by Jeffrey M. Perkel
What distinguishes a liver cell from a neuron? Fundamentally, it's all in the DNA. Not the genetic content per se—the two cells are identical on that score—but rather the way that content is implemented. Such is the purview of epigenetics, the study of those forces that influence gene expression that are not encoded in the DNA.
Epigenetics includes everything from non-coding RNAs, to histone modifications, to chemical modifications of the DNA itself. One well-studied example of this latter class is DNA methylation: specifically, the conversion of cytosine in a CpG dinucleotide to 5-methylcytosine. Often clustered in CpG-rich islands, and associated with gene repression, methylation presents particular challenges to epigenetic researchers because it cannot easily be distinguished from non-methylated DNA in cloning, amplification, and sequencing reactions. (An exception is restriction enzymes; the New England Biolabs catalog is rife with isoschizomers like HpaII and MspI, which can differentially cleave a sequence based on its methylation status.)
Fortunately, researchers and molecular tools vendors have now largely filled this gap. Whether your goal is a global snapshot of methylation across the genome, or probing the status of a few specific CpGs, an assay exists to meet your need.
In fact, according to Sallie Cassel, marketing director for antibodies and immunoassay kits at Millipore, which offers a number of methylation reagents, the abundance of choices can create confusion for newcomers to the field. "[The field] is fragmented," says Cassel. "The challenge isn't so much that the techniques are difficult as that they aren't universal."
That said, many methylation workflows do have one common denominator, she says: bisulfite conversion. "Ninety percent [of researchers] use some method that requires bisulfite conversion," she estimates.
Available in kits from ActiveMotif (MethylDetector™ Bisulfite Modification Kit), BioChain (DNA Methylation Detection Kit), Life Technologies (MethylCode Bisulfite Conversion Kit), Millipore (CpGenome Fast DNA Modification Kit), Qiagen (EpiTect Bisulfite Kits), and Zymo Research (EZ DNA Methylation™ kits), among others, bisulfite conversion is a chemical process that transforms unmethylated cytosines to uracil while leaving methylated cytosines unchanged. Subsequent amplification replaces methylcytosine with cytosine and uracil with thymine—differences that can be distinguished via methylation-specific PCR (MSP), sequencing, or restriction enzyme digestion.
It's not as easy as it sounds, says Marc Van Eden, vice president of business development and marketing at Zymo Research. "Bisulfite modification of DNA has been around since the 1970s," says Van Eden. "But at the same time, it involves a multi-step chemical reaction that can be deleterious to DNA. So the chemistry has to really be fine-tuned to ensure that conversion is reliable and robust."
"It's easy to say but phenomenal to do," he adds.
Another common denominator: many researchers enrich genomic samples for methylated DNA (prior to bisulfite conversion or microarray analysis, for instance). There are two basic approaches, one using an antibody to 5'-methylcytosine to capture methylated DNA (MeDIP, "Methyl-DNA ImmunoPrecipitation"), and one using methyl-CpG binding domain (MBD)-containing proteins to produce the same result.
According to Amy Cuneo, product manager at Life Technologies, one significant difference between the two methods is that, while antibodies used in MeDIP specifically recognize 5-methylcytosine in the context of single-stranded DNA, MBD2, used in the company's MethylMiner™ DNA Enrichment Kit, prefers double-stranded DNA.
"In terms of experiments you might do afterwards, especially when talking about generating libraries for next-generation sequencing, that becomes a major advantage," Cuneo says.
Another advantage: while immunoprecipitated DNA is released from its column en masse using proteinase K, methylated DNA can be gradually released from MBD2 with a salt gradient. "You can discriminate different degrees of methylated DNA," Cuneo says, "rather than the all-or-nothing [elution] that you would get with an antibody."
Other commercial purification offerings include Zymo's Methylated-DNA IP Kit, based on MeDIP, and ActiveMotif's MethylCollector™ Ultra, which couples MBD2b with MBD3L1 to create a heterodimeric protein with enhanced affinity for methyl-CpG—a process called MIRA ("Methylated-CpG Island Recovery Assay"). ActiveMotif also offers a complementary UnMethylCollector™ kit to enrich for non-methylated DNA, using a CXXC binding domain protein.
According to Kyle Hondorp, a product manager at ActiveMotif, the availability of kits selective for both methylated and unmethylated DNA allows researchers to validate their results; many users tend to run them both in parallel. "If you get no results from MethylCollector, you might assume [the sample] is unmethylated," she explains. "With the UnMethylCollector kit, you get a positive result."
Once you have obtained (and possibly bisulfite-converted) your methylated DNA, you need to analyze it. One popular approach uses microarrays. A number of vendors offer such tools, including Agilent Technologies' Human CpG Island Microarray Kit, Illumina's HumanMethylation27 DNA analysis BeadChip, Roche NimbleGen's Human DNA Meth 2.1M Deluxe Promoter Array, and Affymetrix's seven-array GeneChip® Human Tiling 2.0R Array Set.
The process varies from platform to platform. In Agilent's and NimbleGen's case, an aliquot of a genomic DNA sample is subjected to MeDIP, while another is saved as an input DNA control. The two samples are then fluorescently labeled (the input DNA with Cy3, the enriched fraction with Cy5), pooled, and hybridized to the array. The Cy5/Cy3 intensity ratio at each probe on the array indicates the extent of methylation over that region.
"People have a pretty good idea of what areas of the genome are methylated, and arrays are pretty efficient ways of getting to those," says Alicia Burt, Agilent's director of microarray applications.
Agilent's 244,000-element array covers more than 27,000 human CpG islands and "UMRs," or under-methylated regions; NimbleGen's Human 2.1M Deluxe Promoter array also covers more than 27,000 CpG islands and includes more than 27,000 promoter regions, all at 100-bp resolution. Yet neither platform can report methylation status with single-nucleotide resolution (that is, the methylation state of specific cytosines within each region). But Illumina's Infinium and GoldenGate assays can.
Targeting more than 27,000 CpG sites, Illumina's Infinium methylation assay adapts the Illumina Infinium I Whole Genome Genotyping assay for measuring CpG methylation using quantitative "genotyping" of bisulfite-converted genomic DNA.
VeraCode Methylation, in contrast, enables high-throughput validation studies through a solution-based approach in which 96 or 384 user-specified CpG loci are analyzed with the GoldenGate Assay for Methylation. Bisulfite-treated genomic DNA is mixed with assay oligos, one of which is complimentary to either the converted U of the unmethylated site, or the protected C of a methylated site. Following hybridization, primers are extended and ligated to locus-specific oligos to create a template for universal PCR. Finally, labeled PCR primers are used to create a detectable product, with individual assays localizing to specific VeraCode beads by hybridization of address sequences.
In both cases, says Brent Applegate, associate product manager at Illumina, "The biggest advantage is single CpG-site resolution. Rather than settling for the methylation status of a region, you are gaining specific, quantitative measurements of the methylation levels for specific CpG loci."
Kenneth Nephew, professor of cellular and integrative physiology at the Indiana University School of Medicine, takes an alternative approach to genome-scale methylation analysis in his research on breast and ovarian cancers. After successfully using array-based methods, Nephew now uses "MIRA-Seq," a hybrid technique combining MIRA, bisulfite conversion, and massively parallel DNA sequencing using Illumina's Genome Analyzer or Roche's 454 GS FLX sequencer.
According to Nephew, the millions of short reads that result from this approach provide richer data than the microarray-based approaches he formerly used. "You get much more quantitative information, and deeper information, with sequencing," he says.
Sometimes, though, Nephew's team needs to focus on just a few CpG sites. In that case, he uses MethyLight, "a TaqMan® assay on bisulfite-converted DNA." MethyLight uses distinct TaqMan® probes and/or PCR primers to discriminate between the methylated and non-methylated forms of bisulfite-treated DNA. Qiagen offers a handful of premade MethyLight assays in its EpiTect product line.
There are other options, too, including such restriction enzyme-based assays as COBRA ("COmbined Bisulfite Restriction Analysis"), HELP ("HpaII tiny fragment Enrichment by Ligation-mediated PCR"), and McrBC fractionation—and CHARM ("Comprehensive High-throughput Arrays for Relative Methylation"), an array-based approach. Also available is a range of ancillary reagents, from purified methyltransferases to fully methylated (or methylated) DNA controls.
With so many assays to choose from, Cassel advises those new to the epigenetics game to pick an assay format, whatever it is, and stick with it—you don't want to be switching assays mid-stream. Choose an assay from a vendor with experience, she suggests. "Take a look at who's been in this market long enough, and who's going to stick around in this market."
Her other advice: "Do the controls [that come with many of these assay kits]. Until you have gotten your hands wet and have some experience, you need them."