X is for Marking the Spot

Tom James is taking a new approach to halting HIV replication: He's going straight for its genetic code, RNA.

James has found a way to efficiently search for drugs that act on RNA. The strategy, called computational screening, is a series of calculations to model how RNA will interact with potential drug molecules. This virtual testing narrows the field of potentially useful compounds from millions down to a handful of promising ones that can be examined to see if they work as well in reality as the computer predicts. Additionally, James, chair of the department of pharmaceutical chemistry in the UCSF School of Pharmacy, is designing his own compounds with a process called combinatorial chemistry.

Until now, James says, there has been very little drug discovery work targeting RNA and even less using three-dimensional RNA structures. Most current therapies focus on proteins - more proteins are available, more is known about their structure, and people have been thinking about them longer. As more is learned about RNA structure using X-ray crystallography and nuclear magnetic resonance (one of the main focuses of the James lab has been RNA structure determination using NMR), there appear to be unique sites that could be targeted. And that is where his work begins.

"We don't know the rules yet for translating sequence to RNA structure," says James, "but we have unique structures, so we can start to design drugs against some targets on RNA." To prove the utility of their theoretical screen, James and his colleagues started with a well-studied interaction in HIV. They aimed for a region of the HIV RNA called the transactivation response element (TAR). It is known that TAR functions by binding to a protein called Tat, which regulates expression of all viral genes by producing mature, full-length viral RNA. This interaction appears to be a particularly advantageous place to attack HIV because Tat is critical for viral replication. The appeal of effectively extinguishing HIV replication via Tat has led other researchers to pursue disabling Tat directly. James chose instead to block Tat's docking site so it wouldn't matter how many Tat molecules were around; they can't promote HIV replication if they can't link up with TAR.

Big Screen, Small Screen

So TAR became the molecular guinea pig for testing the feasibility of the James virtual screening technology. To identify potential inhibitors of the Tat-TAR interaction, James and his laboratory colleagues screened more than 180,000 random compounds, running them through a barrage of calculations based on molecular characteristics like size, shape, rotation, charge, and flexibility.

The results from the screen allowed them to pare down to about 500 compounds, which they ranked according to a set of probabilities of what would make a good drug, such as general availability and reasonable price. They tested 43 candidates in a series of experiments to see if they inhibited the Tat-TAR interaction. "With 180,000 compounds, we could have been here forever trying to experiment on all of them," says James. "We can save a lot of money and time if we do the computational work first. We chose the top contenders and only experimented on those."

Value of Screening

Although theoretical, computational screening of RNA has proven itself to be a valid predictor. Using NMR as an experimental screening tool, James has verified that some of the top-scoring compounds identified by virtual screening actually do bind to the TAR region. Some surprising news was that the compounds did not have to mimic the way that Tat binds to that region; the calculations predicted some completely new ways of binding that worked to inhibit TAR.

An additional confirmation of the screen was that all of the compounds already known to inhibit Tat-TAR (although none too well) were "re-identified." The 43 highest ranking compounds not previously known to be inhibitors were tested experimentally in collaboration with Matija Peterlin's lab at UCSF. We haven't found a drug yet," James says, "but it's a very good lead for a new way to kill HIV. In this case, it might actually end up working." Some of the compounds get into cells and are fairly nontoxic. The most important verification of the technique was that the structure of a complex formed by one promising compound bound to TAR was determined by NMR in James' lab.

The most promising lead compounds have now been selected for further development by computational and experimental combinatorial chemistry. The computer-aided design method can provide novel compounds that warrant further testing and/or modification by synthesizing related compounds. UCSF collaborator Kip Guy has an unusual piece of equipment for an academic lab - a parallel synthesizer that can perform 96 reactions at a time, enabling them to efficiently synthesize several new related compounds to test. "Although some of the compounds we tested bind quite well," says James, "modifying them should yield even better inhibitors."

by Mitzi Baker

Top photo: Tom James. Photos by Tom James Laboratory.