). Thus, Cenococcum may respond more readily to spruce rootexudates and increase in frequency when roots of spruce proliferatefollowing pine defoliation (9).

The Fungilog experiments indicated significant changes in thephysiological capabilities of the fungal communities. PCA clustered8 of 9 control cores into a distinct group, while most of thetreatment cores were in a looser assemblage that was distinctfrom the control group. These groupings were based primarilyon differences in carbohydrate utilization. The single controlcore that was separate from the main control group was takennear the edge of adjacent treatment and control blocks, andthe hyphae could have crossed over into adjoining blocks. Threeof the nine treatment cores also clustered with the majorityof the control cores. We hypothesize that these cores containedhyphae responding to exudates from spruce, or communities thatwere still in transition, despite the passage of 2 years.

Fungilog results also indicated increased physiological diversityin response to defoliation in the form of increases in boththe range of utilizable substrates and their rate of utilization.This change occurred despite the presence of the second undefoliatedtree species. Thus, quantitative and qualitative differencesin root exudates of the two tree species may play a role ininfluencing fungal-community structure and function.

However, Fungilog results differ from those of the DGGE analysis,which indicated very little change in species composition andrichness. The discrepancy between the two methods may be dueto two factors. First, if a single fungal species has a widerange of carbon substrate usage capabilities, the relativelylarge changes in physiological diversity may result even withonly minor changes in fungal-community composition. Fungi canhave very different niche sizes, defined as the number of carbonsubstrates each species or strain can utilize. Under some conditions,this can range from 18 of 95 Biolog substrates utilized to 92of 95 Biolog substrates utilized (34). Studies of fungal endophytesof conifers are consistent with these observations and indicatethat strains specific to petioles, for example, have a muchbroader substrate utilization capability than do those restrictedto needles (4). Thus, the increase in taxon 8 organisms mightsuffice to produce the Fungilog patterns that we observed. Wewill test this hypothesis by assessing the physiological capabilitiesof the fungus that increased in frequency when or if the fruitingbodies needed for culturing this fungus become available.

A second explanation for the differences between Fungilog resultsand DGGE results could be PCR amplification bias. This biasoccurs when primers are used in multiplex reactions (50), particularlyif the fungi present in the soil lack sequences that correspondexactly to the primers used. Our primers have been tested extensivelyagainst fungi across the Eumycota (15), so we consider thispossibility unlikely. However, even when perfect sequence homologyexists, other factors may inhibit amplification. For example,the secondary structure of the DNA flanking the priming sitecan cause DNA from some species to be preferentially amplified(26), and thus some species might not be detected, even if theyare present in relatively high numbers. We compared the sequencesobtained in this study with those obtained in our previous studyof EM in these study plots (9). None of the EM fungal speciesdetected in our previous study were present in our DGGE analysisof the same soils. Thus, the EM fungi that we detected in ourprevious study form little or no hyphal biomass in soils, orthere was extensive turnover in the EM fungal community. Weplan to test the primer bias hypothesis by repeating the Fungilogassays and then sequencing and identifying the fungi that arecultured on each substrate.

In summary, our data indicate that the defoliation of only pinesin a mixed pine-spruce study site can significantly affect therelative abundance of some soil fungal species, though overallspecies richness and composition are largely unaffected. Furthermore,the defoliation had a significant effect on the physiologicaldiversity of the soil fungal community. However, these resultsindicate that care must be taken when interpreting the dataand that independent assessments of genetic diversity withinFungilog cultures of the type that we plan to perform may beuseful. Given the increasing importance placed on studies ofbiodiversity and physiological diversity, and roles of biodiversityand physiological diversity in ecosystem resilience (e.g., references1, 5, 36, 38, 43, 44, 51, and 52), it is clear that studiesutilizing methods that measure both species composition andspecies physiological diversity are needed.

ACKNOWLEDGMENTS

This work was supported by an NSF grant (DEB/RUI 9420141) toKen W. Cullings and Virgil T. Parker and a NASA-DDF grant toKen W. Cullings.

We thank J. R. Blair for mushroom identifications, Bob Lindstromand Ann Deutch of the Yellowstone Center for Resources, AnnRodman of the Yellowstone Soil Survey, Alan Kerner for logisticalsupport, and Chris Baysdorfer of California State University,Hayward, for DNA sequencing.

FOOTNOTES

* Corresponding author. Mailing address: NASA-Ames Research Center, Mountain View, CA 94035-1000. Phone: (650) 604-2773. Fax: (650) 604-3954. E-mail: kcullings{at}mail.arc.nasa.gov.

REFERENCES

Top