ectomycorrhizal fungi identification in single and pooled root samples: terminal restriction...

Ectomycorrhizal fungi identification in single and pooled

root samples: terminal restriction fragment length polymorphism

(TRFLP) and morphotyping compared

David J. Burkea,*, Kendall J. Martinb, Paul T. Rygiewiczc, Mary A. Topaa

aBoyce Thompson Institute for Plant Research, Tower Road, Ithaca, NY 14853, USAbCenter for Environmental Diagnostics and Bioremediation, University of West Florida, 11000 University Parkway, Pensacola, FL 32514, USA

cUSEPA National Health and Environmental Effects Research Laboratory, 200 SW 35th Street, Corvallis, OR, USA

Received 12 July 2004; received in revised form 2 December 2004; accepted 3 January 2005

Abstract

We describe here the results of a study conducted to evaluate a terminal restriction fragment length polymorphism (TRFLP) approach

targeting rRNA genes for determination of ectomycorrhizal (ECM) communities colonizing the roots of loblolly pine (Pinus taeda L.). Root

tips separated from soil cores were classified according to morphological characteristics and DNA was then extracted from each group of

morphotyped tips. Labeled primers were used to generate terminal restriction fragments (TRF) for molecular fingerprinting of root

colonizing fungi and to determine how well TRFLP could be used to discriminate between ectomycorrhizal types. Morphotypes generally

correlated well with specific TRFs and sequence analysis confirmed that TRFs could be used to discriminate among fungal types. Sequence

analysis indicated that important ECM fungi including Russulaceae, Thelephorales, and Tricholomataceae could be fingerprinted with

TRFLP. In addition, a fixed proportion of the DNA extracted from each morphotype from the same core was used in a pooling experiment

used to assess whether previously identified fungal species types could be distinguished from one another within reconstructed communities.

Since some morphotypes share TRFs, dual analysis of ITS1 and ITS2 was necessary for accurate fingerprinting of fungal types.

Approximately, 5.0G0.3 phylotypes were detected per core with TRFLP-corrected morphotyping as compared to 4.0G0.4 phylotypes using

TRFLP on pooled community samples. TRFLP made on experimental sporocarp communities suggested that reduced ECM richness with

TRFLP may be partly caused by differences in the amount of DNA available for PCR and primer bias. Nonetheless, TRFLP on pooled

morphotypes accounted for more than 93% of colonized root tips. The method can be used to facilitate analysis of mycorrhizal communities

using root tips collected from soil cores.

q 2005 Elsevier Ltd. All rights reserved.

Keywords: Community structure; Ectomycorrhiza; Morphotyping; Pinus taeda; Terminal restriction fragment length polymorphism

1. Introduction

The fungi forming ectomycorrhizal symbioses number

over 5000 species (Molina et al., 2002) spanning all the

phyla of true fungi. Analysis of natural ectomycorrhizal

(ECM) communities has traditionally been a labor-inten-

sive, highly-skilled process with heavy reliance on gross

morphological characterization of the ECM root tips.

Depending on the rigor of the classification protocol, it is

0038-0717/$ - see front matter q 2005 Elsevier Ltd. All rights reserved.

doi:10.1016/j.soilbio.2005.01.028

* Corresponding author. Tel.: C1315 470 6875; fax: C1315 470 6934.

E-mail address: [email protected] (D.J. Burke).

possible to incorrectly assign dissimilar genetic entities into

a morphotype when analyzing individual tips. This results in

the need for multiple samples to be analyzed per

morphotype to allow for mathematical subdivision of the

group into genetic types. Because of the high spatial

variability of naturally occurring ectomycorrhizas, a large

number of samples are needed to effectively address

questions concerning community structure, with many

hundreds of individual tips to be analyzed per sample.

The high labor requirement remains an obstacle to extensive

research efforts, even with the addition of molecular

verification of identified morphotypes using restriction

fragment length polymorphisms (RFLP) which allows

researchers to avoid in-depth microscopic study of root

tips (Horton and Bruns, 2001). A rapid method for whole

Soil Biology & Biochemistry 37 (2005) 1683–1694

www.elsevier.com/locate/soilbio

http://www.elsevier.com/locate/soilbio

D.J. Burke et al. / Soil Biology & Biochemistry 37 (2005) 1683–16941684

community analysis would allow researchers to reasonably

expand research projects on ECM ecology to the level of

replication necessary to overcome these sampling

difficulties.

A number of molecular methods are currently in use for

characterization of microbial communities; these include

restriction fragment length polymorphism analysis (RFLP,

Giovannoni et al., 1990), denaturing gradient gel electro-

phoresis (DGGE, Muyzer et al., 1993), terminal restriction

fragment length polymorphism (TRFLP, Liu et al., 1997)

and length heterogeneity PCR (LH-PCR, Suzuki et al.,

1998). The most promising in terms of ease of use and

resolution appears to be TRFLP (Liu et al., 1997). Since,

first publication in 1997, TRFLP has been shown to be a

highly effective method for analysis of microbial commu-

nities, to be relatively stable to variability in PCR conditions

(Blackwood et al., 2003), and may detect a higher number of

phylotypes than DGGE assays (Marsh et al., 1998;

Moeseneder et al., 1999). In TRFLP analyses, fluorescent

label unique to each PCR primer in a reaction allows for

detection of the terminal fragments of restriction digested

PCR products. These terminal restriction fragments (TRFs)

contain the labeled primer and extend to the first instance of

a restriction site for the enzyme used (Liu et al., 1997).

Using this method, environmental samples can be rapidly

analyzed providing extensive data on the community as

defined by the specificity range of the primers. For

comparison of microbial communities, TRFLP provides a

relatively complete, culture-independent analysis.

TRFLP targeting rRNA genes has effectively been used

to characterize fungal communities in soil including

mycorrhizal communities (Edwards et al., 2004; Edel-

Hermann et al., 2004; Vandenkoornhuyse et al., 2003;

Dickie et al., 2002; Klamer et al., 2002). DNA profiles

developed using TRFLP have been used to assess overall

changes to fungal community structure in soil under

elevated CO2 (Klamer et al., 2002) and with organic matter

amendment (Edel-Hermann et al., 2004). Other researchers

have employed TRFLP as a molecular fingerprinting

method for the identification of specific fungal taxa

(Edwards et al., 2004; Dickie et al., 2002). This approach

has been used to determine the vertical distribution of ECM

in soil (Dickie et al., 2002) and to determine the effect of

fertilization on ECM communities (Edwards et al., 2004).

The use of TRFLP as a fingerprinting technique may

provide researchers with a specific molecular approach for

characterization of ECM communities that overcomes

difficulties attendant with morphotyping. However, no one

to date has examined how accurately TRFLP fingerprinting

predicts ECM species richness of colonized root tips as

compared to morphotyping or the resolution of the resulting

analysis. In effect, are detected TRFs an accurate estimate of

species-type richness or do some TRFs reflect closely-

related fungal species or genera? Since, soil is a hetero-

geneous environment that is expected to contain fungal

species with strategies ranging from mutualistic to

saprophytic, efforts to assess ECM richness from soil

hyphae can be obscured by the presence of additional non-

ECM TRFLP phylotypes (Dickie et al., 2002), making

accurate identification problematic. In addition, seasonally

rare or cryptic species can be missed even when large

numbers of samples are collected (Taylor, 2002).

In the current study, we compared traditional morpho-

typing techniques with molecular approaches for determin-

ing ECM richness of root tips separated from soil cores. Our

purpose was to determine how well a TRFLP fingerprinting

technique could estimate ECM richness compared to

morphotyping and the specificity of TRFLP phylotypes at

the taxonomic level. Our approach was to correlate

morphotype, TRFLP fingerprinting and sequence infor-

mation to develop a relatively accurate portrait of ECM

richness in a small subsample of cores from a natural

system. We then applied this information to analysis of root

tip communities from those same cores.

2. Materials and methods

2.1. Site description and soil sampling

The study site is an 8-year-old loblolly pine (Pinus taeda

L.) genetics plantation located in Scotland County, North

Carolina, USA that is adjacent to the USDA Forest

Service/North Carolina State University Southeastern Tree

Research and Education Site (SETRES). Soils at the site are

excessively-drained sandy loams (O90% sand), with a total

water holding capacity of 12–14 cm in a 2 m soil profile.

The site receives annual precipitation of approximately

120 cm with temperatures that average 268/9 8C in the

summer and winter, respectively. Between November 1993

and January 1994, seedlings of five open-pollinated loblolly

pine families from the North and South Carolina Atlantic

Coastal Plain (ACP) and five drought-hardy Texas families

(TX) were field planted on the site, which had previously

been occupied by an existing 10-year-old loblolly pine stand

that was removed prior to planting. The site is currently

surrounded on three sides by an 18-year-old loblolly pine

stand and on the fourth side by a 40–60-year-old longleaf

pine (Pinus palustris) stand, suggesting the presence of

diverse sources of local ECM inocula. A total of 120 soil

cores were collected from the site between 13 October and

22 November 2002 to a depth of 20 cm with a metal coring

device measuring 15 cm in diameter (3.5 dm3). Soil cores

were collected from both fertilized and non-fertilized family

plots. Each soil core was sieved at the site to separate pine

roots and organic material from soil using a 0.5-cm mesh

screen. Root, rhizosphere soil and organic material retained

by the sieve constituted a sample, which was subsequently

bagged and stored at 4 8C until processed in the laboratory.

D.J. Burke et al. / Soil Biology & Biochemistry 37 (2005) 1683–1694 1685

2.2. Analysis of ECM fungi through morphotyping

We initially morphotyped root tips to produce more

simplified TRFLP patterns of one or a few species per

morphotype. These identified patterns aided in interpreting

the more complex pooled whole-community TRFLP

patterns. Soil core samples were gently mixed with reverse

osmosis water to loosen soil adhering to root surfaces. All

material remaining on a 0.25-mm mesh screen (WS Tyler

Company, Cleveland) was sorted in chilled reverse osmosis

water to remove ECM root tips using a 2! magnification

lens. Viable mycorrhizal tips were subsequently re-exam-

ined under a dissecting scope at a magnification of 40! in

cold physiological saline (8.5 g lK1 NaCl) and morphotyped

according to general procedures outlined in Agerer (2003).

Tips were classified according to main tip and patch color,

patch frequency, branch pattern, tip shape, mantle texture

and luster, and extent of extramatrical hyphae (Agerer,

2003). To determine the validity of the TRFLP approach for

characterizing ECM fungi on tree roots, 10 soil cores with

high ECM richness based on preliminary morphotyping

evidence were chosen for TRFLP typing. Similar tips from

within the same core were combined and enumerated after

completion of the morphotyping procedure; these combined

tips represented a morphotype group (nZ67 for 10 cores).

Morphotyped tips were freeze-dried at K40 8C for 10 days

and stored at K70 8C until DNA extraction.

2.3. Analysis of ECM root tips through PCR-TRFLP

Immediately prior to extraction, all tips within a

morphotype group, which in no case exceeded 50 mg dry

weight, were ground in a 1.5 ml microcentrifuge tube with a

micropestle in 100 ml TNE buffer, and kept on ice. DNA

extraction was performed using a CTAB (cetyltrimethyl-

ammonium bromide) extraction protocol with 0.8% mer-

captoethanol and incubation at 65 8C for 1 h (Baker and

Mullin, 1994). Nucleic acids were purified by phenol/

chloroform extraction and precipitated with polyethylene

glycol 8000 in 2.5 M NaCl. DNA of each morphotype was

resuspended in a final volume of 100 ml TE buffer after

SSU 5.8SITS 1

58A2F/I

NSI1 58A2

ITS1F

Fig. 1. ITS primer map indicating location of primers used in this study (Martin and

58AF and NLB4 were used to amplify ITS2. NSI1 and NLB4 lie outside primer

extraction. Purified DNA was used as template in a PCR

targeting the internal transcribed spacer (ITS) region of the

rRNA gene which has taxonomic significance (White et al.,

1990). Primers NSI1 (GATTGAATGGCTTAGTGAGG-

forward) targeting the Small Subunit (SSU) rRNA gene and

58A2R (CTGCGTTCTTCATCGAT-reverse) targeting the

5.8s rRNA gene were used to amplify the ITS1 region

(Fig. 1). Primers 58A2F (ATCGATGAAGAACGCAG-

forward) targeting the 5.8s RNA gene and NLB4

(GGATTCTCACCCTCTATGAC-reverse) targeting the

Large Subunit (LSU) rRNA gene were used to amplify

the ITS2 region (Fig. 1). These primers have been

extensively tested over 5 years of research on various

topics and have successfully amplified ectomycorrhizal

fungi from more than 2000 single-root-tip DNA extracts

including ascomycetes such as Cenococcum sp. and Tuber

sp. (Martin and Rygiewicz, 1999).

PCR was carried out in 50 ml reaction volumes using

2.0 mM MgCl, 0.2 mM each primer, 0.2 mM dNTP, and 2.0

units (0.4 ml) Faststart Taq DNA polymerase (Roche

Diagnostics GmbH, Mannheim, Germany) on an PTC 100

Thermal cycler (MJ Research, Boston) using a heated lid.

Primers were labeled with fluorochromes 6FAM (forward

primers) and HEX (reverse primers). An initial denaturation

and enzyme activation step of 4 min at 95 8C was followed

by amplification for 35 cycles at the following conditions:

30 s at 95 8C, 60 s at 60 8C, 60 s at 72 8C. A final 5-min

extension at 72 8C completed the protocol. PCR typically

yielded between 50 and 60 ng amplicon mlK1; however,

some tips failed to amplify adequately under these

conditions and necessitated the addition of bovine serum

albumin (BSA) (0.5 mg mlK1) to the reaction mix. In most

instances, tips amplified satisfactorily under these con-

ditions and the small minority of problematic tips that

remained also did not amplify with the more widely used

primer set, ITS1F/ITS4 (Gardes and Bruns, 1993). These

results suggest that problematic tips may have been

senescent at the time of sampling and that the degraded

DNA resulted in low amplicon yield. Preliminary RFLP

analysis of ITS1 and ITS2 labeled amplicon using

endonucleases AluI, HaeIII, MspI and RsaI (Promega,

LSUITS 2

NLB4TS3

R

ITS4

Rygiewicz, 1999). Primers NSI1 and 58A2R were used to amplify ITS1 and

sites targeted by ITS1F and ITS4, respectively.


Madison, WI) found that AluI and HaeIII provided the

greatest number of diagnostic restriction patterns and were

subsequently used to cleave labeled amplicon for TRFLP

analysis. TRFLP analysis was performed on an ABI 310

(Applied Biosystems, Foster City, CA) using 2 ml of labeled

amplicon and the GS500 ROX size standard (Applied

Biosystems, Foster City, CA). TRFs outside the GS500

standard (50–500 bp) were extrapolated beyond the stan-

dard using GeneScan software. These extrapolations were

checked against fragment sizes obtained through RFLP and

through sequence analysis. ITS1 and ITS2 TRFLP patterns

were generated for 67 morphotype groups separated from 10

soil cores. A total of eight TRFs were generated for each

morphotype group (two primers by two amplicons by two

restriction enzymes). TRFs were used to adjust morphotype

classification within each core and were also used to

characterize the diversity of encountered fungal phylotypes

(i.e. a fungal type with a unique set of TRFs). A phylotype

was initially considered to be different if any one of the eight

TRFs differed by more than 10 base pairs.

2.4. Analysis of ectomycorrhizas through sequence analysis

Morphotype extracts that generated unique TRFLPs were

used to sequence the dominant ECM phylotypes and to

confirm the identity of the ECM fungi. Unlabeled amplicon

of the correct size (approximately, 1000 bp for ectomycor-

rhizas using unlabeled SSU and LSU primers) was isolated

on low melting point agarose, excised and purified using a

QIAquick Gel Extraction Kit (QIAGEN, Inc., Valencia,

CA) according to manufacturer’s instructions. Purified

amplicon was cloned using a pGEM-T Easy Vector System

(Promega, Madison, WI), following manufacturers instruc-

tions. Selected colonies were incubated overnight at 37 8C

in LB media, and cultures containing plasmids with

correctly sized inserts were harvested from cell cultures

using a Wizardw Plus SV Minipreps DNA Purification

System (Promega, Madison, WI). RFLP screening was used

to identify sequence types and representative plasmids were

subsequently used as templates for sequencing using Dye

Primer Cycling Kit (Applied Biosystems). Sequences were

generated at the Boyce Thompson Institute sequencing

facility using an Applied BioSystems Prism 3100 Genetic

Analyzer/DNA Sequencer. Generated sequences were

compared to database entries using the FASTA program

(European Bioinformatics Institute). In order to show the

taxonomic affinity of some recovered phylotypes, a multiple

sequence alignment created using Clustal X was used to

construct a distance tree by the neighbor-joining method.

All sequence data between the ITS1F and ITS4 primer sites

was used here for alignment purposes, except for clone

SC8259. Clone SC8259 contained an intron between the 3 0

end of the SSU rRNA gene and the ITS1 region. This intron

was deleted from the sequence used for alignment purposes.

2.5. Analysis of ECM communities through TRFLP

To determine whether TRFLP analysis accurately

predicts community richness, analysis was conducted by

pooling the nucleic acids separately extracted from each

morphotype occurring within a soil core. A total of 10% by

mass of the extracted nucleic acids from each morphotype

recovered from the same soil core were pooled to

reconstruct the fungal communities within that core. All

10 cores were reconstructed in this fashion. The pooled

DNA extract representing each reconstructed core was then

used as template for PCR using labeled primer and standard

reaction conditions with BSA as described previously.

Community richness was determined by identifying TRFs

characteristic for each phylotype. TRFs from the ITS1

region produced with AluI were used as an initial

assessment of phylotype richness. ITS2 TRFs generated

with HaeIII were used to separately assess phylotype

richness within the same cores. For our purposes, the

presence of characteristic TRFs generated from either

primer set confirmed the presence of that phylotype within

our sample. While any one analysis could fail to distinguish

between otherwise distinct types, multiple analyses allow

for a more reliable evaluation of phylotype occurrence in the

mixed communities.

Detection limits of TRFLP were initially determined

using tip numbers of each morphotype group within a soil

core. To further determine TRFLP sensitivity, we con-

structed experimental communities using DNA from

identified sporocarps. The amount of DNA in each

sporocarp extract was determined on a DU 530 spectropho-

tometer (Beckman Instruments, Inc., USA) at a wavelength

of 260 nm (Sambrook et al., 1982). Calculated quantities

were then checked by running 5 ml of each sporocarp extract

into a 2% agarose gel stained with ethidium bromide and

comparing band intensity against a size standard providing

band quantification. The quantity of DNA from five

sporocarps was varied to construct seven different exper-

imental communities with varying amounts of DNA from

each of the five sporocarps. The total amount of DNA in

each experimental community was 125 ng mlK1 and 1 ml

was used for PCR using standard reaction conditions with

BSA as described. Two replicate TRFLPs were made on

each experimental community for each gene region, using

the restriction enzymes described above. These experimen-

tal communities allow us to determine TRFLP sensitivity

independent of the number of morphotyped root tips in a soil

core.

2.6. Statistical analysis

Core morphotype and phylotype richness were analyzed

using ANOVA procedures available through Jump Statisti-

cal Software (SAS Institute, NC).


3. Results


and TRFLP

Ectomycorrhizal richness was found to be significantly

higher with our morphotyping approach (6.7G0.3) than

with TRFLP-corrected morphotyping (5.2G0.4). We found

that morphotyped tips generally gave one dominant TRFLP

pattern, both within a soil core and between soil cores where

the same morphotype was present. Analysis of TRFLP

patterns for both the ITS1 and ITS2 region from all

morphotypes indicated that as many as 28 separate

phylotypes could be detected within 10 selected cores

(Tables 1 and 2). We found that TRFs from either gene

region analyzed alone were not a completely reliable

indicator of phylotype identity; consequently, examination

of both regions was necessary for accurate identification.

3.2. Analysis of ECM sequences

Primers NSI1 (SSU rRNA) and NLB4 (LSU rRNA) were

used to generate sequences which ranged between 900 and

1300 bp in length. Similarity values for most sequences

exceeded 90% (Table 2). Sequence affinity was also

determined through multiple alignment of the entire ITS

region lying between primer sites ITS1F (SSU rRNA) and

ITS4 (LSU rRNA) (Fig. 2). Approximately, 75% of the

phylotypes could be identified to the genus level or higher

through sequence analysis (identity greater than 90%;

Table 2). Some phylotypes provided identity only through

analysis of the 5.8s rRNA gene (three out of 28 TRFLP

types). A small number of phylotypes that provided

consistent TRFLPs (three out of 28 TRFLP types) could

not be properly sequenced as the reaction repeatedly failed

after 100 bp, possibly due to sequence related inhibition.

Recovered sequences represented a wide range of important

ECM families including the Russulaceae (four phylotypes)

and Thelephorales (four phylotypes) (Fig. 2), which

together accounted for 29% of detected TRFLP phylotypes.

In addition, the primers used here also successfully

amplified sequences with high affinity to important ECM

ascomycetes such as Cenococcum geophilum (Table 2,

Fig. 2).

Sequences could be used to accurately predict 92% of

generated TRFs (Table 1), with successfully predicted TRFs

deviating from actual samples by 2.3G0.8% of total TRF

length for the forward primers and 4.4G0.8% for the

reverse primers. This typically resulted in a variation of 5–

15 bp between sample and sequence predicted TRFs,

whereas typical variation in replicate sample TRFs was 1–

2 bp. Cloned sequences could not predict 8% of the TRFs

produced by environmental extracts, even though the cloned

sequences were recovered from the same DNA extracts used

to generate the TRFs (Table 1).

3.3. Analysis of reconstructed and experimental ECM

communities through TRFLP

Phylotype richness of reconstructed ECM communities

was determined through simultaneous analysis of ITS1 and

ITS2 TRFLPs. The appearance of a phylotype-specific TRF

in either region signified the presence of that fungal type

within the core (Table 1). Where a TRF was shared by

different phylotypes in one gene region, the TRFs from the

second region were examined to confirm phylotype identity.

Primers for the ITS1 region provided consistently better

detection and amplification of some fungal types than did

ITS2 primers (Table 3). Phylotype 3 (Inocybe-like)

amplified most consistently with ITS1 primers, whereas

phylotype 5 (Cenococcum-like) amplified most consistently

with ITS2 primers in reconstructed community samples.

Since, phylotypes detected within the reconstructed com-

munity could on occasion be detected with only one primer

set, analysis of both regions was necessary for accurate

estimates of phylotype richness (Table 3). In most cases,

phylotypes that comprised more than 6% of the ECM tips

within a soil core could be detected with at least one of the

two primer sets (Table 3). On two occasions, a phylotype

that comprised approximately 20% of the root tips within a

soil core could not be detected with either primer. However,

phylotypes detected through TRFLP community analysis

accounted on average for 93G3% of all colonized root tips.

Ectomycorrhizal richness within the reconstructed commu-

nities as measured by TRFLP was slightly lower than

richness as determined with TRFLP-corrected morphotyp-

ing (4.0G0.4 and 5.0G0.3 phylotypes per core, respect-

ively; one-way ANOVA, PZ0.07).

In experimental sporocarp communities, we found that as

little as 1 ng of ECM DNA can be detected by TRFLP, but

this depends to some extent on the primer set employed and

the target ectomycorrhiza (Table 4). For example, the MR2

(Tricholoma-like) sporocarp was consistently detected with

either primer set in experimental communities, even when

1 ng of its DNA was present in the PCR reaction. On the

other hand, the MR1 (Russula-like) sporocarp could only be

detected in experimental communities with either primer set

when more than 25 ng of its DNA was present in the PCR

reaction (Table 4). However, the MR1 sporocarp could be

detected with both primer sets when larger relative amounts

of its DNA were added to the experimental sporocarp

community (Table 4).

4. Discussion


and TRFLP

The morphotyping approach used in this study indicated

a significantly higher level of ECM richness than

did TRFLP-corrected morphotyping. However, since

Table 1

Morphotype description of detected phylotypes and corresponding terminal fragment lengths for both the ITS1 and ITS2 regions

Morphotype description PT no. ITS1-Alu1 ITS1-HaeIII ITS2-Alu1 ITS2-HaeIII

NSI1 58A2R NSI1 58A2R 58AF NLB4 58AF NLB4

Black, Cenococcum-like 1 598 62 237 340 376 109 241 250

Yellow orange 2 388 61 ND ND ND ND ND ND

Brown-yellow, white patches 3 290 127 – 418 382 111 – 544

Orange silver tips 4 271 119 42 141 158 112 182 390

Black, Cenococcum-like 5 264 62 ND 310 377 109 240 250

Yellow with white sheen 6 173 166 177 226 244 111 – 582

Yellow amber 7 153 214 2180 371159 158438 112 182506 39049

Yellow amber 8 66 340 – 424 435 117 230 82

Yellow 9 66 289 – 359 215 111 493 89

White, crystal entity 10 65 276 – 432 243 112 480 83

Yellow orange, white patches 11 66 260 2530 156438 282 112 – 562

Golden yellow, shiny 12 66 236 203 280 199 112 188 82

Yellow red, ring tailed 13 66 217 175 214 436 114 183 84

Yellow orange, hairy 14 66 214 175 128 437 112 426 82

Yellow brown 15 66 133 – 383 158 112 183 390

Orange, yellow rhizomorphs 16 65 159 176 115 434 114 364 82

Unknown 17 62 306 197 296 152 189 509 159

Red brown, surface sheen 18 – 640 32220 68670 230 109 435 216

Yellow orange 19 – 427 303 123 209 116 175 409

Orange brown, bristly 20 – 409 – 388 206 112 183 218

Reddish yellow 21 – 406 248 155 215 111 494 89

Gray brown, white 22 – 394 – 334 158 112 183 390

Brown, white patches 23 – 390 32 214354 305 112 183 387

Brown yellow, white patches 24 – 390 32 214342 215117 112167 493156 89429

Medium brown 25 – 383 241 115 138 110 149 84

Light orange 26 – 380 – 380 206 116180 – 612

Yellow amber, white patches 27 – 362 – 359 215430 111 494547 890

Orange brown 28 – 356 ND ND 215 111 494 89

A dashed line (–) in place of a TRF indicates that the amplicon is unrestricted and a similar size peak appears with both labeled primers. The underlined TRFs

were not accurately predicted from recovered sequences. The TRFs predicted from the recovered sequence are indicated as a superscript.


the morphotyping approach used here is likely to split

similar fungal types into distinct groupings, TRFLP-

corrected morphotyping should result in lower phylotype

diversity (Horton and Bruns, 2001). Morphotyping is

subject to bias driven by natural gradations in morphologi-

cal characteristics, which may be the result of phenotypic

expression, tip age, environmental conditions and the skill

level of the individual observer. For these reasons,

morphotyping coupled with molecular analysis such as

RFLP has been adopted to avoid such potential errors and

provide correction of the initial morphotyping effort

(Horton and Bruns, 2001). TRFLP was used here in lieu

of RFLP to correct our morphotyping efforts and avoid

potential errors in identification.

It has become routine to generate RFLP or TRFLP

patterns through amplification of the entire ITS region lying

between the SSU rRNA and LSU rRNA gene for analysis of

ectomycorrhiza (Edwards et al., 2004; Dickie et al., 2002;

Klamer et al., 2002; Horton and Bruns, 2001; Karen et al.,

1997). We applied a different approach and analyzed the

ITS1 and ITS2 regions separately. Our expectation was that

separate analysis of both regions would provide better

phylotype discrimination in extracts containing multiple

species than one analysis of the entire region which would

have generated only two TRFs per phylotype. Through

analysis of TRFLPs from both gene regions we could detect

at least 28 separate phylotypes in 10 cores (Tables 1 and 2).

TRFLP analysis has been shown to distinguish a greater

number of phylotypes than other sensitive techniques such

as DGGE (Moeseneder et al., 1999); consequently, the level

of discrimination shown in this study might not have been

possible if other microbial fingerprinting techniques had

been used.

However, one drawback to using short PCR products for

TRFLP may be the increased likelihood that the generated

amplicon will not contain a cut site for the restriction

enzyme used. For example, some of our phylotypes failed to

cut within the ITS1 region using enzyme AluI; conse-

quently, analysis of both regions and the use of more than

one restriction enzyme was required for proper discrimi-

nation (Table 1). In addition, approximately 10 of our

phylotypes generated a similar sized TRF using primer

NSI1, suggesting the presence of a conserved AluI

restriction site on the SSU rRNA gene. Generation of a

longer amplicon could alleviate this problem through the

potential incorporation of a greater number of restriction

sites. However, closely related phylotypes could still

generate similar TRFs and the choice of restriction enzyme

Table 2

Sequence affinity of detected phylotypes

PT no. Length Sequence similarity Sequence affinity Accession no.

1 1326 93 Cenococcum sp. Z48537

2 957 98 Clavaria sp. AY456373

3 952 97 Inocybe sp. AY456377

4 950 90 Tomentella sp. AJ534912

5 650 99 Cenococcum geophilum AY394919

6 967 96 Tomentellopsis sp. AJ438983

7 240 84 Ascomycetesa AY299228

8 1091 99 Xerocomus sp. AY456374

9 NA NA Unknown NA

10 736 98 Tricholoma ustale AB036894

11 977 96 Laccaria sp. AJ534899

12 1120 98 Tylopilus sp. AY456372

13 924 98 Piloderma sp AY097053

14 923 86 Piloderma sp. AJ534902

15 520 90 Tomentella sp.b AJ534912

16 NA NA Unknown NA

17 1145 96 Rhizopogon sp. AJ515395

18 325 84 Basidiomycetesa AF402139

19 1016 92 Lactarius sp AJ534936

20 1012 90 Lactarius sp. AJ53490

21 977 93 Russula sp. AY061726




25 855 93 Uncultured ECM AY394918

26 977 96 Russula sp. AY061709

27 893 77 Basidiomycetes AF430292

28 NA NA Unknown NA

The length of the sequence used for FASTA comparison in base pairs is shown. Sequence similarity (%) for the individual phylotype as based upon the

indicated length is shown for the best match as determined by FASTA. The accession number and affinity of the best database match to our sequence are

indicated.a Results based on 5.8s rRNA gene only.b Results based on ITS2 region only.


would need to be empirically determined as is the case with

RFLP typing (Wurzburger et al., 2001). The greatest

advantage in using multiple primer sets may be in reducing

PCR bias, thus permitting a more complete portrait of ECM

richness in our system.

4.2. Analysis of ECM sequences

Sequence analysis was used to confirm the separate

identity of ECM phylotypes detected through TRFLP

(Table 2, Fig. 2). Sequences recovered represent important

ectomycorrhiza including the Russulaceae and Thelephor-

ales which have been previously found to colonize loblolly

pine (Edwards et al., 2004). We were able to successfully

correlate TRFLP patterns with specific sequences recovered

from morphotyped tips. However, there were some

discrepancies between TRFs predicted from the recovered

sequences and actual TRFs generated from environmental

DNA (Table 2). These differences could not be improved

through repeated cloning and sequencing of the phylotypes.

Predicted TRFs varied from actual TRFs by approximately

2.3% of the total length for the forward primers and 4.4%

for the reverse primers. These differences may be related to

the effect of the fluorochromes used on TRF size estimates

with GeneScan. Both the fluorochrome 6FAM and choice of

internal size standard can affect TRF sizing as reported by

other researchers (Ritchie et al., 2000). The different effect

of 6FAM and HEX on size estimates could explain the

larger discrepancy between actual and predicted TRF size

noted for the reverse primers.

Some phylotypes did display greater variability between

sequence-predicted and actual TRFs such that 8% of the

TRFs produced by environmental extracts did not match

recovered sequences. These problems principally affected

five of 28 identified phylotypes (Table 1). Pseudo-TRFs

have been partly attributed to the production of single

stranded amplicon during PCR, and spurious TRFs have

been shown to increase in prevalence with increasing cycle

numbers (Egert and Friedrich, 2003). However, this

phenomenon is not reported to be more common with

environmental extracts than cloned DNA (Egert and

Friedrich, 2003). Suboptimal restriction digest conditions,

such as prolonged incubation times or high pH, could

interfere with proper functioning of endonucleases and

result in cutting at secondary sites, but the enzymes used

here should not exhibit star activity (Fisher Technical

3


Reference). An alternative explanation is that the recovered

clones do not represent the dominant sequence within the

environmental extract, but reflect intra-specific sequence

variation within the samples. However, since, a minimum of

10 clones were screened for each morphotype, more than

300 clones in total for this study, it seems unlikely that the

dominant sequence types would not be recovered.

4.3. Analysis of reconstructed and experimental ECM

communities through TRFLP

Ectomycorrhizal ecologists regularly consider ECM tips

as discrete units useful for quantification (Horton and Bruns,

2001). As this is one basis widely used to assess the

presence and function of ECM, we initially used root tip

numbers to determine sensitivity of TRFLP. On average,

phylotypes detected with TRFLP accounted for 93% of

colonized root tips (Table 3) and did provide a more

conservative estimate of within core ECM richness. It is

commonly known that some ECM can produce relatively

more hyphae than others, and develop a thicker mantle

(Smith and Read, 1997). Additionally, as ECM tips are

ephemeral structures, not all tips encountered are at the

same stage of growth and development. Consequently, the

use of tip numbers makes it more difficult to establish

rigorous detection limits of TRFLP since tip number may

not necessarily reflect likelihood of amplification success

(e.g. it does not take into account the amount of target for

PCR). Failure of some morphotypes to amplify in

reconstructed communities may be the result of lower

relative target quantity but could also be due to differences

in primer affinity (Lueders and Friedrich, 2003). To further

establish detection limits of TRFLP, we constructed model

communities using DNA from five identified sporocarps

collected at the field site. Most sporocarps could be detected

in these experimental communities even when represented

by relatively low amounts of DNA (Table 4). However, the

MR1 sporocarp (Russula-like) was more problematic and

could only be detected when larger relative quantities of its

DNA were present within the experimental community.

This certainly indicates that primer bias or differences in

gene copy number could contribute to our failure to detect

some groups of morphotyped tips within the reconstructed

communities as well (Lueders and Friedrich, 2003; Crosby

and Criddle, 2003). Detection of some ECM types may

therefore depend on the interaction between the amount of

target sequence present within the sample, and the affinity of

that sequence for the primers employed.

Fig. 2. Neighbor joining tree visually illustrating the taxonomic affinities of select

not meant to imply relatedness of the various taxa. The tree was created using Clu

between primer sites ITS1F-ITS4. Tuber melanosporum and Cenococcum geop

percentage of bootstrap replicates out of 1000 that support that branch. Values less

(SC: Scotland County) as well as EMBL/GenBank/DDBJ accession numbers. Phy

follows Bruns et al. (1998).

These results suggest that rare or cryptic phylotypes

within a soil core can be overlooked by TRFLP. Rare and

cryptic phylotypes may be important members of ECM

communities, potentially providing both functional redun-

dancy and a reservoir of new recruits that can colonize plant

roots. Rare species may be important colonists on a seasonal

basis depending on climatic conditions, or following a

disturbance (Dahlberg, 2001). In our study in reconstructed

communities, TRFLP provided a conservative estimate of

ECM richness, but in real communities TRFLP might

increase the power of detection. Techniques that are guided

by morphotyping can overlook ECM fungi during early

stages of root colonization, or not distinguish between two

ECM fungi that are similar in physical appearance.

However, these ECM fungi could potentially be detected

by TRFLP. In addition phylotypes that are rare within a

given core would potentially be detected in other cores

where they constitute a larger percentage of colonized tips.

Indeed, we have found in our larger study (120 cores), that

rare phylotypes not detected by TRFLP within 10 cores

presented here, are often detected within reconstructed

community samples of other cores (unpublished data).

Since, ECM are patchy in space, and even small spatial

differences can reveal additional ECM fungal types,

analysis of a large number of samples is necessary for

accurate estimates of system level ECM richness (Dahlberg,

2001; Stendell et al., 1999). TRFLP of pooled ECM tips

without prior sorting into morphotype classes could allow

analysis of a larger number of samples, and increase the

likelihood of detecting rare species.

5. Conclusion

In the current study, we were able to correlate ECM

morphotypes from 10 soil cores with TRFLP patterns.

Sequence analysis confirmed that the identified TRFLP

phylotypes represented distinct ECM fungi including

important groups such as the Russulaceae and Thelephor-

ales. We were able to apply TRFLP to pooled ECM

communities reconstructed from morphotype extracts and

determined that TRFLP can provide an accurate, albeit

conservative, estimate of ECM richness in soil cores. This

conservative estimate may be a function of the interaction

between relative sequence abundance within a soil core (e.g.

the amount of target for PCR) and the affinity of the

sequences for the primers used. Nonetheless, this technique

can be used to expand sample replication such that rare

ed sequences retrieved from TRFLP phylotypes. Please note that the tree is

stal X and is based on multiple alignment of entire sequence region located

hilum are used here as an outgroup. The number at each node indicates

than 50% are not shown. Clone numbers of recovered sequence are shown

lotype number of each clone (PT no.) is also indicated. Taxonomy of groups

Table 3

Detection of phylotypes in pooled ECM communities with primers for ITS1 and ITS2

Core IDa Morphotype group PT no. Tip count Percent tips Detectionb

ITS1 ITS2

AC1 5D1 826 9 160 45 x x

822 10 80 23 x x

823/827 24 60 17 x x

825 1 32 9 x x

824/828 3 11 3 x o

829 28 10 3 x x

AC1 F4R 800/801 13 172 48 x x

796/797/798 23 162 45 x x

799 5 13 4 x x

802 18 11 3 o o

TX2 4B1 768 5 49 34 x x

769 4 27 19 x o

771 23 25 17 x o

766/767 19 18 12 o o

765 8 17 12 x x

770 12 9 6 x x

AC1 4R1 720/721 3 221 48 x o

718/723 7 101 22 x x

724 25 82 18 o o

719 16 54 12 x x

722 14 5 1 x x

AC1 4B1 742/744 20 73 31 x x

741 27 54 23 x x

745 22 48 21 x x

743 5 45 19 o x

740 6 12 5 o o

AC2 3D1 574 6 132 57 x o

576 23 72 31 x x

573 26 45 19 o o

577 5 29 13 o x

575 2 15 6 x o

579 15 8 3 x o

AC1 4B2 789/792 14 145 41 x o

788/790 11 123 35 x o

793 5 76 22 o o

794 Unk. 6 2 o o

TX2 1B1 94/100 10 393 57 x x

95 5 205 30 x x

98 19 42 6 o o

97 7 31 5 x o

96/99 14 15 2 o o

TX1 1B 85 12 368 50 x x

84 20 223 30 x o

83 5 127 17 x x

86/87 13 13 2 x x

82 6 8 1 x o

TX2 1D2 88/89/92 27 142 76 x x

90 5 34 18 x x

93 21 11 6 x x

Group numbers separated by a/indicate that the morphotype was found to be the same through TRFLP correction. Although tips for these groups were extracted

separately, the data is combined here for simplification. For each group, the number of tips indicated was extracted for amplification in this study. For example,

160 tips from group 826 were all extracted together, used for TRFLP and 10% of the DNA was used for the pooling experiment.a Indicates that the core was collected from either an Atlantic Coastal Plain pine genotype (AC) or from a Texas pine genotype (TX).b Boxes marked with an ‘x’ indicate that the phylotype could be detected in the pooled sample with that primer set. Boxes marked with an ‘o’ indicate failure

to detect the phylotype with that primer set in pooled samples.


Table 4

Detection of ECM in experimental sporocarp communities for determination of TRFLP sensitivity

Model no. Sporocarp number and sequence affinitya

MR2 Tricholoma-

like

S9 Tylopilus-like P11 Amanita-like Paxillus-like MR1 Russula-like Total DNA

1 112.5 (C/C) 6.25 (C/o) 2.50 (C/o) 2.50 (C/o) 1.25 (o/o) 125.0

2 100.0 (C/C) 6.25 (C/o) 6.25 (C/o) 6.25 (C/o) 6.25 (o/o) 125.0

3 75.0 (C/C) 25.0 (C/C) 12.5 (C/C) 6.25 (C/o) 6.25 (o/o) 125.0

4 25.0 (C/C) 25.0 (C/C) 25.0 (C/C) 25.0 (C/C) 25.0 (o/o) 125.0

5 6.25 (C/C) 6.25 (C/C) 12.5 (C/C) 25.0 (C/C) 75.0 (C/C) 125.0

6 6.25 (C/C) 6.25 (C/C) 6.25 (C/C) 6.25 (C/C) 100.0 (C/C) 125.0

7 1.25 (C/C) 2.50 (C/C) 2.50 (C/C) 6.25 (C/C) 112.5 (C/C) 125.0

The amount of DNA from each sporocarp was varied to create seven model communities. The amount of DNA (ng) from each sporocarp contained in 1 ml of

the indicated model community is shown. The total amount of DNA in 1 ml of each model community was 125 ng. Symbols (C or o) indicate whether the

sporocarp could be detected in the model community with primers for either the ITS1 or ITS2 region (ITS1/ITS2). A ‘C’ indicates that TRFs specific for that

sporocarp were detected, whereas ‘o’ indicates that the TRFs specific for that sporocarp were not detected with the indicated primer set.a Sequence affinity determined by comparing sequence against database entries using the FASTA program (European Bioinformatics Institute). Percent

identities were greater than 94% to the genera listed except for P11 where identity was 83%.


and cryptic species are better represented in mycorrhizal

community studies. Our examination of TRFLP for soil core

analysis of ECM involved a limited number of samples and

is part of a larger project examining the effect of host

genotype on ECM communities. Future work will expand

our use of TRFLP to answer questions concerning the effect

of plant host genetics and fertilization on ECM fungal

communities.

Acknowledgements

This work was funded by a National Science Foundation

grant from the Ecology and Evolutionary Physiology

Program. The authors thank Matthew Garner, Timothy

Huggins, Paul King, Dr Andreas Nocker, Alberto Stolfi and

Dr Steve McKeand for technical assistance over the course

of the project. We also thank Dr Thomas Horton for advice

and encouragement concerning the molecular techniques

employed here, and for reviewing an earlier version of this

manuscript. The manuscript was subjected to the US

Environmental Protection Agency’s (USEPA’s) peer and

administrative reviews, and it was approved for publication.

Mention of trade names or commercial products in this

paper does not constitute endorsement or recommendation

of use.

References

Agerer, R., 2003. Color Atlas of Ectomycorrhiza. Einhorn-Verlag,

Germany.

Baker, D.D., Mullin, B., 1994. Diversity of Frankia nodule endophytes of

the actinorhizal shrub Ceanothus as assessed by RFLP patterns from

single nodule lobes. Soil Biology & Biochemistry 26, 547–552.

Blackwood, C.B., Marsh, T., Kim, S.-H., Paul, E.A., 2003. Terminal

restriction fragment length polymorphism data analysis for quantitative

comparison of microbial communities. Applied and Environmental

Microbiology 69, 926–932.

Bruns, T.D., Szaro, T.M., Gardes, M., Cullings, K.W., Pan, J.J.,

Taylor, D.L., Horton, T.R., Kretzer, A., Garbelotto, M., Li, Y., 1998.

A sequence database for the identification of ectomycorrhizal

basidiomycetes by phylogenetic analysis. Molecular Ecology 7, 257–

272.

Crosby, L.D., Criddle, C.S., 2003. Understanding bias in microbial

community analysis techniques due to rrn operon copy number

heterogeneity. Biotechniques 34, 790–802.

Dahlberg, A., 2001. Community ecology of ectomycorrhizal fungi: an

advancing interdisciplinary field. New Phytologist 150, 555–562.

Dickie, I.A., Xu, B., Koide, R.T., 2002. Vertical niche differentiation of

ectomycorrhizal hyphae in soil as shown by TRFLP analysis. New

Phytologist 156, 527–535.

Edel-Hermann, V., Dreumont, C., Perez-Piqueres, A., Steinberg, C., 2004.

Terminal restriction fragment length polymorphism analysis of

ribosomal RNA genes to assess changes in fungal community structure

in soils. FEMS Microbiology Ecology 47, 397–404.

Edwards, I.P., Cripliver, J.L., Gillespie, A.R., Turco, R.F., 2004. Long-term

optimal fertilization changes the community structure of basidiomy-

cetes associated with loblolly pine on a nitrogen poor soil. New

Phytologist 162, 755–770.

Egert, M., Friedrich, M.W., 2003. Formation of pseudo-terminal restriction

fragments, a PCR-related bias affecting terminal restriction fragment

length polymorphism analysis of microbial community structure.

Applied and Environmental Microbiology 69, 2555–2562.

Gardes, M., Bruns, T.D., 1993. ITS primers with enhanced specificity for

basidiomycetes—application to the identification of mycorrhizae and

rusts. Molecular Ecology 2, 113–118.

Giovannoni, S.J., Britschgi, T.B., Moyer, C.L., Field, K.G., 1990. Genetic

diversity in Sargasso Sea bacterioplankton. Nature 345, 60–63.

Horton, T.R., Bruns, T.D., 2001. The molecular revolution in ectomycor-

rhizal ecology: peeking into the black-box. Molecular Ecology 10,

1855–1871.

Karen, O., Hogberg, N., Dahlberg, A., Jonsson, L., Nylund, J.-E., 1997.

Inter- and intraspecific variation in the ITS region of rDNA of

ectomycorrhizal fungi in Fennoscandia as detected by endonuclease

analysis. New Phytologist 136, 313–325.

Klamer, M., Roberts, M.S., Levine, L.H., Drake, B.G., Garland, J.L., 2002.

Influence of elevated CO2 on the fungal communities in a coastal scrub

oak forest soil investigated with terminal restriction fragment length

polymorphism analysis. Applied and Environmental Microbiology 68,

4370–4376.

Lueders, T., Friedrich, M.W., 2003. Evaluation of PCR amplification bias

by terminal restriction fragment length polymorphism analysis of


small-subunit rRNA and mcrA genes by using defined template

mixtures of methanogenic pure cultures and soil DNA extracts. Applied

and Environmental Microbiology 69, 320–326.

Liu, W.T., Marsh, T.L., Cheng, H., Forney, L.J., 1997. Characterization of

microbial diversity by determining terminal restriction fragment length

polymorphisms of genes encoding 16S rRNA. Applied and Environ-

mental Microbiology 63, 4516–4522.

Marsh, T.L., Liu, W.T., Forney, L.J., Cheng, H., 1998. Beginning a

molecular analysis of the eukaryal community in activated sludge.

Water Science Technology 37, 455–460.

Martin, K.J., Rygiewicz, P.T., 1999. Fungal-specific PCR primers

developed for analysis of the ITS region of environmental DNA

extracts, Agronomy Abstracts. American Society of Agronomy,

Madison, WI.

Moeseneder, M.M., Arrieta, J.M., Muyzer, G., Winter, C., Herndl, G.J.,

1999. Optimization of terminal-restriction fragment length polymorph-

ism analysis for complex marine bacterioplankton communities and

comparison with denaturing gradient gel electrophoresis. Applied and

Environmental Microbiology 65, 3518–3525.

Molina, R., Caldwell, B.A., Castellano, M.A., Horton, T., Smith, J.E.,

2002. Mycorrhizae: ectomycorrhizal fungi. In: Bitton, G. (Ed.),

Encyclopedia of Environmental Microbiology. Wiley, New York,

pp. 2124–2132.

Muyzer, G., de Waal, E.C., Uitterlinden, A.G., 1993. Profiling of complex

microbial populations by denaturing gradient gel electrophoresis

analysis of polymerase chain reaction-amplified genes coding for 16s

rRNA. Applied and Environmental Microbiology 59, 695–700.

Ritchie, N.J., Schutter, M.E., Dick, R.P., Myrold, D.D., 2000. Use of length

heterogeneity PCR and fatty acid methyl ester profiles to characterize

microbial communities in soil. Applied and Environmental Micro-

biology 66, 1668–1675.

Sambrook, J., Fritsch, E.F., Maniatis, T., 1982. Molecular Cloning—A

Laboratory Manual. Cold Spring Harbor Press, Cold Spring Harbor, NY

pp. 468–469.

Smith, S.E., Read, D.J., 1997. Mycorrhizal Symbiosis. Academic Press,

San Diego.

Stendell, E.R., Horton, T.R., Bruns, T.D., 1999. Early effects of prescribed

fire on the structure of the ectomycorrhizal fungus community in a

Sierra Nevada ponderosa pine forest. Mycological Research 103, 1353–

1359.

Suzuki, M., Rappe, M.S., Giovannoni, S.J., 1998. Kinetic bias in estimates

of coastal picoplankton community structure obtained by measurements

of small-subunit rRNA gene PCR amplicon length heterogeneity.

Applied and Environmental Microbiology 64, 4522–4529.

Taylor, A.F.S., 2002. Fungal diversity in ectomycorrhizal communities:

sampling effort and species detection. Plant and Soil 244, 19–28.

Vandenkoornhuyse, P., Ridgway, K.P., Watson, I.J., Fitter, A.H.,

Young, J.P.W., 2003. Co-existing grass species have distinctive

arbuscular mycorrhizal communities. Molecular Ecology 12, 3085–

3095.

White, T.J., Bruns, T., Lee, S., Taylor, J.W., 1990. Amplification and direct

sequencing of fungal ribosomal RNA genes for phylogenetics. In:

Innis, M.A., Gelfand, D.H., Sninsky, J.J., White, T.J. (Eds.), A Guide to

Methods and Applications. Academic Press, New York, pp. 315–322.

Wurzburger, N., Bidartondo, M.I., Bledsoe, C.S., 2001. Characterization of

Pinus ectomycorrhizas from mixed conifer and pygmy forests using

morphotyping and molecular methods. Canadian Journal of Botany 79,

1211–1216.

ectomycorrhizal fungi identification in single and pooled root samples: terminal restriction...

Documents