Phylogeny of the Genus Agaricus Inferred from Restriction Analysis of Enzymatically Amplified Ribosomal DNA

Home , Agaricus, Agaricus bisporus, Agaricus bitorquis

Fungal Genetics and Biology 20, 243–253 (1996) Article No. 0039

Britt A. Bunyard,* Michael S. Nicholson,† and Daniel J. Royse‡ *USDA-ARS, Fort Detrick, Building 1301, Frederick, Maryland 21701; †Department of Biology, Grand Valley State University, Allendale, Michigan 49401; and ‡Department of Plant Pathology, Pennsylvania State University, 316 Buckhout Laboratory, University Park, Pennsylvania 16802

Accepted for publication November 15, 1996

Bunyard, B. A., Nicholson, M. S., and Royse, D. J. 1996. accounting for 37% of the total world production of Phylogeny of the genus Agaricus inferred from restriction cultivated mushrooms. Although much is known about A. analysis of enzymatically amplified ribosomal DNA. Fungal bisporus, several aspects remain unclear, especially those Genetics and Biology 20, 243–253. The 26S and 5S concerning its genetic life history (Kerrigan et al., 1993a; ribosomal RNA genes and the intergenic region between Royer and Horgen, 1991; Castle et al., 1988, 1987; Spear et the 26S and the 5S rRNA genes of the ribosomal DNA al., 1983; Royse and May, 1982a,b; Elliott, 1972; Raper et repeat of 21 species of Agaricus were amplified using PCR al., 1972; Jiri, 1967; Pelham, 1967; Evans, 1959; Kligman, and then digested with 10 restriction enzymes. Restriction 1943). Fundamental processes such as the segregation and fragment length polymorphisms were found among the 21 assortment of genes during meiosis remain poorly defined putative species of Agaricus investigated and used to (Kerrigan et al., 1993a; Summerbell et al., 1989; Royse and develop a phylogenetic tree of the evolutionary history of May, 1982a). In addition, little information exists about the A. bisporus. The 58 end of the 26S gene showed more evolution and relatedness of species within the genus variability than the 38 end. A. excellens, A. chionodermus, Agaricus (Kerrigan et al., 1993b). and A. caroli represented the species most distantly In the past, strains of A. bisporus having desirable or related to A. bisporus. We present here the first comprehen- improved traits were developed from single- or multispore sive attempt at systematically resolving the entire genus cultures (Kligman, 1943). Suitable parental lines were Agaricus using modern techniques for molecular genetic selected on the basis of growth habit on defined media analysis. Our data indicate that previous taxonomic (Fritsche, 1981). New strains were then developed follow- schemes, based on morphological characters, are in need ing anastomosis between the hyphae of different parental strains grown in the same petri plate (Fritsche, 1981). of revision. r 1996 Academic Press Through continuous selection, those strains with undesired Index Descriptors: mushrooms; phylogeny; RFLPs; ri- characteristics were eliminated. bosomal DNA. Recently, researchers employed allozyme analysis to identify genotypic classes and to follow the segregation of allozyme- encoding alleles during meiosis (Royse and May, 1990, 1982a; INTRODUCTION Kerrigan and Ross, 1989; Royse et al., 1983; Spear et al., 1983) in various wild and commercial mushrooms. Genetic studies The white button mushroom, Agaricus bisporus (Lange) with allozymes have been conducted on more than 30 crop Imbach (5A. brunnescens Peck) is the most important species (Abler et al., 1991). Spear et al. (1983) created the first commercially cultivated mushroom species in the world, map of linkage groups in A. bisporus using allozymes. How-

1087-1845/96 $18.00 Copyright r 1996 by Academic Press All rights of reproduction in any form reserved. 243 244 Bunyard, Nicholson, and Royse ever, the lack of suitable markers resulted in a very limited map. tandem array of transcribed and nontranscribed stretches of Kerrigan et al. (1991) have added to this map of linkage groups DNA (Cassidy et al., 1984; Petes and Botstein, 1977). The large using restriction fragment length polymorphisms (RFLPs)1 subunit ribosomal RNAs (LSU) vary in length from about 3500 and randomly amplified polymorphic DNA. It is important to to 5000 nucleotides in most eukaryotes and contain regions of note that gene linkage maps for A. bisporus, as yet, do not link rapidly and slowly evolving regions. RFLP analysis of genes or markers to any agronomically important characteristics. Other intergenic regions of rRNA may provide new information methods, relying on DNA analysis, used to explore the relation- about evolutionary relationships of many different fungal taxa. ships among and between taxa have been discussed (Royse et As the sequences of the 26S–28S genes are highly conserved, al., 1993). they are useful for comparisons between distantly related RFLPs have been used to define genotypes of hetero- taxonomic groups. The 5S gene is much less conserved and, as karyotic and homokaryotic strains and to confirm crosses such, may change more rapidly over time. The regions between among several isolates (Summerbell et al., 1989; Castle et the genes, or intergenic regions (IGRs), are not transcribed al., 1987, 1988), as well as to investigate the postmeiotic and, therefore, show much lower conservation when compared distribution of nuclei in A. bisporus (Kerrigan et al., 1993a; to genic regions. The 5S gene and IGRs are useful for Summerbell et al., 1989). Conclusions about reproductive elucidating differences between closely related taxa, at the events are scarce due to the small number of known species or race level. Johansen et al. (1992) determined the genetic markers available for genetic analysis of the species time of divergence between the myxomycetes Didymium and (Kerrigan, 1990; Castle et al., 1987), as well as the lack of Physarum based on the amount of variation in the LSU of the an extensive genetic map of linkage groups. Also of interest two genera. Vilgalys and Gonzales (1990) recently used RFLP is the evolutionary history of A. bisporus which remains analysis of the rDNA subunits for the determination of uncertain (Singer, 1986). Although addressed in great genotypic classes for the fungal species Rhizoctonia solani detail by Kerrigan (1990), many questions still exist concern- Ku¨ hn. Using similar methods, other researchers have begun ing the evolutionary origin of A. bisporus, the relationship taxonomic work involving species of Phanerochaete and Sporot- of cultivars to wild populations of this species, and its richum (Raeder and Broda, 1984), Pythium (Martin, 1990), position within the Agaricales. and Sclerotinia (Kohn et al., 1988). Anderson and Stasovski Phylogenetic assessment of Basidiomycetes based on DNA (1992) determined the molecular phylogeny of species of offer many advantages over traditional morphological methods. Armillaria from the Northern Hemisphere using sequence Divergence times between species are directly related to analysis following polymerase chain reaction (PCR) amplifica- changes in the DNA sequence (see pp. 11–18 in Nei, 1987). tion of the intergenic region between the 26S and the 5S rRNA These changes can be determined through direct sequencing genes. Other researchers have used RFLP analysis of PCR- of the DNA or through analysis of RFLPs found by cutting the amplified rDNA to distinguish isolates of the Gaeumannomy- DNA with restriction endonucleases. Researchers conducting ces–Phialophora complex of fungi (Ward and Akrofi, 1994), phylogenetic studies have begun to utilize the variation found strains of Rhizobium (Laguerre et al., 1994), genotypes of in the sequences of the ribosomal RNA genes (rDNA) (for Pleurotus (Bunyard et al., 1996), and species of Morchella review, see Bruns et al., 1991). These variations are used to (Bunyard et al., 1995, 1994). determine relatedness between species, as well as among a The goals of this investigation were to explore novel single species. The use of RFLP analysis (Bunyard et al., 1996, 1994; Cubeta et al., 1991; Hibbett and Vilgalys, 1991; Kohn et methods of genotypic classification of A. bisporus and to al., 1988; Magee et al., 1987) and direct sequencing (White et elucidate the evolutionary history of A. bisporus using al., 1990; Medlin et al., 1988; Woese and Olsen, 1986) to molecular methods of genetic analysis. investigate the genes coding for the production of 16S–18S, 5.8S, 26S–28S, and 5S ribosomal RNA are well established methods for the assessment and comparison of phylogenetic relationships of organisms (Sogin, 1990) over a wide range of MATERIALS AND METHODS taxonomic levels. The rRNA genes of fungi are located on a single chromosome and are present as repeated subunits of a Isolates

1 Abbreviations used: RFLP, restriction fragment length polymor- Fifty-four isolates, comprising 21 putative species of phism; rDNA, ribosomal RNA gene; IGR, intergenic region; LSU, large Agaricus spp. collected from worldwide sources were used subunit ribosomal RNA. in this study. Cultures were grown in liquid potato dextrose

Copyright r 1996 by Academic Press All rights of reproduction in any form reserved. Phylogeny of the Genus Agaricus 245 broth supplemented with yeast extract. Mycelia were Phylogeny Construction filtered and air dried prior to DNA extraction. Isolate numbers from the Pennsylvania State University Mush- Phylogenies were inferred from data collected by RFLP room Culture Collection, species, source, and geographic analysis of the 26S rDNA and the IGR-1/5S rDNA. site are listed in Table 1. PHYLIP, a phylogeny inference package written by J. Felsenstein (1993), was used to determine the amount of variation between isolates and, thus, estimate the evolution- Enzymatic Amplification of Target Area ary divergence between isolates. The degree of genetic Known fungal ribosomal DNA primers found to be divergence between two DNA sequences is correlated conserved among fungal taxa were used to amplify the 38 with the proportion of DNA fragments that they share. and 58 halves of the 26S ribosomal RNA gene and the Within PHYLIP, the MIX program was used to analyze the IGR-1/5S rDNA region using PCR (Mullis and Faloona, discrete character data resulting from the restriction frag- 1987) for the ultimate purpose of RFLP analysis. Primer ment data. MIX finds all the most parsimonious phyloge- sequences (Table 2) were based on the known sequence of netic trees. The CONSENSE program was used to deter- the rDNA repeat from Saccharomyces cerevisiae (Georgiev mine the consensus tree. One thousand bootstrap et al., 1981). No further purification of the amplified replications of the data were analyzed. product was necessary, as a single discreet band was found by electrophoresis following all amplifications. DNA amplified from the 26S and 5S rRNA genes and IGR-1 (Fig. 1) RESULTS was digested using restriction enzymes. Restriction digests were analyzed by gel electrophoresis for the presence of RFLPs between the species of Agaricus. A single 1.4-kb product resulted from PCR amplification Fungal DNA was extracted by the method of Zolan and of genomic DNA using the primers LROR and LR7 for the Pukkila (1986). The PCR was performed using standard 58 end of the 26S rDNA gene (Fig. 1). Likewise, a 1.5-kb conditions and reagents supplied with the Tfl DNA poly- product was amplified for all isolates using the primers merase (Epicentre Technologies, Madison, WI) and cus- ALR7R and LR12 for the 38 end of the 26S rDNA gene tom-made primers (Integrated DNA Technologies, Coral- (Fig. 1). The 5S/IGR-1 product, amplified using primers ville, IA) (final concentration 1 µM each). Products of PCR LR12R and M-1, varied from 1.4 to 1.7 kb, depending on amplification were detected electrophoretically by running the isolate. 4 µl of reaction product through a 1% agarose gel in A number of RFLPs were seen upon digestion of the Tris–acetic acid–EDTA buffer (pH 8.0) at 1.5 V/cm for PCR-amplified DNA using the previously mentioned re- approximately 1 h. Visualization of products was made striction endonucleases. The 58 end of the 26S gene possible by addition of 5 µl ethidium bromide (1 mg/ml) to showed more variability than the 38 end. In fact, many the molten gel and UV illumination following electrophore- restriction digests of the 38 end showed no RFLPs. The sis. most variation among restriction sites was seen with the 5S/IGR-1 product. Using the restriction enzyme MspI, a Restriction Analysis 610-bp band was seen in most isolates, but absent from many others (Fig. 2). In addition, two isolates possessed For restriction digests, an appropriate amount of buffer, unique RFLPs (Fig. 2). With MspI, WC773 had a 260-bp PCR-amplified DNA, and restriction enzyme were com- band, not found in any other isolate. Likewise, a 370-bp bined in a microfuge tube (as specified by the restriction band was unique to WC776. enzyme manufacturer) and incubated (37°C) for 6 h. The Analysis of the 5S/IGR-1 confirmed differences be- restriction enzymes used were HhaI, AluI, HpaII, BstUI, tween isolates that the 26S gene analysis was unable to HaeIII, RsaI, HinfI, HpaI (U.S. Biochemical, Cleveland, resolve. Many individuals were found to be identical based OH), TaqI, and NdeII (Boehringer Mannheim, Indianapo- on RFLP analysis of the 38 and 58 ends of the 26S gene. lis, IN). Restriction fragments were visualized by agarose Isolates of A. bisporus were found to be very similar gel electrophoresis and staining with ethidium bromide (as following analysis of all three regions. Many isolates of A. above). The restriction fragments were scored for each augustus also were found to be nearly identical. isolate by comparing the fragment size with a 123-bp l A phylogenetic tree (Fig. 3) was constructed showing DNA ladder (Gibco BRL, Gaithersburg, MD). the amount of difference among all the isolates examined,

TABLE 1 Culture Numbers for Isolatesa of Agaricus spp. Used in the rDNA RFLP Analysis

Species PSUMCC No. Source/geographic site; collector

A. arvensis Schaeffer:Fries WC145 British Columbia; L. Schisler, 1975 A. arvensis Schaeffer:Fries WC280 Pennsylvania; Fergus, 1979 A. arvensis Schaeffer:Fries WC344 Pennsylvania; L. Schisler A. arvensis Schaeffer:Fries WC345 Pennsylvania; Fergus, 1981 A. augustus Fries WC19 British Columbia; L. Schisler A. augustus Fries WC76 British Columbia; L. Schisler, 1974 A. augustus Fries WC106 British Columbia; L. Schisler, 1974 A. augustus Fries WC129 British Columbia; L. Schisler, 1975 A. augustus Fries WC133 British Columbia; L. Schisler, 1975 A. augustus Fries WC209 British Columbia; L. Schisler, 1975 A. augustus Fries WC210 British Columbia; L. Schisler, 1975 A. augustus Fries WC211 British Columbia; L. Schisler, 1975 A. augustus Fries WC212 British Columbia; L. Schisler, 1975 A. augustus Fries WC213 British Columbia; L. Schisler, 1975 A. augustus Fries WC253 British Columbia; L. Schisler A. augustus Fries WC414 California; ATCC 34767, C. A. Raper Ax1–4, 1981 A. augustus Fries WC415 California; ATCC 34752, C. A. Raper Ax1–3, 1981 A. augustus Fries WC416 California; cross of ATCC 34767 and 34752 A. augustus Fries WC416–26 California; cross of ATCC 34767 and 34752 A. bernardii (Quelet) Saccardo WC772 Maryland; ATCC 52974, J. San Antonio 303, 1993 A. bisporus (Lange) Imbach MC310 MSA culture 1, 1955 A. bisporus (Lange) Imbach MC348 Swayne Spawn Co. 18, 1975 A. bisporus (Lange) Imbach MC404 Ontario; ARP 13, R. W. Kerrigan, 1990 A. bisporus (Lange) Imbach MC406 California; ARP 15, R. W. Kerrigan, 1990 A. bisporus (Lange) Imbach MC428 California; ARP 61, R. W. Kerrigan, 1991 A. bisporus (Lange) Imbach MC429 California; ARP 62, R. W. Kerrigan, 1991 A. bisporus (Lange) Imbach MC430 China; ARP 63, R. W. Kerrigan, 1991 A. bisporus (Lange) Imbach MC436 Pennsylvania; D. J. Royse, 1993 A. bitorquis (Quelet) Saccardo MW14 Formerly A. rodmanii; J. W. Sinden A. bitorquis (Quelet) Saccardo P4 Commercial hybrid; Darlington, 1979 A. campestris Link:Fries MW8 Source unknown; 32, J. W. Sinden A. campestris Link:Fries MW10 Source unknown; 4, J. W. Sinden A. campestris Link:Fries WC11 British Columbia; L. Schisler, 1975 A. campestris Link:Fries WC83 British Columbia, 3, J. Hill, 1984 A. campestris Link:Fries WC136 British Columbia; L. Schisler, 1974 A. campestris Link:Fries WC418 Pennsylvania; L. Schisler A. caroli Pilat WC773 ATCC 60009, J. San Antonio 267, 1993 A. chionodermus Pilat WC774 ATCC 56453, J. San Antonio 251, 1993 A. edulis Vittadini WC775 ATCC 56060, J. San Antonio 242, 1993 A. excellens (Molliard) Molliard WC776 ATCC 56180, J. San Antonio 634, 1993 A. fissuratus (Molliard) Molliard WC777-A Denmark; ATCC 56178, J. San Antonio 612, 1993 A. fissuratus (Molliard) Molliard WC777-B Denmark; ATCC 56178, J. San Antonio 612, 1993 A. hortensis (Cooke) Pilat WC785 Maryland; ATCC 56450, J. San Antonio 221, 1993 A. langei (Molliard) Molliard WC784 ATCC 56103, J. San Antonio 621, 1993 A. macrocarpus (Molliard) Molliard WC778 Denmark; ATCC 56064, J. San Antonio 600, 1993 A. nivescens (Moller) Moller WC779 ATCC 38034, A135, T. J. Elliott, 1993 A. placomyces Peck WC71 British Columbia; 2, L. Schisler, 1974 A. rodmanii Peck WC780 ATCC 56062, J. San Antonio 600, 1993 A. silvaticus Schaeffer ex Secretan sensu Krombholz WC28 British Columbia; L. Schisler, 1975 A. silvicola (Vittadini) Peck WC277 Idaho; 21, L. Schisler, 1979 A. silvicola (Vittadini) Peck WC278 Idaho; 21, L. Schisler, 1979 A. subfloccosus (Lange) Pilat WC781 ATCC 56065, 266, L. Schisler, 1974 A. subrufescens Peck MW17 Source unknown; J. W. Sinden, 1947 A. subrufescens Peck WC782 Taiwan; ATCC 48095, C.-C. Tu, 1993

a Agaricus isolates are maintained as part of the Pennsylvania State University Mushroom Culture Collection (PSUMCC).

TABLE 2 the unrooted tree, constructed for all the data pooled by Primer Name, Primer Sequence, and Location and Direction of Primer species, also clusters most putative species of Agaricus, Extension Used for PCR Amplification of the 26S rRNA Gene (Fig. 1) with A. bisporus being the most distant to A. excellens, A. of Agaricus spp. chionodermus, and A. caroli (Fig. 4). Primer Primer sequence Locationa direction name (58 = 38) of primer extension ALR7R AGATCTTGGTGGTAGTA 1432–1448/ DISCUSSION 25SrDNA = 38end LR0R ACCCGCTGAACTTAAGC 26–42/25S rDNA = 38end LR7 TACTACCACCAAGATCT 1448–1432/25S The current taxonomic status of A. bisporus remains rDNA = 58end LR12 TTCTGACTTAGAGGCGTTCAG 3126–3106/25S uncertain. A. bisporus is placed with the fleshy fungi in the rDNA = 58end order Agaricales. This large group of fungi includes several LR12R CTGAACGCCTCTAAGTCAGAA 3106–3126/25S rDNA = 5S commercially cultivated mushrooms such as Lentinula rDNA edodes, Volvariella volvacea, and Pleurotus ostreatus. The M-1b AACCACAGCACCCAG- 119–97/5S rDNA = nuclear GATTCCC large DNA classification of the Agaricales has posed major taxonomic problems over the years (Alexopoulos and Mims, 1979) a Within 25S rDNA of Saccharomyces cerevisiae (Georgiev et al., with all the members originally placed in a single family, 1981). b Walker and Doolittle, 1982. the Agaricaceae. More recently, Moser (1983) divided the Agaricales into 11 families while Singer (1986) divided the order into 16 families. Similarly, the family Agaricaceae has been modified over the years. Formerly, this family com- using parsimony and 1000 bootstrap replications of the prised a single genus, Agaricus (Alexopoulos and Mims, data. The consensus tree clustered most isolates of Agari- 1979; Smith, 1978). Singer (1986) later divided the family cus closely, suggesting that most species in the genus are into numerous genera, including Agaricus, the type-genus. closely related. Likewise, the data were analyzed using the Members of the genus Agaricus are usually described as computer program RESTSITE (Nei and Miller, 1990) to Basidiomycete fungi having a white or yellow to brown generate distance values (data not shown) based on the fleshy fruiting body, stipe with an annulus, lamellae (gills) different RFLP patterns (by the method of Nei and Li, white, pink, or gray when young; maturing to chocolate 1979). The resulting neighbor-joining tree (Saitou and Nei, brown, and a purple-brown spore print. 1987) supported the finding of close relatedness within The genus Agaricus, sensu Singer (1986) is divided into Agaricus. However, the genus seems to be highly polymor- 42 species grouped into five sections by virtue of exhibiting phic, as several cases are shown where there is greater common subtle characteristics. These traits usually involve intra- (than inter-) species dissimilarity (Fig. 3). Similarly, color change of internal basidiocarp tissue upon exposure

FIG. 1. Primer locations for amplification of the 5S/IGR-1 and 58 and 38 ends of the 26S rRNA gene of Agaricus spp. Complementary strand synthesis in the 58 to 38 direction (shown by arrows).

FIG. 2. Restriction analysis of PCR-amplified 58 end of the 26S rRNA gene of Agaricus spp. digested with MspI. Lanes A1-1, A2-1, B1-1, B1-13, B2-1, and B2-13 are 123 l DNA run as a size marker. Lanes A1-2 through A1-20 include the isolates WC145, MW14, WC213, MW8, WC211, WC83, WC344, WC714, WC71, WC280, WC212, MW17, WC413, WC253, WC345, WC346, WC136, WC133, and MW10. Lanes A2-2 through A2-20 are WC415, WC278, WC19, WC416-26, WC416, WC76, WC418, WC106, WC11, WC277, WC414, WC129, P4, WC28, WC210, WC209, MC348, MC404, and MC310. Lanes B1-2 through B1-12 are WC773, MC428, WC277, MC13, WC774, MC430, WC782, WC781, WC779, WC776, and MC429. Lanes B2-2 through B2-12 are WC784, MC406, WC778, WC780, WC777-A, WC775, WC777-B, MC436, WC772, WC785, and control. Some isolates were not used in this study.

to air, reactions to chemical reagents (KOH, NaOH, days of cultivation of the button mushroom, the scientific name

HNO3, aniline oil, etc.), and morphology of the basidio- for the species has been argued by some and confused by spore. Species in Section 1 typically exhibit a reddening of others. Formerly, the species was called Agaricus campestris, a the pileus (mushroom cap) upon aging, bruising, or expo- species now believed to be only distantly related (Malloch et al., sure to air. Reaction to aniline oil is deep reddish brown on 1987). Isaacs (1967) decided the cultivated mushroom was the the surface of the pileus. Members of Section 2 are similar same species as that of the holotype specimen of Agaricus physiologically to those of Section 1, but typically are found brunnescens, described originally in 1900 by Peck. Malloch growing within wooded locations rather than open areas, as (1976) agreed with Isaacs’ earlier conclusions and argued for A. in Section 1 species. Species of Section 3 typically exhibit brunnescens to be the proper name for the cultivated species. yellowing of the inner tissues when bruised. A yellow color Many researchers did not concur, and Singer (1984) went so far is also given upon reaction to NaOH or KOH. A red as to present evidence that A. bisporus and A. brunnescens are discoloration is seen by reacting with aniline oil plus two separate species! The use of nuclear and mitochondrial

HNO3. Members of Section 4 also yellow upon bruising, DNA RFLPs once again confirmed the conspecificity of A. but exhibit negative reactions when tested with aniline oil bisporus and A. brunnescens (Malloch et al., 1987). In fact, and HNO3. Members of Section 5 exhibit a scaly or woolly Malloch et al. (1987) found the amount of variation between pileus. While members of Section 5 are not considered the ‘‘two’’ was no greater than the amount of variation among edible, most other species of Agaricus are highly sought any two strains of cultivated A. bisporus. Additionally, A. from the wild or are commercially cultivated. Adding to the bisporus has been collected in the wild (Kerrigan et al., 1995; taxonomic confusion are other schemes for classification of Kerrigan and Ross, 1989). So, it seems the debate over the Agaricus (Moser, 1983). proper name may have been resolved in favor of A. brunne- The evolutionary history and taxonomy of A. bisporus has scens, though A. bisporus is likely to remain in use as it has been been in a state of flux historically. The lack of phylogenetic the accepted name in the literature for so long (Kerrigan, information may have been obscured, at times, by the debate 1987). surrounding the proper name of the species. Since the early Additionally, the evolutionary position of A. bisporus within

FIG. 3. Consensus tree of all Agaricus spp. isolates used in this study. The phylogenetic tree is based on 1000 bootstrap replications of data resulting from RFLP analysis of the 38 and 58 halves of the 26S rRNA gene and the IGR-1/5S rDNA region. The phylogenetic tree was constructed using parsimony in the package PHYLIP (Felsenstein, 1993). the Agaricales remains uncertain (Singer, 1986). Kerrigan The phylogenetic tree (Fig. 3) suggests the genus (1990) has addressed this issue in great detail but many Agaricus comprises many closely related species. Further- questions still exist concerning the evolutionary origin of A. more, several cases are shown where there is greater intra- bisporus, the relationship of cultivars to wild populations of this than interspecies dissimilarity. There are several possible species, and its position within the Agaricales. reasons for this. One explanation is that two putatively

Differences in geographic origin have been suggested in other cases where more variation was found between geographically isolated populations of a fungal species than between two distinct species (Bunyard et al., 1995, 1994; Jung et al., 1993; Yoon et al., 1990). Finally, polymorphisms and DNA heterogeneity implying high intraspecies diversity actually may predate the taxa. Slight (17% of the characters) length variations were found in one of the rDNA regions amplified (the 5S/IGR-1 product). While such length variations can increase the distance values, the overall tree topology should not be affected (S. Kumar, personal communication). The models used here for tree building are valid when DNA differences do not exceed 30% (Nei, 1987). Indeed, other authors have found that similar length variation (15%) does not violate assumptions of models for RFLP analysis (Avise et al., 1992, 1987; Bermingham and Avise, 1986). The most variation was seen with the 5S/IGR-1 product. For example, distance data (not shown) from this region alone infer a divergence of 16.5% between A. caroli and all the other species of Agaricus, including A. langei and A. silvaticus. These results were unexpected as A. caroli formerly has been thought closely related to A. langei and A. silvaticus (Singer, 1986; Moser, 1983). FIG. 4. Evolutionary positions of Agaricus species determined using Also of interest are isolates MC406 and MC428. Al- RFLP analysis of the 38 and 58 halves of the 26S rRNA gene and the though these isolates exhibit four-spored basidia, MC406 IGR-1/5S rDNA region. The phylogenetic tree is based on the data from and MC428 have been identified as A. bisporus var. Fig. 1, pooled by species using the computer program RESTSITE (Nei burnettii (Callac et al., 1993). Our data, in support of the and Miller, 1990). The phylogenetic tree was constructed using parsimony taxonomic findings of Callac et al. (1993), found these two in the package PHYLIP (Felsenstein, 1993). isolates most closely group with other isolates of A. bisporus. Furthermore, MC406 and MC428 are reported to cross with other (two-spored) isolates of A. bisporus different species may, in fact, be one species. That the (Kerrigan et al., 1994). This supports our belief that the specific names have been in a constant state of flux (as taxonomic scheme of the genus Agaricus, based solely on pointed out above) demonstrates the morphological simi- morphological characters, may be an artificial one. larities between some ‘‘species.’’ Additionally, cases may Our data indicate that previous taxonomic schemes, exist where isolates were originally misidentified, although based on morphological characters, is in need of revision. analyses should be based on type specimens. It is also In particular, delineation between members of Section 1 possible that different geographic locations or differences and Section 2, sensu Singer (1986), may be artificial, as no in environmental conditions can result in polymorphic strict conformation to these two groups was seen (Fig. 4). appearances in nature. The possibility of environmental As stated above, Singer (1986) separated these members variation has led some authors to utilize DNA analysis into two sections based on occurrence in nature. Likewise, (Bunyard et al., 1995, 1994), immunological relationships members of Section 3 were not all part of a major branch (Jung et al., 1993), and allelic variation (Yoon et al., 1990) on the consensus tree. Additionally, our data suggest that to confirm taxonomic placement based on morphological species previously considered close relatives (e.g., A. characters. Zervakis et al. (1994) found much of the caroli, A. langei, and A. silvaticus) may not be that closely variability of Pleurotus spp. was attributed to geographic related (Fig. 4). The similar physiologies exhibited by these origin. Recently, Bunyard et al. (1996) also found a great distantly related species may be resultant of convergent deal of intra- and interspecific variability within Pleurotus. evolution which could lead to unrelated organisms having

Copyright r 1996 by Academic Press All rights of reproduction in any form reserved. Phylogeny of the Genus Agaricus 251 similar morphological characteristics. Alternatively, other near monoculture and a much reduced genetic heterogene- modes of speciation (e.g., allopatry, parapatry, or sympatry) ity among lines of the species (Royse and May, 1982a). In that result from populations becoming geographically or contrast, wild populations of A. campestris have consider- reproductively isolated might result in morphologically able genetic variability (May and Royse, 1982b). Because similar species. of the lack of variability in A. bisporus, future breeding of Members of Section 2 were scattered on disparate new lines with improved traits will necessitate the location branches of the phylogenetic tree (Fig. 3). The consensus and use of new genetic resources from wild populations tree places A. silvaticus with a large group of members (Anderson, 1993). As the sources for these populations from Section 1. Likewise, A. langei did not group with become more scarce, it is urgent that efforts are directed members of Section 2. However, nearly all members of toward the identification of remaining genetic resources Section 3 were found to group together. All isolates but one before they disappear (Anderson, 1993). Furthermore, of A. arvensis grouped on a major branch with A. silvicola new methods for rapid and reliable genetic analysis of and nearly all isolates of A. augustus (Fig. 3). Interestingly, populations must be developed to expedite the identifica- four isolates of A. augustus were found to be identical and tion of novel, wild genes. One of the best known sources grouped with a single isolate of A. arvensis (WC344). This for new genetic resources of A. bisporus may come from suggests that WC344 may have been misidentified. Over- the western part of the United States, specifically, Califor- all, little difference was seen between the two subgroups of nia (Kerrigan et al., 1993a,b). Kerrigan et al. (1993b) Section 3. While the phylogenetic tree (Fig. 4) presented examined populations composed of wild A. bisporus iso- here does not entirely support the classification scheme of lates, as well as commercial escapes. Kerrigan et al. Singer (1986), much less conformation to that of Moser (1993b) provided much new and important information, (1983) was shown. including evidence of hybridization in nature between wild and commercial isolates. Recently, new isolates of wild A. The data presented here are the first comprehensive bisporus have been found in Canada and the Pacific attempt at systematically resolving the entire genus Agari- Northwest (Kerrigan et al., 1995). Efforts (e.g., Agaricus cus using RFLP analysis of PCR-amplified ribosomal DNA Recovery Program; Kerrigan et al., 1993a) to locate and sequences. The data presented may be less ambiguous preserve wild A. bisporus germ plasm need to be in- than those resulting from traditional morphological exami- creased. nation. However, these data point to the need for further investigation of this taxonomic group. Our study may indicate shortcomings with the current classification scheme of Agaricus, based solely on morphological characters. ACKNOWLEDGMENTS However, we do point out that the current study provides information that should be regarded in addition to the We thank V. Wilkinson for assistance with culture maintenance. We also accepted taxonomic scheme. We realize that our study is greatly appreciate the technical comments of S. Kumar and J. Felsen- based on a limited number of characters and would benefit stein. The authors also thank B. J. Christ and C. P. Romaine for review by the addition of genetic information from other regions and helpful discussion of the manuscript. We thank Gary Smythers of the of the genome. RFLP analysis can be used as a guide to the Frederick Biomedical Supercomputing Center for providing assistance amount of difference among the species of Agaricus. with the computer analyses. Direct sequencing of the genomic regions used above, though more time-consuming, would provide even more genetic information gleened from the genus. It is of equal REFERENCES importance that other regions within the genome are analyzed. Genetic information about A. bisporus undoubt- edly will be of taxonomic importance, as well as of Abler, B. S. B., Edwards, M. D., and Stuber, C. W. 1991. Isoenzyme commercial use. identification of quantitative trait loci in crosses of elite maize inbreds. Before the advent of commercially produced pure Crop Sci. 31: 267–274. Alexopoulos, C. J., and Mims, C. W. 1979. Introductory Mycology. Wiley, culture spawn, growers often collected and maintained New York. their own mycelial cultures for mushroom production. Anderson, J. B. 1993. Molecular tools in breeding Agaricus. In Genetics Therefore, some genetic variability was maintained over and Breeding of Edible Mushrooms (S. T. Chang, J. A. Buswell, and the years. However, commercial spawn use has led to a P. G. Miles, Eds.), Gordon & Breach, Philadelphia.

Anderson, J. B., and Stasovski, E. 1992. Molecular phylogeny of northern Hibbett, D. S., and Vilgalys, R. 1991. Evolutionary relationships of hemisphere species of Armillaria. Mycologia 84: 505–516. Lentinus to the polyporaceae: Evidence from restriction analysis of Avise, J. C., Bowen, B. W., Lamb, T., Meylan, A. B., and Bermingham, E. enzymatically amplified ribosomal DNA. Mycologia 83: 425–439. 1992. Mitochodrial DNA evolution at a turtle’s pace: Evidence for low Isaacs, B. F. 1967. Studies in the genus Agaricus. I. Agaricus cratecellus genetic variability and reduced microevolutionary rate in the Testu- and its relationships. Mich. Bot. 6: 3–12. dines. Mol. Biol. Evol. 9: 457–473. Jiri, H. 1967. Zytologische studien u¨ ber die gattung Agaricus. Mushroom Avise, J. C., Reeb, C. A., and Saunders, N. C. 1987. Geographic Sci. 6: 77–81. population structure and species differences in mitochondrial DNA of Johansen, S., Johansen, T., and Haugli, F. 1992. Extrachromosomal mouthbrooding marine catfishes (Ariidae) and demersal spawning ribosomal DNA of Didymium iridis: Sequence analysis of the large toadfishes (Batrachoididae). Evolution 41: 991–1002. subunit ribosomal RNA gene and subtelomeric region. Curr. Genet. Bermingham, E., and Avise, J. C. 1986. Molecular zoogeography of 22: 305–312. freshwater fishes in the southeastern United States. Genetics 113: Jung, S. W., Gessner, R. V., Keudell, K. C., and Romano, M. A. 1993. 939–965. Systematics of Morchella esculenta complex using enzyme-linked Bruns, T. D., White, T. J., and Taylor, J. W. 1991. Fungal molecular immunosorbent assay. Mycologia 85: 677–684. systematics. Annu. Rev. Ecol. Syst. 22: 525–564. Kerrigan, R. W. 1987. What’s in a name?: The Chaetaceous case of the Bunyard, B. A., Chaichuchote, S., Nicholson, M. S., and Royse, D. J. chaste champignon. In Cultivating Edible Fungi (P. J. Wuest, D. J. 1996. Ribosomal DNA analysis for resolution of genotypic classes of Royse, and R. B. Beelman, Eds.), pp. 141–154. Elsevier, Amsterdam. Pleurotus. Mycol. Res. 100: 143–150. Kerrigan, R. W. 1990. Evidence of genetic divergence in two populations Bunyard, B. A., Nicholson, M. S., and Royse, D. J. 1994. A systematic of Agaricus bisporus. Mycol. Res. 94: 721–733. Morchella assessment of using RFLP analysis of the 28S ribosomal Kerrigan, R. W., and Ross, I. K. 1989. Allozymes of a wild Agaricus RNA gene. Mycologia 86: 762–772. bisporus population: New alleles, new genotypes. Mycologia 81: Bunyard, B. A., Nicholson, M. S., and Royse, D. J. 1995. Phylogenetic 433–443. resolution of Morchella, Verpa, and Disciotis [Pezizales: Morchella- Kerrigan, R. W., Carvalho, B. B., Horgen, P. A., and Anderson, J. B. 1995. ceae] based on restriction enzyme analysis of the 28S ribosomal RNA Indigenous and introduced populations of Agaricus bisporus, the gene. Exp. Mycol. 19: 223–233. cultivated button mushroom, in eastern and western Canada: Implica- Callac, P., Billette, C., Imbernon, M., and Kerrigan, R. W. 1993. tions for population biology, resource management, and conservation Morphological, genetic, and interfertility analyses reveal a novel, of genetic diversity. Can. J. Bot. 73: 1925–1938. tetrasporic variety of Agaricus bisporus from the Sonoran Desert of Kerrigan, R. W., Horgen, P. A., and Anderson, J. B. 1991. A genetic California. Mycologia 85: 835–851. linkage map for Agaricus bisporus. In Proceedings of the First Cassidy, J. R., Moore, D., Luc, B. C., and Pukkila, P. J. 1984. Unusual International Seminar on Mushroom Science. Mushroom Experimen- organization and lack of recombination in the ribosomal RNA genes of tal Station, Horst, the Netherlands, May 14–17, 1991. Coprinus cinereus. Curr. Genet. 8: 607–613. Kerrigan, R. W., Royer, J. C. Baller, L. M., Kohli, Y., Horgen, P. A., and Castle, A. J., Horgen, P. A., and Anderson, J. B. 1987. Restriction Anderson, J. B. 1993a. Meiotic behavior and linkage relationships in fragment length polymorphisms in the mushrooms Agaricus brunne- the secondarily homothallic fungus Agaricus bisporus. Genetics 133: scens and Agaricus bitorquis. Appl. Environ. Microbiol. 53: 816–822. 225–236. Castle, A. J., Horgen, P. A., and Anderson, J. B. 1988. Crosses among homokaryons from commercial and wild-collected strains of the Kerrigan, R. W., Horgen, P. A., and Anderson, J. B. 1993b. The California mushroom Agaricus brunnescens (5A. bisporus). Appl. Environ. population of Agaricus bisporus comprises at least two ancestral Microbiol. 54: 1643–1648. elements. Syst. Bot. 18: 123–136. Cubeta, M. A., Echandi, E., Abernathy, T., and Vilgalys, R. 1991. Kerrigan, R. W., Imbernon, M., Callac, P., Billette, C., and Olivier, J.-M. Characterization of anastomosis groups of binucleate Rhizoctonia 1994. The heterothallic life cycle of Agaricus bisporus var. burnettii and species using restriction analysis of an amplified ribosomal RNA gene. the inheritance of its tetrasporic trait. Exp. Mycol. 18: 193–210. Phytopathology 81: 1395–1400. Kligman, A. M. 1943. Some cultural and genetic problems in the Elliott, T. J. 1972. Sex and the single spore. Mushroom Sci. 8: 11–18. cultivation of the mushroom, Agaricus campestris. Am. J. Bot. 30: Evans, H. J. 1959. Nuclear behaviour in the cultivated mushroom. 745–763. Chromosoma 10: 115–135. Kohn, L. M., Petsche, D. M., Bailey, S. R., Novak, L. A., and Anderson, J. Felsenstein, J. 1993. PHYLIP: Phylogeny Inference Package User’s B. 1988. Restriction fragment length polymorphisms in nuclear and Guide. Frederick Biomedical Supercomputing Center, Frederick, mitochondrial DNA of Sclerotinia species. Phytopathology 78: 1047– MD. 1051. Fritsche, G. 1981. Some remarks on the breeding, maintenance of strains Laguerre, G., Allard, M.-R., Revoy, F., and Amarger, N. 1994. Rapid and spawn of Agaricus bisporus and A. bitorquis. Mushroom Sci. 11: identification of rhizobia by restriction fragment length polymorphism 367–385. analysis of PCR-amplified 16S rRNA genes. Appl. Environ. Microbiol. Georgiev, O. I., Nikolaev, N., Hadjiolov, A. A., Sdryabin, K. G., Zakharyev, 60: 56–63. V. M., and Bayev, A. A. 1981. The structure of the yeast ribosomal RNA Malloch, D. 1976. Agaricus brunnescens: The cultivated mushroom. genes. 4. Complete sequence of the 25 S rRNA gene from Saccharomy- Mycologia 68: 910–919. ces cerevisiae. Nucleic Acids Res. 9: 6953–6958. Malloch, D., Castle, A., and Hintz, W. 1987. Further evidence for

Agaricus brunnescens Peck as the preferred name for the cultivated genetic analysis of diversity in populations of edible mushrooms. In Agaricus bisporus. Mycologia 79: 839–846. Mushroom Biology and Mushroom Products (S. Chang, J. A. Buswell, Magee, B. B., D’Souza, T. M., and Magee, P. T. 1987. Strain and species and S. Chiu, Eds.). Chinese Univ. Press, Shatin, Hong Kong. identification by restriction fragment length polymorphisms in the Royse, D. J., Spear, M. C., and May, B. 1983. Single and joint segregation ribosomal DNA repeat of Candida species. J. Bacteriol. 169: 1639– of marker loci in the shiitake mushroom, Lentinus edodes. J. Gen. Appl. 1643. Microbiol. 29: 217–222. Martin, F. N. 1990. Variation in the ribosomal DNA repeat unit within Saitou, N., and Nei, M. 1987. The neighbor-joining method: A new single-oospore isolates of the genus Pythium. Genome 33: 585–591. method for reconstructing phylogenetic trees. Mol. Biol. Evol. 4: May, B., and Royse, D. J. 1982. Genetic variation and joint segregation of 406–425. biochemical loci in the common meadow mushroom, Agaricus Campes- Singer, R. 1984. Agaricus brunnescens Peck and Agaricus bisporus tris. Biochem. Genet. 20: 1165–1173. (Lange) Imbach. Mycotaxon 20: 479–482. Medlin, L., Elwood, H. J., Stickel, S., and Sogin, M. L. 1988. The Singer, R. 1986. In The Agaricales in Modern Taxonomy, pp. 483–488. characterization of enzymatically amplified eukaryotic 16S-like rRNA- Koeltz, Koenigstein, Germany. coding regions. Gene 71: 491–499. Smith, A. H. 1978. Morphology and classification. In The Biology and Moser, M. 1983. In Keys to Agarics and Boleti, pp. 230–238. English Cultivation of Edible Mushrooms (S. T. Chang and W. A. Hayes, Eds.), translation from the 4th ed., Phillips, London. pp. 3–33. Academic Press, New York. Mullis, K. B., and Faloona, F. A. 1987. Specific synthesis of DNA in vitro Sogin, M. L. 1990. Amplification of ribosomal RNA genes for molecular via a polymerase-catalyzed chain reaction. Methods Enzymol. 155: evolution studies. In PCR Protocols: A Guide to Methods and Applica- 335–350. tions (M. A. Innis, D. H. Gelfand, J. J. Sninsky, and T. J. White, Eds.), Nei, M. 1987. In Molecular Evolutionary Genetics, pp. 64–107. Columbia pp. 307–314. Academic Press, New York. Univ. Press, New York. Spear, M. C., Royse, D. J., and May, B. 1983. Atypical meiosis and joint Nei, M., and Li, W. H. 1979. Mathematical model for studying genetic segregation of biochemical loci in Agaricus brunnescens. J. Hered. 74: variation in terms of restriction endonucleases. Proc. Natl. Acad. Sci. 417–420. USA 76: 5269–5273. Summerbell, R. C., Castle, A. J., Horgen, P. A., and Anderson, J. B. 1989. Nei, M., and Miller, J. C. 1990. A simple method for estimating average Inheritance of restriction fragment length polymorphisms in Agaricus numbers of nucleotide substitutions within and between populations brunnescens. Genetics 123: 293–300. from restriction data. Genetics 125: 873–879. Vilgalys, R., and Gonzalez, D. 1990. Organization of ribosomal DNA in Pelham, J. 1967. Techniques for mushroom genetics. Mushroom Sci. 6: the Basidiomycete Thanatephorus praticola. Curr. Genet. 18: 277–280. 49–64. Walker, W. F., and Doolittle, W. F. 1982. Redividing the basidiomycetes Petes, T. D., and Botstein, D. 1977. Simple Mendelian inheritance of on the basis of 5S rRNA sequences. Nature 299: 723–724. reiterated ribosomal DNA of yeast. Proc. Natl. Acad. Sci. USA 74: Ward, E., and Akrofi, A. Y. 1994. Identification of fungi in the Gaeuman- 5091–5095. nomyces–Phialophora complex by RFLPs of PCR-amplified ribosomal Raeder, U., and Broda, P. 1984. Comparison of lignin degrading white rot DNAs. Mycol. Res. 98: 219–224. fungi Phanerochaete chrysosporium and Sporotrichum pulverulentum White, T. J., Bruns, T., Lee, S., and Taylor, J. 1990. Amplification and at the DNA level. Curr. Genet. 8: 499–506. direct sequencing of fungal ribosomal RNA genes for phylogenetics. In Raper, C. A., Raper, J. R., and Miller, R. E. 1972. Genetic analysis of the PCR Protocols: A Guide to Methods and Applications (M. A. Innis, life cycle of Agaricus bisporus. Mycologia 64: 1088–1117. D. H. Gelfand, J. J. Sninsky, and T. J. White, Eds.), pp. 315–322. Royer, J. C., and Horgen, P. A. 1991. Towards a transformation system for Academic Press, New York. Agaricus bisporus. In Genetics and Breeding of Agaricus (L. J. L. D. Woese, C. R., and Olsen, G. J. 1986. Archaebacterial phylogeny: van Griensven, Ed.), pp. 129–134. Pudoe, Wageningen. Perspectives on the urkingdoms. Syst. Appl. Microbiol. 7: 161–177. Royse, D. J., and May, B. 1982a. Use of isozyme variation to identify Yoon, C., Gessner, R. V., and Romano, M. A. 1990. Population genetics genotypic classes of Agaricus brunnescens. Mycologia 74: 93–102. and systematics of the Morchella esculenta complex. Mycologia 82: Royse, D. J., and May, B. 1982b. Genetic relatedness and its application in 227–235. selective breeding of Agaricus brunnescens. Mycologia 74: 569–575. Zervakis, G., Sourdis, J., and Balis, C. 1994. Genetic variability and Royse, D. J., and May, B. 1990. Interspecific allozyme variation among systematics of eleven Pleurotus ostreatus species based on isozyme Morchella spp. and its inferences for systematics within the genus. analysis. Mycol. Res. 98: 329–341. Biochem. Syst. Ecol. 18: 475–479. Zolan, M. E., and Pukkila, P. J. 1986. Inheritance of DNA methylation in Royse, D. J., Bunyard, B. A., and Nicholson, M. S. 1993. Molecular Coprinus cinereus. Mol. Cell Biol. 6: 195–200.