Elisabeth A. Stevens, Emily J.A. Blakemore & James C. Reeves.
Molecular Biology & Diagnostics Section, National Institute of Agricultural Botany, Huntingdon Road, Cambridge. CB3 0LE.
Emily J.A. Blakemore E-mail: firstname.lastname@example.org
The sequences of the internal transcribed spacers (ITS) of the rRNA genes of 21 Pyrenophora spp. isolates were determined. The size of the ITS1 region varied from 178 to 190 base pairs and the size of the ITS2 region varied from 158-161 base pairs. The results of the ITS analysis for the four barley infecting species, Pyrenophora graminea, Pyrenophora teres, Pyrenophora teres f. sp. maculata and Pyrenophora hordei, indicated that they were closely related with only one base difference in the ITS1 region separating P. graminea from the other Pyrenophora spp. The ITS sequence for the P. hordei isolate, a P. teres f. sp. teres isolate and three P. teres f. sp. maculata isolates were identical. The ITS2 was more conserved than the ITS1 region for all the isolates of Pyrenophora used in this study. There was 15.4-17% ITS1 sequence divergence between the two P. avenae isolates and the other Pyrenophora species. This was greater variation than that found between the Pyrenophora species pathogenic on barley (0.0-0.9%). There was no well-defined boundary between the interspecific and intraspecific variation of the Pyrenophora spp. pathogenic on barley. The high degree of similarity in the ITS regions for the Pyrenophora species pathogenic on barley implies that the taxonomic status of these pathogens needs to be re-evaluated for improved diagnosis and crop protection.
Fungi classified as Pyrenophora spp. are responsible for some agriculturally important diseases of cereals across the world (Porto-Puglia et al., 1986; Duczek et al., 1991). Four species of Pyrenophora are currently known to infect barley and cause disease compared to only one species that is known to infect oats and another for wheat, named Pyrenophora tritici-repentis.
The five Pyrenophora species discussed in this study have been classified on the basis of symptom production, mode of transmission, host plant species and morphological features. Pyrenophora graminea Ito & Kuribay. causes leaf stripe on barley ( Figure 1) and is strictly seed transmitted. Pyrenophora teres Drechs. and Pyrenophora teres Drechs. f. sp. maculata cause net and spot blotch diseases respectively ( Figure 1) and are mainly dispersed by crop debris, although seed-borne inoculum is probably important for introduction to disease free fields. Pyrenophora hordei Wallwork, Lichon & Sivanesan is a newly identified barley pathogen isolated from Australia and its method of transmission is not known (Wallwork et al.,1992). The oat-infecting species, Pyrenophora avenae, causes a range of leaf stripe, blotch and spot symptoms as the disease progresses on host plants. It is primarily seed transmitted, although secondary infection by air-borne conidia does occur. P. hordei has similar shape conidia to the anamorph of P. avenae (Wallwork et al., 1992) despite their different host species.
Although these fungi are classified as separate species, P. graminea and both forms of P. teres have been demonstrated to hybridise with one another and produce fertile progeny (Smedegaard-Petersen, 1983). The morphology of P. graminea, P. teres and P. teres f. sp. maculata is also similar and morphologically indistinct isolates are often assigned to species on the basis of disease symptoms and host plant reactions only. These features mean that it is difficult to delineate accurately the different species and consequently there must be some doubt as to their taxonomic status as species. This also has consequences for modes of transmission based on their species identity.
The difficulties experienced in the accurate identification of these pathogens cause problems for reliable disease diagnosis both in the field and in seed health tests. This subsequently affects the efficacy of disease control strategies and management practices. A need exists for a more thorough understanding of the relationships of these pathogens with each other and of their interactions with the host plants, so that breeding for resistance and disease control may be carried out more effectively. Greater understanding of the taxonomic relationships could aid in increasing the accuracy of delineation of Pyrenophora spp. Many studies have used DNA sequences to infer phylogenetic relationships between species (e.g. Sreenivasaprasad et al., 1996a) as well as between strains of the same species (e.g. Appel & Gordon, 1996). For this purpose a genomic region must be identified which has conserved homologous sequences that may be found in all the isolates under investigation, yet which is sufficiently variable to enable differences to be found between different isolates.
The 18S, 5.8S and 28S ribosomal RNA genes have been well characterised and are known to have conserved sequences. There is unlikely to be sufficient variation in the genic sequences to provide useful phylogenetic data, but the spacer regions between the conserved genes are much more variable and thus have the potential for interspecific comparisons. Studies of the ITS regions have revealed useful phylogenetic data for several fungal species, including Colletotrichum spp. (Sreenivasaprasad et al., 1992), Glomus mossae (Lloyd-Macglip et al., 1996), Phytophthora spp. (Lee & Taylor, 1992) and Verticillium spp. (Morton et al., 1995). In addition, sequence variation can be utilised for the design of species-specific or strain-specific primers for use in diagnostic PCR assays (e.g. Brown et al., 1993; Johanson & Jeger, 1993; Ogorman et al., 1994; Sreenivasaprasad et al., 1996b). If there were sufficient interspecific variation in the ITS regions then primers could also be designed for use in a diagnostic PCR seed health test.
Collection of fungal isolates
The Pyrenophora spp. isolates used in this work were obtained from geographically diverse places, as shown in Table 1 Fungal isolates were cultivated on PDA plates at 200C with a photoperiod of 12 h using near UV light for seven days or until sporulation occurred. Species identity of the isolates was confirmed by the plant host it was isolated from. Morphological features including the following: colony colour; mycelial type; production of vertical erect mycelial tufts (coremia); and production of conidiophores and conidia [ISTA working sheets No. 3, and 6. (Rennie and Tomlin, 1984 and 1984a), CMI descriptions No. 388, 389 and 390]. The identity of P. graminea isolates was confirmed by specific PCR using primers designed from RAPD analysis (e.g. Stevens et al., 1996). A long-term culture collection was maintained on Potato Dextrose Agar (PDA) slants under paraffin oil.
A fungal isolate obtained from wheat seed sample obtained in 1991 for a routine seed health test at the Official Seed Testing Station, NIAB, Cambridge. It was tentatively identified as Pyrenophora sp. from observation of all the morphological characteristics mentioned above except for the production of conidiophores and conidia which were absent.
DNA extraction and PCR conditions
Mycelial production and DNA extraction was performed as described by Stevens et al.,1996. Templates for sequencing were prepared by an initial amplification step using the universal primers ITS4 and ITS5 (White et al., 1990). PCR amplification conditions were as described by Stevens et al., 1996. A single fragment approximately 650-700 bp in length was generated and purified for sequencing using the QIAquick PCR purification kit (Qiagen) according to the manufacturer's instructions.
DNA sequencing and analysis
For sequencing, between 30-100 ng of template DNA was needed, which was typically a 2 ml aliquot of the purified ITS DNA fragment. The sequencing reactions were prepared using the ABI Prismä Dye Terminator Cycle Sequencing Ready Reaction Kit, according to the manufacturer's instructions (Perkin Elmer Applied Biosystems) and run on an ABI 377 sequencer. Four reactions were set up for each isolate, using the primers ITS 1-4 inclusive so that both strands could be sequenced. Subsequent alignment of the sequence data was performed using DNAstarâ software. A phylogenetic tree was obtained from the combined analysis of the ITS 1 and 2 region using maximum likelihood method (DNAML program from Phylip 3.5) and by the neighbour-joining method using TREECON software (Van de Peer, Y. and De Wachter, R. 1994). As the sequence data was relatively short in length, a one-parameter-model for calculating the distance matrices was carried out (Jukes and Cantor, 1969). Bootstrap analysis with one hundred replicates was used to estimate the robustness of the evolutionary trees.
The sequence data were also compared with those of a P. graminea isolate and a P. avenae isolate available through the EMBL database (accession numbers X78124 and X1723 respectively).
For the Pyrenophora isolates used in this study, it was possible to distinguish P. graminea from P. teres f. sp. teres and P. avenae by observing colony morphology and shape of conidia from microscopic analysis. It was difficult to distinguish P. teres f. sp. maculata and P. hordei from P. teres f. sp. teres using morphological characteristics due to their similarity.
An initial PCR amplification using the universal primers ITS4 and ITS5 (White et al., 1990) generated a single DNA fragment that was approximately 650 bp in length from all the isolates that were investigated ( Figure 2).
The entire ITS1 region ( Figure 3) varied in length between 178-180 bp for isolates of P. graminea, P. teres f. sp. teres, P. teres f. sp. maculata and P. hordei. This region was slightly longer for the two P. avenae isolates (188-190 bp). The ITS2 region ( Figure 4) was 161 bp in length for all the isolates except for Pa3828 (P. avenae). A summary of the number of bases in the ITS1 and ITS2 regions is presented in Table 1
The P. hordei isolate and all the P. teres f. sp. maculata isolates investigated had a 180 bp ITS1 region, whilst all the P. teres f. sp. teres isolates except Pt29 had a 178 bp ITS1 region. P. graminea isolates had a more variable ITS1 region, ranging from 178-180 bp in length. The sequence variations causing these inter- and intra- specific differences were confined to two regions which are discussed in more detail below.
A single point mutation at position 42 of the ITS1 alignment (Figure 3) distinguishes P. graminea (G) isolates from P. teres f. sp. teres, P. teres f. sp. maculata, P. hordei and P. avenae (T). There was no other variation in the ITS sequence that distinguished P. graminea from the other Pyrenophora spp.
Variation between positions 144-158 of ITS1 involved changes in A or T repeats (Figure 3). Both inter and intraspecific variation was observed. The interspecific variation was mainly due to the insertion or deletion of nucleotides within a run of a particular base, in this case T. P. teres f. sp. maculata isolates and the P. hordei isolate had nine T's whereas all the P. teres f. sp. teres isolates except Pt29 had seven T's. Two major sequence variations were discovered amongst the P. graminea isolates, which may reflect the fact that they came from a wider variety of sources. However, the differences did not appear to correlate directly with the varying geographical origins of the isolates. For example, at least one UK isolate of P. graminea was present in each of the two major sequence variant groups.
There was little sequence variation within the ITS1 region ( Table 2 for the Pyrenophora spp. pathogenic on barley. Most of the sequence variation was due to differences in the number of repeated T residues in a run of bases (Figure 3), position 144-158). There was, however, much more interspecific sequence variation between Pyrenophora spp. pathogenic on barley and the two P. avenae isolates.
Sequence analysis of the ITS 2 region (Figure 4) revealed that it was identical between all isolates of Pyrenophora spp. pathogenic on barley except for two base differences in isolate Pt18 (P. teres f. sp. teres) and one base difference for isolate Pm1881 and Pm1892 (P. teres f. sp. maculata). The ITS2 region of the P. graminea isolates in this analysis was one base different from that of the same region for the P. graminea isolate published on the EMBL database (accession number X78124).
Pyrenophora spp. are present occasionally on wheat seed samples sent to NIAB for routine seed health testing. The fungal isolate named "Pa91-1b" was found on one of these samples and it was tentatively classified as a Pyrenophora sp. from observation of its dense olive green mycelium and the presence of light grey coremia which are especially characteristic of P. avenae and P. teres. Production of conidia which are used as one of the key diagnostic characters for identification of Pyrenophora are not always present or observed in culture. Morphological characters such as colony colour may be subject to environmental variation and so a molecular approach was used to help identify this isolate. Comparison of the sequence data for the ITS1 and 2 regions (Figs. 3 & (4) from this isolate, and with the Pyrenophora spp. used in this study indicated that there was good homology. It was also observed that the ITS1 and 2 sequence of Pa91-1b was more similar to Pyrenophora spp. pathogenic on barley than to P. avenae ( Figure 5). This shows the powerful ability of ITS sequencing to help classify fungi from an unknown isolate.
The two P. avenae isolates shared 98.3% sequence similarity across the entire ITS1 region, 5.8S gene and ITS2 region. There was 15.4-17% variation in the ITS1 and 2 regions between these two isolates and the isolates of P. graminea, P. teres f. sp. teres, P. teres f. sp. maculata and P. hordei. This level of interspecific variation is in the same range as the 3-58% ITS variation found in other fungi (Gardes et al, 1991, Lloyd-MacGlip et al, 1996 and Gaskell et al, 1997). The intraspecific variation (0.0-0.9%) of the ITS region for the Pyrenophora isolates used in this study was also comparable to the 0.0-5.8 % variation observed in other fungi (Sreenivasaprasad et al, 1996a). The differences were throughout the entire ITS1 and 2 regions (Figs. 3 & (4). A phylogenetic tree was obtained from the combined analysis of the ITS1 and ITS2 regions (Figure 5) using the Neighbour-joining Method from Treecon 1.3b software (Van de Peer and de Wachter, 1994). A phylogenetic tree was also produced by the Neighbour-joining method from Phylip 3.5 software (Felsenstein, 1993) but the result is not shown as it was similar to that in Figure 5. Both methods produced phylogenetic trees, which separated the Pyrenophora spp. into two clades, one which contained the isolates pathogenic on barley and the other pathogenic on oats. These two P. avenae isolates can therefore be identified as being different from the four Pyrenophora species pathogenic on barley, but they are sufficiently similar to conclude that they are related to them.
Identification of interspecific sequence variation in the ITS regions of several genera of fungi has been successfully used to obtain species specific primers for diagnostic tests (Morton et al., 1995 and Chen et al., 1996). This approach was attempted for the development of species specific primers for several Pyrenophora spp. that could then be used as part of a seed health test.
The sequence data analysis of the ITS 1 region indicated that there was one consistent interspecific difference which separated P. graminea from P. teres f. sp. teres, P. teres f. sp. maculata and P. hordei. This molecular data in conjunction with the high level of morphological similarity indicates their close genetic relatedness. The single nucleotide base substitution at position 42 in ITS1 of P.graminea was not sufficient to design effective species specific PCR primers but it could be used in a ligase chain reaction assay (LCR). LCR is a highly sensitive method for distinguishing single base-pair differences between DNA sequences (Barany, 1991). This method was successfully used to identify Erwinia stewartii, by discriminating a single base pair difference in the 16s rRNA gene which is unique to this bacterial plant pathogen (Wilson et al., 1994).
The low intraspecific variation in the ITS1 sequence amongst P. teres f. sp. teres, P. teres f. sp. maculata and P.graminea isolates between positions 144-158 was due to repeated runs of A's and T's. Such mutations in runs of the same base are common and thought to be slippage during DNA replication (Vogler and DeSalle, 1994). The pattern of these mutations does tend to suggest some substructuring of the populations and therefore may be rare recombinations via hybridisation with other species as indicated by Smedegaard-Peterson (1983).
The intraspecific variation present in Pyrenophora may be due to its population biology. A genetic study on five populations of P. teres from the USA and Germany (Peever and Milgroom, 1994) suggested that they were derived from a common source population even though there was a high level of genetic variation within each population from random sexual reproduction. There is little information on the population structure of P.graminea and P. teres f. sp. maculata but the low interspecific ITS variation between P. graminea, P. teres f. sp. teres, P. teres f. sp. maculata and P. hordei also suggests that these species may have coevolved from a common source pathogen. Further research to investigate the population biology and behavioural ecology of these pathogens is needed to fully understand the ITS sequence variation.
There was intraspecific variation in the ITS region but it did not correlate with differences in geographical origin of the isolates. It may be related to the origin of the seed which is transported widely. The variation among isolates from different countries was no greater than those from a narrower geographic range. This is similar to the results of the ITS analysis for some Glomus mossae isolates (Lloyd-MacGlip et al., 1996), where it was concluded that the intraspecific variation evolved on a timescale that was slower than the rate at which these fungi disseminate across the world.
The ITS 2 region was similar for all the four Pyrenophora spp. pathogenic on barley except for three isolates (Pm1881, Pm1892 and Pt18) which each have one or two base changes. This indicates that the ITS2 region is more highly conserved than the ITS1 region as has been shown previously for other fungal species (Gaskell et al., 1997, Sreenivasaprasad et al., 1996a). The ITS2 region was therefore not suitable for designing species specific primers of Pyrenophora spp.
ITS sequence has been in many cases a good molecular tool for distinguishing fungal species (White et al, 1990). ITS mutations generally occur at a rate approximating that of speciation processes, and will only become "fixed" in a large-scale population if that form of the pathogen is widespread and reproductively or geographically isolated from other populations without that mutation. Mutations in the ITS regions are neutral markers that can give an insight into the epidemiology but is never responsible for those changes. The greater interspecific variation between P. avenae and the other Pyrenophora spp. can be explained by their difference in host specificity leading to isolation. This interspecific variation between P. avenae and the other Pyrenophora spp. may allow the design of PCR primers which will amplify a specific DNA fragment from P. avenae but not from other Pyrenophora species.
There were no differences between the ITS sequences of P. hordei and four of the P. teres f. sp. maculata isolates which suggests a close relationship. The intraspecific divergence for P. teres f. sp maculata (0-0.31 %) is greater than the divergence between P. teres f. sp maculata and P. hordei (0-0.29%). Classification of P. hordei as a distinct species was made on the basis of smaller ascospore size (Wallwork et al., 1992). Further ITS sequence data from other isolates of P. hordei are needed in addition to data from other biochemical methods and morphological characterisation to give support to its potential synonymity with P. teres f. sp. maculata. Morphological characters of fungi are known to be influenced by environmental factors. Therefore they may not always be an accurate way of identification unless the morphological statistics are based on a large population taken from diverse origin and environment (Smedegaard-Petersen, 1983). The problem with using morphological characters for taxonomy is illustrated by Pyrenophora japonica S. Ito & Kurib. which was originally described as a pathogen causing spot symptoms on barley in Japan and South Africa (Scott, 1991). It was later discovered from a high homology in DNA banding patterns and mating studies that P. japonica was a synonym of P. teres f. sp. maculata (Crous et al., 1995).
The low interspecific variation in the ITS regions of the four Pyrenophora spp. pathogenic on barley supports other evidence for their close phylogenetic similarity. For example it has been demonstrated that they are able to cross hybridise and produce fertile progeny (Smedegaard-Petersen, 1983). Attempts to develop species specific monoclonal antibodies for P. graminea also provide evidence for taxonomic similarity (Burns and George, 1995). Three different immunisation techniques were used in an attempt to produce specific cell lines for this pathogen. Seven antibodies were generated but all exhibited isolate variability with respect to P. graminea and there was cross-reactivity with some P. teres isolates. The antibodies did not however cross react with other genera. This genus specificity but difficulty in generating species specific antibodies emphasises the close association of the Pyrenophora spp.
This phylogenetic analysis (Figure 5) of Pyrenophora spp. pathogenic on barley indicates a need for clarification of the taxonomic status of this genus. The divergence between P. teres f. sp. teres and P. teres f. sp maculata (0.0-0.6%) or P. graminea (0.3-0.9%) and P. hordei (0.0-0.3%) overlaps the amount of divergence within P. teres f. sp. teres (0.0-0.3%) or P. graminea (0.0-0.3%) respectively. This overlap of interspecific and the intraspecific variation shows that the species boundaries are not well defined. It may be appropriate, therefore, to reclassify P. graminea, P. teres f. sp. teres, P. teres f. sp. maculata and P. hordei as variants of the same species rather than different ones as presently classified. This is similar to reclassification of Colletotrichum spp. which was also partly based on lack of ITS sequence divergence (Sreenivasaprasad et al, 1996).
There is little information on how these four pathogens interact with each other and whether they have their own distinct ecological niche. It is reported that both forms of P. teres are capable of rapid adaptation and differentiation into physiological races (Brandl and Hoffman, 1991). An example of this was in Saskatchewan where there was a rapid change from absence of P. teres f. sp. maculata to it representing 35 % of the P. teres population in thirteen years (Tekauz, 1990). This has important consequences for barley breeding programs. It is not known if P. graminea and P. avenae are also capable of such rapid adaptation to change in environment and host varieties.
A greater range of Pyrenophora isolates should be investigated using other molecular tools such as AFLP (Majer et al, 1996) to help obtain an accurate taxonomy for this genus. Research on the diversity of this pathogen complex and the interactions of its components with each other are needed for effective crop protection and disease resistance breeding programmes.
This research was supported by MAFF. We would like to thank Karen Husted, Haakon Magnus, Andrei Tekauz and IMI for providing some of the Pyrenophora spp. isolates. I appreciate the help and expertise for identification and maintenance of fungal cultures from Jane Thomas, David Kenyon, Angela Rutherford and Christine Lang of the Seed Diagnostics Unit at NIAB.
Brown AE, Muthumeenakshi S, Sreenivasaprasad S, Mills PR, Swinburne TR, 1993. A PCR primer-specific to Cylindrocarpon heteronema for detection of the pathogen in apple wood. FEMS Microbiology Letters 108, 117-120.
Chen WD, Gray LE, Grau LE, 1996. Molecular differentiation of fungi associated with brown stem rot and detection of Phialophora-gregata in resistant and susceptible soybean cultivars. Phytopathology 86, 1140-1148.
Gandebouef D, Dupre C, Roeckel-drevet P, Nicolas P and Chevalier G, 1997. Typing Tuber ectomycorrhizae by polymerase chain reaction of the internal transcribed spacer of rDNA and the sequence characterised amplified region markers. Canadian Journal of Microbiology 43, 723-728.
Gardes M , White T J, Fortin J A, Bruns T D and Taylor J W, 1991. Identification of indigenous and introduced symbiotic fungi in ectomycorrhixae by amplification of nuclear and mitochondrial ribosomal DNA. Canadian Journal of Botany 69, 180-190.
Gaskell GJ, Carter DA, Britton WJ, Tovey ER, Benyon FHL, Lovborg U, 1997. Analysis of the internal transcribed spacer regions of ribosomal DNA in common airborne allergenic fungi. Electrophoresis 18, 1567-1569.
Lloyd -Macglip SA, Chambers SM, Dodd JC, Fitter AH, Walker C, Young JPW, 1996. Diversity of the ribosomal internal transcribed spacers within and among isolates of Glomus mosseae and related mycorrhizal fungi. New Phytologist 133, 103-111.
Morton A, Carder JH, Barbara DJ, 1995. Sequences of the internal transcribed spacers of the ribosomal RNA gene and relationships between isolates of Verticillium alboatrum and V. dahliae. Plant Pathology 44, 183-190.
Ogorman D, Xue B, Hsiang T, Goodwin PH, 1994. Detection of Leptosphaeria korrae with polymerase chain reaction and primers from the ribosomal internal transcribed spacers. Canadian Journal of Botany 72, 342-346.
Sreenivasaprasad S, Brown AE, Mills PR, 1992. DNA sequence variation and interrelationships among Colletotrichum species causing strawberry anthracnose. Physiological and Molecular Plant Pathology 41, 265-281.
Stevens EA, Alderson J, Blakemore EJA, Reeves JC, 1996. Development of a multiplex PCR seed health test to detect and differentiate three pathogens of barley. In: Marshall G, ed. Diagnostics in Crop Protection symposium proceedings no. 65, 1996. University of Warwick, UK: BCPC, 99-104.
Tekauz A, 1990. Characterisation and distribution of pathogenic variation in Pyrenophora teres f. teres and Pyrenophora teres f. maculata from western Canada. Canadian Journal of Plant Pathology 12, 141-148.
Van de Peer Y, De Wachter R, 1994. TREECON for Windows: a software package for the construction and drawing of evolutionary trees for the Microsoft Windows environment. Computer Applications Bioscience 10, 569-570.
White TJ, Bruns T, Lee S, Taylor J, 1990. Amplification and direct sequencing of fungal ribosomal RNA genes for phylogenetics. In: Innis MA, Gelfand DH, Sninsky JJ, White TJ, eds. PCR Protocols: A guide to methods and applications. USA: Academic press inc., 315-322.