Molecular Ecology (1998) 7, 257272

A sequence database for the identification of

ectomycorrhizal basidiomycetes by phylogenetic analysis
T. D . B R U N S , * T. M . S Z A R O , * M . G A R D E S , K . W. C U L L I N G S , J . J . PA N , D . L . TAY L O R , * *
T. R . H O RT O N , A . K R E T Z E R , M . G A R B E L O T T O , * a n d Y. L I
*Department of Plant and Microbial Biology, University of California, Berkeley, 947203102, USA, CESAC/CNRS, Universit
Paul Sabatier/Toulouse III, 29 rue Jeanne Marvig, 31055 Toulouse Cedex 4, France, NASA-Ames Research Center, MS-239-4,
Moffett Field, CA 94035-1000, Department of Biology, Indiana University, Bloomington IN 47405, USA, **Department of Ecology,
Evolution, and Marine Biology, University of California, Santa Barbara, CA 93106, USA, USDA Forest Service, Forest Science
Laboratory, Corvallis, OR 97331, USA, Dept of Botany and Plant Pathology, Corvallis, OR 97331, USA, Fox Chase Cancer
Center, 7701 Burholme Ave, Philadelphia, PA 19911, USA

Abstract
We have assembled a sequence database for 80 genera of Basidiomycota from the
Hymenomycete lineage (sensu Swann & Taylor 1993) for a small region of the mitochondrial large subunit rRNA gene. Our taxonomic sample is highly biased toward known
ectomycorrhizal (EM) taxa, but also includes some related saprobic species. This gene
fragment can be amplified directly from mycorrhizae, sequenced, and used to determine
the family or subfamily of many unknown mycorrhizal basidiomycetes. The method is
robust to minor sequencing errors, minor misalignments, and method of phylogenetic
analysis. Evolutionary inferences are limited by the small size and conservative nature of
the gene fragment. Nevertheless two interesting patterns emerge: (i) the switch between
ectomycorrhizae and saprobic lifestyles appears to have happened convergently several
and perhaps many times; and (ii) at least five independent lineages of ectomycorrhizal
fungi are characterized by very short branch lengths. We estimate that two of these groups
radiated in the mid-Tertiary, and we speculate that these radiations may have been caused
by the expanding geographical range of their host trees during this period. The aligned
database, which will continue to be updated, can be obtained from the following site on
the WorldWide Web: http://mendel.berkeley.edu/boletus.html.
Keywords: Agaricales, Aphyllophorales, Boletales, evolution, mitochondrial LrRNA gene, molecular
clock
Received 22 May 1997; revision received 17 September 1997; accepted 14 October 1997

Introduction
The diversity of ectomycorrhizal (EM) fungi is huge.
Thousands of species are known on worldwide or regional
scales and tens of species are frequently encountered even
within monoculture forests of 0.1 ha (Bruns 1995). This
diversity alone would represent an intimidating factor for
many ecological studies, but the difficulty in dealing with
EM fungi is compounded by the fact that most species are
identifiable only by their fruiting structures.
Much effort has been made to remedy this problem, but
all of the existing methods still leave significant numbers
of unknowns. Morphological approaches have resulted in
Correspondence: T. D. Bruns. Fax: +01-510-642-4995; E-mail:
boletus@socrates.berkeley.edu
1998 Blackwell Science Ltd

beautifully illustrated manuals (Agerer 1987; Ingleby et al.

1990), but the number of species described in this way are
relatively few and many common types are essentially
described as imperfect states with unknown affinities
(Agerer 1987; Ingleby et al. 1990; Agerer 1994). Molecular
methods currently available enable one to match restriction fragment length polymorphisms (RFLP) patterns of
unknown mycorrhizae to known fungi (Gardes et al. 1991;
Henrion et al. 1992; Gardes & Bruns 1993; Krn et al. 1997),
or to use DNA probes to test for specifically characterized
taxa or genotypes (Marmeisse et al. 1992; Bruns & Gardes
1993). These methods offer three main advantages over
morphological methods: (i) some can be directed at strainlevel or at least subspecific-level identification; (ii) they
require less time to learn than morphological methods;

258

T. D . B R U N S E T A L .

and (iii) large numbers of samples can be dealt with more

easily because the initial visual sorting process is relatively
fast and mycorrhizal tips can then be freeze-dried and
stored indefinitely. Unfortunately, RFLP or probe-based
methods have the same limitation as morphological
approaches: unmatched types remain unknown.
In this study we present a sequence database that helps
to place unknowns into smaller, essentially family or subfamily sized, monophyletic groups. The region chosen for
this database is a small ( 400 bp) fragment of the mitochondrial large subunit rRNA gene. The fungal sequences
for this region can be amplified directly from individually
extracted mycorrhizal root tips using fungal-specific
primers, and the resulting products can be sequenced and
compared to known sequences in the database. This
method, in combination with the other available molecular methods, enables us to identify virtually all mycorrhizal samples we have encountered to some meaningful
taxonomic level (Cullings et al. 1996; Gardes & Bruns
1996; Taylor & Bruns 1997).

Materials and methods

Extractions of DNA were made from herbarium samples,
cultures of identified basidiomycetes and freeze-dried
mycorrhizae collected in nature (Table 1) by methods

described previously (Bruns et al. 1990; Gardes & Bruns

1993). The unknown mycorrhizae were derived from
Pinus muricata, members of the Monotropoideae, and
orchids; details of these studies have been reported elsewhere (Cullings et al. 1996; Gardes & Bruns 1996; Taylor &
Bruns 1997). Diluted crude extracts were used as templates for 35 amplification cycles using the ML5 and ML6
primers (White et al. 1990), an annealing temperature of
53 C or 55 C, and other cycling parameters as previously
described (Bruns et al. 1990).
Sequences of the PCR products were determined
manually with S35 labelling using single-stranded templates generated from asymmetric reactions as described
(Bruns et al. 1990), or by cycle sequencing of doublestranded products using fluorescent dideoxy-terminators and an ABI 377 automated sequencer according to
the manufacturers instructions (Applied Biosystems,
Prism kit). The primers ML5 and ML6 were used as
sequencing primers. Sequences were determined for
both strands, compared, and corrected for the 67 taxa
indicated (Table 1). All other sequences were determined
in only a single direction but error correction in these
was facilitated by comparison to closely related
sequences and the original data for all variant positions
were re-examined and confirmed or corrected.
An initial alignment of 60 taxa was made with the

Table 1 Taxa and specimens in the database

Tree location*

Taxon

Isolate

ds/ss

Accession no.**

1415
9
7
9
7
9
6
12
12
12
12
12
12
12
12
12
11
11
1
1314
1
1
1
1
1

Agaricus brunnescens
Albatrellus ellisii
Albatrellus flettii
Albatrellus peckianus
Albatrellus skamanius
Albatrellus syringae
Alpova olivaceotinctus
Amanita calyptrata
Amanita francheti
Amanita gemmata
Amanita magniverrucata
Amanita muscaria
Amanita pachycolea
Amanita pantherina
Amanita phalloides
Amanita silvicola
Armillaria albolanaripes
Asterophora lycoperdoides
Austroboletus betula
Bolbitius vitellinus
Boletellus ananas
Boletellus chrysenteroides
Boletellus russellii
Boletus affinis
Boletus edulis

SAR88/411
TDB-1493
TRH264
DAOM-216310 *
JL 9289
DAOM-216918
JMT-5376
TDB-1498
TDB-928
TDB-1523
TDB-1514
TDB-1513
TDB-1508
BRECKON306
TDB-1639
TDB-1506
TDB-1404
TDB-1227
RV-9.2*
SAR 84100
HDT-6597
TDB-513
TDB-800
TDB-538
TDB-1002

ds
ss
ds
ds
ds
ds
ds
ds
ds
ds
ds
ds
ds
ds
ds
ds
ss
ss
ss
ds
ds
ss
ss
ss
ss

S:1333156
S:1333174
S:1333189
S:1333202
S:1333243
S:1333261
S:1333271
S:1333278
S:1333291
S:1333302
S:1333348
S:1333370
S:1333371
S:1333372
S:1333373
S:1333374
S:1333375
S:1333376
S:1333377
S:1333378
S:1333379
S:1333380
S:1333381
S:1333382
S:1333383

1998 Blackwell Science Ltd, Molecular Ecology, 7, 257272

E C T O M Y C O R R H I Z A L B A S I D I O M Y C E T E D ATA B A S E

259

Table 1 Continued
Tree location*

Taxon

Isolate

ds/ss

Accession no.**

1
1
1
1
1
1
1
89
6
7
17
17
17
2
1
6
4
4
4
13
13
13
1
16
6
16
16
5
5
13
89
15
3
15
15
15
6
13
16
14
8
8
1
1
1
11
6
13
13
89
2
3
2
3
5
1
1415

Boletus flaviporus
Boletus mirabilis
Boletus pallidus
Boletus satanas
Boletus smithii
Boletus subglabripes
Boletus viridiflavus
Bondarzewia montana
Brauniellula albipes
Byssoporia terrestris
Cantharellus cibarius
Cantharellus cinnabarinus
Cantharellus tubaeformis
Chalciporus piperatoides
Chamonixia ambigua
Chroogomphus vinicolour
Coniophora arida
Coniophora puteana
Coniophora puteana
Cortinarius ponderosus
Cortinarius vanduzerensis
Cortinarius violaceus
Gastroboletus citrinibrunneus
Gautieria monticola
Gomphidius glutinosus
Gomphus clavatus
Gomphus floccosus
Gyrodon merulioides
Gyroporus cyanescens
Hebeloma crustuliniforme
Heterobasidion annosum
Hygrocybe cantharellus
Hygrophoropsis aurantiaca
Hygrophorus pudorinus
Hygrophorus sordidus
Hygrophorus speciosus
Hymenogaster sublilacinus
Inocybe sororia
Kavinia alboviridis
Laccaria laccata
Lactarius piperatus
Lactarius volemus
Leccinum holopus
Leccinum manzanitae
Leccinum rubropunctum
Leucopaxillus amarus
Melanogaster tuberiformis
Naematoloma aurantiaca
Nolanea sericea
Panus conchatus
Paragyrodon sphaerosporus
Paxillus atrotomentosus
Paxillus involutus
Paxillus statuum
Phaeogyroporus portentosus
Phylloporus rhodoxanthus
Piloderma croceum

TDB-1008
TDB-1306
TDB-1231
TDB-1000
TDB-970
TDB-634
TDB-1236
TDB-1471
F-2431
Z-14*
TDB-1427
TDB-389
TDB-1434
TDB-973
HS-2021
TDB-1010
FP-104367-SP*
FP-102011*
MAD515*
HDT-53966
TRH281
TDB-1320
HDT-40189
SNF-115
TDB-957
TDB-1583
TDB-1310
TDB-532*
TDB-1214
TRH277
KV-340
TDB-334
TDB-585
TDB-1557
TDB-727
TDB-650
F2250
TDB-1427
SNF-284
HDT 53791
TDB-1223
TDB-1225
DJM-592
TDB-969
TDB-1203
TDB-1336
TDB-1042, JMT-26
TDB-585*
SAR 88415
TDB-1049
TDB-420*
TDB-782*
TDB-642*
REH-5904
HDT-42534
TDB-540*
CBS 294.77

ss
ss
ss
ds
ss
ss
ss
ss
ds
ss
ds
ss
ss
ss
ds
ds
ss
ss
ss
ds
ds
ds
ds
ds
ds
ss
ss
ds
ds
ds
ds
ds
ds
ds
ds
ds
ds
ss
ds
ds
ss
ss
ds
ss
ss
ss
ss
ds
ss
ss
ds
ss
ds
ds
ds
ds
ds

S:1333384
S:1333385
S:1333386
S:1333387
S:1333388
S:1333389
S:1333390
S:1333391
S:1333392
S:1333393
S:1333394
S:1333395
S:1333396
S:1333397
S:1333398
S:1333399
S:1333400
S:1333401
S:1333402
S:1333403
S:1333404
S:1333405
S:1333406
S:1333407
S:1333408
S:1333409
S:1333411
S:1333414
S:1333416
S:1333418
S:1333421
S:1333423
S:1333425
S:1333428
S:1333437
S:1333488
S:1333489
S:1333490
S:1333491
S:1333492
S:1333493
S:1333494
S:1333495
S:1333496
S:1333497
S:1333498
S:1333499
S:1333500
S:1333501
S:1333502
S:1333503
S:1333504
S:1333505
S:1333506
S:1333507
S:1333508
S:1333509

1998 Blackwell Science Ltd, Molecular Ecology, 7, 257272

260

T. D . B R U N S E T A L .

Table 1 Continued
Tree location*

Taxon

Isolate

ds/ss

Accession no.**

5
9
10
1
16
16
6
6
6
6
6
6
8
8
10
18
3
3
3

Pisolithus arrhizus
Polyporoletus sublividus
Pseudotomentella tristis
Pulveroboletus ravenelii
Ramaria araiospora
Ramaria conjunctipes
Rhizopogon truncatus
Rhizopogon evadens
Rhizopogon ochraceorubens
Rhizopogon subcaerulescens
Rhizopogon villosulus
Rhizopogon vinicolour
Russula laurocerasi
Russula rosacea
Sarcodon imbricatum
Sebacina sp.
Serpula himantioides
Serpula himantioides
Serpula incrassata
Serpula incrassata
Strobilomyces floccopus
Suillus cavipes
Suillus ochraceoroseus
Suillus sinuspaulianus
Suillus tomentosus
Tapinella panuoides
Thelephora sp.
Thelephora terrestris
Tomentella atrorubra
Tomentella cinerascens
Tomentella lateritia
Tomentella sublilacina
Tricholoma flavovirens
Tricholoma manzanitae
Tricholoma pardinum
Truncocolumella citrina
Tulasnella irregularis
Tylopilus alboater
Waitea circinata
Xerocomus chrysenteron
Xerocomus subtomentosus
1MR (Pinus) Amanita gemmata
2MR (Pinus) Amanita francheti
3MR (Pinus) Tomentella sublilicina
4MR (Pinus) Tomentella sublilicina
5MR (Pinus) Tomentella sublilicina
6MR (Pinus) Tomentella sublilicina
7MR (Pinus)
8MR (Pinus)
9MR (Pinus)
10MR (Pinus)
11MR (Pinus) Rhizopogon subcaerulescens
12MR (Pinus) Russula xerampelina
13MR (Pinus) Russula xerampelina
14MR (Pinus)
15MR (Pinus)
16MR (Pinus) Xerocomus chrysenteron

TDB-1051,1052
DAOM 194363
LT-60
TDB-1307
TDB-1414
TDB-1479
AHS-68359
TDB-1303
TDB-1015
F-2882
AHS-65445
AHS-68595
TDB-1222
TDB-895
LT-2
UAMH6444*
Bud-205-A*
FP-94342-R*
L-11504-SP*
MAD563
TDB-1213
TDB-645
SAR-84137*
DAOM-66996*
TDB-661*
RLG-12933-SP
TDB-1504
S-142*,1542
LT64
LT66
LT56
TDB-2015
TDB-1395
KMS 194
TDB-1032
AHS-30164
UAMH-574*
TDB-1206
GA-846*
TDB-365*
TDB-991
935F2 ML5
995AA ML5
935E2 ML5
935BR ML5
SEEDLING19
930C-ML6
939B ML5
942C2R ML5
945 A2 ML5
936F2R ML5
995AB ML5
944B ML5
942B2 ML5
936AR ML5
935E ML5
996 BC2R ML

ss
ds
ds
ds
ss
ss
ss
ss
ss
ds
ss
ss
ss
ds
ds
ds
ds
ss
ss
ss
ss
ds
ds
ds
ss
ds
ds
ss
ds
ds
ds
ds
ss
ds
ss
ds
ds
ss
ds
ds
ss
ss
ss
ss
ss
ss
ss
ss
ss
ss
ss
ss
ss
ss
ss
ss
ss

S:1333510
S:1333511
S:1333512
S:1333513
S:1333514
S:1333515
S:1333516
S:1333517
S:1333518
S:1333519
S:1333520
S:1333521
S:1333522
S:1333523
S:1333524
S:1333525
S:1333526
S:1333527
S:1333528
S:1333529
S:1333530
S:1333531
S:1333532
S:1333533
S:1333534
S:1333535
S:1333536
S:1333537
S:1333538
S:1333539
S:1333540
S:1333541
S:1333542
S:1333543
S:1333544
S:1333545
S:1333546
S:1333547
S:1333645
S:1333649
S:1333651
L46376
L46377
L46378
L46379
S:1333657
L46380
L46381
L46382
L46383
L46384
L46385
L46386
L46387
L46388
L46389
L46390

1
6
6
6
6
3
10
10
10
10
10
10
11
11
11
6
18
1
1415
1
1
12
12
10
10
10
10
10
10
1
1
6
8
8
17
17
1

1998 Blackwell Science Ltd, Molecular Ecology, 7, 257272

E C T O M Y C O R R H I Z A L B A S I D I O M Y C E T E D ATA B A S E

261

Table 1 Continued
Tree location*

Taxon

Isolate

ds/ss

Accession no.**

1
1
6
14
10
8
10
10
6
6
12
8
6
8
8

17MR (Pinus)
18MR (Pinus)
19MR (Pinus)
20MR (Pinus) Laccaria amethysteo-occidentalis
21MR (Pinus)
22MR (Pinus) Russula brevipes
23MR (Pinus)
24MR (Pinus)
25MR (Monotropa hypopithys)
26MR (Monotropa hypopithys)
27MR (Hemitomes)
28MR (Monotropa uniflora)
29MR (Monotropa hypopithys)
30MR (Monotropa uniflora)
31MR (Monotropa uniflora)

930 B PATT
945 B PATT
927 B PATT
996 BA ML5
915 R2 ML5
SEEDLING B
SD 41 FALL
SD 49 FALL
4M 3 12 92
C1 & Lake grant2
spoint
mich1
lake wtII
23
22 28 32

ss
ss
ss
ss
ss
ss
ss
ss
ss
ss
ss
ss
ss
ss
ss

L46391
L46392
L46393
L46394
S:1333667
L46395
S:1386427
S:1386428
S:1386429
S:1386430
S:1386431
S:1386432
S:1386433
S:1386434
S:1386435

*Position on tree in Fig. 2, two numbers are given for taxa located between numbered groups.
MR taxa are field-collected ectomycorrhizae.
Cultures are indicated*; all others are from either dried fruit body collections or field collected mycorrhizae for MR collections.
ds, sequence determined in both directions; ss, sequence determined in a single direction.
**S, accesssion numbers are for the Genome Sequence Database; others are GenBank.

multiple alignment program (Higgins et al.

1992) on a Sun Sparc station.
This alignment was examined and adjusted manually
using Microsoft Word on a large-screen Macintosh computer. Manual alignment was facilitated by the use of a
colour font. Sample names were temporarily removed during manual alignment to avoid bias. Additional taxa were
aligned in small groups and added into the large alignment
manually by the same method. After viewing the results of
initial phylogenetic analyses, the alignment within wellsupported monophyletic groups was re-examined and
adjusted. This was done to ensure that all identical or
nearly identical sequences were aligned in the same way.
The final alignment, which excludes the unalignable 5
portion of the ML5/ML6 fragment (Fig. 1), was formatted
as a PA U P file (Swofford 1993). All base positions included
in the analysis were written in upper case, a small internal
region that could not be aligned was written in lower case,
and these latter bases were then ignored by equating them
to missing data. Most gaps that were introduced for alignment purposes were also treated as missing data, but a few
were coded as fifth character states X. The criteria used
for these codings are the same as those described previously (Bruns et al. 1992). Exclusion sets of taxa were setup
for convenient analysis of subsets of the data. P H Y L I P formatted files, used for distance analysis, were derived from
the PA U P file. Other than format they also differed from
the latter in that all gaps were treated as missing data.
Neighbour-joining analysis was initially conducted on
all known and unknown taxa using Kimura 2-parameter
C L U S TA L V

1998 Blackwell Science Ltd, Molecular Ecology, 7, 257272

distances; the programs D N A D I S T and N E I G H B O R from

P H Y L I P 3.4 (Felsenstein 1995) were used to generate the
distance matrix and to produce the tree. Confidence in the
branches of the neighbour-joining tree was assessed by
bootstrap analysis (Felsenstein 1985), using 500 replicates.
The programs S E Q B O O T , D N A D I S T , N E I G H B O R , and C O N S E N S E in the P H Y L I P package (Felsenstein 1995) were
used for this purpose. All P H Y L I P programs were run on a
Sun Sparc station.
Parsimony analyses were conducted with PA U P 3.1.1
on subsets of taxa to assess the effects of method of analysis. The taxa used were selected based on results from the
neighbour-joining analysis such that multiple representatives from all major clades and divergent taxa not clearly
placed in such groups were included. The representatives
selected were chosen to maximize sequence differences
within major clades. A total of 10 random addition
sequences were run using the heuristic search option in
PA U P 3.1.1 (Swofford 1993) on a Macintosh Quadra 800.
Small batches of unknown taxa were analysed with this
subset of knowns to compare placements between parsimony and neighbour-joining methods.
Later analyses were conducted with a beta version of
PA U P (4.0d52) written by David Swofford using neighbour-joining of the patristic distance matrix on a Power
Macintosh 7500/100. Confidence in branches was
assessed with 1000 bootstrap replicates.
Molecular clock estimates were made for 1137 aligned
positions from the 5 and central portions of the nuclear
small subunit (Nu-SSU) rRNA gene in the following way.

262

T. D . B R U N S E T A L .

Fig. 1. A. Diagrammatic representation of sequenced portions of

the ML5ML6 region. Total size of the region inclusive of primers
is 416 bp in Suillus sinuspaulianus. Most other taxa have very
similar-sized fragments. The cross-hatched regions indicate sizevariable portions of the fragment. The 5-most region, which
contains the ML5 primer sequence, was excluded from the analysis because it varies as much as 100 bp in size and the sequences
are too different to be aligned. The database starts immediately
after this region with the conserved sequence indicated. Most
taxa have a sequence that is nearly identical to the top line. The
lower line shows the variant positions that are typical of
members of the Russulaceae. The shaded region contains small
inserts of less than 15 bp in some agarics; it is included in the
database, but portions of it are excluded from the analysis as
described in the Materials and methods. The 3-most portion of
the region is not determined for most taxa in the database,
because of technical difficulties associated with its proximity to
the ML6 primer. B. Five introns encountered infrequently in the
taxa sampled. Approximate locations of know introns is shown
(vertical arrows, ae). Primers that work around these are indicated (horizontal arrows) and their names and sequences are
given. Flanking regions of the gene that are outside of the target
fragment are shown with dashed lines.

The shortest parsimony tree was found for the following

13 selected taxa (GenBank accession number): Tremella
moriformis (U00977), T. globospora (U00976), Spongipellis
unicolour (M59760), Boletus satanas (M94337), Xerocomus
chrysenteron (M94340), Phylloporus rhodoxanthus (M90825),
Paragyrodon sphaerosporus (M90826), Paxillus atrotomentosus (M90824), Chroogomphus vinicolour (M90822),
Gomphidius glutinosus (M90823), Rhizopogon subcaerulescens (M90827), Suillus cavipes (M90828), Suillus
sinuspaulianus (M90829). All 24 trees that were three steps
longer or less were compared with maximum likelihood

(DNAML, no clock assumption, Felsenstein 1995), and

those that were significantly worse than the best tree were
rejected. The remaining trees were examined visually and
those that included Paragyrodon within, rather than as the
sister group to, the boletoid group were rejected; this criterion was used because four rRNA genes examined to
date (Mt-LSU, Mt-SSU Nu-SSU, and Nu-LSU) all depict
Paragyrodon as outside the boletoid group and collectively
show strong support for this relationship even though the
Nu-SSU gene does not specify this relationship strongly
(Bruns & Szaro 1992; T. D. Bruns, unpublished data). The
remaining 12 trees, which differed only in the branching
orders within the boletoid group and in the placement of
Paxillus relative the suilloid and boletoid groups, were
each submitted to maximum likelihood with a clock constraint (D N A M L K , Felsenstein 1995); this program forces
the constraint that each terminal branch is equidistant
from the root and thus corrects for the rate differences on
a given topology. Calibration for the root of the tree was
based on Berbee and Taylors estimate of 220 Ma for the
divergence of Tremella from polypores, agarics and boletes
(Berbee & Taylor 1993). A 100 Ma error was allowed by
also using estimates of 270 and 170 Ma.

Results and Discussion

Phylogenetic analyses and placement of unknowns
A total of 152 sequences were determined for the
ML5ML6 region of the mitochondrial large subunit. Of
these sequences, 121 were derived from identified samples and 31 from initially unidentified mycorrhizae
(Table 1). The known samples include representatives of
80 genera from 17 families. We were able to determine
sequences for virtually all samples. The only consistent
exceptions can be attributed to taxa that contained introns
in this region. We know that at least five introns can be
present and these can dramatically increase the size of the
region and in some cases disrupt the ML5 primer site (Li
1995); this can make it difficult to amplify or sequence the
region from DNA templates. Among the taxa we have
sampled, however, introns were rarely encountered and
were only found within a subset of species of Albatrellus,
Byssoporia,
Coniophora,
Heterobasidion,
Hydnellum,
Hygrophorus, Kavinia, Macrolepiota, Rhizopogon, and
Suillus. Because these introns were fairly rare and variably
present even within species, we did not try to use them
for identification purposes, but instead tried to work
around them. Obtaining full-length sequence in species
containing multiple introns was often difficult and not
always achieved. Primers that avoid the introns and
amplify the flanking pieces of the structural gene (Fig. 1)
helped us to obtain at least partial sequences in all but
two species (Albatrellus ovinus and Hydnellum peckii).
1998 Blackwell Science Ltd, Molecular Ecology, 7, 257272

E C T O M Y C O R R H I Z A L B A S I D I O M Y C E T E D ATA B A S E
Alternatively, rRNA templates and reverse transcriptase
PCR could be used, but this approach may necessitate different preservation and extraction methods to ensure that
RNA templates are not degraded.
We were able to determine sequences directly from
field-collected mycorrhizae in virtually all cases.
Although contaminating soil fungi must have been present on most or all of these samples, they did not appear
to contribute to the sequences determined. We say this
because the sequence clarity was usually very good and
the unknown sequences were placed into well-defined
lineages of EM fungi (Fig. 2). Furthermore, many of these
placements have been confirmed by internal transcribed
spacer region (ITS) RFLP matches or oligonucleotide
probing (Cullings et al. 1996; Gardes & Bruns 1996). These
two other methods use different primers and target different regions and thus are independent of the ML5/ML6
sequence analyses (Bruns & Gardes 1993; Gardes & Bruns
1993; Bruns 1996)
After exclusion of the unalignable 5 portion of the
fragment and 15 bp of internal positions (Fig. 1), the
remaining sequence was represented by 339 aligned positions within the database; 40 bp of the 3 end was not
determined in the majority of the taxa. In total, the number of variable positions is 181, and 143 of these are cladistically informative (i.e. they contain variant states that are
shared by two or more taxa).
A neighbour-joining tree based on patristic distances
generated from PAUP is shown and branches supported by
more than 50% of the bootstrap replicates are indicated
(Fig. 2). A tree based on Kimura 2-parameter distances
generated by PHYLIP contained all but one of the major
groupings shown in Fig. 2a. The only exception was that
the position of the Hygrophoraceae (group 15, Fig. 2) was
shifted and was no longer monophyletic. In both trees all
of the unknowns were placed within the same groups
indicated.
The large number of taxa, the relatively low number of
informative characters, and the many near-zero branch
lengths made the number of equally parsimonious trees
very high and the computational time too long to allow
for a complete analysis of the entire dataset with parsimony. However, even very short (< 10 min) and incomplete parsimony runs using the whole dataset resulted in
the same placement of all of the unknown taxa into the
same numbered family or subfamily groups indicated
(Fig. 2). Longer runs with subsets of taxa resulted in very
similar trees to the one shown in Fig. 2; all the branches
that were supported by more than 60% of the neighbourjoining bootstrap replicates and also many of the lesser
supported branches were also found with the partial parsimony analyses.
Even partial sequences resulted in fairly confident and
apparently accurate placements at the family or subfamily
1998 Blackwell Science Ltd, Molecular Ecology, 7, 257272

263

level. The placements of partial sequences of Gomphus clavatus, Ramaria conjunctipes, and Kavinia alboviridis are good
examples, as all were placed with other known members of
the Gomphaceae (group 16, Fig. 2d) with high confidence
(99% bootstrap). Thelephoroid unknowns MR-1, 4 and 7 are
also good examples; these have recently been confirmed to
be closely related to Thelephora and Tomentella by sequence
analysis of the internal transcribed spacer region (D. L.
Taylor, unpublished results). Partial sequences, however,
often yielded artifactually long-terminal branches especially if the missing data were in highly conserved regions.
This was true if neighbour-joining distances were displayed, and it is the reason that we chose to display character changes (i.e. parsimony distances) on the tree shown.
Sequence error and minor misalignments also appear
to have little effect on placement of unknowns. The major
effect was that the terminal branch lengths were exaggerated. We did not test this in a rigorous way, but we have
observed this effect from preliminary analysis of many
unknowns in which the sequences were initially incomplete, poor in quality, or misaligned. Yet all such
sequences were placed correctly by the phylogenetic analysis as judged by later analysis of completed and accurate
sequences, or by ITS-RFLP matches to species within the
groups. The phylogenetic resolution was low in many
parts of the tree as judged by bootstrap analysis (i.e. those
< 70% in Fig. 2). Fortunately the low phylogenetic resolution had almost no effect on the family or subfamily placement of the unknown mycorrhizal fungi we encountered.
This apparent contradiction is true because the unknowns
we encountered and tested turned out to be members of
groups that were well sampled and strongly supported
by phylogenetic analyses. The strong support is due to the
fact that very few sequence differences occur within most
major mycorrhizal lineages sampled, while sequence
variation between these groups and other taxa is moderate to large (Table 2). Indeed, many closely related species
and genera have identical or nearly identical sequences in
this region. For example, among the nine species of
Amanita sampled, six have identical sequences and only
A. francheti differs by more than 2%. Similarly, within the
suilloid group, some species of Suillus, Rhizopogon and
Gomphidiaceae have identical sequences, and all others
placed within this group differ at only a few positions.
Placements within the boletoid (Fig. 2, group 1) and
suilloid groups (group 6), the Russulaceae (group 8), the
Thelephorales (group 10), and Amanita (group 12) were
typically unequivocal because these groups are both well
sampled in our database and have very distinct and relatively uniform ML5/ML6 sequences. High bootstrap values define all of these lineages except the suilloid group
(group 6), and even in this case the bootstrap value was
moderately high (78%) and within the range that can be
considered as strong (Hillis & Bull 1993). Furthermore,

264

T. D . B R U N S E T A L .

1998 Blackwell Science Ltd, Molecular Ecology, 7, 257272

E C T O M Y C O R R H I Z A L B A S I D I O M Y C E T E D ATA B A S E

1998 Blackwell Science Ltd, Molecular Ecology, 7, 257272

265

266

T. D . B R U N S E T A L .

1998 Blackwell Science Ltd, Molecular Ecology, 7, 257272

E C T O M Y C O R R H I Z A L B A S I D I O M Y C E T E D ATA B A S E

267

Fig. 2 a, b, c and d. Phylogenetic placement of unknowns. The phylogram is based on a neighbour-joining analysis of patristic distances.
Horizontal distance is based on number of inferred substitutions (i.e. parsimony criteria). Vertical distance is arbitrary. Numbers indicate
percentage of bootstrap replicates from a sample of 1000 that support the indicated branches; unlabelled branches have values less than
50% or are in parts of the tree where the branch lengths are too small to label. All sequences other than unknown mycorrhizae were
derived from identified herbarium collections or cultures (Table 1). Sebacina sp. may not conform to the current circumscription of that
genus, but we give the name that was originally reported for it (Currah et al. 1990). Groups named in lower-case letters are not currently
recognized as formal taxa; we use them here for convenient reference to apparently monophyletic lineages. OMR, cultured from orchid
mycorrhizae. PS, partial sequence. OP, confirmed as suilloid (group 6) by oligonucleotide probing (Bruns & Gardes 1993); MR, unknown
mycorrhiza, preceded by a unique number, followed by plant host in parentheses and, if the type has been matched by ITS-RFLP, the fungal species name is given; *, taxa that are nonmycorrhizal; *?, suspected to be nonmycorrhizal; ?, unknown ecology.

virtually all unknowns that we initially identified as suilloid by phylogenetic analysis were also confirmed with
oligonucleotide probes, ITS-RFLP matches, or ITSsequence analysis (Bruns & Gardes 1993; Gardes & Bruns
1993, 1996; Cullings et al. 1996).
1998 Blackwell Science Ltd, Molecular Ecology, 7, 257272

Cantharellus, Hygrophorus, and the Gomphaceae also

have very distinct sequences that should allow unknowns
from these groups to be recognized easily, even though
our sample remains fairly small. Interestingly, there is one
group of unknown mycorrhizal sequences (14MR and

268

T. D . B R U N S E T A L .

Table 2 Kimura 2-parameter distances expressed as percentage

base substitutions for selected parts of the dataset

Distances range
within group*

Distance to
next closest
known taxon*

Amanitaceae
8 species, 6 sections

04.3

3.2

Boletoid group
24 spp., 1112 genera

02.0

3.1 (8.6)

Cantharellaceae
3 spp., 1 genus

00.04

25

Cantharelloid group
Cantharellus plus
unknowns 14, 26

012

25

Gomphaceae
3 spp., 3 genera

04.3

12.3

Hygrophoraceae
3 spp., 2 genera

0.45.9

8.0

Russulaceae
4 spp. 2 genera

1.93.2

7.8

Suilloid group
17 spp., 911 genera

00.5

2.5

mycorrhizae, both of which were found to have specific

EM associations that were previously unknown (Cullings
et al. 1996; Taylor & Bruns 1997). Another advantage of
this database is that it could potentially be used for identification of EM fungi in nonmycorrhizal states such as
mycelial mats and rhizomorphs.
Sequence-based family identifications can also be used
to narrow the search for species-level identification by
other molecular methods such as ITS-RFLP matching.
Indeed, many of our successful RFLP matches in previous
studies were facilitated by phylogenetic placement of
unknown ML5/ML6 sequences (Cullings et al. 1996;
Gardes & Bruns 1996; Taylor & Bruns 1997). One previously unmatched type, 20MR (Gardes & Bruns 1996), has
now been identified as Laccaria amethysteo-occidentalis. It
was initially missed because its ITS-RFLP differed from
the isolate of the species with which it was compared, but
after adding Laccaria laccata to the database it became clear
that this unknown was likely to be a Laccaria. This encouraged us to try more extensive ITS-RFLP comparisons, and
these revealed that some isolates of Laccaria amethysteooccidentalis were perfect ITS-RFLP matches to the
unknown.

Thelephorales
56 species, 5 genera,
and unknowns

03.7

5.4

Assessing placement of unknowns

Taxon

*taxa represented by partial sequences are excluded from these

distance measures, because their values are inflated by the
missing data; see text.
3.1 is the minimum distance from members of the boletoid
clad to Chalciporus, Paxillus involutus or Paragyrodon. The nextclosest relatives are the other members of the Paxillaceae and
Serpula at 8.6%.

15MR) that is part of a strongly supported monophyletic

group with Cantharellus, but the sequences are quite distinct from the three known sequences of Cantharellus that
we sampled (Fig. 2). These unknowns were associated
with Pinus muricata in coastal California. We have also
found nearly identical sequences associated with Sarcodes
sanguinea in the Sierras, but in neither case were we able
to match these unknowns to any fruiting species (Cullings
et al. 1996; Gardes & Bruns 1996). We know that they are
closely related to Clavulina cristata.
Use of this database enables one to identify the fungal
component of many unknown mycorrhizae to the level of
family or subfamily. This is an improvement over the
current state of affairs in which many mycorrhizal fungi
are not assignable to any meaningful taxonomic group. It
is particularly useful when the morphology of the interaction is atypical due to the influence of the plant host.
The best examples are the monotropoid and some orchid

To use this database for identification it is important to

realize that confidence of the placement within the tree is
an important criterion. Two types of evidence can be
used: (i) internal confidence, and (ii) external independent evidence. To access internal confidence we have
used bootstrap analysis (Felsenstein 1985), but this may
not be necessary to use each time a new unknown is analysed. From the analyses we report here it seems safe to
say that placements of unknowns will be strongly supported within any of the groups listed in Table 2 if their
sequence differences from other members of the group
fall within the range listed. The suilloid group is a minor
exception; placements into this group are likely to be only
moderately supported by bootstrap but, as discussed
above, are very likely to be correct. Placements within
Cortinariaceae, Tricholomataceae, and related families
(lineages 11, 13, and 14) are more difficult to interpret,
because relationships among these taxa are not resolved
well by these data.
Independent evidence is the strongest confirmation.
We have used ITS-RFLP and oligonucleotide probe
analysis where feasible. The large number of candidate
species that one may need to compare in order to find an
ITS-RFLP match will remain a problem, but extensive
ITS-RFLP databases should help to solve this (Krn et al.
1997). Oligonucleotide probes will probably have a limited value in the foreseeable future because few are currently available and they require significant effort to test
1998 Blackwell Science Ltd, Molecular Ecology, 7, 257272

E C T O M Y C O R R H I Z A L B A S I D I O M Y C E T E D ATA B A S E
thoroughly. Confirmation by careful morphological
characterization is also feasible for distinctive and well
described types (Agerer 1987; Ingleby et al. 1990).

Evolutionary implications
The limited phylogenetic resolution of this region results
in low confidence in many of the major branches of the
tree shown, and the highly biased selection of taxa toward
EM species would be another problem if phylogenetic
estimation were the main goal. Nevertheless, two interesting evolutionary patterns transcend these problems and
are worth noting: (i) saprobic and EM taxa are intermixed
throughout the tree, and (ii) all of the EM groups for
which we have large samples exhibit very short withingroup branch lengths relative to other branches in the
tree.
The first pattern can be seen in several parts of the tree.
In the Boletales (Fig. 2a, all groups) the wood-decaying
species of the Coniophoraceae (groups 3 and 4) and
Paxillaceae (group 3) appear to be the close relatives of the
boletoid and suilloid groups, the two largest samples of
mycorrhizal fungi in our database. This connection of the
Coniophoraceae and Paxillaceae to the Boletaceae is also
supported by secondary chemistry (Gill & Steglich 1987).
At the base of the Russulaceae and the Thelephorales are
three wood-decaying taxa: Bondarzewia, Heterobasidion
and Panus. Bondarzewia has been hypothesized to be
related to the Russulaceae, based on morphological characters (Singer 1986) and this hypothesis is also suggested
by independent sequence data from the mitochondrial
small subunit rRNA gene (Hibbett & Donoghue 1995).
The latter work by Hibbet and Donoghue also placed
other wood-decaying taxa (Auriscalpium, Lentinellus,
Echinodontium and Gloeocystidiellum) into the clade that
includes the Russulaceae. Finally, within the central area
of the tree (Fig. 2c), saprobic and mycoparasitic taxa such
as Agaricus, Asterophora, Bolbitius, and Nematoloma are
intermixed with EM taxa such as Tricholoma, Inocybe, and
Cortinarius. The exact relationships within this loose
group are not clear from these data, as judged by multiple
equally parsimonious trees, short internodal branches,
and weak bootstrap values; nevertheless, it is clear that
the sequences of both nonmycorrhizal and EM taxa are
very similar to each other within this cluster. Collectively
these examples show that the switch between saprobic
and EM lifestyles probably happened convergently several and perhaps many times. These examples suggest
that different lineages of EM basidiomycetes may well
have different biochemical capacities which in turn may
relate to their ability to degrade litter and extract mineral
nutrients (Bruns 1995).
The second pattern, that of the short branches, can be
best seen in the boletoid (Fig. 2, group 1) and suilloid
1998 Blackwell Science Ltd, Molecular Ecology, 7, 257272

269

groups (group 6), the Russulaceae (group 8), the

Thelephorales (group 10) and the Amanitaceae (group
12). In these lineages the samples are large and diverse
enough that taxon selection is unlikely to be the reason for
the short branch lengths. If read from a molecular clock
perspective, the short branches suggest that these five
groups, and perhaps several others, represent relatively
recent radiations. We can not directly address the time
scale of the radiation with the ML5/ML6 data because
this molecule has not previously been used for molecular
clock estimates. Thus, no estimated rate of change nor any
dated branches have been determined. Furthermore, the
small size of the molecule would limit the resolution of
time estimates. For these reasons we turned to the nuc
SSU rRNA gene, which has been previously used for
molecular clock estimates in the fungi (Berbee & Taylor
1993; Simon et al. 1993), and for which data are available
for representatives of two of these five raditions: the suilloid and boletoid groups. We estimated divergence times
for these two groups by submitting 12 possible topologies
to a maximum likelihood molecular clock model
(DNAMLK), using Berbee and Taylors estimate of
220 Ma, and allowing an error of 50 Ma on this estimate
(Fig. 3). This approach yielded estimates ranging from 60
to 35 Ma for the suilloid group and 9431 Ma for the boletoid group. The larger range and greater age of estimates
for the boletoid group is probably caused by the inclusion
of Phylloporus, which is know to have an accelerated rate
of change in the nuc SSU rRNA gene (Bruns & Szaro
1992). In any case both sets of estimates essentially fall in
the early to mid Tertiary period and coincide with the
only fossil ectomycorrhiza found to date a 50 Ma old
middle Eocene ectomycorrhiza that appears to be suilloid
(Lepage et al. 1997). The upper end of our range of estimates also overlaps the Eocene-Oligocene transition.
During this time the earths climate became cooler and
more temperate, and trees in the Pinaceae and Fagales,
both obligate ectomycorrhizal taxa, came to dominate the
temperate forests (Berggren & Prothero 1992). Combining
these facts and estimates with our original observation of
short branch lengths in at least five lineages of ectomycorrhizal basidiomycetes, we speculate there has been a convergent radiation of several groups of EM fungi in
response to the expanding geographical ranges of their
shared plant hosts. This hypothesis may later be rejected
if many of the nonectomycorrhizal fungal groups also
appear to radiate during a similar time period, but our
current EM-biased taxomonic sample does not allow us to
address this issue.

This database in relation to current and future needs

We view this database as a working version that we will
continue to develop over the next several years. Updated

270

T. D . B R U N S E T A L .

Fig. 3 Estimated divergence times for the Boletales (1), boletoid group (2), and suilloid group (3) based on maximum likelihood analysis
of nuclear small subunit rRNA gene sequences. Berbee & Taylors (1993) estimated divergence of 220 Ma (b) for the node indicated (*) is
assumed. Estimated times are given graphically for the Boletales and suilloid group and in tabular form for all three lineages. Range of
estimates is derived by allowing 50 Ma (a & c) variation from the Berbee and Taylor date and through analysis of 12 other topologies that
differ slightly from the one shown. All 12 trees shared the internal branches indicated in bold and were not significantly different from
each other based on Kishino & Hasegawa (1989) tests. The tree is drawn to the geological time scale shown. Epochs of the Tertiary: P,
Palaeocene; E, Eocene; O, Oligocene; M, Miocene; unmarked, Pliocene.

versions will be posted on our website as they become

available. In its current state it is a useful supplement to
existing methods such as ITS-RFLP analysis and detailed
morphotyping. It fills a need for identification of types
that cannot be determined by other methods. The main
advantages of the ML5-ML6 fragment are: its small size,
its alignability, and the availability of fungal-specific
primers to amplify it.
Although most major EM groups are included in the
database, there are several significant omissions. The most
obvious is the lack of any ascomycetes. To date we have
not been able to amplify these well with the ML5/ML6
primers. This is not too surprising given that known
ascomycetous sequences (e.g. Saccharomyces, Aspergillus,
Neurospora and Podospora) are very divergent and essentially unalignable relative to the basidiomycetous
sequences in the database. Within the Basidiomycota there
are also some important omissions. These include non-

thelephoroid resupinate taxa (e.g. Amphinema), gasteroid

fungi of uncertain taxonomic placement (e.g. Leucogaster,
Leucophleps, and others), Coltricia (Hymenochaetaceae)
and additional taxa within the Tricholomataceae and
Cortinariaceae.
Other than such omissions, which can be corrected
over time, this database still has one major disadvantage:
its lack of resolution among closely related genera. For
this reason we expect that this database will probably be
replaced by one based on sequences from the ITS region.
The main advantages of ITS sequences are their much
greater resolution among closely related species and genera (Bruns 1996). ITS is such a popular target for phylogenetic analysis that sequences of it are rapidly
accumulating; these include sequences from some EM
groups such as Cortinarius, Dermocybe, Suillus, Tricharina,
and Wilcoxina (Liu et al. 1995; Egger 1996; Kretzer et al.
1996; Kretzer & Bruns 1997; Liu et al. 1997). Currently,
1998 Blackwell Science Ltd, Molecular Ecology, 7, 257272

E C T O M Y C O R R H I Z A L B A S I D I O M Y C E T E D ATA B A S E
many other important EM groups remain unsampled.
Thus, until many more ITS sequences are available the
database presented here will remain a useful tool for the
identification of EM fungi.

Acknowledgements
We thank David Jacobson for partial determination of Thelephora,
Melanogaster, and Bondarzewia sequences, and David Hibbet for
his detailed review and suggestions. Support from this work
came from and NSF grant DEB-9307150 to T.D.B.

References
Agerer R (1987) Colour Atlas Of Ectomycorrhizae. SchwbischGmnd, Einhorn-Verlag.
Agerer R (1994) Index of unidentified ectomycorrhizae III.
Names and identifications published in 1992. Mycorrhiza, 4,
183184.
Berbee ML, Taylor JW (1993) Dating of the evolutionary radiations of the true fungi. Canadian Journal of Botany, 71,
11141127.
Berggren WA, Prothero DR (1992) Eocene-Oligocene Climatic
and Biotic Evolution (Prothero DR, Berggren WA). Princeton,
Princeton University Press.
Bruns TD (1995) Thoughts of the processes that maintain local
species diversity of ectomycorrhizal fungi. Plant and Soil, 170,
6373.
Bruns TD (1996) Identification of ectomycorrhizal fungi using a
combination of PCR-based approaches. In: Fungal
Identification Techniques. (Rossen L, Rubio V, Dawson MT,
Frisvad J, eds). Barcelona, Spain, European Commission of
Science Research and Development, pp. 116123.
Bruns TD, Szaro TM (1992) Rate and mode differences between
nuclear and mitochondrial small-subunit rRNA genes in
mushrooms. Molecular Biology and Evolution, 9, 836855.
Bruns TD, Gardes M (1993) Molecular tools for the identification of ectomycorrhizal fungitaxon-specific oligonucleotide probes for suilloid fungi. Molecular Ecology, 2,
233242.
Bruns TD, Fogel R, Taylor JW (1990) Amplification and sequencing of DNA from fungal herbarium specimens. Mycologia, 82,
175184.
Bruns TD, Vilgalys R, Barns SM et al. (1992) Evolutionary relationships within the Fungi: analyses of nuclear small subunit
rRNA sequences. Molecular Phylogenetics and Evolution, 1,
231241.
Cullings KW, Szaro TM, Bruns TD (1996) Evolution of extreme
specialization within a lineage of ectomycorrhizal epiparasites. Nature, 379, 6366.
Currah RS, Smreciu EA, Hambleton S (1990) Mycorrhizae and
mycorrhizal fungi of boreal species of Platanthera and
Coeloglossum (Orchidaceae). Canadian Journal of Botany, 68,
11711181.
Egger KN (1996) Molecular systematics of E-strain mycorrhizal
fungi: Wilcoxina and its relationship to Tricharina (Pezizales).
Canadian Journal of Botany, 74, 773779.
Felsenstein J (1985) Confidence intervals on phylogenies: an
approach using the bootstrap. Evolution, 39, 783791.
Felsenstein J (1995) PHYLIP phylogenetic inference package,
version 3.5c. Computer programs distributed by the author.
1998 Blackwell Science Ltd, Molecular Ecology, 7, 257272

271

Seattle WA USA, Department of Genetics, University of

Washington.
Gardes M, Bruns TD (1993) ITS primers with enhanced specificity for basidiomycetes application to the identification of
mycorrhizae and rusts. Molecular Ecology, 2, 113118.
Gardes M, Bruns TD (1996) Community structure of ectomycorrhizal fungi in a Pinus muricata forest: Above- and
below-ground views. Canadian Journal of Botany, 74,
15721583.
Gardes M, White TJ, Fortin J, Bruns TD, Taylor JW (1991)
Identification of indigenous and introduced symbiotic
fungi in ectomycorrhizae by amplification of nuclear and
mitochondrial ribosomal DNA. Canadian Journal of Botany,
69, 180190.
Gill M, Steglich W (1987) Pigments of fungi (Macromycetes). In:
Progress In The Chemistry Of Organic Natural Products. (Herz
W, Grisebach H, Kerby GW, Tamm C, eds) New York,
Springer Verlag, 51, pp. 1317.
Henrion B, Le Tacon F, Martin F (1992) Rapid identification of
genetic variation of ectomycorrhizal fungi by amplification of
ribosomal RNA genes. New Phytologist, 122, 289298.
Hibbett DS, Donoghue MJ (1995) Progress toward a phylogenetic
classification of the Polyporaceae through parsimony analysis of mitochondrial ribosomal DNA sequences. Canadian
Journal of Botany, 73, S853S861.
Higgins DG, Bleasby AJ, Fuchs R (1992) C L U S TA L V : Improved
software for multiple sequence alignment. Computers and
Applied Bioscience, 8, 189191.
Hillis DM, Bull JL (1993) An empirical test of bootstrapping as a
method for assessing confidence in phylogenetic analysis.
Systematic Biology, 42, 182192.
Ingleby K, Mason PA, Last FT, Fleming LV (1990) Identification of
Ectomycorrhizas. London, Institute for Terrestrial Ecology,
Natural Environmental Research Council.
Krn O, Hgberg 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, 313315.
Kishino H, Hasegawa M (1989) Evaluation of the maximum likelihood estimate of the evolutionary tree topologies from DNA
sequence data, and the branching order in Hominoidea.
Journal of Molecular Evolution, 29, 170179.
Kretzer A, Bruns TD (1997) Molecular revisitation of the genus
Gastrosuillus. Mycologia, 89, 586589.
Kretzer A, Li Y, Szaro T, Bruns TD (1996) Internal transcribed
spacer sequences from 38 recognized species of Suillus sensu
lato: Phylogenetic and taxonomic implications. Mycologia, 88,
776785.
Lepage BA, Currah RS, Stockey RA, Rothwell GW (1997) Fossil
ectomycorrhizae from the middle Eocene. American Journal of
Botany, 4, 410412.
Li Y (1995) Molecular Characterization And Evolution Of
Mitochondrial Introns Within The Large Subunit Ribosomal RNA
Genes Of Suillus. PhD Thesis. University of California,
Berkeley, CA, USA.
Liu YJ, Rogers SO, Ammirati JF (1997) Phylogenetic relationships
in Dermocybe and related Cortinarius taxa based on nuclear
ribosomal DNA internal transcribed spacers. Canadian Journal
of Botany, 75, 519532.
Liu YJ, Rogers SO, Ammirati JF, Keller G (1995) Dermocybe,
Section Sanguineae: A look at species within the Sanguinea
complex. Beiheft zur Sydowia, 10, 142154.

272

T. D . B R U N S E T A L .

Marmeisse R, Debaud JC, Casselton LA (1992) DNA probes for

species and strain identification in ectomycorrhizal fungus
Hebeloma. Mycological Research, 96, 161165.
Simon L, Bousquet J, Levesque RC, LaLonde M (1993) Origin and
diversification of endomycorrhizal fungi and coincidence
with vascular land plants. Nature, 363, 6769.
Singer R (1986) The Agaricales in Modern Taxonomy. Koenigstein,
Koeltz Science Books.
Swann EC, Taylor JW (1993) Higher taxa of basidiomycetes: An
18S rRNA gene perspective. Mycologia, 85, 923936.
Swofford DL (1993) PA U P : phylogenetic analysis using parsimony, version 3.1.1. Champaign, Il. USA, Illinois Natural
History Survey.
Taylor DL, Bruns TD (1997) Independent, specialized invasions

of the ectomycorrhizal mutualism by two non-photosynthetic

orchids. Proceedings of the National Academy of Science, USA, 94,
45104515.
White TJ, Bruns TD, Lee SB, Taylor JW (1990) Amplification and
direct sequencing of fungal ribosomal RNA genes for phylogenetics. In: PCR Protocols A Guide To Methods And
Applications. (Innis MA, Gelfand DH, Sninsky JJ, White TJ,
eds). New York, Academic Press, pp. 315322.

The authors are interested in the ecology of ectomycorrhizal

fungi. The development of molecular identification methods has
been a necessary step in exploring this interest.

1998 Blackwell Science Ltd, Molecular Ecology, 7, 257272