Josh P. Roberts has an M.A. in the history and philosophy of science, and he also went through the Ph.D. program in molecular, cellular, developmental biology, and genetics at the University of Minnesota, with dissertation research in ocular immunology.
There was a time when DNA methylation meant that the “5” position of cytosine was modified with a methyl group (5-mC), that it was found concentrated at CpG islands and genetic regulatory elements and that the associated chromatin was usually silent. This is still largely the case. But as researchers looked closer, they found subtleties in terms of methylation’s genomic location, modification density and tissue distribution.
Even methylation’s chemical identity held surprises. In 2009, a portion of what was thought to be 5-mC was found to be 5-hydroxymethylctyosine (5-hmC). The latter has a different overall distribution than 5-mC—being concentrated more in exons, for example—and thus likely a distinct biological significance. And that’s not all: It’s now known that the "ten eleven translocation" (TET) enzymes that convert 5-mC to 5-hmC will convert it further to 5-formylcytosine (5-fC) and then to 5-carboxycytosine (5-caC). In other words, what once was thought to be a single modification is actually at least four, and the “gold standard” technique for pinpointing 5-mC—bisulfite sequencing—cannot individuate them.
But a new suite of tools can. Some of these already have been incorporated into commercially available kits and reagents. Others can be reproduced from published protocols. Given the rising significance of epigenetic modifications in disease and development, there’s a good chance you’ll be needing these in your own research. Here’s what you need to know.
Enrichment
Modified DNA often is concentrated in discrete regions, and researchers are keen to know where those regions are located. One approach is methylated DNA immunoprecipitation (MeDIP).
Researchers can build their own MeDIP protocols using commercially available or homemade antibodies. But several vendors, including Active Motif and Diagenode, now offer complete MeDIP and hMeDIP kits that use 5-mC- or 5-hmC-specific antibodies to enrich for modified DNA. Kits include both positive and negative control sequences as well as control primer pairs. The resultant DNA can then be sequenced or examined on a per-locus basis by (q)PCR.
Alternatively, users can perform protein pull-downs. Active Motif’s magnetic bead-based MethylCollector Ultra™ kit uses a heterodimeric complex of methyl-binding proteins (MBD2b and MBD3L1), which, says product manager Kyle Hondorp, enriches methylated DNA better than an antibody and “has better affinity than just an MBD protein alone.”
Unfortunately, says Peng Jin, professor of human genetics at Emory University, all these methods suffer from the same basic problem: “There is always a bias toward highly modified regions.” That is, highly methylated sequences will be enriched over regions that are more sporadically modified.
An alternative strategy enlists beta-glucosyltransferase to add a modified glucose moiety specifically to 5-hmC, and uses that modified sugar to capture or detect the 5-hmC-containing DNA. The glucose can be radiolabeled, for example, or conjugated with a reagent such as biotin. “This is starting to be widely used,” says Jin. “There has been no bias.” Active Motif has commercialized the technology in its Hydroxymethyl Collector™ kit, which uses streptavidin magnetic beads to capture biotin-labeled fragments.
New England Biolabs’ (NEB’s) EpiMark® 5-hmC and 5-mC Analysis Kit uses the beta-glucosyltransferase reaction to turn a cleavable MspI restriction site into a noncleavable one, enabling the enzyme to discriminate between 5-mC and the glucosylated 5-hmC residues in the context of CCGG.
The ability of certain restriction enzymes to cut at modified or nonmodified cytosines is a well established way to discriminate between the two. NEB has recently discovered a family of novel modification-dependent endonucleases—exemplified by MspJI—which recognizes modified cytosine and cleaves at a fixed distance away from the modified site.
“The most exciting property of MspJI is that for a single symmetrically methylated CpG site, it can essentially cleave on both sides of the site and extract a DNA fragment of approximately 31 bp containing the methylated CpG,” notes Sriharsa Pradhan, division head, RNA biology and epigenetics at NEB.
Sequencing
The problem Jin sees with many of the above approaches is that “the resolution is very low”—on the order of tens of bases.
To achieve base-pair resolution, DNA typically is reacted first with sodium bisulfite, which converts cytosines (and 5-fC and 5-caC) to uracils but leaves the 5-mC and 5-hmC intact. When the DNA is sequenced, those modified cytosines are read as cytosines and the newly converted uracils as thymines. Thus any base that reads as a cytosine must have been modified.
Yet bisulfite sequencing cannot differentiate 5-mC and 5-hmC. Chuang He, director of the University of Chicago’s Institute for Biophysical Dynamics, and Jin developed TET-assisted bisulfite sequencing (TAB-Seq) to overcome that problem. Before bisulfite conversion, the DNA is treated with beta-glucosyltransferase to protect 5-hmC. TET is then used to oxidize the remaining cytosines (including 5-mC) to 5-caC, which are then converted to uracil by the bisulfite and read as thymine. TAB-Seq is now available as a kit from WiseGene.
(A conceptually similar method, oxidative bisulfite sequencing (OxBS-Seq), which oxidizes 5-hmC to 5-fC, can also distinguish 5-mC from 5-hmC. Again, comparison to a parallel, standard bisulfite sequencing reaction reveals the original identity of the bases.)
He and Jin have modified the TAB-Seq method to detect 5-fC and 5-caC at single-base resolution, but according to He, “it’s not practical yet.” These modifications are so rare—perhaps 1,000 to 10,000 fold less abundant than 5-mC—that it’s difficult to obtain a reliable sequence.
Cross-talk
Methylated DNA comprises only part of a cell’s epigenetic complexity, and efforts are underway to determine how it and other aspects of transcriptional control work together to define a cell’s unique fate and identity.
Active Motif recently released its Nucleosome Occupancy and Methylome sequencing (NOMe-Seq) kit, which combines bisulfite sequencing with an assay that identifies which sequences are bound by nucleosomes. Similarly, two different groups in 2012 published new methods combining chromatin immunoprecipitation (ChIP) with bisulfite sequencing [1,2].
But as you use these techniques, remember: DNA modifications are dynamic—and reversible. It’s now thought that 5-fC and 5-caC are intermediates in the cycle between methylated and unmethylated cytosine. “So whenever you deal with epigenetic reversible modification of DNA, you have to think about how to reverse it—you have to study these intermediates,” says He.
Whether 5-fC and 5-caC also have functions of their own, complete with cellular “reader” proteins that can recognize them, is still an open question. Now that researchers have tools to study them, it likely won’t remain so for long.
References
[1] Brinkman, AB, et al., “Sequential ChIP-bisulfite sequencing enables direct genome-scale investigation of chromatin and DNA methylation cross-talk,” Genome Res, 22(6):1128-38, 2012. [PubMed]
[2] Statham, AL, et al., “Bisulfite sequencing of chromatin immunoprecipitated DNA (BisChIP-seq) directly informs methylation status of histone-modified DNA,” Genome Res, 22(6):1120-7, 2012. [PubMed]