Microb Ecol (2010) 59:744–756 DOI 10.1007/s00248-010-9652-3

PLANT MICROBE INTERACTIONS

Genetic Diversity and Structure of Neotyphodium Species and Their Host Achnatherum sibiricum in a Natural Grass– System

Xin Zhang & Anzhi Ren & Huacong Ci & Yubao Gao

Received: 8 June 2009 /Accepted: 26 February 2010 /Published online: 30 March 2010 # Springer Science+Business Media, LLC 2010

Abstract Achnatherum sibiricum () is a perennial even when their gene flows do not match each other. bunchgrass native to the Inner Mongolia Steppe of China. Furthermore, we suggested that the same genotype of This grass is commonly infected by epichloë endophyte as well as host should be confirmed if the with high-infection frequencies. Previously, we identified objective of the study is to know the influence of endophyte two predominant Neotyphodium spp., N. sibiricum and N. or host genotype on their symbiotic relationship, instead of gansuense. In the present study, genetic diversity and just considering whether the plant is infected by an structure were analyzed for the two predominant Neo- endophyte or not, since endophytes from the same host typhodium spp. as well as the host grass. We obtained 103 species could exhibit high levels of genetic diversity, which fungal isolates from five populations; 33 were identified as is likely to influence the outcome of their symbiotic N. sibiricum and 61 as N. gansuense. All populations relationship. hosted both endophytic species, but genetic variation was much higher for N. gansuense than for N. sibiricum. The majority of fungal isolates were haploid, and 13% of them Introduction were heterozygous at one SSR locus, suggesting hybrid origins of those isolates. Significant linkage disequilibrium The Epichloë spp. and their asexual descendants Neo- of fungal SSR loci suggested that both fungal species typhodium spp.—collectively called epichloë endophytes— primarily propagate by clonal growth through plant seeds, comprise a group of filamentous fungi (, whereas variation in genetic diversity and the presence of ) that form symbiotic infection within many hybrids in both endophytic species revealed that although cool-season grasses [20]. The Epichloë species typically clonal propagation was prevalent, occasional recombination exhibit a sexual stage with the development of stromata might also occur. By comparing genetic differentiation around the base of plant flag leaf followed by the among populations, we found around 4–7-fold greater reproductive sterilization of the host, often known as plant differentiation of endophyte populations than host popula- choke disease [53, 54]. However, in the asexual stage of tions, implying more restricted gene flow of endophytes Epichloë spp. and all Neotyphodium spp., the does than hosts. We proposed that endophyte infection of A. not produce any external structure and is transmitted sibiricum might confer the host some selective advantages through seeds of infected plants [43]. under certain conditions, which could help to maintain The symbiosis between epichloë endophytes and their high-endophyte-infection frequencies in host populations, hosts has been suggested to range from antagonistic to mutualistic [32, 38]. Some obligate horizontally transmitted Epichloë spp. are more likely to be antagonistic to the host Xin Zhang and Anzhi Ren contributed equally to this work. : : : because the reproduction of the host is aborted upon X. Zhang A. Ren H. Ci Y. Gao (*) formation of stromata, whereas many vertically transmitted Department of Plant Biology and Ecology, Neotyphodium spp. are proposed as plant mutualists since College of Life Sciences, Nankai University, Tianjin 300071, People’s Republic of China the fitness of the endophyte and the host is closely related e-mail: [email protected] [42]. Early studies of grass–endophyte associations using Diversity and Structure of Endophytes and Host 745 two economically important grass species, perenne order to predict their symbiotic relationship, the ecological and L. arundinaceum (= Schedonorus arundinaceus= dynamics of the plant communities as well as their ) suggested that endophytes were evolutionary processes. plant-defending mutualists [11, 41]. The benefits conferred In China, there have been few studies on endophytic to host by the endophyte are well understood, which fungi associated with native grasses, despite the fact that typically include protection from mammalian and insect many grass species have been found to be infected by herbivory [5, 27], disease resistance [25], stress tolerance endophytes [51]. The most studied grass is Achnatherum [2] and improved growth and reproduction [29, 35]. For inebrians, a widely distributed endophyte-infected grass native grass–endophyte associations, however, their inter- notorious for its toxicity to livestock [6]. In our previous actions are more variable. In some grass species, endophyte extensive survey in the Inner Mongolia Steppe of China, infection could benefit their hosts under stress conditions another species, A. sibiricum (L.) was also found to be [6, 12, 24, 29, 33, 44], while in other studies, endophyte- commonly infected by epichloë endophytes with high- infected grasses do not appear to be competitive over those infection frequency [51]. In a previous study, N. sibiricum that are endophyte free [17, 18]. In natural populations, the and N. gansuense were identified as two predominant costs and benefits of harboring endophytes may vary with endophytic species in A. sibiricum populations [56]. The genotype of both plant and endophyte, as well as with objective of the present study is to further evaluate the environmental conditions [32, 38]. Among those factors genetic diversity and population genetic structure of the two that may influence the outcome of endophyte–grass predominant Neotyphodium spp. in different A. sibiricum interactions, endophyte genotype and the interacting host populations, as well as the relationship of genetic differenti- genotype has proved to be most critical [30]. ation and geographical distance between populations. Further- It is reported that in native grass populations, endophytic more, we compared the genetic structure of the host and fungal communities may be more diverse and variable than endophytes to better understand the movement of genes and that of agronomic and economically important species [45]. how this movement may influence population genetic struc- This reflects on the endophyte species diversity, and also ture and therefore the co-evolution of the host and endophytes. the endophyte morphological and genetic diversities [9, 10, 13, 45]. For example, only one fungal genotype was transmitted through an economically important cultivar of Materials and Methods tall fescue, Kentucky 31 (KY-31) [39]. In contrast, considerable morphological, physiological and biochemical Sampling of the Plant Materials variations were observed among endophytic isolates grown from native tall fescue collected worldwide [10, 34]. In In a previous extensive survey, A. sibiricum was found at only another endophyte-infected native grass, Arizona fescue eight geographical locations in the Inner Mongolia Steppe (Festuca arizonica), >400 different fungi were isolated [51]. In the present study, A. sibiricum from five natural from five Arizona fescue populations, which might rival populations, IMGERS-CAS, Xi Ujimqin Qi, Huolingol, fungal diversity in woody plants [45]. In the study of ArshanandHailarinthemid-andeasternInner endophyte genetic diversity, marked genetic variation has Mongolia Steppe were chosen (Fig. 1). Populations were been detected in several epichloë species studied, including selected from the western most distribution of A. sibiricum E. festucae from F. rubra and Festuca subg. Festuca, N. to the east and covered various plant community types, lolii from L. perenne and N. starrii from F. arizonica [3, 14, including meadow steppe, typical steppe and forest edge. 46, 50]. There are several processes by which endophytes Plant material was collected in July 2006. Within each could obtain their genetic variation, which typically include population, 22 (Xi Ujimqin Qi) to 31 (Hailar) plant sexual reproduction of Epichloë spp., interspecific hybrid- individuals were randomly chosen along a line transect. ization of Neotyphodium spp., as well as mitotic or The distance between sampled plant individuals was at parasexual recombination and mutation accumulation [13, least 5 m to minimize the probability of sampling ramets 28, 46, 49, 50]. Moreover, studies of their population belonging to the same genet. Leaves of each plant were genetic structure detected much higher levels of population harvested and stored in silica gel in locked bottles before differentiation of E. festucae than N. starrii. The occasional transporting back to the laboratory for DNA extraction. sexual reproduction and sporulation may facilitate gene Seeds from each plant were collected in September 2006 flow among endophyte populations and thus reduce and deposited in separate envelopes for fungal endophyte population differentiation level of E. festucae [3, 46, 50]. detection and isolation. Only one stroma each was found Considering the influence of endophyte genotype on the for the Xi Ujimqin Qi, Huolingol and Arshan populations, performance of the symbiont, it is important to determine separately, after screening several hundred plant individuals in the diversity of the endophyte and the interacting grasses in each population. 746 X. Zhang et al.

Figure 1 Locations of the studied A. sibiricum populations. The five Approximate distances between sites are: Hailar to Arshan, 215 km; sampling sites are marked with solid squares. Location of Inner Arshan to Huolingol, 184 km; Huolingol to Xi Ujimqin Qi, 170 km; Mongolia Autonomous Region within China is showed on top left. Xi Ujimqin Qi to IMGERS-CAS, 147 km

Endophyte Isolation, DNA Extraction and Amplification DNA. All reagents used above were purchased from of Microsatellites Sangon (Shanghai, China). PCR reactions were performed on a Biometra T1 thermocycler (Göttingen, Germany) with Seeds from each A. sibiricum individual were used for an initial incubation at 95°C for 5 min, followed by 30 endophyte isolation following the method of Wei et al. [52]. cycles with denaturation at 94°C for 30 s, annealing at 56°C Fungal isolates were identified as N. sibiricum or N. for 30 s, and extension at 72°C for 1 min; this was followed gansuense according to their colonial morphology on PDA by a final extension reaction at 72°C for 10 min. medium after growing at 25°C in the dark for 4 weeks, as Plant genomic DNA from silica-gel-dried leaves was well as by comparing their tub2 and actG sequences with extracted using a modification of the protocol of standard BLAST searches in GenBank. Isolates other than N. phenol–chloroform method [23]. The primers used in micro- sibiricum and N. gansuense were excluded from analysis. satellite amplification in plant were synthesized according to Finally, 17–20 isolates from each population were used for Chen et al. [7, 8]. Four primer pairs (Asi003, Ain026, Ain041 the microsatellite analysis. Among the populations, there and Ain069) were finally selected. The PCR reactions were were 2–15 isolates of N. sibiricum and 2–18 of N. gansuense. carried out using 10 mmol L−1 PCR buffer, 1.5 mmol L−1 −1 −1 Fungal genomic DNA was prepared following the MgCl2,300μmol L of each dNTP, 300 nmol L of each method of Guo et al. [22]. Four pairs of published primer, 0.5 U Taq DNA polymerase and 20 ng of plant microsatellite primers: NCESTA1DH04, NCESTA1RD04, genomic DNA in a 20 µL reaction volume. Amplification NCESTA1LB02 and NCESTA1LF05 [13] were synthe- was performed as described by Chen et al. [7, 8]. sized by Sangon Biological Engineering Technology & Five microliters of PCR products amplified from Services Co. Ltd (Shanghai, China) for amplification. PCR genomic DNA of both endophyte and plant was visualized reactions were carried out in a final volume of 20 µL, by electrophoresis on a 1.5% agarose gel and stained with −1 −1 containing 10 mmol L PCR buffer, 1.5 mmol L MgCl2, ethidium bromide to check the results of PCR amplification 200 μmol L−1 of each dNTP, 200 nmol L−1 of each primer, and determine the approximate size of the amplified 1 U Taq DNA polymerase and 50 ng of fungal genomic fragments. The PCR products were then denatured in Diversity and Structure of Endophytes and Host 747 formamide loading buffer at 95°C for 5 min before absence (0) of bands for each locus to generate a matrix. separating on a 6% polyacrylamide denaturing gels of Genetic similarity based on simple matching coefficient 410×330×3.5 mm (length×width×thickness) in size for was calculated using the SIMQUAL module. The generated approximately 3 h at 1500 V. Following electrophoresis, the similarity matrix was then analyzed using the UPGMA gel was silver-stained according to the protocol of Bassam clustering method. In addition, the neighbor-joining (NJ) et al. [4]. After staining, the gel was air-dried and photo- method was used to estimate the dendrograms. All dendro- graphed with a digital camera (Nikon E5000). grams were created with the TREE program of NTsys.

Data Analysis Results The amplification results were analyzed using the Gene Genius BioImaging System (SynGene), and the size of Genetic Diversity of Endophytes fragments was estimated with the φX 174-Hae III digest DNA marker from Takara Biotechnology (Dalian, China). Of the 103 isolates obtained from five populations, 33 were Because of the co-dominant nature of SSR markers, the identified as N. sibiricum and 61 as N. gansuense (Table 1). amplified SSR bands representing different alleles of All populations hosted both endophytic species. The individuals at each locus were scored as different geno- remaining nine isolates were non-epichloë isolates and types. Bands were coded as diploid data and loci with were excluded from the microsatellite analysis. single band were considered as homozygous. Genetic The four SSR markers had 4–7 alleles per locus across diversity is defined here as the probability that two all isolates. Within 94 fungal isolates analyzed, two alleles randomly chosen haplotypes are different in the sample. at locus NCESTA1DH04 were detected in 12 (six of N. This value is zero if every individual is the same and one if sibiricum and six of N. gansuense) isolates, whereas only a every individual is different. Diversity statistics for N. single allele per locus was detected in the remaining sibiricum, N. gansuense and A. sibiricum in each popula- isolates (Table 1). Isolates with multi-alleles per locus were tion and in all populations, including percentage of unique only found in the Huolingol and Hailar populations. haplotypes, percentage of polymorphic loci and genetic Therefore, in the following analysis, each genotype was diversity were calculated separately using Arlequin soft- coded as diploid data, where loci with single allele were ware version 3.1 [16]. In addition, analysis of molecular treated as homozygous. Overall, 32 unique genotypes were variance (AMOVA) of haplotypic data was also conducted identified in the sample of 94 fungal isolates (Table 1). The using Arlequin to reveal genetic structure among and within sample of 33 individuals of N. sibiricum which accounted populations. The FST coefficient obtained from the AMOVA for 35.11% of the total fungal isolates had seven (21.87%) was used to estimate genetic differentiation among popula- haplotypes, whereas the 61 individuals of N. gansuense tions. For both the host and the two endophytic species, we (64.89% of the total fungal isolates) had 25 (78.13%) carried out a Mantel test between the matrices of pairwise haplotypes (Table 2).

FST and pairwise geographical distance to test the isolation The IMGERS-CAS population was dominated by N. by distance. Pairwise geographical distances were calculated sibiricum, whereas the Huolingol, Arshan and Hailar as linear distance between populations. populations were dominated by N. gansuense. In the Xi

The index of association (IA) used to test for multilocus Ujimqin Qi population, approximately the same number of linkage disequilibrium, i.e. the non-random association of N. sibiricum and N. gansuense were obtained (Table 2). All alleles among loci, was calculated using Multilocus version the parameters were calculated separately for the two 1.3 [1], within each population and across all populations. endophytic species, as well as for all fungal isolates pooled

IA is a generalized measure of linkage disequilibrium and together. The number of haplotypes, frequency of unique IA=0 if there is no association between the loci [26]. haplotypes and genetic diversity all indicated that isolates Moreover, an alternative standardization for the covariances of N. gansuense had higher genetic diversity than isolates

(rd) was inferred using the index of association modified to of N. sibiricum (Table 2). The single-allelic genotype H2 remove the dependency of the number of loci analyzed. was a common genotype in N. sibiricum; it was dominant Tests of significance by 1,000 randomizations of the data in three of the five populations investigated and had a set were performed to determine if IA and rd differed frequency of 75.76% of all genotypes in N. sibiricum. significantly from zero. Moreover, genetic diversity of N. sibiricum varied greatly Relationships among individual fungal isolates (isolates among populations. In the IMGERS-CAS and Xi Ujimqin of N. sibiricum and N. gansuense were pooled together) Qi populations, all fungal isolates shared the same genotype were assessed based on SSR markers using NTsys program (H2) and thus there was no genetic variation within these ver. 2.1 [36]. SSR data were scored as presence (1) or populations. However, in the Huolingol and Hailar popula- 748 X. Zhang et al.

Table 1 Allele coding used for the definition of Neotyphodium genotypes from five A. sibiricum populations

Haplotypes Number of isolates Allele sizes of each locus

N. sibiricum N. gansuense NCESTA1DH04 NCESTA1RD04 NCESTA1LB02 NCESTA1LF05

H1 2 227 378 326 239 H2 25 227 378 323 215 H3 4 227 378 326 206 H4 2 227 378 321 215 H5 4 227 381 326 206 H6 7 227 384 329 206 H7 5 227 381 329 206 H8 1 227 378 329 206 H9 2 227 381 326 215 202 H10 1 227 381 321 215 202 H11 1 227 381 323 215 202 H12 1 227 384 323 215 199 H13 5 224 378 326 206 H14 1 224 378 328 206 H15 5 224 381 329 206 H16 5 224 381 323 206 H17 1 224 375 326 206 H18 1 224 384 331 206 H19 1 224 384 329 206 H20 1 224 378 329 206 H21 1 224 386 326 206 H22 1 224 381 326 206 H23 1 227 384 323 206 H24 1 227 384 326 206 H25 1 233 375 326 206 H26 3 230 378 329 206 H27 1 230 381 329 206 H28 1 230 381 326 206 H29 1 230 378 329 218 H30 1 239 378 326 206 H31 6 224 381 323 206 199 H32 1 224 381 323 215 199 Total 33 61

tions, although the number of fungal isolates was relatively Population Genetic Structure of Endophytes small, many had different genotypes and thus genetic diversity was very high. In contrast, except for the All markers were polymorphic across the two sampled IMGERS-CAS population, genetic diversity of N. gansuense endophytic species (Table 3). The percentage of poly- did not vary much among populations and had a much morphic markers among N. gansuense isolates (2.6 out of higher average genetic diversity than that of N. sibiricum. four markers; 65%) was remarkably higher than that among Diversity and Structure of Endophytes and Host 749

Table 2 Polymorphism and genetic diversity of microsatellite markers for two Neotyphodium species in the five A. sibiricum populations

Endophytes Populations Number of Number of unique Frequency of unique Genetic haplotypes haplotypes (%) diversity Isolates Haplotypes

N. sibiricum IMGERS-CAS 15 1 0 0 0.0000 Xi Ujimqin Qi 8 1 0 0 0.0000 Huolingol 2 2 1 50 1.0000 Arshan 3 2 0 0 0.6667 Hailar 5 4 4 100 0.9000 Total 33 7 5 71.43 0.5133 N. gansuense IMGERS-CAS 2 1 1 100 0.0000 Xi Ujimqin Qi 11 6 3 25 0.8364 Huolingol 18 5 3 83.33 0.7908 Arshan 17 5 4 94.12 0.7500 Hailar 13 13 10 76.92 1.0000 Total 61 25 21 84.00 0.6754

N. sibiricum isolates (1.2 out of four markers; 30%). At in the calculation, significant linkage disequilibrium was population level, estimates of linkage disequilibrium were detected only in the Hailar population in N. sibiricum,and made only for populations with no less than two in two of the five populations in N. gansuense.Inother polymorphic loci and three fungal isolates. Significant populations, however, there were insufficient polymorphic linkage disequilibrium was detected when isolates from loci for the analysis. five populations of N. sibiricum or N. gansuense were In the dendrogram inferred from SSR data, all isolates pooled together, as well as when isolates from two identified as N. sibiricum were clearly separate from Neotyphodium spp. were pooled together, indicating that isolates of N. gansuense with a genetic similarity of ≈0.70 both endophytic species were largely clonal across all (Fig. 2). The cluster analysis also showed that most N. populations sampled. While significant for both Neotypho- sibiricum isolates were identical in all SSR loci analyzed. dium spp., IA was much higher in N. sibiricum than N. Within the cluster of N. gansuense, intra-specific variation gansuense. When isolates from one population were used was evident with several separated clusters observed. In

Table 3 Number of polymorphic loci, index of association between loci (IA and its modification rd) and their shared P values of N. sibiricum and N. gansuense in five A. sibiricum populations

Endophytes Populations Number/percentage (%) of polymorphic loci Index of association

IA rd P

N. sibiricum IMGERS-CAS 0 (0) –– – Xi Ujimqin Qi 0 (0) –– – Huolingol 2 (50) –– – Arshan 1 (25) –– – Hailar 3 (75) −0.1270 −0.0646 0.6960 Total 1.2 (30) 0.9528 0.5017 <0.001 N. gansuense IMGERS-CAS 0 (0) –– – Xi Ujimqin Qi 3 (73) 0.0330 0.0165 0.4530 Huolingol 3 (75) 0.7237 0.3799 <0.001 Arshan 3 (75) 0.4067 0.2135 0.0090 Hailar 4 (100) −0.1279 −0.0433 0.8410 Total 2.6 (65) 0.1393 0.04834 0.0030 Total (two Neotyphodium species) 4 (100) 0.3393 0.1132 <0.001

Statistically significant P values (P<0.05) are marked with bold numbers 750 X. Zhang et al.

I7(G) I13(G) X56(G) X72(G) X79(G) H36(G) E92(G) X5(G) X13(G) X28(G) X35(G) E44(G) X66(G) E32(G) E56(G) X19(G) X22(G) E43(G) A21(G) A78(G) A97(G) E33(G) E91(G) A1(G) A32(G) A45(G) A50(G) A72(G) A82(G) A83(G) E10(G) E47(G) E58(G) H71(G) H74(G) H82(G) H83(G) A12(G) A19(G) A20(G) A28(G) A36(G) X93(G) H99(G) H100(G) A16(G) E27(G) E97(G) H1(G) H7(G) H39(G) H44(G) H57(G) H59(G) H9(G) H20(G) H43(G) H95(G) H98(G) E94(S) A98(G) E30(G) I14(S) I15(S) I17(S) I22(S) I23(S) I29(S) I30(S) I33(S) I35(S) I36(S) I38(S) I40(S) I41(S) I42(S) I43(S) X34(S) X38(S) X47(S) X48(S) X53(S) X69(S) X83(S) X85(S) A61(S) A67(S) H5(S) A14(S) E64(S) H14(S) E25(S) E45(S) E35(S)

0.70 0.75 0.80 0.85 0.90 0.95 1.00

Genetic similarity

Figure 2 Dendrogram generated by UPGMA clustering of Dice genetic similarity coefficients of Neotyphodium isolates from five populations of A. sibiricum. I IMGERS-CAS, X Xi Ujimqin Qi, H Huolingol, A Arshan, E Hailar, S N. sibiricum, G N. gansuense addition, isolates from the same population tended to Genetic Diversity and Structure of Host A. sibiricum cluster together, forming several small geographical groups in the dendrogram. Overall, we detected 44 different genotypes from the 134 The five fungal populations showed high differentiation samples of A. sibiricum (Table 5). Genotypes were coded as

(Table 4). For N. sibiricum, FST=0.6799 (P<0.001); and diploid data because of the diploid nature of the plant. for N. gansuense, FST=0.3490 (P<0.001). No isolation by There were 3–6 polymorphic alleles per locus with an distance was observed for fungal populations (Mantel test: average of 4.25. Statistics for polymorphism and genetic P=0.128 for N. sibiricum and P=0.651 for N. gansuense). diversity of A. sibiricum at the population level are Diversity and Structure of Endophytes and Host 751

Table 4 Analysis of molecular variance from five populations of N. sibiricum and N. gansuense

Endophyte Source of variation d.f. Variance components Percentage of variation FST

N. sibiricum Among populations 4 0.38440 67.99 0.67993** Within populations 28 0.18095 32.01 Total 32 0.56535 N. gansuense Among populations 4 0.40484 34.90 0.34901** Within populations 56 0.75513 65.10 Total 60 1.15997 Two Neotyphodium species Among populations 4 0.43152 32.98 0.32978** Within populations 89 0.87699 67.02 Total 93 1.30850

**Significant with P<0.001 summarized in Table 6. The average percentage of unique ization was proposed as the common source of genetic haplotypes belonging to each population was low, whereas variation of vertically transmitted grass endophytes [19, the genetic diversity was much higher than that of the 49]. This process might also contribute to the genetic endophytes. variation observed in the endophytes symbiotic with A. Multilocus linkage disequilibrium analyses revealed sibiricum. Furthermore, contagious spread of Neotypho- significant levels of non-random association among dium endophytes through asexual conidia was previously microsatellite markers of A. sibiricum. When isolates from proposed [31, 47]. The presence of two Neotyphodium spp. one population were used in the calculation, significant in the same population in this study might facilitate this linkage disequilibrium was detected in four of the five recombination process. populations (Table 7). It is not known whether sexual or parasexual recombi-

In A. sibiricum populations, FST=0.0924 (P<0.001; nation occurred in endophyte-infected A. sibiricum;ifit Table 8), indicating that 9.24% of the total genetic variation did, the higher level of genetic variation observed in N. was among populations. Geographical separation did not gansuense compared to N. sibiricum might be the result of appear to limit gene flow since no isolation by geographical more frequent recombination in N. gansuense. There are distance was found (Mantel test, P=0.889). several possible processes that might explain the difference in genetic variation between the two endophytic species. First, the frequency of sexual reproduction might occur at Discussion different rates for the two endophytic species. Second, the efficiency of endophyte transmission has been suggested to Genetic Diversity of Two Endophytic Species influence endophyte-infection frequency, the distribution of endophytes and therefore the particular associated geno- Marked genetic variation in the populations of N. sibiricum and types [21, 37]. Endophyte-transmission efficiency differs N. gansuense symbiotic with A. sibiricum was detected in this among host populations and species, as well as endophyte study. We suspect that if more isolates were analyzed, then genera [1], which is likely to contribute at least to some average genetic diversity of N. sibiricum would be much extent to the selection on some specific genotypes in this lower than observed. The small sample size and therefore study. Another possible explanation that might account for high-intrapopulation polymorphism detected in the Huolingol, the different genetic variations of the two endophyte species Arshan and Hailar populations caused by low-isolation is local adaptation [46, 50]. In our previous study, we found frequency of N. sibiricum in these populations should account that infection frequency of N. sibiricum tended to increase for the deviation. Thus, genetic diversity of N. gansuense was with decreased precipitation, and we suspected that infec- much higher than that of N. sibiricum (Table 2). tion of N. sibiricum might be associated with drought In the present study, we found that the majority of fungal resistance of the grass [56]. Low-intrapopulation variation isolates were haploid, as only one allele was observed in all and the prevalence of only one genotype in N. sibiricum SSR loci, whereas only 13% of fungal isolates were populations observed in the present study suggest that there heterozygous at one SSR locus, suggesting hybrid origins might be strong selection for particular fungal genotypes of the isolates. This observation supports our previous under drought stress. In contrast, we did not find a result that non-hybrid strains are prevalent in A. sibiricum dominant haplotype in all examined populations of N. populations [56]. In previous studies, interspecific hybrid- gansuense. The N. gansuense was mostly isolated from A. 752 X. Zhang et al.

Table 5 Allele coding used for the definition of A. sibiricum genotypes from five populations

Haplotypes Number of isolates Allele sizes of each locus

Locus Ain041 Locus Ain026 Locus Ain069 Locus Asi003

G1 7 122 199 296 180 G2 3 122 199 288 180 G3 12 122 199 278 180 G4 1 122 201 296 182 G5 2 122 197 296 178 G6 2 122 197 278 178 G7 1 122 197 268 178 G8 2 122 199 278 180 118 G9 1 122 199 268 180 118 G10 1 122 199 296 180 118 278 G11 1 122 199 288 180 118 G12 7 122 199 278 180 116 G13 1 124 199 268 180 116 G14 3 122 199 268 180 116 G15 2 122 199 296 180 116 G16 2 122 197 278 178 116 G17 1 122 197 268 178 116 G18 2 124 199 278 180 116 G19 1 120 199 296 180 116 278 G20 1 122 199 296 180 116 278 G21 1 122 199 326 180 278 G22 4 122 199 296 180 278 G23 1 124 197 268 178 G24 1 122 197 296 178 278 G25 1 120 199 296 180 G26 1 120 199 278 180 G27 3 118 199 296 180 G28 1 118 199 288 180 G29 7 118 199 278 180 G30 3 118 199 268 180 G31 3 118 199 296 180 278 Diversity and Structure of Endophytes and Host 753

Table 5 (continued)

Haplotypes Number of isolates Allele sizes of each locus

Locus Ain041 Locus Ain026 Locus Ain069 Locus Asi003

G32 4 118 197 278 178 G33 1 118 197 268 178 G34 1 118 197 296 178 288 G35 1 118 197 296 178 278 G36 10 116 199 268 180 G37 7 116 199 296 180 278 G38 1 116 197 296 178 278 G39 2 116 199 288 180 G40 17 116 199 278 180 G41 8 116 199 296 180 G42 1 116 197 296 178 288 G43 1 116 197 278 178 G44 1 114 201 288 182 Total 134

sibiricum growing in more favorable environmental con- other, with a genetic similarity of ≈0.70, suggesting that ditions where the selection for particular genotypes might SSR is informative for fungal identification in A. sibiricum. not be as strict as that for N. sibiricum. Moreover, Within the N. sibiricum cluster, most isolates were identical genotypic diversity of endophytes observed in the present in all SSR loci analyzed, suggesting clonal reproduction study suggests that native forage grasses potentially harbor through plant seeds predominated in N. sibiricum. This is more diverse and novel endophyte resources compared to confirmed by significant linkage disequilibrium of fungal that associated with agronomic grasses [40, 41]. SSR loci. In contrast, genetic variation of N. gansuense appeared to be much more diverse. Several clusters of Population Genetic Structure of Two Endophytic Species highly similar isolates were also detected. Most of these clusters include isolates from the same population. More- In the dendrogram generated from SSR data, isolates of N. over, isolates from different populations were also found sibiricum and N. gansuense were clearly separate from each with high similarity within the same cluster, indicating that

Table 6 Polymorphism and genetic diversity of microsatellite markers for A. sibiricum in five populations

Populations Number of Number of unique haplotypes Percentage of unique haplotypes (%) Genetic diversity

Isolates Haplotypes

IMGERS-CAS 26 17 8 38.46 0.9662 Xi Ujimqin Qi 22 18 8 40.91 0.9827 Huolingol 27 11 2 7.41 0.8689 Arshan 28 14 6 42.86 0.9392 Hailar 31 13 4 12.90 0.9097 Total 134 44 28 27.61 0.9333 754 X. Zhang et al.

Table 7 Number of polymorphic loci, index of association Populations Number/percentage (%) of polymorphic loci Index of association between loci (IA and its modifi- I r P cation rd) and their shared P A d values of A. sibiricum in five populations IMGERS-CAS 4 (100) 0.6140 0.2063 <0.001 Xi Ujimqin Qi 4 (100) 0.5872 0.1987 <0.001 Huolingol 2 (50) −0.0755 −0.0755 0.7820 Arshan 4 (100) 0.6717 0.2251 <0.001 Statistically significant P values Hailar 4 (100) 0.6463 0.2162 <0.001 (P<0.05) are marked with bold Total 4 (100) 0.6289 0.2100 <0.001 numbers clonal lineages were important in N. gansuense popula- structure of the host plant and fungal endophytes could be tions, and that gene flow among populations also occurred. explained by several processes. Among them, reproductive Even though stromata formation on A. sibiricum was only system and migration patterns are two very important occasionally observed in native populations, it could be factors that influence the genetic population differentiation quite efficient or the transition from asymptomatic inflor- [15]. Achnatherum sibiricum is transmitted either by clonal escences to stromata might occur sporadically under certain growth or pollen-mediated wind dispersal, for which the environment conditions [46]. gene flow is much greater than that of predominantly seed- We found significant differentiation among populations transmitted clonal growth of the endophyte. of both endophyte species (FST=0.6799 in N. sibiricum, When the gene flow of endophytes is lower than that of FST=0.3490 in N. gansuense). The degree of differentiation their hosts, they are proposed to be at a coevolutionary observed in this study is comparable to that observed in disadvantage, thus less able to infect grasses, and their Neotyphodium spp. in Arizona fescue. In both cases, infection frequency should decrease over time [15]. However, endophytes exhibited low levels of gene flow between this does not really happen in A. sibiricum populations. populations, consistent with the observed dominance of clonal Instead we observed consistently high-endophyte-infection reproduction. Genetic differentiation of E. festucae detected frequency in A. sibiricum during a 5-year continuous in fine fescue (FST=0.0814) and red fescue (FST=0.197) investigation. Differential gene flow rate between endophyte populations was much lower than what we observed in N. and host populations has also been reported in Neotypho- sibiricum and N. gansuense [3, 50], indicating more frequent dium–Arizona fescue association, in which mismatch of gene communication among E. festucae populations than among flow is inferred based on the theory of the geographic mosaic populations of N. sibiricum and N. gansuense.Occasional of co-evolution by Thompson [46, 48]. In the present study, sexual reproduction and sporulation of E. festucae may we proposed that endophyte-transmission efficiency might facilitate genetic communication among populations [46, be quite high or the infection of A. sibiricum might confer 55]. No isolation by distance was found for the two fungal the host some selective advantages under certain conditions, populations in the present study, therefore seed dispersal does which could help to maintain high-endophyte-infection not explain the differentiation between fungal populations. frequencies in host populations [21]. Detailed studies of the outcomes of endophyte infection on A. sibiricum in different Comparison of Population Differentiation of Host biotic and abiotic environments in the future may help and Endophytes explain their coevolutionary processes. Previous studies of symbiotic relationships of endophyte– Here, we compared the genetic structure of two predomi- grass associations usually compared the difference in perfor- nantly vertically transmitted endophytes and a pollen- mance of endophyte-infected and endophyte-free grasses. In mediated, perennial host grass. We found around 4–7-fold the present study, however, we found endophytes with high greater differentiation of endophyte populations than host levels of genetic diversity, which has been shown to influence populations, implying more restricted gene flow of endo- the outcome of endophyte–grass associations [30]. Therefore, phytes than hosts. This large difference between population we suggest that same genotype of endophyte as well as host

Table 8 Analysis of molecular variance from five populations Source of variation d.f. Variance components Percentage of variation FST of A. sibiricum Among populations 4 0.09746 9.24 0.09241** Within populations 129 0.95725 90.76 Total 133 1.05471 **Significant with P<0.001 Diversity and Structure of Endophytes and Host 755 should be confirmed if the objective of the study is to know 17. Faeth SH (2002) Are endophytic fungi defensive plant mutualists? the influence of endophyte or host genotype on their Oikos 98:25–36 symbiotic relationship, instead of only considering whether 18. Faeth SH, Sullivan TJ (2003) Mutualistic asexual endophytes in a native grass are usually parasitic. Am Nat 161:310–325 a plant is infected by an endophyte or not. 19. Gentile A, Rossi MS, Cabral D, Craven KD, Schardl CL (2005) Origin, divergence, and phylogeny of epichloë endophytes of native Argentine grasses. Mol Phylogenet Evol 35:196–208 Acknowledgements This research was financially supported by the 20. Glenn AE, Bacon CW, Price R, Hanlin RT (1996) Molecular National Key Basic Research Program of China (No. 2007CB106802) phylogeny of and its taxonomic implications. and the National Natural Science Foundation of China (No. 30770348 Mycologia 88:369–383 and 30970460). 21. Gundel PE, Batista WB, Texeira M, Martínez-Ghersa MA, Omacini M, Ghersa CM (2008) Neotyphodium endophyte infection frequency in annual grass populations: relative importance of mutualism and – References transmission efficiency. Proc R Soc B 275:897 905 22. Guo LD, Hyde KD, Liew ECY (2000) Identification of endophytic fungi from Livistona chinensis based on morphology and rDNA 1. Afkhami ME, Rudgers JA (2008) Natural history note. Symbiosis sequences. New Phytol 147:617–630 lost: imperfect vertical transmission of fungal endophytes in 23. Hillis D, Moritzc C, Mable B (1996) Molecular systematics. grasses. Am Nat 172:405–416 Sinauer Associates, Sunderland, MA, p 655 2. Arachevaleta M, Bacon CW, Hoveland CS, Radcliffe DE (1989) 24. Kannadan S, Rudgers JA (2008) Endophyte symbiosis benefits a Effect of the tall fescue endophyte on plant response to rare grass under low water availability. Funct Ecol 22:706–713 environmental stress. Agron J 81:83–90 25. Kimmons CA, Gwinn KD, Bernard EC (1990) Nematode 3. Arroyo García R, Martínez Zapater JM, García Criado B, reproduction on endophyte-infected and endophyte-free tall Zabalgogeazcoa I (2002) Genetic structure of natural populations fescue. Plant Dis 74:757–761 of the grass endophyte Epichloë festucae in semiarid grasslands. 26. Maynard Smith J, Smith NH, O'Rourke M, Spratt BG (1993) Mol Ecol 11:355–364 How clonal are bacteria? Proc Nat Acad Sci USA 90:4384– 4. Bassam BJ, Caetano-Anollés G, Gresshoff PM (1991) Fast and 4388 sensitive silver staining of DNA in polyacrylamide gels. Anal 27. Miles CO, Menna ME, Jacobs SWL, Garthwaite I, Lane GA, Biochem 196:80–83 Prestidge RA, Marshall SL, Wilkinson HH, Schardl CL, Ball 5. Breen JP (1994) Acremonium endophyte interactions with en- OJ, Latch GCM (1998) Endophytic fungi in indigenous hanced plant resistance to insects. Annu Rev Entomol 39:401–423 Australasian grasses associated with toxicity to livestock. Appl 6. Bruehl GW, Kaiser WJ, Klein RE (1994) An endophyte of Environ Microb 64:601–606 Achnatherum inebrians, an intoxicating grass of northwest China. 28. Moon CD, Craven KD, Leuchtmann A, Clements SL, Schardl CL Mycologia 86:773–776 (2004) Prevalence of interspecific hybrids amongst asexual fungal 7. Chen N, Yang Y, Li CJ, Nan ZB (2008) Isolation and endophytes of grasses. Mol Ecol 13:1455–1467 characterization of polymorphic microsatellite loci in Achnathe- 29. Morse LJ, Faeth SH, Day TA (2002) Effect of Neotyphodium rum sibiricum (Poaceae). Conserv Genet 9:1699–1701 endophyte infections on growth and leaf gas exchange of Arizona 8. Chen N, Yang YZ, Yang XL, Zhang XX, Li CJ (2008) Twelve fescue under contrasting water availability regimes. Environ Exp polymorphic microsatellite loci for Achnatherum inebrians (Poa- Bot 48:257–268 ceae). Conserv Genet 9:961–963 30. Morse LJ, Faeth SH, Day TA (2007) Neotyphodium interactions 9. Christensen MJ, Latch GCM, Tapper BA (1991) Variation within with a wild grass are driven mainly by endophyte haplotype. isolates of Acremonium endophytes from perennial ryegrasses. Funct Ecol 21:813–822 Mycol Res 95:918–923 31. Moy M, Belanger F, Duncan R, Freehoff A, Leary C, Meyer W, 10. Christensen MJ, Latch GCM (1991) Variation among isolates of Sullivan R, White JF Jr (2000) Identification of epiphyllous Acremonium endophytes (A. coenophialum and possibly A. mycelial nets on leaves of grasses infected by Clavicipitaceous typhinum) from tall fescue (Festuca arundinacea). Mycol Res endophytes. Symbiosis 28:291–302 95:1123–1126 32. Müller CB, Krauss J (2005) Symbiosis between grasses and 11. Clay K (1988) Fungal endophytes of grasses: a defensive asexual endophytes. Curr Opin Plant Biol 8:450–456 mutualism between plants and fungi. Ecology 69:10–16 33. Pan JJ, Clay K (2004) Epichloë glyceriae infection affects carbon 12. Clement SL, Wilson AD, Lester DG, Davitt CM (1997) Fungal translocation in the clonal grass Glyceria striata. New Phytol endophytes of wild barley and their effects on Diuraphis noxia 164:467–475 population development. Entomol Exp Appl 82:275–281 34. Piano E, Bertoli FB, Romani M, Tava A, Riccioni L, Valvassori 13. de Jong EVZ, Guthridge KM, Spangenberg GC, Forster JW M, Carroni AM, Pecetti L (2005) Specificity of host–endophyte (2003) Development and characterization of EST-derived simple association in tall fescue populations from Sardinia, Italy. Crop sequence repeat (SSR) markers for pasture grass endophytes. Sci 24:1456–1463 Genome 46:277–290 35. Rice JS, Pinkerton BW, Stringer WC, Undersander DJ (1990) 14. de Jong EVZ, Dobrowolski MP, Bannan NR, Stewart AV, Smith Seed production in tall fescue as affected by fungal endophyte. KF, Spangenberg GC, Forster JW (2008) Global genetic diversity Crop Sci 30:1303–1305 of the perennial ryegrass fungal endophyte Neotyphodium lolii. 36. Rohlf FJ (2000) NTSYSpc, numerical and multivariate Crop Sci 48:1487–1501 analysis system, ver.2.1. Exeter Software, Setauket, NY 15. Delmotte F, Bucheli E, Shykoff JA (1999) Host and parasite 37. Rudgers JA, Afkhami ME, Rúa MA, Davitt AJ, Hammer S, population structure in a natural plant–pathogen system. Heredity Huguet VM (2009) A fungus among us: broad patterns of 82:300–308 endophyte distribution in the grasses. Ecology 90:1531–1539 16. Excoffier L, Laval G, Schneider S (2005) Arlequin ver.3.0: an 38. Saikkonen K, Faeth SH, Helander M, Sullivan TJ (1998) Fungal integrated software package for population genetics data analysis. endophytes: a continuum of interactions with host plants. Annu Evol Bioinf Online 1:47–50 Rev Ecol Syst 29:319–343 756 X. Zhang et al.

39. Saikkonen K (2000) Kentucky 31, far from home. Science endophytes of tall fescue grass by hybridization with Epichloë 287:1887a species. Proc Natl Acad Sci USA 91:2542–2546 40. Saikkonen K, Wäli P, Helander M, Faeth SH (2004) Evolution of 50. Wäli P, Ahlholm J, Helander M, Saikkonen K (2007) Occurrence endophyte–plant symbioses. Trends Plant Sci 6:275–280 and genetic structure of the systemic grass endophyte Epichloë 41. Saikkonen K, Lehtonen P, Helander M, Koricheva J, Faeth SH festucae in fine fescue populations. Microb Ecol 53:20–29 (2006) Model systems in ecology: dissecting the endophyte–grass 51. Wei YK, Gao YB, Xu H, Su D, Zhang X, Wang YH, Lin F, Chen literature. Trends Plant Sci 11:428–433 L, Nie LY, Ren AZ (2006) Occurrence of endophytes in grasses 42. Schardl CL, Liu J-S, White JF, Finkel RA, An Z, Siegel MR native to northern China. Grass Forage Sci 61:422–429 (1991) Molecular phylogenetic relationships of nonpathogenic 52. Wei YK, Gao YB, Zhang X, Su D, Wang YH, Xu H, Lin F, Ren grass mycosymbionts and clavicipitaceous plant pathogens. Plant AZ, Chen L, Nie LY (2007) Distribution and diversity of Syst Evol 178:27–41 Epichloë/Neotyphodium fungal endophytes from different pop- 43. Schardl CL (1996) Epichloë species: fungal symbionts of grasses. ulations of Achnatherum sibiricum (Poaceae) in the Inner Annu Rev Phytopathol 34:109–130 Mongolia Steppe, China. Fungal Divers 24:329–345 44. Schardl CL (2001) Epichloë festucae and related mutualistic 53. White JF (1994) Endophyte–host associations in grasses. XX. symbionts of grasses. Fungal Genet Biol 33:69–82 Structural and reproductive studies of Epichloë amarillans sp. 45. Schulthess FM, Faeth SH (1998) Distribution, abundance, and nov. and comparisons to E. typhina. Mycologia 86:571–580 association of the endophytic fungal community of Arizona fescue 54. White JF Jr (1994) Taxonomic relationships among the members (Festuca arizonica). Mycologia 90:569–578 of the Balansieae (Clavicipitales). In: Bacon CW, White JF Jr 46. Sullivan TJ, Faeth SH (2004) Gene flow in the endophyte (eds) Biotechnology of endophytic fungi of grasses. CRC Press, Neotyphodium and implications for coevolution with Festuca Boca Raton, FL, pp 3–20 arizonica. Mol Ecol 13:649–656 55. Zabalgogeazcoa I, Vázquez de Aldana BR, García Criado B, 47. Tadych M, Bergen M, Dugan FM, White JF Jr (2007) Evaluation García Ciudad A (1999) The infection of Festuca rubra by the of the potential role of water in spread of conidia of the fungal endophyte Epichloë festucae. Mediterranean permanent Neotyphodium endophyte of Poa ampla. Mycol Res 111:466–472 grasslands. Grass Forage Sci 54:91–95 48. Thompson JN (1994) The coevolutionary process. University of 56. Zhang X, Ren AZ, Wei YK, Lin F, Li C, Liu ZJ, Gao YB (2009) Chicago Press, Chicago Taxonomy, diversity and origins of symbiotic endophytes of 49. Tsai HF, Liu JS, Staben C, Christensen MJ, Latch GCM, Siegel Achnatherum sibiricum in the Inner Mongolia Steppe of China. MR, Schardl CL (1994) Evolutionary diversification of fungal FEMS Microbiol Lett 301:12–20