Researchers at Duke University Medical Center have developed an assay that is 1000 times more sensitive at distinguishing drug-resistant HIV mutations than commercially available tests. Along with its ability to highlight a single mutated HIV lurking among 10 000 copies of the virus, the technique also detects when a virus has more than one mutation—a trick that only clonal or single genome sequencing methods can achieve. (Nat Methods 2007, 4:123–125; doi:10.1038/nmeth995).
To develop the parallel allele-specific sequencing assay, or PASS, investigators led by Feng Gao modified the polony technique. Developed in the late 1990s, the polony method enables researchers to suspend DNA fragments in a thin film of polyacrylamide gel and then carry out solid-phase PCR. The result is in thousands of discrete clumps - each made up of millions of copies of a single DNA molecule - mounted on a microscope slide. These polymerase colonies, or polonies, are tagged with fluorescent molecules that attach to specific DNA mutations and then counted by a micro-array scanner.
Co-author Jun Zhu was using the polony technique to study pre-RNA splicing. With the procedure he could profile, in one fell swoop, the thousands of RNA existing in a cell. After hearing a talk by Zhu, Gao thought the technique might be adapted to fish for HIV that have drug-resistance mutations, thus eliminating the labor-intensive, time-consuming and costly clonal sequencing procedure normally used to study drug resistance.
The hunch proved correct. The team first performed a series of experiments using a 1.1 kb pol gene fragment that contained sites of all the major drug-resistance mutations of reverse transcriptase and protease. Those experiments showed that the assay could easily detect mutations present at 0.01-0.1%, and that it worked on 27 viral genomes. The procedure discerned more than 20 primary drug-resistant mutations. And because the fluorescent tags are applied and washed off one after another, the researchers could see whether each polony, or individual viral genome, carried multiple mutations.
Next, the researchers evaluated the assay on HIV from infected patients. They analyzed 13 plasma samples from three different groups of patients: those who had never received antiviral treatment, those who had been treated previously but were currently not taking antivirals, and patients who were being treated but not responding. When the team looked for the M184V mutation, which causes resistance to reverse transcriptase inhibitors, they detected no such mutation in treatment-naive patients, minor populations (less than 2%) in two of six previously treated patients and major drug-resistant populations (36-100%) in patients whose antiviral treatment was not working. The assay spotted triple- or quadruple-drug resistant HIV strains in four patients.
The researchers have now tested the assay on more than 100 patients and verified its reliability compared to the best available assays. “Basically we can get a little bit more sensitivity than even real-time PCR based analysis,” Gao says. “And we can do the linkage analysis that real-time PCR genotype can't do and only clonal or single genome sequencing methods can do.”
Several companies are interested in licensing the technology, Gao says. A main goal now is to simplify and automate the priming and sequencing step. Currently it takes 1.5-2 h to apply, scan and remove the fluorescent tag for each mutation. According to Gao, designing a drug cocktail for a treatment-naive patient would probably require initially screening for eight to 10 key mutations.
© 2007 Lippincott Williams & Wilkins, Inc.