Comprehensive identification of 5-methylcytosines in Frankia genomes

To investigate restriction-modification systems operating in Frankia, we comprehensively identified 5-methylcytosines (5mCs) in Frankia strains ACN14a and CcI3 by bisulfite genome sequencing. We identified 7344 and 3934 5mCs in the ACN14a and CcI3 genomes, respectively. These cytosine methylations fell into two categories in which the rate of methylated molecules contained within the libraries differed. In category 1, cytosines were methylated in most of the chromosome molecules contained in the library. In category 2, however, cytosines in only some of the chromosome molecules were methylated. In the ACN14a genome, both categories of methylation were found, whereas in the CcI3 genome only category 2 methylation was detected. We found evidence of consensus sequences around the three types of 5mC: 5′-GGC5mCGGCC-3′ (ACN14a, category 1), 5′-GGC5mCGGCT-3′ (ACN14a, category 2), and 5′-CGGC5mCGGC-3′ (CcI3, category 2). The consensus sequence for the ACN14a category 1 5mCs was identical to the recognition sequence of a type II methyltransferase M.FseI, which was isolated from Frankia sp. NRRL 18528. A Frankia ACN14a gene, FRAAL0689, showed strong homology to M.FseI, indicating that the gene is responsible for category 1 methylation in ACN14a. Consensus sequences for category 2 5mCs were not identical to any recognition sequences of known restriction endonucleases and DNA methyltransferases.