Abstract

Many Ceratocystis species cause cankers and unique vascular wilt diseases, often on a broad and unpredictable range of plant hosts. Characteristic necrosis of xylem parenchyma cells and dark staining of surrounding tissue is typically evident, especially in woody hosts. The molecular basis for this unique pathogenicity and host range remains unclear, but bacterial-type phosphatidylinositol phospholipase C (bPI-PLC) genes were recently identified in unusually high copy number in multiple Ceratocystis species, and the PLCs may play a role in host membrane disruption. We produced a high-quality long-read genome assembly of the rapid [O]hi'a death pathogen, Ceratocystis lukuohia, and identified 81 partial or complete PLC-like genes, each with a unique DNA sequence, encoding signal peptides and a PLC-X domain. The putative translations mostly ranged from 300 to 500 amino acids that differed markedly from the fungal and prokaryotic bPI-PLCs at sites conferring phosphatidylinositol specificity, suggesting a novel family of secreted PLCs (Cer-PLCs). Remarkably, 73 of the full or partial Cer-PLC genes reside in a single 543 kb gene cluster in C. lukuohia. Comparison to an available long-read genome assembly of C. fimbriata revealed a similar Cer-PLC cluster of 61 genes, with a gene order and arrangement broadly similar to that of the C. lukuohia cluster, except for a large inversion at the beginning of the cluster. Differences suggest that the cluster is dynamic, with many apparent indels involving multiple Cer-PLCs. We compared 40 newly-assembled genomes of Ceratocystis strains and eight publicly available genomes and found that the Cer-PLCs comprise a gene family present in all Ceratocystis species but differing greatly in number (26 to 92), with 64 to 92 in species of the highly aggressive Latin American Clade. The two closest relatives of Ceratocystis have Cer-PLCs but not in the gene cluster: Chalaropsis spp. have only one Cer-PLC, and Berkeleyomyces basicola has 25 related Cer-PLCs scattered across multiple contigs. No Cer-PLC was detected in the more-distant members of the Ceratocystidaceae. The unique cluster in Ceratocystis apparently arose through insertion of Cer-PLCs within an ancestral gene cluster with a CeGAL transcription factor, followed by repeated duplications and rapid diversification of Cer-PLCs, perhaps driven by unequal crossover events. This extraordinary expansion, diversification, and maintenance of Cer-PLCs may have played a major role in the evolution of aggressiveness and host range in Ceratocystis.