69 S. SÁNCHEZ MÁRQUEZ 1 , G.F. BILLS 2 and I. ZABALGOGEAZCOA 1 1 Instituto de Recursos Naturales y Agrobiologia, CSIC, Salamanca, Spain 2 Centro de Investigación Básica de España, Merck, Sharp & Dohme, Madrid, Spain izabalgo@usal.es Abstract Morphological and molecular methods were used to identify the endophytic mycobiota of the grass Dactylis glomerata. Fungal endophytes belonging to 109 different species were isolated from asymptomatic plants sampled in different ecosystems in Spain. Species accumulation curves showed that most species commonly infecting this grass have been identified, but the number of singleton species occasionally infecting the plants is likely to increase with more sampling effort. A large endophytic assemblage consisting of fungi with diverse ecological roles, and potentially unknown species was found in a small number of plants. Keywords: endophytes, Dactylis glomerata, diversity, abundance Introduction Evidence accumulated in surveys suggests that all plant species are associated with fungal endophytes. High species diversity per plant species is a characteristic of the endophytic mycobiota (Stone et al. 2004). In grasses, systemic endophytes belonging to the Epichloë and Neotyphodium genera are well known. However, not much is known about other endophytic species associated with grasses. This paper describes a wide range of endophytic species associated with Dactylis glomerata, a grass common in many different ecosystems in Spain. Methods Plant collection and endophyte isolation Asymptomatic plants of Dactylis glomerata were sampled at 10 locations in the province of Salamanca, one in Ávila, one in Cáceres, and two in La Coruña. These locations represent different habitats (e.g. semiarid grasslands, coastal meadows, roadsides, etc.). To isolate endophytes, leaf and stem pieces were surface- disinfected with a solution of 1% active chlorine for 10 minutes. Root fragments were disinfected with a 5 minute rinse in ethanol, followed by treatment with a 1% active chlorine solution for 15 minutes, and 2 minutes in ethanol. The treated fragments were rinsed in sterile water and plated on potato dextrose agar (PDA) containing 200 mg/L of chloramphenicol. To induce sporulation in sterile isolates in PDA medium, fungi were cultured in three other media: malt extract agar, water agar, and water agar containing autoclaved pieces of leaves of Dactylis glomerata. DNA amplification and molecular taxonomy DNA was extracted from mycelium with a commercial kit (RedExtract-N-Amp Plant PCR, Sigma Aldrich). The ITS1- 5.8S rRNA-ITS2 region was amplified with a polymerase chain reaction (PCR) using primers ITS4 and ITS5 (White et al. 1990). Only one strand of the PCR amplicon was sequenced in a reaction started with primer ITS4. In subset of 12 isolates, both strands of the amplicons were sequenced; these complete sequences were used to analyse the reliability of the taxonomic information obtained with the corresponding partial one-sided sequences. The FASTA algorithms (Pearson 1990) were used to interrogate the EMBL/Genbank database of fungal nucleotide sequences. Because for most fungal species the range of intraspecific variation in ITS sequences is unknown (Taylor et al. 2000), sequences with a similarity greater than 97% were considered to belong to the same species. This arbitrary distance has been used in other studies (O’Brien et al. 2005; Neubert et al. 2006). Quantification of fungal diversity Species accumulation curves, showing the relationship between the number of plants sampled and the number of fungal species identified, were made using EstimateS 7.5 software (Colwell 2005). To estimate the possible total number of endophytic species which could be associated to Dactylis glomerata, the incidence- based coverage estimator (ICE), and the Chao2 estimator of total species richness were calculated (Chazdon et al. 1998). Shannon’s index of diversity (H’) was estimated from the relative abundance of each taxon identified (Zak & Willig 2004). Results and Discussion Isolation and identification of endophytes From 120 plants, a total of 316 fungal isolates were identified. On the average, 2.63 species were identified on each plant; only 13 plants did not yield any endophytes. Using morphological and molecular characteristics for identification, 91 different species of fungi belonging to 63 genera could be identified (Table 1). An additional set of sterile fungi belonging to 18 different species could not be identified because their sequences were different to any entry from the EMBL fungal database, or were similar to database entries corresponding to unidentified fungi. In total, 53% of all endophytic species could be identified by morphological characters. If only the isolates that could be identified at the genus or species level are considered (Table 1), 66% of them could be identified with the use of phenotypic characters. The partial sequences obtained contained the complete nucleotide sequence of ITS1 and 5.8S rRNA, but most were incomplete at the 3’end of the ITS2 region. On the average, these sequences contained about 92% of the total ITS2 sequence. When 12 partial sequences of these characteristics were compared to their corresponding complete sequences for identification purposes, in all cases, the entry retrieved with FASTA from the Genbank database was the same using a partial or a complete sequence. This result suggests that partial sequences missing information at the 3’ end may be as reliable as the complete versions for approximating identification. Further evidence of the value of these partial sequences comes from the fact that there was agreement in the molecular and morphological identification, at least to genus rank, for all isolates whose identities were similar to database entries and were greater than 95% (Table 1). Although limited in value for rigorous phylogenetic analysis, partial sequences derived from single sequencing reactions can be useful for database interrogation where large numbers of isolates are processed.