CpG islands—regions of the genome with higher-than-expected frequency of CpG dinucleotides—are found in the promoters of more than half of all mammalian genes. Unlike CpG dinucleotides outside of these islands, CpG islands are normally protected from cytosine methylation and the transcriptionally repressive chromosome structure associated with it. Yet there are many instances, ranging from genomic imprinting and X-chromosome inactivation to the repression of tumor suppressor genes, in which such epigenetic changes are found.
Researchers can query the DNA methylation status of specific sequences within a CpG island using methylation-specific PCR (MSP), a powerful, quick and cost-effective method accessible to virtually any PCR-equipped lab. The principle is simple. First, sodium bisulfite is used to deaminate unmethylated cytosines and convert them to uracils while leaving methylated cytosines intact. Two PCR reactions are then performed on the bisulfite-treated DNA. One reaction uses primers designed to amplify a template originally unmethylated (the U primer pair), and the other reaction, using the M primer pair, amplifies a template originally methylated.
The design of the primers is extremely critical, says James Gordon Herman, M.D., professor of oncology at Johns Hopkins Medicine, first author on the 1996 PNAS paper introducing the technique [1]. “That’s actually where most people get into trouble with it,” he says. Below are Herman’s tips for successful MSP primer design.
Design primers to the bisulfite-converted template
Don’t just design the primers and then (virtually) convert them. It sounds obvious, but you need to remember that when the C’s are converted to U’s, the DNA strands are no longer complementary. “So a primer designed for the upper strand is different than the lower strand. Most people don’t understand that until they draw it out by hand and see that,” Herman says.
Make sure there are non-CpG C’s in the target sequence
There should be C’s that are converted, regardless of methylation status—non-CpG C’s fit the bill because they’re never methylated. Herman advises, “If you don’t have enough C’s that are non-CpG’s, then unconverted DNA and methylated DNA look the same.”
Aim for six to seven CpG’s
After performing 5,000 or so MSPs over the years, Herman has found that what tends to correlate with gene expression and gene silencing is having between four and eight CpG’s between the two primers, with six or seven being ideal.
“If you go lower than that, specificity is reduced,” he explains. “And you can get nonspecific amplification (unmethylated or methylated) that would be the wrong read, because it’s less discriminating.”
Having more CpG’s works wonderfully on homogenously methylated cell lines “but tends to give inaccurate reads in primary tumors where there’s heterogeneity,” Herman adds.
Keep the targets short
The process of bisulfite conversion damages the template DNA, so you’ll get the best results by trying to amplify shorter stretches—less than 150 base pairs for the amplicons, if possible, as long as your primers don’t overlap. “You can go up to 200 if you need to stretch it, but since you’re actually doing a PCR that’s dependent on the primer sequences, there’s no advantage to going longer,” Herman says.
If you’re going to run the products on a gel, be sure you can resolve the M-primed amplicons from those primed with the U primers. And if you’re going to do real-time studies, make sure you leave space between primers for the probe.
Obey the rules of PCR
All of the considerations for designing PCR primers apply to MSP.
Some are obvious. For example, make sure the primers are compatible with each other in terms of melting temperature, don’t form duplexes with themselves or the other primer and are specific for the target sequence (i.e., not promiscuous). If possible, design the M and U primer-reactions to have the same conditions so that they can be run simultaneously.
Less obvious is that converting C’s to U’s (and thus T’s) reduces the complexity (increasing the possibility of promiscuous binding) and lowers the annealing temperature, meaning that MSP generally requires longer primers than standard PCR to achieve the same specificity.
Use the appropriate computer program
PCR primers can be designed with pen and paper, but—especially with the added concerns of MSP—“it just becomes inherently easier to do that with software,” says Herman, whose lab developed an application, MSPprimer, specifically to design primers for MSP [2]. MSPprimer incorporates all the standard PCR concerns, plus “bisulfite specificity that makes sure you don’t amplify an unconverted DNA. And then there’s methylation specificity to make sure that you can actually distinguish between methylated and unmethylated” CpG’s, Herman explains.
Scientists generally investigate CpG islands’ methylation status because of its relationship to gene function. If the promoter of a particular gene is heavily methylated in a particular disease, for example, that’s an indication that dysregulation of that gene may be involved in the pathology. So Herman’s final piece of advice is to focus on regions critical to the biology. “You could design primers to look anywhere,” he says, but it’s not that important to find “methylation in some regions where it doesn’t regulate gene function or gene expression.”
References
[1] Herman, JG, Graff, JR , Myohanen, S, Nelkin, BD, Baylin, SB, “Methylation-specific PCR: a novel PCR assay for methylation status of CpG islands,” PNAS, 93(18): 9821–9826, 1996.
[2] Brandes, JC, Carraway, H, Herman, JG, ”Optimal primer design using the novel primer design program: MSPprimer provides accurate methylation analysis of the ATM promoter,” Oncogene, 26(42): 6229-6237, 2007.