The methylomes of six bacteria

11450–11462 Nucleic Acids Research, 2012, Vol. 40, No. 22 doi:10.1093/nar/gks891 Published online 2 October 2012 The methylomes of six bacteria Iain A. Murray1, Tyson A. Clark2, Richard D. Morgan1, Matthew Boitano2, Brian P. Anton1, Khai Luong2, Alexey Fomenkov1, Stephen W. Turner2, Jonas Korlach2,* and Richard J. Roberts1,* 1 New England Biolabs, 240 County Road, Ipswich, MA 01938 and 2Pacific Biosciences, 1380 Willow Road, Menlo Park, CA 94025, USA Received August 1, 2012; Revised August 31, 2012; Accepted September 3, 2012 ABSTRACT INTRODUCTION Six bacterial genomes, Geobacter metallireducens GS-15, Chromohalobacter salexigens, Vibrio breoganii 1C-10, Bacillus cereus ATCC 10987, Campylobacter jejuni subsp. jejuni 81-176 and C. jejuni NCTC 11168, all of which had previously been sequenced using other platforms were re-sequenced using single-molecule, real-time (SMRT) sequencing specifically to analyze their methylomes. In every case a number of new N6-methyladenine (m6A) and N4-methylcytosine (m4C) methylation patterns were discovered and the DNA methyltransferases (MTases) responsible for those methylation patterns were assigned. In 15 cases, it was possible to match MTase genes with MTase recognition sequences without further sub-cloning. Two Type I restriction systems required sub-cloning to differentiate their recognition sequences, while four MTase genes that were not expressed in the native organism were sub-cloned to test for viability and recognition sequences. Two of these proved active. No attempt was made to detect 5-methylcytosine (m5C) recognition motifs from the SMRTÕ sequencing data because this modification produces weaker signals using current methods. However, all predicted m6A and m4C MTases were detected unambiguously. This study shows that the addition of SMRT sequencing to traditional sequencing approaches gives a wealth of useful functional information about a genome showing not only which MTase genes are active but also revealing their recognition sequences. We are becoming accustomed to the ever-increasing speed and reduced cost with which DNA can be sequenced. However, what is often lost in this frenzy of sequencing is the fact that DNA consists of more than just four bases. In eukaryotes, we have known for a long time about the epigenetic role of 5-methylcytosine (m5C), sometimes called the fifth base, and more recently it has been found that 5-hydroxymethylcytosine, 5-formylcytosine and 5-carboxylcytosine are also present (1–4). However, two more modified bases, N6-methyladenine (m6A) and N4-methylcytosine (m4C), are also common in bacterial genomes, where they function as components of restriction–modification (RM) systems (5). Until recently, these have usually been ignored because of the lack of simple methods to determine their locations. However, with the advent of single-molecule, real-time (SMRT) sequencing (6–8), it has suddenly become possible to detect these modified bases as a part of the routine sequencing procedure. The methylated bases that are found in bacterial and archaeal genomes serve important functions as part of RM systems, where they protect the host chromosome against the otherwise deleterious action of the partner restriction enzyme(s), which are needed to destroy unwanted incoming transmissible DNA elements such as phages (9). However, in some cases these methyltransferases (MTases) also serve regulatory roles as with the Dam MTase of Escherichia coli, which introduces m6A residues that play a key role in DNA repair and also have important effects during the initiation of replication (10). Several studies have also implicated MTases in regulating gene expression, phase variation and pathogenicity (11,12). Given the many DNA MTases that are typically found in prokaryotic genomes, it seems likely that they will have hitherto undocumented effects aside from their *To whom correspondence should be addressed. Tel: +978 380 7405; Fax: +978 380 7406; Email: Correspondence may also be addressed to Jonas Korlach. Tel: +650 521 8006; Fax: +650 323 9420; Email: jkorlach@pacificbiosciences.com ß The Author(s) 2012. Published by Oxford University Press. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0/), which permits unrestricted, distribution, and reproduction in any medium, provided the original work is properly cited. Nucleic Acids Research, 2012, Vol. 40, No. 22 11451 key role in RM systems. To date, there has been no genome-wide assessment of the extent of DNA methylation by known MTases such as E. coli Dam (10) and Dcm (13) or the cell cycle MTase, CcrM, of Caulobacter crescentus (14). It is not known if their methylation specificities are as precise as the customary recognition sequences suggest or whether the enzymes are promiscuous. This is particularly interesting to know for RM systems as there are no obvious selective constraints on MTase specificity provided that the core recognition sequence of the restriction enzyme is fully modified. Recently, we have shown that by cloning an individual MTase gene into a plasmid and propagating it in an otherwise methylation-deficient strain of E. coli, it is easily possible through SMRT sequencing to detect all of the bases modified on the plasmid (15). Precise recognition sequences were convincingly demonstrated and mostly matched that of the cognate restriction enzyme when the MTase was part of an RM system. However, some promiscuous methylation was observed, with the Dam gene of E. coli being a particularly striking example. There was one caveat to this interpretation though: because the MTase genes in that study were cloned on a multi-copy number plasmid (50–200 copies per cell), it could be that the observed promiscuity arose because of overexpression. Given that the results for the plasmids were very clear, it seemed that it might be possible to perform a direct analysis of bacterial genomes using the SMRTsequencing method and thus obtain an accurate estimate of the extent of methylation in the native organism. By then, comparing a bioinformatic analysis of the RM systems with the direct measurement of just what was methylated, it should be possible to assign recognition sequences to individual MTase genes. Of particular interest in this sort of analysis are the Type I and Type III RM systems, which have generally been very difficult to analyze by previous, more tedious techniques (16). In both of these kinds of systems, the specificity comes from a single subunit of the enzyme—the S subunit of the Type I enzymes and the M subunit of the Type III enzymes (16). Thus, it seemed likely that recognition sequences for both types of MTases could be discovered relatively easily. To demonstrate the feasibility of this approach, we chose initially to analyze six genomes with relatively few RM systems before moving on to more complicated cases. MATERIALS AND METHODS the culture (...truncated)