Multiple gene genealogies and species recognition in the ectomycorrhizal fungus Paxillus involutus

mycological research 112 (2008) 965–975

journal homepage: www.elsevier.com/locate/mycres

Multiple gene genealogies and species recognition in the ectomycorrhizal fungus Paxillus involutus Jenny HEDH1, Peter SAMSON1, Susanne ERLAND, Anders TUNLID* Department of Microbial Ecology, Lund University, Ecology Building, SE-223 62, Lund, Sweden

article info

abstract

Article history:

Paxillus involutus (basidiomycetes, Boletales) is a common ectomycorrhizal fungus in the

Received 11 October 2007

Northern Hemisphere. The fungus displays significant variation in phenotypic characters

Received in revised form

related to morphology, physiology, and ecology. Previous studies have shown that P. invo-

14 January 2008

lutus contains several intersterility groups and morphological species. In this study, we

Accepted 24 January 2008

have used concordance of multiple gene genealogies to identify genetically isolated species

Corresponding Editor: Karl-Henrik

of P. involutus. Fragments from five protein coding genes in 50 isolates of P. involutus col-

Larsson

lected from different hosts and environments in Europe and one location in Canada were analysed using phylogenetic methods. Concordance of the five gene genealogies

Keywords:

showed that P. involutus comprises at least four distinct phylogenetic lineages: phyloge-

Boletales

netic species I (with nine isolates), II (33 isolates), III (three isolates), and IV (five isolates).

Cryptic species

The branches separating the four species were long and well supported compared with

ITS sequences

the species internodes. A low level of shared polymorphisms was observed among the

Mycorrhizas

four lineages indicating a long time since the genetic isolation began. Three of the phylo-

Phylogenetic species

species corresponded to earlier identified morphological species: I to P. obscurosporus, II to

Species concepts

P. involutus s. str., and III to P. validus. The phylogenetic species had an overlapping geographical distribution. Species I and II differed partly in habitat and host preferences. ª 2008 The British Mycological Society. Published by Elsevier Ltd. All rights reserved.

Introduction Paxillus involutus is one of the most well-studied ectomycorrhizal (ECM) fungi. The fungus is widespread in the Northern hemisphere and forms ectomycorrhizal with numerous coniferous and deciduous tree species (Wallander & So¨derstro¨m 1999). The unusually broad host range, relatively fast-growing mycelium, and rapid colonization of roots can explain why P. involutus is one of the most commonly used ECM fungi in laboratory experiments. Between 1996 and 2007, more than 300 research articles containing information on P. involutus were published (ISI Web of Science). For example, studies of this species have contributed to our understanding of the mechanisms of nutrient assimilation, carbon transfer, weathering, and heavy metal tolerance by ECM fungi. P. involutus is

also an important model for molecular studies of the ECM symbiosis, and it was recently approved for genome sequencing by the US Department of Energy (DOE) and the Joint Genome Institute (JGI). Several studies have reported a large variability in the morphology, physiology, and ecology between different isolates of P. involutus (e.g. Gafur et al. 2004; Laiho 1970). The large phenotypic variation raises the question whether P. involutus consists of several, genetically isolated species. Based on mating-type tests of specimen collections from the surroundings of Uppsala, Sweden, Fries (1985) reported the existence of three intercompatibility groups (biological species) in what was considered to be P. involutus. Sporocarps from the first group were mainly found in coniferous and deciduous forests (forest group), while those from the second and third groups

* Corresponding author. Tel.: þ46 46 222 3757. E-mail address: [email protected] 1 These authors have contributed equally to this work. 0953-7562/$ – see front matter ª 2008 The British Mycological Society. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.mycres.2008.01.026

966

were mainly found in park and garden areas (park groups). However, no distinct morphological differences were identified between sporocarps of these groups. Accordingly, it was suggested that the intersterility groups represented three cryptic species of P. involutus (Fries 1985). More recently, Hahn & Agerer (1999) made an extensive morphological and anatomical study of strains assigned to P. involutus collected in Europe and North America. Based on these data, it was proposed that the P. involutus species group should be divided into six morphological species: P. involutus s. str., P. validus, P. obscurosporus, P. vernalis, P. albidulus, and P. rubicundulus (syn. P. filamentosus). The division of P. involutus into several species has also been indicated by phylogenetic analyses of DNA sequence data generated from the nuclear encoded ITS region (Jarosch & Bresinsky 1999; Le Que´re´ et al. 2004). However, the support values for several clades were weak. Furthermore, data from a single locus cannot be used for delineating genetically isolated species (Avise & Wollenberg 1997). Taylor et al. (2000) have promoted the use of multiple gene genealogies to recognize boundaries of such lineages, referred to as phylogenetic species (PS). According to this approach, trees of multiple genes have the same topology due to fixations of previously polymorphic loci following genetic isolation. Conflicts among independent gene topologies can be caused by recombination between individuals within a species, and transition from concordance to conflict determines the limits of species (Taylor et al. 2000). This approach has been used to identify genetically isolated lineages and species in a number of different fungi including basidiomycetes (Johannesson & Stenlid 2003; Kauserud et al. 2006). Many of these studies have shown that the PS approach often recognizes additional genetically isolated species that have not been defined using mating tests (biological species) or phenotypic characters (morphological species) (Taylor et al. 2000; Taylor et al. 2006). In this study, we have used concordance of multiple gene genealogies to identify genetically isolated species in P. involutus. Fragments from five protein coding genes in 50 strains of the P. involutus s. lat. group collected from different hosts and environments in Europe and one location in Canada were analysed.

Materials and methods Isolates The study included 50 strains of what was assumed to be Paxillus involutus collected in Sweden, several other countries in Europe and one from Canada (Table 1). In addition, two specimens of P. filamentosus (syn. P. rubicundulus) and herbarium specimens of P. obscurosporus, P. rubicundulus and P. validus were included. Isolates included in this study obtained as mycelium (Table 1) are kept in a culture collection at the Department of Microbial Ecology, Lund University.

PCR and DNA sequencing DNA was extracted from fruit bodies (dried or frozen) or cultures using the E.Z.N.A. Fungal DNA Kit (Omega Bio-Tek).

J. Hedh et al.

Regions from five nuclear, protein-encoding genes previously isolated from the Paxillus involutus strain ATCC 200175 were amplified: actin (actA, GenBank accession number AY586027, 725 bp encompassing 482 bp exons and 243 bp introns), btubulin (b-tubA, AY586022, 569 bp encompassing 316 bp exons and 253 bp introns), glucose 6-phosphate isomerase ( gpiA, AY585998, 526 bp encompassing 424 bp exons and 102 bp introns), small GTPase protein (rabA, AY585950, 801 bp encompassing 485 bp exons and 316 bp introns), and hydrophobin A (hydA, DQ646583, 463 bp encompassing 355 bp exons and 108 bp introns) (Le Que´re´ et al. 2006; Rajashekar et al. 2007). In addition, the ITS region encompassing ITS1, ITS2 and 5.8S rRNA (AY585913, 555 bp) was analysed. The genes were PCR amplified using the primers in Table 2. PCR was performed in 25 ml reactions containing 19.4 ml dH20, 0.1 ml Easy-A polymerase (5 U/ml; Stratagene), 2.5 ml 10 Easy-A buffer, 0.5 ml dNTPs (10 mM each), 0.5 ml of each primer (10 mM) and 1.5 ml of genomic DNA. Templates that did not amplify with the Easy-A enzyme, were amplified using the AmpliTaq Gold DNA polymerase (Applied Biosystems). The 20 ml reactions contained 13.3 ml dH20, 2 ml 10 PCR buffer, 2.4 ml MgCl2 (25 mM), 0.4 ml dNTPs (10 mM each), 0.08 ml Taq polymerase (5 U/ml), 0.4 ml of each primer (10 mM) and 1.0 ml genomic DNA. PCR and cycle sequencing reactions were performed in a GeneAmp PCR system 9700 thermocycler. Cycling conditions for Easy-A polymerase were as follows: 2 min at 95  C followed by 30 cycles of 95  C for 40 s, 54–60  C for 30 s and 72  C for 1.5 min, and a final 7 min extension step at 72  C. Cycling conditions for Taq polymerase were as follows: 7 min at 94  C followed by 31 cycles of 94  C for 30 s, 54–60  C (depending on gene) for 30 s and 72  C for 2 min and 30 s, and then finally a 7 min extension step at 72  C. Amplification products were electrophoresed in a 1 % agarose gel with ethidium bromide. PCR products were purified with isopropanol and used as templates for sequencing. The amplification products were used as starting material for DNA sequencing using the Big Dye Terminator kit (Applied Biosystems) and template-specific primers (Table 2). A standard cycling protocol was followed (Applied Biosystems). The products were purified by ethanol precipitation and finally loaded onto an ABI 3100 DNA sequencer (Applied Biosystems). Overlapping sequences were aligned and trimmed using the Sequencher 3.1.1b4 program (Gene Codes Corporation). In the majority of the amplicons, both strands were sequenced.

Phylogenetic analysis The sequences were aligned using MUSCLE (version 3.6) (Edgar 2004). The ends were manually trimmed and ambiguous sites were removed using JalView (Clamp et al. 2004). Initial analysis showed that the gaps in the alignment were phylogenetically un-informative and they were removed from the analyses. Maximum parsimony (MP) and statistical-parsimony genealogies were constructed using PAUP (Swofford 2002) and TCS v. 1.21 (Clement et al. 2000), respectively. Bayesian analyses were performed using MrBayes v. 3.1.2 (Ronquist & Huelsenbeck 2003). The evolutionary model was selected by using the program MrModeltest v. 2.2 (Nylander JAA, 2004. MrModeltest v2. Program distributed by the author, Evolutionary Biology

Isolate

Tissuea Originb

Location

Habitat

Paxillus involutus s. lat. SE03083102 M, F SE03090704 M, F SE03091104 F SE03091215 F SE03092501 M, F SE03071001 M, F

SE SE SE SE SE SE

Ska˚ne, Ska˚ne, Ska˚ne, Ska˚ne, Ska˚ne, Ska˚ne,

SE03071622

M, F

SE

Ska˚ne, Lund

Public lawn

PAO03090701 IA04090201 AT04083001 SE04071201 Pi17PL SE04091101 SE04091102 Pi08BE

M, F M, F M, F F M F F M

SE NO NO SE PL SE SE BE

Ska˚ne, Stoby ˚ kershus A ˚ kershus A Blekinge, Gro¨nsla¨tt Chrzanow Halland, Hult Halland, Hult Maatheide

Wetland, sandfill Garden lawn Spruce forest Forested old pasture Unknown Spruce forest Spruce forest Unknown

Can (LH-145-03) Pi14BE ATCC 200175 SE03092803 SE04091903 SE04091904 HW04092801 Pi11SE KBE02067 KB04090601 KB04090602 KB04090603 SE03100903 PAO03091402 SE03092808 KBE02053 KBE02062 KBE02064 KBE02065 Pi01SE Pi9SE SE05091402 SE05091403 EL04090401

M, F M M, F F F F F M F F F F M, F F F F F F F M M F F F

CA BE GB SE SE SE SE SE SE SE SE SE SE SE SE SE SE SE SE SE SE SE SE NO

Ontario, Thunder Bay Paal Scotland Sma˚land, Bokliden Ska˚ne, Ellebo¨ke Ska˚ne, Ellebo¨ke Ska˚ne, Fulltofta Ska˚ne, Ho¨gana¨s Ska˚ne, Ho¨ja Ska˚ne, So¨dera˚sen Ska˚ne, So¨dera˚sen Ska˚ne, So¨dera˚sen Ska˚ne, Torna Ha¨llestad Ska˚ne, Vankiva Sma˚land, Vik Ska˚ne, Va¨sby Ska˚ne, Va¨sby, Ska˚ne, Va¨sby, Ska˚ne, Va¨sby, Ska˚ne, Va¨sby Ska˚ne, Va¨sby, Sma˚land, Parisma˚la Sma˚land, Parisma˚la ˚ kershus A

Conservation area Unknown Unknown Spruce forest Spruce forest Spruce forest Spruce forest Wetland Unknown Beech forest Beech forest Beech forest Beech forest Pine forest Spruce forest Sandy pine forest Sandy pine forest Sandy pine forest Sandy pine forest Sandy pine forest Unknown Spruce forest Spruce forest Garden lawn

Hja¨rup Hja¨rup Hja¨rup Hja¨rup Hja¨rup Lund

Garden lawn Garden lawn Public lawn Public lawn Garden lawn Public lawn

Host speciesc

actA

gpiA

hydA

rabA

b-tubA

ITS

PSd

Corylus, Betula Betula Betula Betula, Salix Betula Populus, Cornus, Tilia, Fagus, Quercus, Sorbus, Acer Populus, Tilia, Acer, Fagus, Quercus, Sorbus Populus, Betula Tilia Picea, Betula Betula, Quercus Unknown Picea, Pinus, Fagus, Quercus, Betula Picea, Pinus, Fagus, Quercus, Betula Pinus, Populus, Piceaf Betula, Populus Unknown Betula Picea Picea Picea Picea Fraxinus Pinus Quercus Quercus Quercus Sorbus, Fagus, Betula, Quercus Pinus, Betula Pinus, Picea, Betula Pinus, Betula, Quercus Pinus, Betula Pinus Pinus Pinus Betula, Pinus, Quercus Betula, Picea, Pinus Betula, Picea, Pinus Quercus

EF625863 EF625864 EF625865 EF625866 EF625867 EF625861

EU084597 EU084598 EU084599 EU084600 EU084601 EU084595

EU084641 EU084642 EU084643 EU084644 EU084645 EU084639

EF695379 –e EF695380 EF695381 EF695382 EF695377

EU084554 EU084555 EU084556 EU084557 EU084558 EU084553

EU078709 EU078710 EU078711 EU078712 EU078713 AY585910

I I I I I I

EF625862

EU084596

EU084640

EF695378

–

AY585911

I

EF625853 EF625844 EF625836 EF625874 – EF625875 EF625876 AY586029

EU084588 EU084579 EU084571 EU084607 EU084593 EU084608 EU084609 AY586000

EU084630 EU084621 EU084614 EU084651 EU084636 EU084652 EU084653 DQ646591

EF695369 EF695360 EF695353 EF695388 – EF695389 EF695390 AY585952

EU084544 EU084535 EU084527 EU084564 EU084550 EU084565 EU084566 AY586024

EU078714 EU078715 EU078716 EU078717 EU078718 EU078719 EU078720 AY585919

I I II II II II II II

EF625837 – AY585923 EF625868 EF625877 EF625878 EF625843 EF625859 EF625852 EF625845 EF625846 EF625847 EF625873 EF625855 EF625869 EF625848 EF625849 EF625850 EF625851 AY586028 EF625860 EF625879 EF625880 EF625841

EU084572 – AY585998 EU084602 EU084610 EU084611 EU084578 EU084592 EU084587 EU084580 EU084581 EU084582 EU084606 EU084590 EU084603 EU084583 EU084584 EU084585 EU084586 AY585999 EU084594 EU084612 EU084613 EU084576

EU084615 EU084635 DQ646583 EU084646 EU084654 EU084655 EU084620 EU084634 EU084629 EU084622 EU084623 EU084624 EU084650 EU084632 EU084647 EU084625 EU084626 EU084627 EU084628 DQ646590 EU084638 EU084656 EU084657 EU084618

EF695354 EF695375 AY585950 EF695383 EF695391 EF695392 EF695359 EF695374 EF695368 EF695361 EF695362 EF695363 – EF695371 EF695384 EF695364 EF695365 EF695366 EF695367 AY585951 – EF695393 EF695394 EF695358

EU084528 EU084549 AY586022 EU084559 EU084567 EU084568 EU084534 EU084548 EU084543 EU084536 EU084537 EU084538 EU084563 EU084546 EU084560 EU084539 EU084540 EU084541 EU084542 AY586023 EU084552 EU084569 EU084570 EU084532

EU078721 EU078722 AY585913 EU078723 EU078724 EU078725 EU078726 EU078727 EU078728 – – – AY585916 EU078729 AY585920 EU078730 – EU078731 EU078732 AY585918 EU078733 EU078734 EU078735 EU078736

II II II II II II II II II II II II II II II II II II II II II II II II

Ectomycorrhizal fungus Paxillus involutus

Table 1 – Isolates included in the study. Information about the isolates geographical origin, habitat and possible host species are given together with GenBank accession numbers for the sequences of the five nuclear genes (actA, gpiA, hydA, rabA, b-tubA) and the ITS-region

(continued on next page)

967

968

Table 1 – (continued) Isolate

Tissuea Originb

Location

Host speciesc

Habitat

actA

gpiA

hydA

rabA

b-tubA

ITS

PSd

EF625838 EF625840 EF625839 – EF625871 AY586030 AY586031 EF625854 EF625870 EF625872

EU084573 EU084575 EU084574 – EU084605 AY586001 AY586002 EU084589 EU084604 –

EU084616 – EU084617 EU084637 – DQ646594 DQ646595 EU084631 EU084648 EU084649

EF695355 EF695357 EF695356 EF695376 EF695386 AY585953 AY585954 EF695370 EF695385 EF695387

EU084529 EU084531 EU084530 EU084551 EU084562 AY586025 AY586026 EU084545 EU084561 –

EU078737 EU078738 EU078739 EU078740 EU078741 AY585917 AY585915 EU078742 EU078743 AY585921

II II II II III III III IV IV IV

EF625842

EU084577

EU084619

–

EU084533

AY585922

IV

AY585949 EU084658

AY585946 DQ646596 EU084659 EU084660

EU084663 EU084662

AY585948 EU084661

EU084664 –

CBS 477.89 CBS 674.97 CBS 591.89 Pi5BE SE03100403 May Nau PAO03090703 SE03100401 SE03100501

M M M M M, F M M F M, F M, F

CZ NL NL BE SE FR FR SE SE SE

Unknown Unknown Unknown Unknown Ska˚ne, Kjugekull Unknown Unknown Ska˚ne, Ha¨ssleholm Ska˚ne, Kjugekull Ska˚ne, Kjugekull

Unknown Unknown Unknown Unknown Pasture Unknown Unknown Public lawn Pasture Pasture

HW03092501

M, F

SE

Ska˚ne, Torna Ha¨llestad

Unmanaged lawn

Unknown Unknown Unknown Unknown Quercus, Betula Populusg Quercusg Betula Fraxinus, Populus,Corylus Fagus, Fraxinus, Betula, Quercus, Corylus Picea, Betula

Germany Ska˚ne, Haga ka¨rr

Unknown Wetland

Alnus Alnus, Fraxinus

DE

Bavaria

Unknown

Tiliah

–

–

–

–

–

EU084665

I

DE

Bavaria

Unknown

Tiliah

–

–

–

–

–

EU084666

III

GB

Scotland

Unknown

–

–

–

–

–

EU084667

P. filamentosus (syn. P. rubicundulus) Pf01DE M DE KBE02056 F SE Reference materialh P. obscurosporus CH 290/98 F P. validus CH 243/97 F P. rubicundulus Orton 2905 F

a M, mycelium; F, fruitbody (sporocarp). b BE, Belgium; CA, Canada; CZ, Czechoslovakia; DE, Germany; FR, France; GB, Great Britain; NO, Norway; NL, The Netherlands; PL, Poland; SE, Sweden. c Woody plants within a 10 m radius. d Phylogenetic species recognized by concordance of the gene genealogies of the five nuclear genes (c.f. Fig 2). e –, no sequence obtained. f According to Blaudez et al. (1998). g According to Gafur et al. (2004). h Sequences obtained from herbarium type specimens described by Hahn & Agerer (1999).

J. Hedh et al.

Ectomycorrhizal fungus Paxillus involutus

969

Table 2 – Primers used for PCR amplification and DNA sequencing of actin (actA), glucose-6-phosphate isomerase ( gpiA), hydrophobin A (hydA), small GTPase protein (rabA), ß-tubulin (ß-tubA), and ITS in various Paxillus strains Gene actA gpiA hydA rabA ß-tubA ITS

a b c d e

Directiona

Primer ID b

P164 P165b P530c P531c P365d P366bd P522b P523b P168b P169b ITS 1e ITS 4e

Primer sequence 0

F R F R F R F R F R F R

5 -ATGGAAGATGAAGTTGCCGCC 50 -GGTCAAATCACGTCCAGCGAGA 50 -AGTCGAACGGAAAATTCGTC 50 -TTCCTGTTCTTCTGATAGTAGTGAA 50 -CACATACGTCTTTGCTCTTGAC 50 -GTACAGGACCGGATGTTGACGAG 50 -TGGCTGAAGGTAGTAACTAC 50 -CAAGAGCCGTTCTTAGTGGG 50 -GGTTTTGGAAAATGCTTCTCGC 50 -GCAACTTCGCCAACAGAAGGGATCC 50 -TCCGTAGGTGAACCTGCGG 50 -TCCTCCGCTTATTGATATGC

F and R indicate forward and reverse primers, respectively. From Le Que´re´ et al. (2006). This study. From Rajashekar et al. (2007). From White et al. (1990).

Centre, Uppsala University). For the Bayesian analyses, eight chains of Markov chain Monte Carlo (MCMC) were used and the analysis ran for 1M generations with a sample frequency of 5. The burn in was set to 100K based on visual inspection of the stationary phase of the MCMC. Phylogenetic analyses were performed for each gene region and for the combined data. The five genes were concatenated using SNAP Workbench (Aylor et al. 2006). To confirm the concatenated tree based on the ‘total evidence’ (i.e. all five genes) a supertree was constructed with Rainbow v1.2 beta (http://genome.cs.iastate.edu/Rainbow/). The supertree was constructed using matrix representation with parsimony (Baum 1992, Ragan 1992), which combines individual trees derived from multiple datasets, and thus overcoming the missing data. Each of the five gene trees was used to construct binary data files describing the topology in each tree. These binary data files were concatenated with SNAP Workbench resulting in a binary data file describing a supertree. On this binary data file a MP analysis as described above was performed. In the MP analysis the characters were treated unordered with equal weights using the heuristic search option, with the tree bisection–reconnection (TBR) branch swapping algorithm and random addition sequence option with 100 replicates to find multiple islands in order to find the most parsimonious tree(s). All other settings were default. BS support values for branching topologies were determined with the same parameters using 1K search replicates. To assess incongruence

between the topologies in the single gene trees when the five genes were concatenated, Farris test (Farris et al. 1995) were performed with PAUP. In the statistical parsimony analysis, done with TCS, the connection limit was set to 95 % and all other settings were default values. Nucleotide diversity, p, was calculated with DnaSP v4.10.9 (Rozas et al. 2003). Polymorphism analysis was performed with SITES (Hey & Wakeley 1997).

Results Polymorphism data The sequence variation of five protein encoding genes in 50 isolates of the Paxillus involutus s. lat. group was analysed in this study (Table 1). After removal of ambiguous sites and gaps, the combined data for the five gene regions encompassed 1893 sites of which 335 were variable (18 %; Table 3). The loci differed in the level of variation. The most polymorphic gene region was hydA (p ¼ 0.054), and the least polymorphic region was actA (p ¼ 0.016). The five genes also differed in the number of exons/introns ranging from gpiA (2/1) to rabA (5/6).

Analyses of single locus datasets The genealogy of each of the five gene regions was constructed (Fig 1). Relationships among the clades were

Table 3 – Summary of the five DNA sequence alignments Gene

actA

gpiA

hydA

rabA

b-tubA

Number of strains Length of alignment (bp) Number of removed ambiguous sites and gaps (bp) Length of final alignment (bp) Numbers of variable nucleotides Nucleotide diversity (p)

49 349 14 335 45 0.016

49 374 29 345 39 0.018

50 379 29 350 105 0.054

47 480 17 463 66 0.030

50 434 34 400 80 0.039

Concatenated 52 2016 123 1893 335

970

J. Hedh et al.

actA

1.00/97

rabA

IV

1.00/98

I II

0.74/64

1.00/100

II

0.99/87

1.00/100

1.00/100

1.00/100

IV 1.00/100

I

1.00/85

1.00/95

1.00/100

III

0.01

III

0.1

gpiA

-tubA

II

1.00/69

II

1.00/100

1.00/100

IV

1.00/99

I

1.00/95 0.89/85

1.00/100

1.00/99

III

1.00/95

0.1

IV 1.00/100

1.00/98 0.1

I III

hydA

II

1.00/100

1.00/100

0.70/54

1.00/100

0.97/73 1.00/100 0.1

IV

1.00/100

I III

Fig 1 – Phylogenies of Paxillus. involutus derived from actA, gpiA, hydA, rabA, and ß-tubA using Bayesian analysis. The Bayesian PPs above 0.5 are given by the branches and BS values from MP above 50 are also given, the latter after ‘‘/’’. The tree was rooted using sequences from P. filamentosus. The scale denotes expected changes per nucleotide. Four different phylogenetic species are identified which are framed in boxes labelled I–IV. Strains and accession numbers to genes are presented in Table 1. Three of the PS corresponded to earlier identified morphological species: I to P. obscurosporus, II to P. involutus s. str., and III to P. validus.

Ectomycorrhizal fungus Paxillus involutus

topologically identical using MP and Bayesian tree analyses. Therefore, the Bayesian tree is shown with the MP BS values included for the well-supported branches. Four well-supported clades were identified in four of the single-locus trees (rabA, gpiA, b-tubA, and hydA). Three out of the four clades were clearly identified in the actA tree. As previously reported, two different copies of the hydA (designated hydA1 and hydA2) gene were amplified in the isolates May and Nau (Rajashekar et al. 2007). All other strains contained only one copy of hydA. Trees were constructed using either hydA1 or hydA2 from Maj and Nau. The topologies of the two trees were slightly different, but both of them clustered the hydA sequences into four clades containing the same sets of isolates. All trees shown (Figs 1–3), are based on the analyses of the hydA2 gene copy of May and Nau.

Combined dataset and identification of phylogenetic species The Farris test did not reveal any significant incongruence between the single locus trees ( p ¼ 0.07) and, therefore, it was concluded that there were no impediments to the construction of a concatenated gene tree. When the combined dataset was evaluated, four distinct clades were identified with Bayesian PPs of 1 (MP BS values of 100; Fig 2): Phylogenetic species

971

(PS) I encompassing nine isolates; PS II having 33 isolates, PS III with three isolates, and PS IV containing five isolates. Furthermore, the analysis showed that the species were separated into two groups, one containing PS II and IV, and one with PS I and III, respectively. The clades corresponding to PS I to IV were located at a large distance from the sequences of Paxillus filamentosus. Overall the species internodes were short and poorly supported as compared with branches leading to the species (Fig 2). The nucleotide diversity (p) was significantly lower within as compared with between the identified phylospecies (PS; Table 4). Furthermore, the level of shared polymorphisms among the four lineages was low (0–0.05 %) compared with the polymorphisms occurring within each phylogenetic species (0.2–1.3 %; Supplementary Material Table S2). The topology of the concatenated tree was verified by constructing a supertree (Supplementary Material Fig S1). This approach, which is constructed based on the topology of each single locus tree, produced the same four distinct clades as were observed in the concatenated tree. This indicates that the missing data in Table 1 does not pose any problems in delineating the four phylogenetic species within the P. involutus s.l. species group. Recombination produces networks of sequences rather than bifurcating evolutionary trees, which should be revealed

Fig 2 – Phylogeny of Paxillus involutus derived from the concatenated sequences of actA, gpiA, hydA, rabA, and ß-tubA using Bayesian analysis. The Bayesian PPs above 0.5 are given by the branches and BS values from MP above 50 are also given, the latter after ‘‘/’’. The tree was rooted using sequences from P. filamentosus. The scale denotes expected changes per nucleotide. Four phylogenetic species are recognized (I to IV). Strains are described in Table 1. Three of the PS corresponded to earlier identified morphological species: I to P. obscurosporus, II to P. involutus s. str., and III to P. validus.

972

J. Hedh et al.

Fig 3 – Phylogenetic networks of Paxillus involutus derived from the concatenated gene for two of the phylogenetic species, PS I (A) and PS II (B). The open circles denotes a nucleotide changes. The networks were created with the statistical parsimony analysis using the TCS software (Clement et al. 2000).

Ectomycorrhizal fungus Paxillus involutus

973

Table 4 – Numbers of variable nucleotides and nucleotide diversity (p) (in brackets) between and within the four phylospecies (PS) of Paxillus involutus s. lat.a

PS PS PS PS

I II III IV

PS I

PS II

PS III

11 (0.0018)

121 (0.020) 114 (0.0006)

141 (0.0085) 179 (0.024) 25 (0.010)

PS IV 144 181 124 76

(0.0083) (0.025) (0.029) (0.0009)

a The number of compared nucleotide sites is 1893.

in an analysis of the combined data. Analyses of the sequences within PS I and PS II provided some evidences for network relationships (Fig 3).

support values for the branches separating the species were weaker compared with those in the combined tree of the five protein-encoding genes (Fig 2). The analysis based on the ITS sequences also included reference material from several of the morphological species described by Hahn & Agerer (1999) and strains analyzed by Jarosch & Bresinsky (1999). Based on the clustering of these sequences, PS I corresponds to Paxillus obscurosporus, PS II to P. involutus s. str. and PS III to P. validus. Species IV was not supported by any of the analysed reference material. In total, the ITS region (including the 5.8 S gene) contained 28 variable sites (Supplementary Material Table S1). Eleven of the polymorphisms occurred in more than one of the species, and 17 were specific for one of the lineages including one for PS I, seven for PS II, three for PS III, and six for PS IV. Ten out of the 11 polymorphisms were common for PS II and PS IV.

ITS sequences and comparisons between phylogenetic species and morphological species

Habitat and host preferences

The lineages representing the four PS were also recognized in a tree constructed from ITS sequences (Fig 4). However, the

All four phylogenetic species were found within the county of Ska˚ne in the south of Sweden. Species IV was only

Pf01DE P. rubicundulus*

0.86/51 0.97/57 0.61/61

1.00/100

III I IV

1.00/99 0.88/51

0.80/91

II

0.1

Fig 4 – Phylogeny derived from the ITS tree using Bayesian analyses. The Bayesian PPs above 0.5 are given by the branches and BS values from MP above 50 are also given, the latter after ‘‘/’’. The scale denotes expected changes per nucleotide. PS 1 to 4 are framed in boxes labelled I to IV. Asterisks (*) indicate sequences obtained from herbarium specimens. Accession numbers and strains are presented in Table 1. Dots (C) indicate sequences obtained from GenBank: P. vernalis, strain Pv2 (accession number AF167689); P. involutus strains Pi3 (AF167695), Pi9H (AF167696), PiM3 (AF167700), Pi2 (AF167699), Pi11 (AF167698), Pi10H (AF167697) (Jarosch & Bresinsky 1999). The tree was rooted using the sequence from P. filamentosus. Three of the PS corresponded to earlier identified morphological species: I to P. obscurosporus, II to P. involutus s. str., and III to P. validus.

974

represented in Ska˚ne, all other groups included collections from different countries in Europe and PS II even included a collection from Canada (Table 1). There were clear habitat differences between PS I and PS II. The PS I collections all came from public lawns and gardens, there were no potential coniferous host plants registered within this group. The sporocarps often occurred in aggregates, the size range of fully developed sporocarps was 13–23 cm diam. Sporocarps from PS II were mostly found in established forests of different kinds. There were often coniferous trees registered as possible host plants. The sporocarps were never found in aggregates. The fully developed sporocarps ranged in size from 5–14 cm diam. The number of collections was too low to draw any firm conclusions about the sizes of sporocarps from group III and IV. However, the few collections analysed from these groups came from unmanaged grassland or active pasture with some trees, a type of habitat that was under-sampled compared with lawns and forests in this study. Even in monoculture forests and plantations, the fruit bodies mostly appeared at the occasional spot where there was a mixture of tree species. Betula was a common host plant for all four phylogenetic species.

Discussion The analyses of the five sequenced loci clearly identified four distinct genetic lineages within the Paxillus involutus s.lat. species group. The five gene genealogies were highly concordant. The branches separating the four phylogenetic species were long and well supported compared with the branches occurring within the species. Evolutionary theory predicts that polymorphisms shared by sibling species decrease with time since the onset of isolation (Taylor et al. 2000). Thus, the very low proportion of shared sequence polymorphism between the Paxillus lineages indicates that the four PS have been genetically isolated for a long time. Species delimitation by genealogical analyses of multiple loci is dependent on identifying the nodes where transitions from concordance to incongruity among branches occur. Such conflicts among the gene trees produce network relationships due to recombination between individuals within the species (Taylor et al. 2000). Network relationships were indicated between the sequences within PS I and II of P. involutus, suggesting recombination. Alternatively, the networks can also be due to incomplete lineage sorting of polymorphic loci (Maddison 1997). Lineage sorting of polymorphic sites is an ongoing process following genetic isolation and the proportion of loci for which polymorphisms are shared changes in inverse proportion to the time since genetic isolation began (Taylor et al. 2000). As discussed above, PS I and PS II have been genetically isolated for a long time, which suggests that a large part of shared polymorphisms have been lost. Larger numbers of isolates and/or more polymorphic gene regions are needed for detecting possible reticulate relationships among individuals of PS III and IV. Furthermore, sampling of isolates from other continents besides Europe was limited in our study. P. vernalis, that is restricted to North America (Watling 1969), was represented by only one ITS

J. Hedh et al.

sequence. In the ITS tree, the sequence of P. vernalis was positioned close, but outside the PS III clade. Data from more collections are needed to determine whether P. vernalis comprises a distinct phylogenetic species within the P. involutus species group. On a global scale, it is therefore possible that the P. involutus s. lat. contains more than four genetically isolated lineages. Hahn & Agerer (1999) recognized six morphospecies within the P. involutus s. lat. species group: P. involutus s. str., P. validus, P. obscurosporus, P. vernalis, P. albidulus, and P. rubicundulus. Previous phylogenetic analyses based on ribosomal sequences separates P. rubicundulus from strains and species of P. involutus (Jarosch & Bresinsky 1999). This is clearly confirmed by our analyses of protein encoding genes. According to the clustering of ITS sequences (cfr Fig 4), PS I, II, and III corresponded to P. obscurosporus, P. involutus s. str. and P. validus, respectively. We were not able to amplify any sequences from herbarium material of P. albidulus. This species has so far only been reported from Czechoslovakia (Sˇutara 1991). P. albidulus lacks the typical brown pigmentation of other species within the P. involutus species group (Hahn & Agerer 1999). We did not record any white sporocarps within our collections and, therefore, exclude the possibility that PS IV could correspond to P. albidulus. Thus, PS IV is most likely a previously not described species within the P. involutus s. lat. species group. The overlapping geographical distributions of PS I to IV and the apparent lack of recombination between them, indicates that these phylogenetic species are reproductively isolated. Three of the phylogenetic species most probably correspond to the biological species identified by Fries (1985). Fries’ intersterility group I, which contained the largest group of isolates, showed a preference to coniferous trees and was found in forests, and the group most likely corresponds to PS II. Intersterility groups II and III with a preference for deciduous trees and mainly collected in park and garden lawns, might correspond to PS I and III. The records of size, growth patterns and habitats of PS I and PS II could be compared with those of P. obscurosporus and P. involutus s. str. Hahn & Agerer (1999) reported Tilia, Corylus, Quercus, and Abies as host plants for P. obscurosporus. We found all except Abies among our possible host plants in PS I, in addition to several other broad-leaf trees (see Table 1). A preference for loamy and neutral soils from tree nurseries could also be comparable with the soil in parks and gardens in Ska˚ne. Sporocarps were described to grow in aggregates in both cases. However, the size of the sporocarps described by Hahn & Agerer were 7–40 cm in diameter, which is larger than found for PS I collections in our study. Hahn & Agerer (1999) listed many tree species, both coniferous and broad leaf as host plants for P. involutus s. str., which was in line with our findings for PS II. The size of the sporocarps of P. involutus s. str. as reported by Hahn & Agerer was similar to that of the PS II specimens collected in our study. Furthermore, the sporocarps of P. involutus s. str. (Hahn & Agerer 1999) and PS II as recorded in this study did not occur in aggregates. A reliable and detailed phylogeny is a prerequisite for addressing a number of questions regarding the ecology and evolution of symbiosis in P. involutus. Analyses of ITS sequences have become a standard tool for identifying species of ECM fungi in environmental studies (Horton & Bruns

Ectomycorrhizal fungus Paxillus involutus

2001; Ko˜ljalg et al. 2005). As a result of this study, researchers could now recognize different species within the P. involutus s. lat. based on ITS sequence similarities.

Acknowledgements This study was supported by grants from the Swedish Research Council for Environment, Agricultural Sciences and Spatial Planning (FORMAS), the Swedish Research Council (VR) and Kungliga Fysiografiska Sa¨llskapet in Lund. P.S. was supported by grants from the Research School in Genomics and Bioinformatics. DNA sequencing was performed at the SWEGENE Center of Genomic Ecology at the Ecology Building in Lund, supported by the Knut and Alice Wallenberg Foundation through the SWEGENE consortium. We thank Eva Friman for help with DNA sequencing and Karl-Henrik Nilsson for advice on the phylogenetic analyses.

Supplementary material Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.mycres.2008.01.026

references

Avise JC, Wollenberg K, 1997. Phylogenetics and the origin of species. Proceedings of the National Academy of Sciences, USA 94: 7748–7755. Aylor DL, Price EW, Carbone I, 2006. SNAP: Combine and Map modules for multilocus population genetic analysis. Bioinformatics 22: 1399–1401. Baum BR, 1992. Combining trees as a way of combining datasets for phylogenetic inference, and the desirability of combining gene trees. Taxon 41: 3–10. Blaudez D, Chalot M, Dizengremel P, Botton B, 1998. Structure and function of the ectomycorrhizal association between Paxillus involutus and Betula pedula. II. Metabolic changes during mycorrhiza formation. New Phytologist 138: 543–552. Clamp M, Cuff J, Searle SM, Barton GJ, 2004. The Jalview Java alignment editor. Bioinformatics 20: 426–427. Clement M, Posada D, Crandall KA, 2000. TCS: a computer program to estimate gene genealogies. Molecular Ecology 9: 1657– 1659. Edgar RC, 2004. MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Research 32: 1792– 1797. Farris JS, Ka¨llersjo¨ M, Kluge AG, Bult C, 1995. Testing significance of incongruence. Cladistics 10: 315–319. Fries N, 1985. Intersterility groups in Paxillus involutus. Mycotaxon 24: 403–410. Gafur A, Schu¨tzendu¨bel A, Langenfeld-Heyser R, Fritz E, Polle A, 2004. Compatible and incompetent Paxillus involutus isolates for ectomycorrhiza formation in vitro with poplar (Populus canescens) differ in H2O2 production. Plant Biology 6: 91–99. Hahn C, Agerer R, 1999. Studium zum Paxillus involutus Formenkreis. Nova Hedwigia 69: 241–310. Hey J, Wakeley J, 1997. A coalescent estimator of the population recombination rate. Genetics 145: 833–846.

975

Horton TR, Bruns TD, 2001. The molecular revolution in ectomycorrhizal ecology: peeking into the black box. Molecular Ecology 10: 1855–1871. Jarosch M, Bresinsky A, 1999. Speciation and phylogenetic distances within Paxillus s. str. (basidiomycetes, Boletales). Plant Biology 1: 701–706. Johannesson H, Stenlid J, 2003. Molecular markers reveal genetic isolation and phylogeography of the S and F intersterility groups of the wood-decay fungus Heterobasidion annosum. Molecular Phylogenetics and Evolution 29: 94–101. Kauserud H, Stensrud O, Decock C, Shalchian-Tabrizi K, Schumacher T, 2006. Multiple gene genealogies and AFLPs suggest cryptic speciation and long-distance dispersal in the basidiomycete Serpula himantioides (Boletales). Molecular Ecology 15: 421–431. Ko˜ljalg U, Larsson KH, Abarenkov K, Nilsson RH, Alexander IJ, Eberhardt U, Erland S, Hoiland K, Kjoller R, Larsson E, Pennanen T, Sen R, Taylor AF, Tedersoo L, Vralstad T, Ursing BM, 2005. UNITE: a database providing web-based methods for the molecular identification of ectomycorrhizal fungi. New Phytologist 166: 1063–1068. Laiho O, 1970. Paxillus involutus as a mycorrhizal symbiont of forest trees. Acta Forestalia Fennica 106: 1–73. Le Que´re´ A, Astrup EK, Rajashekar B, Schu¨tzendu¨bel A, Canba¨ck B, Johansson T, Tunlid A, 2006. Screening for rapidly evolving genes in the ectomycorrhizal fungus Paxillus involutus using cDNA microarrays. Molecular Ecology 15: 535–550. Le Que´re´ A, Schu¨tzendu¨bel A, Rajashekar B, Canba¨ck B, Hedh J, Erland S, Johansson T, Tunlid A, 2004. Divergence in gene expression related to variation in host specificity of an ectomycorrhizal fungus. Molecular Ecology 13: 3809–3819. Maddison WP, 1997. Gene trees in species trees. Systematic Biology 46: 523–536. Ragan MA, 1992. Phylogenetic inference based on matrix representation of trees. Molecular Phylogenetics and Evolution 1: 53–58. Rajashekar B, Samson P, Johansson T, Tunlid A, 2007. Evolution of nucleotide sequences and expression patterns of hydrophobin genes in the ectomycorrhizal fungus Paxillus involutus. New Phytologist 174: 399–411. Ronquist F, Huelsenbeck JP, 2003. MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19: 1572– 1574. Rozas J, Sanchez-DelBarrio JC, Messeguer X, Rozas R, 2003. DnaSP, DNA polymorphism analyses by the coalescent and other methods. Bioinformatics 19: 2496–2497. Sˇutara J, 1991. Paxillus albidulus, a new species of the family Pax ´ Mykologie 45: 129–133. illaceae. Ceska Swofford DL, 2002. PAUP*: phylogenetic analysis using parsimony (*and other methods). Sinauer Associates, Sunderland, MA. Taylor JW, Jacobson DJ, Kroken S, Kasuga T, Geiser DM, Hibbett DS, Fisher MC, 2000. Phylogenetic species recognition and species concepts in fungi. Fungal Genetics and Biology 31: 21–32. Taylor JW, Turner E, Townsend JP, Dettman JR, Jacobson D, 2006. Eukaryotic microbes, species recognition and the geographic limits of species: examples from the kingdom Fungi. Philosophical Transactions of the Royal Society of London, Series B, Biological Sciences 361: 1947–1963. Wallander H, So¨derstro¨m B, 1999. Paxillus. In: Cairney JWG, Chambers SM (eds), Ectomycorrhizal Fungi: key genera in profile. Springer-Verlag, Berlin, pp. 231–252. Watling R, 1969. New fungi from Michigan. Notes from the Royal Botanic Garden Edinburgh 29: 59–66. White TJ, Bruns T, Lee S, Taylor J, 1990. Amplification and direct sequencing of fungal ribosomal RNA genes for phylogenetics. In: Innis MA, Gelfand DH, Sninsky JJ, White TJ (eds), PCR Protocols: a guide to methods and applications. Academic Press, San Diego, pp. 315–322.