This Article Abstract Full Text (PDF) Services Similar articles in this journal Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Moon, C. D. Articles by Schardl, C. L. Search for Related Content PubMed Articles by Moon, C. D. Articles by Schardl, C. L. Agricola Articles by Moon, C. D. Articles by Schardl, C. L.

The evolutionary origins of three new Neotyphodium endophyte species from grasses indigenous to the Southern Hemisphere¹

     Department of Plant Pathology, University of Kentucky, Lexington, Kentucky 40546-0091, USA

Christopher O. Miles

     AgResearch, Ruakura Research Centre, Private Bag 3123, Hamilton, New Zealand

Ulla Järlfors
Christopher L. Schardl ²

     Department of Plant Pathology, University of Kentucky, Lexington, Kentucky 40546-0091, USA

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS TAXONOMY DISCUSSION LITERATURE CITED

 

Members of the genus Neotyphodium are asexual, seedborne, protective fungal endophytes of cool season grasses that have likely evolved either directly from sexual Epichloë; species, or by the interspecific hybridization of distinct lineages of Epichloë; and Neotyphodium. We investigated the evolutionary origins of Neotyphodium endophytes from several grasses that are indigenous to the Southern Hemisphere using a multiple-gene phylogenetic approach. Intron regions of the genes encoding ß-tubulin (tub2), translation elongation factor 1- (tef1) and actin (act1) were amplified by polymerase chain reaction and sequenced. Phylogenetic analyses of these sequences, aligned with homologous sequences from Epichloë; spp., revealed the evolutionary origins of the Southern Hemisphere endophytes, where one lineage of apparently non-hybrid origin, and three lineages of unique interspecific hybrid origin were identified. On the basis of morphology, host range and evolutionary history, we propose three new species of Neotyphodium. Neotyphodium aotearoae was isolated from Echinopogon ovatus populations from New Zealand and Australia, and comprised a unique, apparently non-hybrid lineage within the Epichloë; species phylogeny. In contrast, an interspecific hybrid lineage was identified from two Australian Ec. ovatus populations, whose ancestry apparently involved lineages closely related to extant E. festucae and an E. typhina genotype similar to that of isolates from Poa pratensis. Endophytes infecting South African Melica racemosa and M. decumbens (dronkgras) appeared to be hybrids of E. festucae and N. aotearoae or close relatives. The names N. australiense and N. melicicola are proposed for these two hybrid lineages, respectively. The origin of N. tembladerae, an established endophyte species from South American Poa and Festuca spp., was also investigated. Neotyphodium tembladerae appeared to be of hybrid origin, involving E. festucae and an E. typhina genotype similar to that of isolates from Poa nemoralis. The results of this study highlight the widespread occurrence of interspecific hybrid Neotyphodium lineages on a global scale, and the extent of endophyte gene-flow between the Northern and Southern Hemispheres.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS TAXONOMY DISCUSSION LITERATURE CITED

 
Neotyphodium Glenn et al endophytes are asexual, filamentous fungi that colonize the apoplasts of cool season grasses (subfamily Poöideae), and are disseminated clonally in the seed. These endophytes and their grass hosts share mutualistic relationships where the endophyte is provided with nutrition and a means of propagation within the host. In return, the host may be protected from a number of biotic and abiotic stress factors that include grazing by animals (Bacon et al 1977, Fletcher and Harvey 1981), diseases (Gwinn and Gavin 1992), and drought conditions (Arechevaleta et al 1989). The protective activity against grazing animals is perhaps the most important feature of the grass-endophyte association to the agricultural industry. Endophytes can produce a number of classes of alkaloid compounds, some of which are highly toxic to grazing livestock. The ergot alkaloids are causative factors in fescue toxicosis (Bacon et al 1986) and indole-diterpenoids, such as lolitrems, are responsible for ryegrass staggers (Gallagher et al 1984). However, certain alkaloids are known to play a beneficial role in pasture persistence. Lolines and peramine confer significant deterrent and toxic activity against insect pests (Johnson et al 1985, Prestidge et al 1985, Rowan and Gaynor 1986, Bush et al 1993). Thus Neotyphodium endophytes are economically important components of pastoral systems.

Neotyphodium endophytes are closely related to sexual Epichloë; species, the grass choke pathogens (Schardl et al 1991). These two groups share a similar growth habit during the vegetative stage of the host life cycle, and like Neotyphodium endophytes, some species of Epichloë; may be clonally transmitted through the seed. However, Epichloë; species may also undergo a sexual cycle. This involves the formation of fungal fruiting structures (stromata) on immature inflorescences, followed by heterothallic mating, resulting in the production of perithecia and ascospores. Infection of uninfected plants and seed is achieved via the contagious spread of airborne ascospores that have been forcibly ejected from perithecia (Chung and Schardl 1997a, Brem and Leuchtmann 1999). The endophyte sexual cycle is undertaken to the detriment of host reproduction, as affected flowering tillers are sterilized (choked) in the process. The degree to which the sexual cycle is expressed in Epichloë; species varies from 100% seed transmission to complete sterilization of the host, and appears to involve both host and endophyte genotypes, as well as environmental conditions, though details of the factors involved are currently not well understood (Bucheli and Leuchtmann 1996, Leuchtmann and Schardl 1998). Nonetheless, Epichloë; species that are able to undergo both modes of transmission may have co-evolved closely with their grass hosts at the level of the grass tribe (Schardl et al 1997). Endophyte-host associations have been classified into three types based on the level of stromatal expression (White 1988). In type 1 associations the host is completely sterilized by the endophyte, whereas in type 2 associations, only a fraction of flowering tillers are choked. In contrast, type 3 associations describe cases where endophytes are solely seed transmitted (White 1988). To date, ten species of Epichloë; have been formally described, based on characteristics such as morphology, natural host range, mating tests, isozymes, and more recently, by DNA sequences (White 1993, Leuchtmann et al 1994, Leuchtmann and Schardl 1998, Schardl and Leuchtmann 1999).

That Neotyphodium endophytes have evolved from Epichloë; species has been firmly established from analyses of molecular data (Schardl et al 1994, Tsai et al 1994, Moon et al 2000, Craven et al 2001a). From these studies, two main evolutionary processes are hypothesized to be responsible for the formation of Neotyphodium species. The first is by direct evolution from a single Epichloë; species, presumably by loss of the sexual state. Neotyphodium lolii Latch et al and N. typhinum var. canariense Moon et al are thought to have evolved in this manner from E. festucae Leuchtm. et al (Schardl et al 1991, Leuchtmann et al 1994) and E. typhina (Pers. : Fr.) Tul. (Moon et al 2000), respectively. The second mode of Neotyphodium evolution is via the interspecific hybridization of distinct Epichloë; and Neotyphodium lineages (Schardl et al 1994, Tsai et al 1994, Moon et al 2000, Craven et al 2001a). Hybrid lineages are commonly recognized by the presence of multiple gene copies using techniques such as isozyme, microsatellite, and Southern blot analyses (Leuchtmann and Clay 1990, Schardl et al 1994, Tsai et al 1994, Moon et al 1999). If highly discriminative methods are employed, such as sequencing orthologous genes followed by phylogenetic analyses, the ancestral lineages that contributed to the hybrid endophyte may be readily traced. The origins of several hybrid endophytes from Lolium L. grasses have been identified using these approaches (Schardl et al 1994, Tsai et al 1994, Moon et al 2000, Craven et al 2001a). These include N. coenophialum (Morgan-Jones et W. Gams) Glenn et al [symbiotic with tall fescue, L. arundinaceum (Schreb.) Darbysh. = Festuca arundinacea Schreb.] of E. typhina, E. festucae, and E. baconii White et al ancestry (Tsai et al 1994), and N. uncinatum (W. Gams et al) Glenn et al [symbiotic with meadow fescue, L. pratense (Huds.) Darbysh. = F. pratensis Huds.] of E. typhina and E. bromicola Leuchtm. et Schardl ancestry (Craven et al 2001a). Evidence from isozyme and microsatellite data also suggests that endophytes from some Festuca L., and Hordeum L. are hybrid, but these techniques alone are not sufficiently discerning to firmly ascertain the ancestry of these endophytes (Leuchtmann 1994, Moon 1999).

To date, the majority of Neotyphodium and Epichloë; endophytes that have been analyzed have been from the Northern Hemisphere, mainly Europe and North America. However, endophytic fungal infections have been reported since the early twentieth century in grasses native to the Southern Hemisphere (see Cabral et al 1999, Miles et al 1998). Furthermore, several of these grasses have been associated with toxicoses to grazing animals, including a ryegrass staggers-like disorder of stock grazing Echinopogon spp. in Australia (Seddon and Carne 1926, Everist 1981), and huecú toxicosis of animals grazing Festuca and Poa L. spp. in South America (see Pomilio et al 1989, Cabral et al 1999). A drunken-like behavior has also been associated with animals grazing Melica decumbens Thunb. (dronkgras) in South Africa (see Gibbs Russell and Ellis 1982). Previous characterizations of these endophytes from morphology, serological tests, alkaloid analyses, and DNA sequences suggest that these are Neotyphodium endophytes, and recently, the endophyte associated with huecú toxicosis was named Neotyphodium tembladerae Cabral et J. F. White (Cabral et al 1999).

In this study we report on the evolutionary origins of Neotyphodium endophytes from several grasses that are indigenous to the Southern Hemisphere, including Echinopogon ovatus (G. Forst) P. Beauv. from New Zealand and Australia, Melica decumbens and M. racemosa Thunb. from South Africa, and Poa huecu Par. (symbiotic with N. tembladerae) from South America. From phylogenetic analysis of DNA sequences from intron regions of the genes encoding ß-tubulin (tub2), translation elongation factor 1- (tef1) and actin (act1), we have identified the likely ancestry of these endophytes and propose three new Neotyphodium species: N. aotearoae Moon, Miles et Schardl, N. australiense Moon et Schardl, and N. melicicola Moon et Schardl, on the basis of genetic and morphological characteristics.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS TAXONOMY DISCUSSION LITERATURE CITED

 
Biological materials – The endophyte-infected grasses and fungal isolates used in this study are listed in Table I. All grasses listed, apart from Poa huecu, were grown from seed and maintained in the greenhouse. A culture of Neotyphodium tembladerae, from P. huecu, was obtained from the American Type Culture Collection (ATCC; Manassas, Virginia), accession 200844. Cultures were deposited at Centraalbureau voor Schimmelcultures (CBS; Utrect, The Netherlands). Herbarium specimens were prepared using the method of Pollack (1967) and deposited in the Cornell University Plant Pathology Herbarium (CUP).

Examination of endophyte morphology in culture – For the macroscopic examination of endophyte cultures, small agar blocks (ca 1 mm²) were cut from the margins of actively growing fungal colonies and used to inoculate PDA plates. Plates were incubated in the dark at 22 C. Morphological observations and colony diameter measurements were taken after 8 wk. Cultures for the microscopic examination of fungal structures were prepared by grinding ca 5 mm² of mycelium from a growing colony margin in 500 µL sterile water. Aliquots of 100 µL were spread on 1.5% water agar plates and incubated at 22 C until growth was evident and fruiting structures, if produced, were present. Agar blocks were mounted on slides under a coverslip and examined by light microscopy. Dimensions of hyphae, conidia, and conidiogenous cells were recorded. At least 20 measurements were taken from each isolate.

Electron microscopy – Pieces, ca 1 x 2 mm, of endophyte-infected leaf sheaths were prefixed under vacuum in 3% glutaraldehyde in 0.1 M cacodylate buffer, pH 7.2, overnight, then briefly rinsed in 0.1 M cacodylate buffer. The specimens were postfixed in 1% OsO₄ in cacodylate buffer at 4 C for 1.5 h, then rinsed several times in distilled water. The tissue was left overnight at room temperature in a solution of saturated aqueous uranyl acetate that had been diluted to half concentration. The material was dehydrated in an ethanol series followed by acetone and embedded in Spurr's resin. Sections, ca 80 nm, were cut on an LKB Ultrotome III, further stained with uranyl acetate as well as 0.5% lead citrate, ca pH 12, and examined in a Philips 400 electron microscope.

Assessment of epiphyllous growth – Endophyte-infected grasses were examined for epiphyllous endophyte growth as described previously (Craven et al 2001a). From each plant, three tillers were collected and endophyte infection of mature leaves was confirmed by aniline blue staining of epidermal peels from the adaxial surface of the leaf sheath followed by light microscopy. From infected leaves a 2-cm transverse section of the leaf blade was cut directly above the ligule. Leaf blade sections were stained in Calcofluor White M2R (0.02% 4,4'-bis-(4-anilino-6-diethyl-amino-s-triazin-2-ylamino)-2,2'-stilbene disulfonic acid; Polysciences, Inc., Warrington, Pennsylvania) for 10 min. Both the abaxial and adaxial surfaces of the blade were then examined for the presence of hyphae using fluorescent microscopy.

Examination of hybrid endophytes – To ensure that the multiple gene copies detected in endophyte isolates were not the result of mixed or contaminating cultures, endophytes were single-conidiospore isolated by three rounds of single-spore streaking on PDA (Craven et al 2001a). Genetic analyses were repeated for single-conidiospore isolated cultures. To investigate the possibility that endophyte isolates were heterokaryons, the nuclear status of conidiospores was assessed by 4',6-diamidino-2-phenylindole (DAPI) staining, as previously described (Craven et al 2001a).

DNA isolation – Total genomic endophyte DNA was isolated using the method of Al-Samarrai and Schmid (2000) with slight modification. Thirty milligrams of freeze-dried mycelium (prepared from cultures grown on PDA plates overlaid with sterile cellophane) was ground to a fine powder in a mortar with a pestle and liquid nitrogen. Mycelium was resuspended in 1 mL of lysis buffer [40 mM tris hydroxymethylaminomethane (Tris) acetate, 20 mM sodium acetate, 1 mM disodium ethylenediaminetetraacetic acid (EDTA) and 1% sodium lauryl sulfate, pH 7.8]. Two microliters of 10 mg mL^-1 RNase A (Boehringer Mannheim, Mannheim, Germany) was added, the mixture was then pipetted until frothy, and transferred to a 2 mL microcentrifuge tube. A one-third volume of 5 M NaCl was mixed into the sample by inversion, then the mixture was centrifuged at 15 000 g for 12 min at 4 C. The supernatant was transferred to a new tube and sequentially extracted with equal volumes of 25:24:1 phenol:chloroform:isoamyl alcohol (v/v), and 24:1 chloroform:isoamyl alcohol (v/v), with centrifugation at 15 000 g for 10 min between extractions. DNA was precipitated with 2 volumes of 95% ethanol, and pelleted by centrifugation at 15 000 g for 5 min. The DNA pellet was drained briefly and resuspended in 500 µL lysis buffer by pipetting repeatedly. A one third volume of 5 M NaCl was added, then the sample was extracted with an equal volume of 24:1 chloroform:isoamyl alcohol (v/v). DNA was precipitated from the supernatant by addition of 2 volumes of 95% ethanol. DNA was pelleted by centrifugation at 15 000 g for 5 min, washed four times in ice cold 70% ethanol and left to air-dry. DNA was redissolved in 200 µL of water that had been purified by Milli-Q (Millipore Corp., Bedford, Massachusetts).

For waxy, slow-growing endophyte colonies, total genomic DNA was isolated from fresh mycelium using a modified microwave small scale DNA extraction method (Moon et al 2000). DNA samples were quantified by fluorometry as described previously (Moon et al 1999).

DNA amplification – Gene regions for use in phylogenetic analyses were amplified by polymerase chain reaction (PCR) using primers (listed in Table II) that annealed to conserved exon sequences. Specifically, the 5' regions of tub2, including the first three introns, were amplified using primers tub2-exon1d–1 and tub2-exon4u–2. The 5' regions of tef1, including introns 1–5, were amplified with primers tef1-exon1d–1 and tef1-exon5u–1. The 5' regions of act1, including introns 1–5, were amplified using primers act1-exon1d–1 and act1-exon6u–1. The approximate sizes of the amplified products were 980 bp, 860 bp, and 1300 bp for tub2, tef1, and act1, respectively.

DNA sequencing – PCR products to be used as templates in DNA sequencing reactions were purified using a QIAquick PCR Purification Kit (Qiagen Inc., Valencia, California) according to the manufacturer's instructions. The concentration of purified products was estimated from the intensity of sample bands relative to a quantitative 100 bp ladder (Panvera) that had been electrophoresed in agarose gels followed by ethidium bromide staining.

Sequencing reactions were carried out using a BigDye Terminator Cycle Sequencing Kit (PE Applied Biosystems). Both strands of the purified products were sequenced using amplification primers and primers that annealed to internal regions of the amplified product (Table II). Sequence products were separated by electrophoresis on an ABI Prism model 310 or 377 genetic analyzer (PE Applied Biosystems) and analyzed with Sequencing Analysis Version 3.0 software (PE Applied Biosystems).

Occasionally multiple gene copies in individual sequencing reactions were apparent by the presence of multiple dye-terminator peaks at single nucleotide positions. Thus, primers were designed to selectively amplify each gene copy based on conserved sequences with the 3'-end nucleotide based on a polymorphic site. The selective primers are listed in Table II. Specifically, from isolates CBS 109340, 109341 and 109342, multiple tub2 gene copies were selectively amplified using tub2-exon4u–2 and either tub-intron1d-1-selA or C; act1 genes were selectively amplified using act1-exon1d–1and either act-exon5u-2-selC or G. From isolates CBS 109346 and 109347 and ATCC 200844, tub2 genes were selectively amplified using tub2-exon4u–2 and either tub2-intron1d-2-selA or C; tef1 genes were selectively amplified using tef1-exon6u–1 and either tef-exon1d-2-selT or G. Actin genes from ATCC 200844 were selectively amplified using act1-exon1d–1 in combination with either act-intron4u–1selA or C. The resulting selectively amplified gene copies were unambiguously sequenced. Unique tub2, tef1, and act1 gene sequences were identified and deposited into GenBank and accession numbers are listed in Table III.

Gene trees were generated from aligned sequences using PAUP* version 4 (Swofford 1998). Alignment gaps and ambiguous characters were treated as missing information. Maximum parsimony (MP) trees were inferred using the branch-and-bound search algorithm. Characters were unordered and unweighted. Tree roots were estimated by mid-point rooting. One thousand bootstrap trees were generated under the maximum parsimony criteria on using the heuristic search option with simple stepwise addition of sequences and tree bisection-reconnection branch swapping. Where 1000 bootstrap replications could not be completed due to heavy computational requirements, 150 replications were performed instead.

Neighbor-joining (NJ) trees were constructed using both Jukes-Cantor (one-parameter) and Kimura two-parameter distance matrices (transition to transversion ratio was set to 2.0). Bootstrap consensus trees of 1000 replicates were generated using the NJ option.

Alignments and trees are deposited in TreeBASE under accession numbers M1107–M1110 and S696.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS TAXONOMY DISCUSSION LITERATURE CITED

 
Endophyte cultural morphology – Endophytes were cultured from surface-sterilized leaf sheaths and blades of the infected grasses listed in Table I, except N. tembladerae isolate 200844, which was obtained from ATCC. Colony radial growth rates and conidia dimensions are shown in Table IV.

View larger version (56K): [in this window] [in a new window]    FIGS. 1–10. Photographs of Southern-Hemisphere endophyte cultures showing colony and conidial morphologies. Cultures were grown on PDA and are 8 wk old. Figs. 1–4. Various colony morphologies of N. aotearoae isolates: ATCC MYA-1234 (Fig. 1), ATCC MYA-1231 (Fig. 2), CBS 109345 (Fig. 3), and ATCC MYA-1232 (Fig. 4). Figs. 5 and 6. N. australiense isolates CBS 109346 (Fig. 5) and CBS 109347 (Fig. 6). Figs. 7 and 8. N. melicicola isolates CBS 109341 (Fig. 7) and CBS 109340 (Fig. 8). Figs. 9 and 10. N. tembladerae isolate ATCC 200844. Scale bar for colony pictures is 10 mm, and for conidia pictures, 10 µm

Pure cultures were obtained from the sporulating Ec. ovatus endophytes, CBS 109346 and CBS 109347, as well as the N. tembladerae isolate, by repeated rounds of single-spore streaking on PDA. This procedure was unfeasible for the endophytes from Melica because of their very low levels of spore production on PDA. Streaking was attempted from cultures grown on water agar, but subsequent colony growth on water agar was too limited to perform further rounds of streaking. Conidia from all isolates were determined to be uninucleate by DAPI staining followed by fluorescent microscopy.

Electron microscopy – Leaf sheath tissue from endophyte-infected Echinopogon ovatus, plants 908 and 938, was examined by transmission electron microscopy (Figs. 11 and 12 respectively). Host cells adjacent to endophyte hyphae appeared normal and host defense responses, such as papillae, were not observed. Endophyte-infected Melica spp. were not examined by electron microscopy because difficulties were encountered when sectioning this material, which was quite brittle. Neotyphodium tembladerae-infected plant tissue was not available for examination.

View larger version (201K): [in this window] [in a new window]    FIGS. 11 and 12. Electron micrographs showing transverse leaf sheath sections of endophyte-infected Ec. ovatus. Fig. 11. N. aotearoae isolate ATCC MYA-1234 infecting plant 908. Fig. 12. N. australiense isolate CBS 109346 infecting plant 937. Annotations are as follows: hypha (H), endophyte nucleus (EN), host nucleus (N), chloroplast (Ch), host cell wall (W), Golgi apparatus (G), plasmodesmata (), and endoplasmic reticulum (). Bars correspond to 1 µm

From the morphological data, and more significantly, from the following genetic data, we consider the endophytes examined in this study to be new Neotyphodium species, and we propose the names: Neotyphodium aotearoae Moon, Miles et Schardl for the non-sporulating endophytes from Ec. ovatus, Neotyphodium australiense Moon et Schardl for the sporulating endophytes capable of epiphyllous growth from the Australian Ec. ovatus populations, and Neotyphodium melicicola Moon et Schardl for the endophytes from the South African Melica spp.

DNA amplification and sequencing – PCR amplifications of tub2, tef1, and act1 genes from endophyte genomic DNA yielded products of the approximate sizes expected for Neotyphodium isolates. The amplification reactions from several isolates yielded multiple DNA fragments of approximately similar size. These PCR products, when sequenced directly, gave ambiguous signal that is indicative of multiple templates in the sequencing reaction. Thus, individual gene copies were amplified separately with copy-specific primers, then sequenced.

From N. aotearoae isolates, single copies of tub2, tef1, and act1 were amplified and sequenced. Sequences at each gene locus were identical from all isolates. Two copies of each gene were obtained from the N. tembladerae isolate. From the two N. australiense isolates, CBS 109346 and CBS 109347, two copies of tub2 and tef1, but only one act1 gene was amplified, and orthologous gene copies were identical between the isolates. Conversely, two copies of tub2 and act1, but only one copy of tef1 was amplified from the N. melicicola isolates, CBS 109340, CBS 109341, and CBS 109342. For all isolates, the tef1 and act1 copies were identical, but single nucleotide changes were observed in the tub2 copies from isolate CBS 109341, as compared to their counterparts from isolates CBS 109340 and CBS 109342.

Phylogenetic analyses – To determine the evolutionary history of the Southern Hemisphere endophytes, phylogenetic analyses were carried out using tub2, tef1, and act1 sequences from these endophytes, as well as from representative isolates of Epichloë; spp. In general, the topologies of each gene tree indicated Epichloë; relationships similar to those found by Craven et al. (2001b), and the topology of MP trees (Figs. 13–16) found by branch-and-bound search were consistent with NJ trees (not shown).

View larger version (30K): [in this window] [in a new window]     FIG. 13. Phylogram of one of 12 most parsimonious trees resulting from a branch-and-bound search on tub2 gene sequences from representative isolates of Epichloë; spp. and Southern Hemisphere endophytes (bold type). The tree is 126 steps in length, consistency index = 0.8095, retention index = 0.9164, rescaled consistency index = 0.7418, and the left-hand edge is the inferred midpoint root. Most parsimonious trees were fairly similar and branches present in their strict consensus are indicated in bold. The percentage (if >50%) that each clade was represented in 1000 parsimony bootstrap replications is indicated at its respective branch. Host taxon abbreviations following isolate numbers are as follows: Ah (Agrostis hiemalis), Ao (Anthoxanthum odoratum), As (Agrostis stolonifera), At (Agrostis tenuis), Bb (Bromus benekenii), Be (Bromus erectus), Bee (Brachyelytrum erectum), Bps (Brachypodium sylvaticum), Bpp (Br. pinnatum), Cv (Calamagrostis villosa), Dg (Dactylis glomerata), Evr (Elymus virginicus), Ec (Elymus canadensis), Eco (Echinopogon ovatus), Fl (Festuca longifolia), Frr (Festuca rubra subsp. rubra), Gs (Glyceria striata), Hl (Holcus lanatus), Lp (Lolium perenne), Mr (Melica racemosa), Ph (Poa huecu), Php (Phleum pratense), Pn (Poa nemoralis), Pp (Poa pratensis), Ps (Poa silvicola), and So (Sphenopholis obtusata). The number of asterisks (*) indicate the numbers of additional isolates that had identical sequence to the taxon shown

From the tef1 and act1 gene phylogenies (Figs. 14 and 15), the Southern Hemisphere endophytes were grouped consistently with the same Epichloë; and Neotyphodium isolates, as in the tub2 phylogenies. MP analysis on the aligned tef1 sequences yielded 78 most parsimonious trees of 185 steps in length (Fig. 15), and MP analyses on the aligned act1 sequences yielded 106 most parsimonious trees of 164 steps in length (Fig. 16). Both analyses placed genes from N. aotearoae isolates into a distinct clade that was basal to the E. typhina complex (Craven et al 2001b) with 100% bootstrap support. The placement of genes from endophytes containing multiple gene copies was remarkably consistent with that from tub2 analyses. The two tef1 and act1 gene copies from N. tembladerae again grouped with E. festucae genes and the E. typhina genotype from the P. nemoralis isolate. Notably, the N. tembladerae tef1 gene sequence was identical to the E. festucae gene copies. The two tef1 gene copies from the N. australiense isolates were grouped with E. festucae genes and the E. typhina genotype from the P. pratensis isolate. The single N. australiense act1 gene was also placed in a clade with the act1 gene from the same E. typhina isolate from P. pratensis. From N. melicicola, the two act1 gene copies respectively grouped with E. festucae isolates and N. aotearoae, and the single tef1 gene grouped with N. aotearoae. All these relationships were well supported by bootstrap.

View larger version (26K): [in this window] [in a new window]     FIG. 14. Phylogram of one of 78 most parsimonious trees resulting from a branch-and-bound search on tef1 gene sequences from representative isolates of Epichloë; spp. and Southern Hemisphere endophytes (bold type). The tree is 185 steps in length, consistency index = 0.8703, retention index = 0.9520, rescaled consistency index = 0.8285, and the left-hand edge is the inferred midpoint root. Most parsimonious trees were very similar and branches present in their strict consensus are indicated in bold. The percentage (if >50%) that each clade was represented in 1000 parsimony bootstrap replications is indicated at its respective branch. Host taxon abbreviations are as for Fig. 13. The number of asterisks (*) indicate the numbers of additional isolates that had identical sequence to the taxon shown.

View larger version (27K): [in this window] [in a new window]    FIG. 15. Phylogram of one of 106 most parsimonious trees resulting from a branch-and-bound search on act1 gene sequences from representative Epichloë; isolates Southern Hemisphere endophytes (bold type). The tree is 164 steps in length, consistency index = 0.8780, retention index = 0.9424, rescaled consistency index = 0.8274, and the left-hand edge is the inferred midpoint root. Most parsimonious trees were fairly similar and branches present in their strict consensus are indicated in bold. The percentage (if >50%) that each clade was represented in 150 parsimony bootstrap replications is indicated at its respective branch. Host taxon abbreviations are as for Fig. 13. The number of asterisks (*) indicate the numbers of additional isolates that had identical sequence to the taxon shown

View larger version (25K): [in this window] [in a new window]    FIG. 16. Phylogram of one of 12 most parsimonious trees resulting from a branch-and-bound search on combined tub2, tef1, and act1 gene sequences from representative Clavicipitaceous endophytes including Neotyphodium aotearoae (bold type). The tree is 1389 steps in length, consistency index = 0.8647, retention index = 0.8296, rescaled consistency index = 0.7173, and the left-hand edge is the inferred midpoint root. The MP trees were fairly similar and branches present in their strict consensus are indicated in bold. The percentage (if >50%) that each clade was represented in 1000 parsimony bootstrap replications is indicated at its respective branch. Host taxon abbreviations are as for Fig. 13. The number of asterisks (*) indicate the numbers of additional isolates that had identical sequence to the taxon shown

	TAXONOMY

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS TAXONOMY DISCUSSION LITERATURE CITED

 
Neotyphodium aotearoae

C. D. Moon, C. O. Miles et C. L. Schardl, sp. nov. Figs. 1–4.

Coloniae 4–11 mm diametro aetate 5 hebdomadum ad 22 C in PDA, elevatae, moderate vel maxime convolutae, eburneae vel brunneae, plerumque ceraceae aliquando modice coactae. Coloniae reversae brunneae in centro, eburneae ad marginem. Hyphae septatae, 1.5–2.5 µm latae. Conidia in cultura non observata. In superficie hospitis non crescit.

Colony diam 4–11 mm after 5 wk at 22 C on PDA (Figs. 1–4); colonies raised from agar surface, moderately to highly convoluted, light tan to dark brown, usually waxy though occasionally lightly felted, colony margins distinct. Occasional sectors of white felted growth observed (Fig. 2). Colony reverse brown centrally to light tan at margins and fracturing of agar common. Hyphae septate, 1.5–2.5 µm wide. Sporulation in culture not observed. No external growth on host observed.

Etymology. In reference to the Maori name for New Zealand, Aotearoa.

HOLOTYPE. NEW ZEALAND, Waingaro infecting Echinopogon ovatus, Jan. 1995, leg. C. O. Miles, CUP 65623.

Specimens examined. ATCC MYA-1229 from Ec. ovatus, NEW ZEALAND. Northland, Matai Bay, Sept 1998, leg. S. Finch, CUP 65626; ATCC MYA-1230 from Ec. ovatus, Dargaville, Maungaraho Rock, Sep. 1997, leg. P. J. de Lange, CUP 65627; ATCC MYA-1231 from Ec. ovatus, Hunua, Wairoa Falls, Sep. 1997, leg. P. J. de Lange, CUP 65628; ATCC MYA-1232 from Ec. ovatus, Wellington, Wainuiomata stream, Oct. 1997, leg. R. Smith & T. Silbery, CUP 65629; ATCC MYA-1233 from Ec. ovatus, Poor Knights Is., Puweto Valley, Oct. 1997, leg. P. J. de Lange; ATCC MYA-1234 from Ec. ovatus, Taihape, Paengaroa Scenic Reserve, Sep. 1997, leg. P. J. de Lange, CUP 65630; CBS 109344 from the HOLOTYPE, CUP 65623; CBS 109345 from Ec. ovatus, AUSTRALIA. New South Wales, Tallaganda State Forest, Feb. 1997, leg. S. W. L. Jacobs, CUP 65624; CBS 109352 from Ec. ovatus, New South Wales, Mt. Annan Botanical Garden, Jan. 1997, leg. S. W. L. Jacobs, CUP 65625.

Neotyphodium australiense

C. D. Moon et C. L. Schardl, sp. nov. Figs. 5 and 6.

Coloniae 13–19 mm diametro aetate 5 hebdomadum ad 22 C in PDA, elevatae, leviter convolutae in centro, complanatae peripheriam versus, modice coactae, albidae vel eburneae, annulis concentricis pigmentiferis; coloniae reversae albidae. Hyphae septatae, 2–3 µm latae. Conidia in cultura moderate abunda; cellulae conidiogenae 11–25 µm longae, ca 2 µm latae ad basim, ca 1 µm angustatae ad apicem, non septatae ad basim; conidia ellipsoidea vel lunata, 5–7 x 3–4 µm. Mycelium epiphyllum sparsum superficiei folii insidiens. Affinitas genetica ad Epichloëm festucae Leuchtm. Schardl & Siegel et Epichloë typhinam (Pers. : Fr.) Tul. ex Poa pratensi L.

Colony diam 13–19 mm after 5 wk at 22 C on PDA (Fig. 5); colonies raised, slightly convoluted in center, and flattening toward perimeter; lightly felted, off-white to light tan, with pigmented concentric rings; colony margins superficial; colony reverse off-white with some fracturing of agar medium. Hyphae septate, 2–3 µm wide, though occasionally bulging to 4–5 µm wide. Sporulation in culture moderately abundant; conidiogenous cells 11–25 µm long, ca 2 µm wide at base, tapering to ca 1 µm, basal septum not observed; conidia ellipsoidal to lunate, 5–7 x 3–4 µm. Sparse epiphyllous mycelium present on leaf surface. The species shows genetic similarity to Epichloë festucae Leuchtm. Schardl et Siegel and Epichloë typhina (Pers. : Fr.) Tul. isolates from Poa pratensis L.

Etymology. Referring to the geographic origin of Australia.

HOLOTYPE. AUSTRALIA. New South Wales, New England National Park, infecting Echinopogon ovatus, Jan. 1999, leg. A. Martyn, S. W. L. Jacobs & J. Dalby, CUP 65631.

Specimens examined. CBS 109346 from the HOLOTYPE, CUP 65631; CBS 109347 from Ec. ovatus, New South Wales, New England National Park, Jan. 1999, leg. A. Martyn, S. W. L. Jacobs & J. Dalby, CUP 65632.

Neotyphodium melicicola

C. D. Moon et C. L. Schardl, sp. nov. Figs. 7 and 8.

Coloniae 10–19 mm diametro aetate 5 hebdomadum ad 22 C in PDA, leviter vel moderate convolutae, coactae, albae; coloniae reversae flavidae. Hyphae septatae, 2–3.5 µm latae. Conidia in cultura sparsae; cellulae conidiogenae 6–35 µm longae, ca 2.5 µm latae ad basim, ca 1 µm angustatae ad apicem, aliquando septatae ad basim; conidia ellipsoidea vel lunatae, 5–8 x 3.5–5 µm. In superficie hospitis non crescit. Affinitas genetica ad Neotyphodium aotearoae Moon, Miles & Schardl et Epichloëm festucae Leuchtm., Schardl & Siegel.

Colony diam 10–19 mm after 5 wk at 22 C on PDA (Fig. 7); colonies slightly to moderately convoluted, white, felted, margins distinct to superficial; colony reverse light yellow. Hyphae septate, 2–3.5 µm wide. Sporulation in culture sparse; conidiogenous cells 6–35 µm long, ca 2.5 µm wide at base, tapering to ca 1 µm at tip, basal septum observed occasionally; conidia ellipsoidal to lunate, 5–8 x 3.5–5 µm. No external growth on host observed. Shows genetic similarity to Neotyphodium aotearoae Moon, Miles et Schardl and Epichloë festucae Leuchtm., Schardl et Siegel.

Etymology. Referring to the host genus Melica.

HOLOTYPE. SOUTH AFRICA. infecting Melica racemosa, Aug. 1999, leg. C. D. Moon, CUP 65633.

Specimens examined. CBS 109340 from the HOLOTYPE, CUP 65633; CBS 109341 from M. racemosa, SOUTH AFRICA. Aug. 1999, leg. C. D. Moon, CUP 65634; CBS 109342 from M. decumbens, leg. C. D. Moon, CUP 65635.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS TAXONOMY DISCUSSION LITERATURE CITED

 
Three new Southern Hemisphere Neotyphodium species are proposed – In this study the evolutionary origins of Neotyphodium endophytes from a range of grasses that are indigenous to the Southern Hemisphere were investigated using a multiple-gene phylogenetic approach. A phylogenetic species concept was employed to complement morphological characterizations of the endophytes, and three new Neotyphodium endophyte species were identified and described. Two species were associated with Australasian Echinopogon ovatus populations: N. aotearoae, an apparently non-hybrid species, was found in both New Zealand and Australia, and N. australiense, an apparent interspecific hybrid of E. festucae and E. typhina ancestry, was detected in Australia. An endophyte species described as N. melicicola infecting South African Melica racemosa and M. decumbens, appeared to be a hybrid of E. festucae and N. aotearoae or close relatives. Additionally, the evolutionary origin of N. tembladerae, an endophyte from Poa huecu in South America, was also investigated, and this fungus was discovered to be a hybrid. Similar to N. australiense, the ancestors of N. tembladerae likely involved E. festucae and E. typhina, though in this case the E. typhina-like sequences were more similar to those of isolates from Poa nemoralis whereas those from N. australiense showed greater affinity to genes of isolates from Poa pratensis.

Evidence for interspecific hybrid status – Interspecific hybrid endophytes were identified by the presence of multiple tub2, tef1, and act1 gene copies that consistently grouped with disparate Epichloë; and Neotyphodium clades within the Epichloë; phylogeny. We have substantial evidence that the hybrids identified in this study are true nuclear hybrids, and that the multiple gene copies detected are not the result of heterokaryosis or mixed cultures. The conidia of hybrid endophytes were confirmed as uninucleate by nuclear staining. We also ensured that cultures were pure by multiple rounds of single-spore isolation. In the case of N. melicicola, single-spore isolations were not performed due to technical difficulties. However, we believe it is highly unlikely that these cultures were mixed. Within plants infected with multiple endophyte strains, individual strains generally segregate out at the level of the tiller (Meijer and Leuchtmann 1999, Wille et al 1999, Christensen et al 2000) and in this study endophytes were isolated from single leaves. Additionally, the three N. melicicola isolates analyzed had almost identical genotypes, though they had originated from three geographically distinct M. racemosa and M. decumbens populations. It would therefore seem unlikely that the same combination of strains would be present in plants of each of these populations.

Another line of evidence that lends support to the true hybrid status of N. australiense, N. melicicola, and N. tembladerae is conidium size, which has been suggested as an indicator of nuclear content (Kuldau et al 1999). It is expected that hybrid endophytes would have larger genomes than their ancestral counterparts as a result of polyploidy or aneuploidy (Kuldau et al 1999). Indeed, in previous studies, the conidia of hybrid Neotyphodium species have been larger than those of ancestral Epichloë; spp. (Kuldau et al 1999, Craven et al 2001a). Likewise, in this study we observed that conidia of the hybrid Southern Hemisphere Neotyphodium species (average size 6.5 ± 0.7 µm x 3.9 ± 0.5 µm) were larger than those of their ancestral species (5.3 ± 0.4 µm x 2.9 ± 0.3 µm; Craven et al 2001a). Interestingly, the non-hybrid, N. aotearoae, was the only species in this study from which conidia were not observed. Compared with other non-hybrid Neotyphodium species, such as N. lolii, which infrequently produces conidia (Christensen et al 1993), and N. typhinum var. canariense, from which conidia have not been observed (Moon et al 2000), there appears to be a correlation between non-hybrid origin and sparse, or absent, conidia production in Neotyphodium endophytes.

The use of the tub2, tef1, and act1 genes for phylogenetic analyses within the grass endophyte group has been shown to be highly informative (Craven et al 2001a, b). A wealth of sequence data from these genes is available and phylogenies constructed from each gene have elucidated patterns of evolution between Epichloë; spp. (Craven et al 2001a). Phylogenies based on each gene were fairly concordant with one another, with the exception of a few reticulating clades (Craven et al 2001a). Additionally, the tub2, tef1, and act1 genes are all single copy in non-hybrid endophytes (Craven et al 2001a), and are, therefore, suitable for the characterization of hybrid endophytes. Indeed, these genes have been used together to determine the evolutionary origins of the hybrid Neotyphodium species N. uncinatum and N. siegelii (Craven et al 2001b).

Hybridization is common in the generation of Neotyphodium species – The discovery of interspecific hybrid Neotyphodium species from the Southern Hemisphere is fascinating, and suggests that interspecific hybridization among Epichloë; and Neotyphodium endophytes is a global phenomenon. Including those found in this study, ten Neotyphodium spp. of interspecific hybrid origin have been reported so far (Schardl et al 1994, Tsai et al 1994, Moon et al 2000, Craven et al 2001a), and several others have been identified (Moon 1999, CD Moon and KD Craven unpubl). Currently, only three strictly anamorphic Neotyphodium spp. that have been analyzed by multi-gene phylogenetics appear to be non-hybrid: N. lolii, N. typhinum, and N. aotearoae. This observation suggests that interspecific hybridization is likely the most common mechanism for the evolution of asexual species of these grass endophytes.

The mechanics of interspecific hybridization in endophyte evolution are unknown, but conceivably involve superinfection of an infected grass plant, followed by anastomosis (hyphal fusion) of the resident and immigrant endophytes, then karyogamy (nuclear fusion) (Schardl et al 1994). Indeed, vegetative compatibility between different Epichloë; spp. has been demonstrated in vitro (Chung and Schardl 1997b). Following karyogamy, the new polyploid genome may lose chromosomes or chromosomal segments containing redundant genetic information, and give rise to an aneuploid genome (Schardl et al 1994). This loss of genetic information is likely to explain the presence of only single copies of genes in hybrid endophytes. For example, in N. australiense there was only one act1 gene copy, and only a single tef1gene was detected from N. melicicola, though both hybrids contained two copies of the other two genes analyzed.

That most or all interspecific hybrid endophytes are nonpathogenic and asexual is very interesting. The growth regulation of Epichloë; spp. during stroma development presumably involves important signaling events that are dependent on the physiological state of the host, and requires precise coordination with development of the inflorescence (Kirby 1961). Such an intricate interaction may be disrupted or aborted in the case of hetero- or aneuploid hybrid endophytes where differences in gene dosage, compared to the haploid sexual ancestor, may be a critical factor during this process. However, even if stromata were produced, other obstacles would need to be overcome for successful mating. A compatible endophyte of the opposite mating type must be available. Furthermore, chromosomal pairing between the hybrid and partner endophyte would need to be accomplished for meiosis. Thus, it is likely that the sexual cycle was lost as a result of the hybridization.

Evolutionarily, interspecific hybridization processes in asexual endophytes may be significant factors in counteracting Muller's ratchet (Schardl 1996), whereby without sexual recombination, an asexual species will eventually accumulate deleterious mutations that would cause a loss of fitness (Muller 1964). Interspecific hybridization may provide increased fitness by imparting fitness-enhancing properties to the endophyte and symbiosis, such as new or additional pathways for the biosynthesis of protective alkaloids. Furthermore, the presence of multiple gene copies may provide a buffer against the disruption of gene function, as deleterious mutations would be required in all gene copies.

A possibly common evolutionary origin for biosynthesis of tremorgenic compounds – The tremorgenic indole-diterpenoid mycotoxin, lolitrem B, is documented as the major causative factor of ryegrass staggers in livestock grazing N. lolii infected perennial ryegrass (Lolium perenne L.) (Fletcher and Harvey 1981, Gallagher et al 1981). Unidentified indole-diterpenoid compounds have also been detected from endophyte infected Ec. ovatus, M. decumbens, and P. huecu (Miles et al 1998, Lane et al 2000), which likewise have all been associated with staggers-like maladies of grazing animals. A common evolutionary link between the endophyte species of these grasses (i.e., N. lolii, N. australiense, N. melicicola, and N. tembladerae) is that each has an apparent E. festucae ancestor. Additionally, lolitrem B has been detected in tall fescue associated with a Neotyphodium sp. that has been referred to as "Festuca arundinacea taxonomic group two" (FaTG-2) (Christensen et al 1993, Lane et al 2000). The ancestry of FaTG-2 is hybrid, and involves E. festucae as well as E. baconii lineages (Tsai et al 1994). In contrast, the only sexual Epichloë; species known to produce lolitrems is E. festucae (Siegel et al 1990, Lane et al 2000). Thus, it is tempting to speculate that genes encoding enzymes for the biosynthesis of indole-diterpenoids may have originated from an E. festucae lineage. However, indole-diterpenoid compounds have been detected from endophyte-infected Festuca versuta (Siegel et al 1990), Achnatherum inebrians (Miles et al 1996) and Echinopogon ovatus associated with N. aotearoae in New Zealand (Miles et al 1998). The evolutionary origins of the F. versuta endophyte have not been reported. However, both N. aotearoae and the endophyte of A. inebrians do not appear to involve E. festucae (Figs. 13–16; Moon, unpubl obs). The potential associations of indole-diterpenoid biosynthesis with E. festucae ancestry may be assessed once endophyte genes involved in indole-diterpenoid biosynthesis are identified. Indeed, genes involved in paxilline biosynthesis (a presumed precursor to lolitrem B) have been isolated from Penicillium paxilli (Young et al 2001), and orthologues of these genes have recently been cloned from N. lolii (Scott 2001).

In this study, we identified two Neotyphodium spp., N. australiense and N. aotearoae, that infect Australian Ec. ovatus populations. It is not known which of these endophytes were associated with Ec. ovatus that caused staggers in Australia (Seddon and Carne 1926), and the possibility that this disorder could be attributed to another endophyte species cannot be excluded. However, we have since determined that the species infecting the Australian Ec. ovatus plants examined by Miles et al (1998) were N. aotearoae, and indole-diterpenoid compounds were not detected from these associations. In contrast, unidentified indole-diterpenoids were detected from one N. aotearoae-infected Ec. ovatus population from New Zealand (Miles et al 1998). Whether tremorgenic compounds are produced by N. australiense-infected Ec. ovatus has not yet been determined.

Apart from indole-diterpenoids, very little has been reported on the presence of other main classes of secondary metabolites in endophyte-infected grasses of the Southern Hemisphere (Lane et al 2000). However, loline alkaloids have been detected in New Zealand Ec. ovatus associated with N. aotearoae (Miles et al 1998).

An Epichloë; endophyte lineage indigenous to the Southern Hemisphere – The endophyte lineage that gave rise to N. aotearoae and the N. aotearoae-like genes of N. melicicola possibly represents an Epichloë; or Neotyphodium lineage that is unique to the Southern Hemisphere. The consistently basal placement of this lineage within Epichloë; gene phylogenies suggests that it had diverged relatively early in the evolution of the grass endophytes. We consider two alternate possibilities to explain the observed distribution of the N. aotearoae and N. aotearoae-like genotypes. The first possibility is that the N. aotearoae lineage co-evolved with either the Melica or Echinopogon lineages, and was subsequently horizontally transmitted to the other. This scenario implies that the N. aotearoae lineage was capable of contagious spread, most likely via sexual processes. Whether such a sexual lineage still exists is unknown, but the question may possibly be resolved with more intensive sampling in South Africa and Australasia. It has also been postulated that contagious spread may be possible via conidia produced from epiphyllous hyphae on the host epidermis (White et al 1996). With N. aotearoae, this scenario seems unlikely as epiphyllous growth was not detected on hosts infected with either N. aotearoae, nor N. melicicola. A second possibility to explain the distribution of the N. aotearoae and N. aotearoae-like genotypes is that the N. aotearoae lineage is confined to vertical transmission. If this were so, the lineage must have been present in a grass ancestral to Melica and Echinopogon. Given that these two genera are thought to be from evolutionarily disparate grass tribes (tribes Meliceae and Aveneae, respectively) within the Poöideae (Hsiao et al 1995, Catalán et al 1997), and that there are little or no differences in the endophyte genes examined over this course of time, this scenario would appear unlikely. Thus, from the geographic distribution of the N. aotearoae and N. aotearoae-like genes, it seems likely that transoceanic gene-flow has occurred some time in the past.

Another point of interest is that the nine N. aotearoae isolates examined in this study had identical gene sequences at each of the tub2, tef1, and act1 genes, despite the substantial morphological variability between isolates, and the broad geographic range of the sample population. The significance of this extreme genetic similarity by N. aotearoae may indicate very recent geographic distribution of infected Ec. ovatus, or possibly horizontal transmission of isolates.

Other genotypes found in the Southern Hemisphere are closely related to those from Northern Hemisphere – In contrast to the N. aotearoae lineage, other genotypes detected from the Southern Hemisphere endophyte isolates were closely related to lineages found in the Northern Hemisphere. In particular, E. festucae-like genes were found in all three hybrid Southern Hemisphere species, whereas E. festucae nor any other Epichloë; spp. have been reported to be indigenous to the Southern Hemisphere. Additionally, E. typhina-like genes, closely related to those from European E. typhina isolates from Poa nemoralis and P. pratensis, were detected in N. tembladerae and N. australiense, respectively. These results suggest that gene-flow between the hemispheres is not rare, contrary to the notion that tropical regions may serve as a barrier to gene-flow for cool-season grasses (Cabral et al 1999). Indeed, cool-season grasses occur in montane forests of Central America (Darbyshire and Warwick 1992). Furthermore, from a phylogenetic and biogeographical study of the grass genus Poa, at least two groups of Poa species were likely dispersed from North to South America (Soreng 1990). Additionally, the radiation of a group of Poa spp. from North America to Australasia was supported by both genetic and morphological data (Soreng 1990). Other mechanisms that would possibly allow transequatorial and transoceanic gene-flow may include transport of endophyte-infected grass plants or seeds by rafting, or carriage by migratory birds. To further understand the evolution of Southern Hemisphere endophytes and mechanisms of endophyte gene flow on a global scale, it is clear that more intensive endophyte sampling and genotyping need to be performed from this part of the world.

	ACKNOWLEDGMENTS

 
We wish to thank A. Leuchtmann for reviewing the manuscript and writing Latin descriptions, W. Gams and an anonymous reviewer for reviewing the manuscript, and J. Dalby, S. Finch, S.W.L. Jacobs, F. Kellerman, P.J. de Lange, A. Martyn, R. Smith, and T. Silbery and for the collection of seed from Echinopogon ovatus and Melica spp. We also wish to thank W. Hollin and A.D. Byrd for able assistance. This research was supported by United States National Science Foundation grant DEB-9707427.

	FOOTNOTES

 
¹ This is contribution number 02–12–012 of the Kentucky Agricultural Experiment Station, published with approval of the Director.

² Corresponding author, Email: schardl{at}uky.edu

Accepted for publication February 1, 2002.

	LITERATURE CITED

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS TAXONOMY DISCUSSION LITERATURE CITED

 
Al-Samarrai TH, Schmid J., 2000 A simple method for extraction of fungal genomic DNA. Lett Appl Microbiol 30:53-56[Medline]

Arechevaleta M, Bacon CW, Hoveland CS, Radcliffe DE., 1989 Effect of the tall fescue endophyte on plant response to environmental stress. Agron J 81:83-90[Abstract/Free Full Text]

Bacon CW, Lyons PC, Porter JK, Robbins JD., 1986 Ergot toxicity from endophyte-infected grasses: a review. Agron J 78:106-116[Abstract/Free Full Text]

———, Porter JK, Robbins JD, Luttrell ES., 1977 Epichloë typhina from toxic tall fescue grasses. Appl Environ Microbiol 34:576-581[Abstract/Free Full Text]

Brem D, Leuchtmann A., 1999 High prevalence of horizontal transmission of the fungal endophyte Epichloë sylvatica. Bulletin of the Geobotanical Institute ETH 65:3-12

Bucheli E, Leuchtmann A., 1996 Evidence for genetic differentiation between choke-inducing and asymptomatic strains of the Epichloë; grass endophyte from Brachypodium sylvaticum. Evolution 50:1879-1887

Bush LP, Fannin FF, Siegel MR, Dahlman DL, Burton HR., 1993 Chemistry, occurrence and biological effects of saturated pyrrolizidine alkaloids associated with endophyte-grass interactions. Agriculture, Ecosystems and Environment 44:81-102

Cabral D, Cafaro MJ, Saidman B, Lugo M, Reddy PV, White JF Jr., 1999 Evidence supporting the occurrence of a new species of endophyte in some South American grasses. Mycologia 91:315-325

Catalán P, Kellogg EA, Olmstead RG., 1997 Phylogeny of Poaceae subfamily Pooideae based on chloroplast ndhF gene sequences. Mol Phylogenet Evol 8:150-166[Medline]

Christensen MJ, Leuchtmann A, Rowan DD, Tapper BA., 1993 Taxonomy of Acremonium endophytes of tall fescue (Festuca arundinacea), meadow fescue (F. pratensis) and perennial rye-grass (Lolium perenne). Mycol Res 97:1083-1092

———, Simpson WR, Al Samarrai T., 2000 Infection of tall fescue and perennial ryegrass plants by combinations of different Neotyphodium endophytes. Mycol Res 104:974-978

Chung K-R, Schardl CL., 1997a Sexual cycle and horizontal transmission of the grass symbiont, Epichloë typhina. Mycol Res 101:295-301

———, ———. 1997b Vegetative compatibility between and within Epichloë; species. Mycologia 89:558-565

Craven KD, Blankenship JD, Leuchtmann A, Hignight