MYCOTAXON Volume 94, pp. 137-147 October-December 2005

Molecular evidence for the taxonomic status of Metarhizium taii and its teleomorph, taii (, C/avicipitaceae)

Bo HUANG" CHUNRU Li" RICHARD A. HUMBER" KATHIE T. HODGE3, MEIZHEN FAN' & ZENGZHI L('

zzliepahau. edu. cn i Anhui Provincial Key Laboratory for Microbial Pest Control, Anhui Agricultural University, Hefei 230036, China 2 USDA Plant, Soil & Nutrition Laboratory Tower Road, Ithaca, NY 14853, USA 3 Department of Plant Pathology, Cornell University Ithaca, NY 14853, USA Abstract-ITSl-5.8S-ITS2 rDNA was amplified, cloned and sequenced from the stroma of Cordyceps taii, mycelium of Metarhizium taii and cultures of M. anisopliae var. anisopliae from China. Analysis of integrated data of different strains and varieties of M. anisopliae and M. flavoviride from GenBank showed that Metarhizium taii should be treated as a synonym of M. anisopliae var. anisopliae. The ITSl-5.8S-ITS2 rDNA sequences of C. taii and M. taii are identical, indicating that the teleomorph of M. anisopliae var. anisopliae is Cordyceps taii. Key words-molecular systernatics, anamorph-teleomorph connection, synonymy Cordyceps brittlebankisoides

Introduction Concern about adverse effects of chemical insecticides in the environment and the growing problem of insecticide resistance has led to considerable interest in the use of biological agents for control of insect pests (Leal et al. 1997). Many species of entomogenous fungi play important roles in pest management in many countries. var. anisopliae (abbreviation: M.a. anisopliae) is one of the most important entomogenous fungi, and has been used to control pests in America, Brazil, Australia and China, with distinct economic and social benefits. Green muscardine disease, one of the most common fungal diseases of insects, is caused by several taxa in the c1avicipitaceous anamorphic genus Metarhizium. Since this genus was established, several new species or varieties have been described based on seemingly minor morphological differences from M. anisopliae, the tye of the genus/These similarities have led. some to believe that the of this genus was

* Author for correspondence 138

badly confused (Milner et al. 1994). Rombach et aL (1986) recognized three species: M. anisopliae (Metschn.) Sorokin, M. jlavoviride W. Gams & Rozsypal, and M. album readil) Petch. Tulloch (1976) segregated the tye species of the genus into two varieties, M. a. analysi anisopliae and M. anisopliae var. majus (J.R. Johnst.) M.C. Tulloch (abbreviation: M.a. Du majus) primarily based on conidial length, and she accepted M. jlavoviride. Rombach collect et aL (I986) segregated M. jlavoviride into the varieties M. jlavoviride var. jlavoviride isolatei origin¡ (abbreviation: Mj jlovoviride) and, by virtue of its smaller conidia, M. jlavoviride var. minus Rombach et aL (abbreviation: Mj minus). Rath et aL (1995) suggested (Liang was to (but did not validly publish) the 'variety' M. anisopliae var. frigidum based on the abilty of conidia to germinate at low temperatures and on patterns of hydrocarbon clarify utilzation. The isolates assigned to this variety were found by Driver et aL (2000) to other 1 be taxonomically heterogeneous and were characterized based on ITS sequences as belonging in two clades of M. jlavoviride. Guo et aL (I986) described three new species, M. pingshaense Q.T. Chen & H.L. Guo, M. guizhouense Q.T. Chen & H.L. Guo and M. Fungal cylindrosporae Q.T. Chen & H.L. Guo. Metarhizium biformisporae A.Y. Liu et aL (I989) noctuic should be treated as a heterotyic (facultative) synonym of M. cylindrosporae. Liu et aL China. (1989) described two tyes of conidia being produced by M. cylindrosporae - an abilty ascospc not noticed by Guo et aL (1986) - and it appears that the description of M. biformisporae isolate( is best treated as an amplified description of M. cylindrosporae. In any event, the latter Those i species was transferred into the genus , another c1avicipitaceous anamorphic from G genus, by Tzean et al. (1993). Liang et aL (1991) named M. taii isolated from the new DNA¡: species, Cordyceps taii Z.Q. Liang & A.Y. Liu, and postulated a connection between C. specim taii and M. taii based on micro cyclic conidiation. Liu et aL (2001) collected a single were ea specimen and described it as a new species, Cordyceps brittlebankisoides Zuo, Y. Liu et ofbenz aI., whose anamorph (confirmed by the identical ITS sequences from the teleomorphic stored i stroma and in vitro culture) was identified as M. a. majus. solutio! Recently developed molecular techniques, and particularly DNA sequencing, provide TIie valuable tools for studying phylogenetic relationships and clarifying the taxonomic by Mav position of confusing species. Molecular techniques have been used by several authors the folli to study genetic variation and taxonomy of Metarhizium. Liu et aL (1994) determined gelatin, partial sequences for selected regions from small (18S) and large (28S) subunit rRNAs of 2.5 unit species of Metarhizium. Curran et aL (1994) sequenced PCR products from ITSI-5.8S- 500 ~ili ITS2 rDNA regions of31 strains ofMetarhizium, and found 19 kinds ofITSl-5.8S-ITS2 cycler ( rDNA regions. Rakotonirainy et aL (1994) sequenced the D1 and D2 expansion region denatur of 28S rDNA from 42 strains of Metarhizium. The above results collectively confirmed anneali: (i) the monophyly of the genus Metarhizium; (ii) the separation of M. anisopliae and 72 0 C fc M. jlavoviride; and (iii) the need to subdivide M. anisopliae into several varieties or product species. ethidiui Driver et aL (2000) partially resolved the taxonomy of Metarhizium using sequence Am! data and RAPD patterns; ITS sequence data were recognized as the most important Purifica taxonomic criterion. Based on distinctive ITS sequences, M. jlavoviride and M. anisopliae DNj were recognized to include five and four clades, respectively. In addition to recognizing were cle M. album as a distinct species, the classification of Driver et al. (2000) divided M. used to anisopliae into four varieties and M. jlavoviride into four named varieties plus one other an alkal clade nO;Jreated as taxonomically distinct. The four species of Metarhizium described TIie: from China (Guo et al. 1986, Liang et aL 1991) have apparently not been deposited in the didi 139

:cognized three species: readily accessible culture collections and, therefore, were not included in the molecular ~ozsypal, and M. album analyses by Driver et aL (2000). into two varieties, M. a. During investigation of the Cordyceps resources in China, a C. taU specimen was loch (abbreviation: M.a. collected in the Dabeishan Mountains, and its presumptive anamorph, M. taii, was isolated from the specimen. The relationship between Cordyceps and Metarhizium was \.. jlavoviride. Rombach cyclic conidiogenesis and ascospores ivoviride var. jlavoviride originally determined by comparison of micro conidia, M. jlavoviride (Liang et aL 1991), but this important result needed more proof. The aim of this study et aL (1995) suggested was to providegenetic verification of the connection between C. taU and M. taii and to clarify the taxonomic status of M. taii by comparison of its ITS sequences with those of frigidum based on the Jatterns of hydrocarbon other Metarhizium isolates already included in GenBank. y Driver et aL (2000) to Materials and methods ed on ITS sequences as :ribed three new species, Fungal isolates-A specimen of M. taii (HS0006050l) occurring as a pathogen of a :hen & H.L. Guo and M. noctuid (lepidopteran) larva was collected from Huoshan County, Western Anhui, lTae A.Y. Liu et al. (1989) China. A culture originally identified as M. taii (RCEF0772) was isolated from its cylindrosporae. Liu et aL ascospores discharged onto glass slides. A culture of M.a. anisopliae (RCEF0386) was ylindrosporae - an ability isolated from a noctuid larval cadaver collected in Yux County, Western Anhui, China. ption of M. biformisporae Those isolates that are included in this study only by means of sequence data obtained e. In any event, the latter from GenBank are listed in Table 1. 'icipitaceous anamorphic DNA preparation and amplification-A portion of a stroma from the field-collected 1ii isolated from the new co spores) specimen and from cultured mycelia of M. taii (isolated from discharged as a connection between C. were each finely ground under liquid nitrogen, and genomic DNA was extracted by use (200 I) collected a single of benzyl chloride according to the method ofZhu et al. (1994). The extracted DNA was Jankisoides Zuo, Y. Liu et stored in 100 i-I TE buffer (10 mM Tris-HCI, pH 8.0; 1 mM EDTA) at 4"C, and DNA es from the teleomorphic solutions were diluted 10-fold with TE for use in PCR reaction. The nuclear ITSl-5.8S-ITS2 regions were amplified with the primer pair described DNA sequencing, provide clarifyng the taxonomic by Mavridou & Typas (1998). The PCR reactions were performed in 50 i-l volume with n used by several authors the following components: 10 mM Tris-HCI (pH 8.3), 50 mM KCI, 2 mM MgCI2, 0.001 % gelatin, 200 i-M of each dNTP, 20-100 ng genomic DNA, 12 pmole of each primer, and Let aL (1994) determined 2.5 units Taq DNA polymerase (Sangon, China). The reactions were prepared on ice in ge (28S) subunit rRNAs of products from ITSl-5.8S- 500 i-I miClocentrifuge tubes, overlaid with 20 i-l mineral oil, and placed in a thermal cycler (Techine, UK). Cycling parameters were programmed as following: an initial 9 kinds ofITSl-5.8S-ITS2 denaturation at 95" C for 3min, followed by 35 cycles of denaturation at 94" for I min, and D2 expansion region ¡Its collectively confirmed annealing at 54" C for 1 min and extension at 72" C for 2 min, with a final extension at .lion of M. anisopliae and 72'C for 10 min. Effciency of amplification was monitored by running 8 i-I of each product through 1.2 % agarose gels using TAE buffer and visualizing PCR products with e into several varieties or ethidium bromide. Amplification products were purified by use of a Wizard™ PCR Preps DNA :tarhizium using sequence ed as the most important Purification System Kit (Promega Co., France) and methods provided by the company. DNA cloning and sequencing of amplified its region-Purified ITS PCR products ivoviride and M. anisopliae In addition to recognizing were cloned into T-Vectors, which we prepared from SK plasmids. These vectors were . et al. (2000) divided M. used to transform Escherichia coli XLI-Blue, and the plasmid DNA was extracted using an alkaline lysis method (Sambrook et aL 1989). Led varieties plus one other " The primers T7/T3 (forward and reverse) were used to sequence both strands using of Metarhizium described / the dideoxy-nucleotide chain termination on an ABI 3700 automated sequencer at ntly not been deposited in 140

Shanghai Genecore Biotechnologies Company. Sequence data of C. taii (HS00060501), M. taii (RCEF0772) and M.a. anisopliae (RCEF0386) have been deposited in the Hu GenBank genome sequence database with accession numbers AF348393, AF348394 ger and AF348395, respectively. (bi tre Table 1. Previously deposited GenBank sequences used in this study. *

GenBank Host insect / order Country Fl# Fungal taxon of origin sequence (or other source substrate) Ch AF134150 1027 M.a. anisopliae Oxya multidentata I Orthopt. Pakstan C01 AF135210 1029 M.a. anisopliae Schistocerca gregaria I Orthopt. Eritrea are AF135211 1031 M.a. anisopliae Teleogryllus commodus I Orthopt. Australia mi AF135212 1034 M.a. anisopliae Patanga succineta I Orthopt. Thailand 231 AF135213 1045 M.a. anisopliae Dermolepida albohirtum I Coleopt. Australia AF135214 1091 M.a. anisopliae Teleogryllus commodus I Orthopt. Australia Pa: AF135215 1114 M.a. anisopliae Soil, Maquarie Island Australia AF135216 1156 M.a. anisopliae (immunocompromised human) Australia AF136375 163 M.a. anisopliae Kalotermes sp. I Isopt. Australia AF136376 203 M.a. anisopliae Inopus rubriceps I Dipt. Australia AF136928 208 M.a. anisopliae Phaulacridium vittatum I Orthopt. Australia AF137054 114 M.a. anisopliae Antitrogus parvulus I Coleopt. Australia AF137055 23 M.a. anisopliae Aenolamia albofasciata I Hemipt. Mexico New AF137056 700 M.a. anisopliae Costelytra zealandica I Coleopt. Zealand AF137057 328 M.a. anisopliae Inopus rubriceps I Dipt. Australia AF137058 379 M.a. anisopliae Inopus rubriceps I Dipt. Australia Soil (near mound of Coptotermes AF137059 775 M.a. anisopliae Australia lacteus I Isoptera) AF137062 987 M.a. acridum Ornithacris cavroisi I Orthopt. Niger AF137065 147 M.a. lepidiotae Lepidiota consobrina I Coleopt. Australia AF137061 389 M.a. majus Oryetes rhinoceros I Coleopt. Indonesia AF138270 38 Mf flavoviride Adoryphorus couloni I Coleopt. Australia AF138272 1172 Mf minus Niliparvata lugens I Homopt. Philippines M.f novo- AF139853 1125 Soil Australia zealandicum AF139850 72 Mf pemphigi Pemphigus treherni I Homopt. UK AF137067 MaF M. album Nephotettix virescens I Homopt. Philippines Cordyceps Aj309332 Scarabaeidae I Coleoptera China brittlebankisoides 'The fungal taxonomy follows that ofDriver et at. (2000). FI = CSIRO collection of entomopathogenic fungal cultures (Canberra, Australia). M.a. = Metarhizium anisopliae; Mf = . Note that FI- 1029 is derived from the neotye specimen of M. anisopliae. Fii Alignments and analysis-DNA sequences generated by us and downloaded from str GenBank were aligned using Clustal X 1.81 (Thompson et al. 1997), and the alignment was refined by eye. (Bals.-Criv) Vuil, a related c1avicipitacean anamorph, was used as an outgroup. Parsimony analysis was performed in PAUP* fla versioriblO (Swofford 2002) using a heuristic search. A starting tree was obtained via to a "closest" addition sequence, tree-bisection-reconnection was used as the branch- (l swapping algorithm, and MULTREES was off. Gaps were treated as missing data, and cy 112 ambiguously aligned characters were excluded from the analysis. Bootstrap values were /calculated based on 100 replicates of a heuristic fast stepwise addition search. Ci 141

)f C. taii (HS0006050I), Bayesian analysis was carried out using MrBayes 3.0M (Huelsenbeck 2000, been deposited in the Huelsenbeck et al 2001). We used a six parameter model to run four chains for 500,000 s AF348393, AF348394 generations, saving a tree every 100 generations. The first 500 trees were discarded (burn in), and the remaining trees were saved to a file. A 50% majority rule consensus ,tudy. * tree (Fig. 3) was then calculated using PAUP*.

Country Results of oriRIn Characterization of specimen HS00060501 and its isolate, RCEF0772- The host is Pakistan covered with yellow mycelium. Three stromata emerge from the head of the host, they pt. Eritrea hopt. Australia are cylindric, 38 x 3.5-6 mm broad at stalk, with a yellowish fertile head of20-35 x 2-5 Thailand mm. Perithecia flask-shaped with curved neck, obliquely immersed, 810-(950)-1120 x oleopt. Australia 230-380 llm. Asci cylindric, 300-490 x 3.4-5.7 llm, with ascus cap 3.1 llm in diameter. hopt. Australia Part ascospores cylindrical, 20.0-(28.6)-34.6 x 1.4-(2.5)-3.1 llm. Australia an) Australia Australia Australia thopt. Australia t. Australia nipt. Mexico New pt. Zealand Australia Australia ~mes Australia t. Niger pt. Australia Indonesia pt. Australia Philippines Australia UK pt. Philippines China n of entomopathogenic fungal flavoviride. Note that PI- 1029 Figures 1-2. Cordyceps taii. Lepidopteran larvae from Anhui Province, China, bearing ascigerous and downloaded from stromata (immature in Fig. 2) without its associated Metarhizium anamorph. 197), and the alignment related c1avicipitacean Colonies on Czapek's agar attaining a diameter of 35-43mm with 14 days at 25°C, ; performed in PAUP* flat, with little aerial mycelium, dull gray green, reverse center deep green, sub center g tree was obtained via to margin light sandy yellow. Cylindrical conidiogenous cells with short neck, 7.2- s used as the branch- (14.4)-21.6 x 1-(1.4)-1.8 llm, borne in a palisade on mycelium or conidiophore. Conidia ~d as missing data, and cylindric, (5.0- )7.0-9.0( - 10.0) x 2.4-(2.8)-3.2 llm, two-celled conidia absent. alysis. Bootstrap values Based on the long part spores and other important morphological characters, the se addition search. /Cordyceps specimen (HS00060501) from Huoshan County was identified as Cordyceps 142 taii. The morphological characters of the conidiogenous structures and conidia arising Table 2. Nuclee from the part spores agree with the original description of M. taii (Liang et al. 1991). ¡M.a.) an Two more recently collected specimens of Cordyceps taU collected in Anhui Province are Isolatt ilustrated in Figures 1 and 2. M. taii, 18 isola Clarification of taxonomic position of Metarhizium taii-Regions of the 5.8S rRNA lvI.ll. miisopliae and the complete internal transcribed spacer (ITS) regions from M. taU (GenBank ¡vi. taii, FI-I02S AF348394) and M.a. anisopliae from China were sequenced; their total sizes were 449 M. taii, M.a. Ie! and 451 nucleotides, respectively. These sequences and ITS sequences obtained from lvI. tIJii, M.a. J1 GenBank for 20 strains and 4 varieties of M. anisopliae were subjected to phylogenetic lVI.il. IIWjuS, 18 analysis (Fig. 3). The aligned sequences of the new Chinese isolates of M. taii and M. anisopliae showed a high degree of similarity with those of the 17 strains of M. a. ~Genetic di\'er¡ anisopliae from Driver et al. (2000). There are 22 mutations in total, half of which were Relationships transitions (7 C/T and 4 A/G); the remaining mutations included three C/ A, two T/G, total size of th one AfT as well as five putative insertions or deletions. The nature of these mutations identical to th: suggests that these sequences may have diverged only recently. Among these mutations, ITSI-5.8S-ITS 11 occurred in ITSI and 11 in ITS2; the sequences of the 5.8S rRNA gene among that C. taii an these 18 isolates of M. a. anisopliae (a set of sequences including that from FI-1029, (completely ce which was derived from the IMI ex neotype culture for M. anisopliae) and that of M. lliisopliiie are i taii were identicaL. Divergence in 5.8S rRNA gene sequences between M. taU and other the teleomorp Metarhizium species or varieties varies from 0- 1.8% (Tab. 2). The parsimony search found 2864800 equally parsimonious trees, each of 376 steps and with CI=0.811 and RI=0.808. Of the 683 included characters, 465 were constant, and 106 were parsimony-informative. The strict consensus tree is shown in Fig. 3. The The new variE tree resulting from the Bayesian analysis has an identical topology. on the basis e The large number of trees obtained via the parsimony approach reflects a lack of ITS2 sequeiw variation among isolates of a single variety. In general, the ITS locus provided little until morpho resolution at the varietal level, but did provide support for the recognition of all but one means to ide) of the varieties (M. a. majus). The Bayesian analysis generally agreed with the parsimony sequence dat¡ consensus tree in that each of the varieties described by Driver et al. (2000) was supported In the pr by tree topology except for M.a. majus. Metarhizium taii groups with all strains of M.a. flavoviride cl anisopliae with 99% bootstrap support; if the 85th character of the alignment is excluded, iVIetarhiziwn then strain FI379 is included in this clade with 99% bootstrap support. Furthermore, teleomorph e all varieties of M. anisopliae recognized by Driver et al. (2000)-M.a. anisopliae, M.a. and conidia; I majus, M.a. lepidiotae Driver & Milner, and M.a. acridum Driver & Milner-cluster i'letarliiziwn together in a clade with 100% bootstrap support and differ distinctly from M. album their and from all described varieties of M. jlavoviride. These results suggest that M. taU overall genet should properly be treated either as a variety of M. anisopliae or as being taxonomically of NI.a. /1W;U

indistinguishable from M.a. anisopliae. flavoviride. S Table 2 shows that the range of variation in nucleotide divergence among M. taii and genetically-b 18 isolates of M.a.anisopliae from ITS1, ITS2 and both combined falls within the scope originally idi of variation among M.a. anisopliae isolates, and is distinctly greater than that among variety in a d M. taii, M.a. lepidiotae and M.a. acridum. The range of variation among M. taii and 18 TI1e taxa isolates of M.a. anisopliae from ITSl +ITS2 regions is a little less than among M.a. majus unverified si and 18 isolates of M.a. anisopliae. and the ex t) / 'lIisopliae is(

\ 143 ctures and conidia arising Table 2. Nucleotide divergence within and among species for isolates of M. anisopliae 'v. taU (Liang et aL 1991). (M.a.) anisopliae, M.a.lepidiotae, M.a. acridum, M.a. majus and M. taU. * :ted in Anhui Province are Nucleotide divergence (%) Isolates being compared 5.SS ITSI ITS2 ITSI+ITS2 l-egions of the 5.8S rRNA M. taii, IS isolates of M.a. anisopliae 0 0.7-3.0 1.-2.8 1.0-2.3 ¡ from M. taU (GenBank M.a. anisopliae isolates 0 0-4.4 0-2.8 0.3-2.6 their total sizes were 449 M. taii, FI-1029 (ex type of M.a. anisopliae) 0 1.6 2.3 2.0 sequences obtained from M. taii, M.a.lepidiotae, M.a. acridum 0-0.6 9.6-20.5 9.1-10.6 9.3-15.1 subjected to phylogenetic M. taii, M.a. majus 0 3.0 1.7 2.2 ¡e isolates of M. taii and M.a. majus, is isolates of M.a. anisopliae 0 2.2-4.4 0.6-2.3 1.-2.5 of the 17 strains of M. a. 'Genetic divergence (%) = (number of substitutions/total base pairs) x 100. 1 total, half of which were ided three CIA, two T/G, Relationships between C. taii and M. anisopliae var. anisopliae (RCEF0772)- The 1ature of these mutations total size of the ITSI-5.8S-ITS2 region (449 bp) for C. taii (GenBank AF348393) was . Among these mutations, identical to that of M.a. anisopliae (RCEF0772). The sizes ofITSl, 5.8S and ITS2 in the 5.8S rRNA gene among ITSI-5.8S-ITS2 region are 134 bp, 158 bp, 167 bp, respectively. The alignment showed ~ding that from FI-1029, that C. taii and M.a. anisopliae isolated from this Cordyceps specimen had identical nisopliae) and that of M. (completely congruent) ITSI-5.8S-ITS2 sequences. This suggests that C. taii and M.a. ietween M. taii and other anisopliae are genetically identicaL. We suggest that Cordyceps taii and M.a anisopliae are the teleomorphic and anamorphic phases of a single organism. is trees, each of 376 steps cters, 465 were constant, Discussion ~e is shown in Fig. 3. The logy. The new varieties M.a. acridum and M.a. lepidiotum Driver & Milner were recognized 'proach reflects a lack of on the basis of their distinctive ITS sequence data (Driver et aL 2000). The ITS1 and ITS locus provided little ITS2 sequence data were the only characters included in their Latin diagnoses; therefore, recognition of all but one until morphological or other characters are correlated with these genotyes, the only ~reed with the parsimony means to identify varieties of Metarhizium anisopliae within the classification is with taL. (2000) was supported sequence data (Driver et al. 2000). JS with all strains of M.a. In the phylogenetic tree, Cordyceps brittlebankisoides falls in the Metarhizium :ie alignment is excluded, jlavoviride clade. However, its purported anamorph M.a. majus belongs to the p support. Furthermore, Metarhizium anisopliae clade. Liu et al. (2001) concluded that C. brittlebankisoides is the O)-M.a. anisopliae, M.a. teleomorph of M.a. majus based on the morphological characteristics of its phialides Jriver & Milner-cluster and conidia; they did not compare the ITS sequences of their fungus with those of other istinctly from M. album Metarhizium taxa. These authors also noted that the spore sizes and cultural color of ults suggest that M. taii their fungus were similar to those of M.f jlavoviride (Liu et aL 2001). Based on the r as being taxonomically overall genetic evidence, we believe that C. brittlebankisoides is not the teleomorph of M.a. majus, but may instead be the teleomorph of a stil undescribed variety of M. gence among M. taii and jlavoviride. Such a genetically-based reclassification within Metarhizium echoes the .ed falls within the scope genetically-based decision to transfer grasshopper- and locust-pathogenic isolates ~reater than that among originally identified on morphological and cultural bases as M. jlavoviride to a new m among M. taU and 18 variety in a different species as M.a. acridum (Driver et aL 2000). than among M.a. majus The taxonomic status of M. taU as a species has remained both uncertain and unverified since its description. Because the nucleotide divergence between M. taii / and the ex type isolate for M. anisopliae falls within the range of variation among M.a. anisopliae isolates, and very little divergence was found in the ITSI and ITS2 sequences 144 between M. taii and 18 isolates of M.a. anisopliae, we propose that M. taii is a synonym of M.a. anisopliae; there is not even enough distinctive genetic variation between M. taii and M.a. anisopliae to justify treating M. taii as a new variety of M. anisopliae. Tulloch (1976) separated the varieties of M.a. anisopliae and M.a. majus mainly by the shape and lengths of their conidia, (3.3-)5.0-8.0(-9.0) ¡.m and (9.0-)10.0-14.0(-18.0) ¡.m, 83 respectively. The anamorph of Cordyceps taii was recognized as a new species mainly based on the dimensions of its conidia ((4.8- )7.8-9.6( - 14) ¡.mJ and the (apparently very infrequent) formation of two-celled conidia (Liang et al. 1991). In this study, the conidia of our M. taii strain (RCEF0772) are clearly within the limits of M.a. anisopliae defined by Tulloch: the conidia of the strain isolated by Liang fell partly in the range of M.a. anisopliae. Longer than usual conidia in M.a. anisopliae have been described previously in strains from China with lengths of 7.5-8.0 ¡.m (Guo et al. 1986) and in strains from Australia at 8-9 ¡.m (Yip et al. 1992). The lengths of conidia from M. taii fall between those of M.a. anisopliae and M.a. majus, but in view of the high genetic similarity demonstrated here the slight difference between M.a. anisopliae and M. taii, and in conidial length between M.a. anisopliae and M.a. majus, can be regarded as differences among strains. Liang et al. (1991) found differences in the abilty to form synnemata and two-celled conidia compared with other species or varieties of Metarhizium. Comparable morphological differences among isolates of other fungal entomopathogens have not been treated as taxonomically significant. For example, isolates of many diverse entomopathogenic species of Paecilomyces mayor may not form synnemata, with the abilty to do so varying on an isolate-by-isolate basis. Similarly, Li et al. (2001) reported on an isolate of Beauveria bassiana able to produce large synnemata, a behavior not generally known from other isolates of this globally distributed species. The combined molecular and morphological evidence presents compellng support for the conclusion that M. taii cannot be retained as an independent taxon at either the specific or varietal level and that it must be synonymized with M.a. anisopliae. Some studies have characterized the range of variation in nucleotide divergence in ITS I and ITS2 regions from different strains in the other fungal groups. Epicoccum nigrum and Phoma epicoccina are synanamorphs of the same biological species, and but the ranges of divergences in the sequences of their ITS 1 and ITS2 genes have been found to be 0-2.8% and 0-2.6%, respectively (Arenal et al. 2000). As a standard for comparison, other studies have found infraspecific variabilty in ITSI sequences for Hymenoscyphus ericae, Trichoderma harzianum, and Ganoderma lucidum, to be 1.2%-3.5%, 0-2%, and 0-4%, respectively (Egger & Sigler 1993, Grondona et al. 1997). Even higher infraspecific divergence has been found for other fungi such as Gaeumannomyces graminis, with 1.5- 12.0% for ITSl and 0-2.9% for ITS2 (Bryan et al. 1995); and Colletotrichium acutatum 100 I with 0-6% in the ITSI (Sreenivasaprasad et al. 1994). M. a. anisopliae is one of the most common and important species of entomopathogenic fungi. Like other such globally distributed and very important entomopathogens as Beauveria bassiana and Beauveria brongniartii, the teleomorph of this hyphomycete has ~ remained uncertain or unknown. Schaerffenberg (1959) claimed that the teleomorph of M. anisopliae belonged to the order Sphaeriales (), but this hypothesis was never well supported either by Schaerffenberg or by any subsequent workers. Figure 3. 5t Liang et at. (l991) provided the first credible evidence that the teleomorph of any data from !J Metarhizium species is a clavicipitaceous fungus belonging in the genus Cordyceps; probabilitie

\ y 145

¡at M. taU is a synonym iriation between M. taU f M. anisopliae. Tulloch us mainly by the shape M.a1bum - )io.0-14.0( - 18.0) i.m, 83 M.t.novozealandicum s a new species mainly 81 M.f.minus nd the (apparently very M.t.pemphigi n this study, the conidia 84 84 M.a. anisopliae defined M.f.tlavovindes ly in the range of M.a. 87 C.brittlebankisoides ~n described previously 86) and in strains from AF135212 99 im M. taii fall between AF137059 high genetic similarity 100 ae and M. taii, and in AF135215 regarded as differences 91 AF134150 lity to form synnemata M.a.a. ex 1ype ieties of Metarhizium. 82 91 AF135214 mgal entomopathogens 82 isolates of many diverse AF136376 m synnemata, with the AF137055 "i et at. (2001) reported iemata, a behavior not 94 AF136928 species. The combined 76 95 AF137054 port for the conclusion AF137057 . the specific or varietal 79 AF135213

nucleotide divergence AF135211 ¡gal groups. Epicoccum logical species, and but AF135216 99 genes have been found 93 AF137056 mdard for comparison, 100 96 M.taii ces for Hymenoscyphus I.%-3.5%, 0-2%, and M.a.majus 100 \Ten higher infraspecific AF348395 vces graminis, with 1.5- 100 AF136375 illetotrichium acutatum 100 AF137058 ~s of entomopathogenic 100 M.a.lepidiotae t entomopathogens as M.a.acndum f this hyphomycete has ~d that the teleomorph B.bassiana a), but this hypothesis 'r subsequent workers. Figure 3. Strict consensus of 2864800 equally parsimonious trees based on ITS region sequence he teleomorph of any data from Metarhizium spp. Values above the branches indicate bootstrap support; postenor the genus Cordyceps; /' probabilities resulting from the Bayesian analysis are shown below the branches. 146 this connection was based on studies of secondary (microcyclic) conidiogenesis by Liu ZW, Guo H discharged part spores. The data presented here show that M. taii cannot be maintained ribosomal I as an independent species and that the most reasonable treatment for this species is as a Liu ZY, Liang Z facultative (heterotyic) synonym of M.a. anisopliae: of glLbs an Metarhizium anisopliae (Metschn.) Sorokin var. anisopliae, (Plant Parasites of Man and Mavridou A, T) revealed by Animals as Causes ofInfectious Diseases) 2: 267 (1883) (in Russian) 1241. Synonym: Metarhizium taU Z.Q. Liang & A.Y. Liu, Acta Mycol. Sinica 10: 260 (1991). Milner RJ, Dri Because the present molecular and cultural evidence confirms unambiguously that the problems 'I teIeomorph of Metarhizium taii is Cordyceps taii, it must also now be recognized that the 6'h Internal teIeomorph of M. anisopliae var. anisopliae is C. taii. 110. Rath AC, Car Acknowledgments carbohydn Rakotonirainy We are grateful to Drs. Wenying Zhuang and Yijian Yao for reviewing the manuscript. This work within the was supported partly by the National Natural Science Foundation of China (Grant No. 30300004, Res 98: 22~ 30330500), the Key Laboratory of Systematic Mycology and Lichenology (Grant No. 912.), the Anhui Provincial Science Foundation for Excellent Youth and prograrn for "NCET". Rombach MC a pathogen 27: 87-92. Literature cited Sainbrook J, F Arenal F, Platas G, Monte E, Pelaez F. 2000. ITS sequencing support for Epicoccum nigrum and York; Cold Phoma epicoccina being the same biological species. Mycol Res 104: 301-303. Schaerftenber¡ Bryan GT, Daniels MJ, Osbourn AE. 1995. Comparison of fungi within the Gaeumannomyces- aiiisopliae Phialophora complex by analysis of ribosomal DNA sequences. Appl Environ Microbiol praktische 61: 681-689. Sreenivasapra~ Curran J, Driver F, Ballard JWO, Milner RJ. 1994. Phylogeny of Metarhizium: analysis of ribosomal identificat DNA sequence data. Mycol Res 98: 547-552. Swofford D. 2( Driver F, Milner RJ, Trueman WHA. 2000. Taxonomic revision of Metarhizium based on a Sunderlan phylogenetic analysis ofrDNA sequence data. Mycol Res 104: 134-150. Tulloch M .19: Egger K N, Sigler 1. 1993. Relatedness of the ericoid endophytes Scytalidium vaccinii and 1110mpson JL Hymenoscyphus ericae inferred from analysis of ribosomal DNA. Mycologia 85: 219-230. multiple s Grondona I, Monte E, Garcia-Acha Y, Sutton B. 1997. Pyrenochaeta dolichi: an example of a weight in~ confusing species. Mycol Res 101:1405-1408. Tzean SS, H! Guo HL, Ye BL, Yue YY, Chen QT, Fu CS. 1986. Three new species of Metarhizium. Acta Mycol 85: 514-51 Sinica 5: 177-184. 111ompson II Huelsenbeck JP. 2000. MrBayes: Bayesian inferences of phylogeny (softare). New York: University interÜice: of Rochester. Nucl Acid Huelsenbeck JP, Ronquist F, Nielsen ES, Bollback JP. 2001. Bayesian inference of phylogeny and its Yip HI, Ratf ',.impact on evolutionary biology. Science 294:2310- 2314. Tasinania Leal SCM, Bertioli DJ, Butt TM. 1997. Amplification and restriction endonuclease digestion Scarabaei of the Pr 1 gene for the detection and characterization of Metarhizium strains. Mycol Res Zhu H, Qu B 101: 257-265. M)'col Sir Li ZZ, Li CR, Huang B, Fan MZ. 2001. Discovery and demonstration of the teleomorph of Beauveria bassiana (Bals.) Vuill., an important entomogenous fungus. Chinese Sci Bull 9: 751-753. Liang ZQ, Liu AY, Liu JL. 1991. A new species of the genus Cordyceps and its Metarhizium anamorph. Ac;-a Mycol Sinica 10: 257-262. Liu AY, Liang ZQ, Cao 1. 1989. Study on the identification of Metarhizium biformisporae. J Guizhou Agr Univ 2: 27-31.

" ""

\ 147

:onidiogenesis by Liu ZW, Guo HL, Kurtzman CPo 1994. Phylogeny of Metarhizium species as determined by partial 110t be maintained ribosomal RNA sequences. Acta Mycol Sinica 13: 139-151. this species is as a Liu ZY, Liang ZQ, Whalley AJS, Yao YJ, Liu AY. 2001. Cordyceps brittlebankisoides, a new pathogen of grubs and its anamorph, Metarhizium anisopliae var. majus. J Invert Pathol 78: 178- 182. Mavridou A, Typas MA. 1998. Intraspecific polymorphism in Metarhizium anisopliae var. anisopliae irasites of Man and revealed by analysis of rDNA gene species complex and mtDNA RFLPs. Mycol Res 102: 1233- 1241. 260 (1991). Milner RJ, Driver F, Curran J, Glare TR, Prior C, Bridge PD, Zimmermann G. 1994. Recent biguously that the problems with the taxonomy within the genus, Metarhizium, and a possible solution. Abstracts, ecognized that the 6th Internat. Colloq. Invertebr Pathol Microb Control, (Montpeller, France), voL. 2, pp. 109- LLO. Rath AC, Carr q, Graham B R. 1995. Characterization of Metarhizium anisopliae strains by carbohydrate utilization (AP150CH). J Invert Pathol65: 152-161. Rakotonirainy MS, Dutertre M, Cariou ML, Brygoo Y, Riba G. 1994. Phylogenetic relationships anuscript. This work within the genera Metarhizium based on 28s rRNA sequences and isozyme comparison. Mycol 3rant No. 30300004, Res 98: 225-230. 3rant No. 912.), the Rombach MC, Humber RA, Roberts DW. 1986. Metarhizium flavoviride var. minus var. nov., rCET". a pathogen of plant-and leafhoppers on rice in the Philippines and Solomon Islands. Mycotaxon 27: 87-92. Sambrook J, Fritsch EF, Maniatis T. 1989 Molecular Cloning: A Laboratory Manual, 2nd ed. New icoccum nigrum and , York; Cold Spring Harbor Laboratory Press. ,03. Schaerffenberg, B. 1959. Zür Biologie und Ökologie des insektentötendend Pilzes Metarrhizium e Gaeumannomyces- anisopliae (Metsch.) Sorok. (Entwicklung, Kultur, Lebensanspruche, Infektionsverlauf, I Environ Microbiol praktische Bedeutung). Z Angew Entomol44: 262-271. Sreenivasaprasad S, Mils PR, Brown AE. 1994. Nucleotide sequence of the rDNA spacer 1 enables analysis of ribosomal identification of isolates of Colletotrichum as C. acutatum. Mycol Res 98: 186- 1 88. Swofford D. 2002. PAUP*: Phylogenetic analysis using parsimony (*and other methods), version 4. rhizium based on a Sunderland, Massachusetts: Sinauer Associates. Tulloch M.1976. The genus Metarhizium. Trans Brit Mycol Soc 66: 407-411. progressive Ûidium vaccinii and Thompson JD, Higgins DG, Gibson TJ. 1994. Clustal W: improving the sensitivity of gia 85: 219-230. multiple sequence alignment through sequence weighting, position-specific gap penalties and ;hi: an example of a weight matrix choice. Nucl Acids Res 22: 4673-4680. Tzean SS, Hsieh LS, Chen JL, WU WJ. 1993. Nomuraea cylindrospora comb. novo Mycologia irhizium. Acta Mycol 85: 514-519. Thompson JD, Gibson TJ, Plewniak F, Jeanmougin F, Higgins DG. 1997. The Clustal X windows New York: University interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucl Acids Res 24:4876-4882. ~ of phylogeny and its Yip HI, Rath AC, Koen TB. 1992. Characterization of Metarhizium anisopliae isolates from Tasmanian pasture soil and their pathogenicity to redheadedcockchafer (Coleoptera: donuclease digestion Scarabaeidae: Adoryphorus couloni). Mycol Res 96: 92-96. n strains. Mycol Res Zhu H, Qu F, Zhu LH. 1994. Isolation of genomic DNAs from fungi using benzyl chloride. Acta Mycol Sinica 13: 41-47. ~omorph of Beauveria 3ull9: 751-753. and its Metarhizium / Jrmisporae. J Guizhou