The ascomycete and causative agent of maize anthracnose and stem rot, Colletotrichum graminicola, differentiates melanized infection cells called appressoria that are indispensable for breaching the plant cell wall. High concentrations of osmolytes accumulate within the appressorium, and the internal turgor pressure of up to 5.4 MPa provides sufficient force to penetrate the leaf epidermis directly. In order to assess the function of melanin in C. graminicola appressoria, we identified and characterized the polyketide synthase 1 (CgPKS1) gene which displayed high similarity to fungal polyketide synthases (PKS) involved in synthesis of 1,3,6,8-tetrahydronaphthalene, the first intermediate in melanin biosynthesis. Cgpks1 albino mutants created by targeted gene disruption were unable to penetrate intact leaves and ruptured frequently but, surprisingly, were able to penetrate ultrathin polytetrafluoroethylene membranes mimicking the plant surface. Nonmelanized Cgpks1 appressoria were sensitive to externally applied cell-wall-degrading enzymes whereas melanized appressoria were not affected. Expression studies using a truncated CgPKS1 fused to green fluorescent protein revealed fluorescence in immature appressoria and in setae, which is in agreement with transcript data obtained by RNA-Seq and quantitative polymerase chain reaction. Unexpectedly, surface scans of mutant and wild-type appressoria revealed considerable differences in cell-wall morphology. Melanization of appressoria is indispensable for successful infection of intact leaves. However, cell collapse experiments and analysis of the appressorial osmolyte content by Mach-Zehnder interferometry convincingly showed that melanin is not required for solute accumulation and turgor generation, thus questioning the role of melanin as a barrier for osmolytes in appressoria of C. graminicola.

BackgroundAn annotated genomic sequence of the corn anthracnose fungus Colletotrichum graminicola has been published previously, but correct identification of gene models by means of automated gene annotation remains a challenge. RNA-Seq offers the potential for substantially improved gene annotations and for the identification of posttranscriptional RNA modifications, such as alternative splicing and RNA editing.ResultsBased on the nucleotide sequence information of transcripts, we identified 819 novel transcriptionally active regions (nTARs) and revised 906 incorrectly predicted gene models, including revisions of exon-intron structure, gene orientation and sequencing errors. Among the nTARs, 146 share significant similarity with proteins that have been identified in other species suggesting that they are hitherto unidentified genes in C. graminicola. Moreover, 5′- and 3′-UTR sequences of 4378 genes have been retrieved and alternatively spliced variants of 69 genes have been identified. Comparative analysis of RNA-Seq data and the genome sequence did not provide evidence for RNA editing in C. graminicola.ConclusionsWe successfully employed deep sequencing RNA-Seq data in combination with an elaborate bioinformatics strategy in order to identify novel genes, incorrect gene models and mechanisms of transcript processing in the corn anthracnose fungus C. graminicola. Sequence data of the revised genome annotation including several hundreds of novel transcripts, improved gene models and candidate genes for alternative splicing have been made accessible in a comprehensive database. Our results significantly contribute to both routine laboratory experiments and large-scale genomics or transcriptomic studies in C. graminicola.Electronic supplementary materialThe online version of this article (doi:10.1186/1471-2164-15-842) contains supplementary material, which is available to authorized users.