When the AIDS virus sets out to infect a new victim, it recruits acellular accomplice to do the inside job of integrating the viral DNAwith the target cell's family jewels _ its genome.

HIV's kit of burglar tools includes one key enzyme, integrase, whichlinks up with its co-conspirator protein in the target cell, thenactually joins that cell's DNA with the viral DNA. Foiling thisunholy genomic alliance is an obvious goal of drug-discoveringresearchers trying to develop anti-AIDS therapeutics.

Their efforts received two separate but related boosts in the currentissue of Science, dated Dec. 23. One came up with the three-dimensional crystal structure of the integrase molecule's core region;the other cloned the gene that encodes the host cell's protein,corrupted by the viral enzyme to let it take over the cell for its ownreplicative purposes.

A team of X-ray crystallographers and retrovirologists in theNational Institute of Diabetes, Digestive and Kidney Diseases(NIDDK) report "Crystal structure of the catalytic domain of HIV-1integrase: similarity to other polynucleotidyl transferases."

And a group of researchers at Columbia University's Department ofBiochemistry and Biophysics published "Binding and stimulation ofHIV-1 integrase by a human homolog of yeast transcription factorSNF5." A lead author of this paper, retrovirologist Stephen Goff,commented to BioWorld Today anent the NIDDK crystallographyresult: "Their structure is helpful to all of us in the field, in thinkingabout how the protein works, how it binds DNA, and how it bindsfactors like ours."

NIDDK's retrovirologist Robert Craigie told BioWorld, that untilnow only two of the three enzymes that HIV deploys to insert itsDNA into target calls, HIV protease and reverse transcriptase, havebeen elucidated by X-ray crystal structures. "Knowing the structureor properties of integrase, the third enzyme, is one more potentialtarget for drug intervention."

Craigie added that "Maybe more important is the possibility it offersof finding inhibitors that could be used against all three enzymetargets in combination." An accompanying editorial in Sciencesuggests that "such a combined attack would help reduce the chancesthat HIV could mutate and become resistant to drug therapy in allthree enzymes at once."

Solving integrase's crystal structure required a trick of geneticengineering. Crystallizing a protein for X-raying is a little likemaking old-fashioned rock candy from a supersaturated sugarsolution. The devil, for crystallographers, is in that word, "solution."

Unlike sugar, native integrase is virtually insoluble. With trial anderror, the NIDDK researchers hit upon one particular point mutationout of 30 they tried. It replaced one amino acid with another, thusrendering the mutant enzyme soluble, hence solvable.

What they solved by the X-ray structural analysis was the central orcore catalytic domain of integrase, namely, 162 of its 288 aminoacids. Now, Cragie said, "We are trying to get the crystal structurefor the full length of the integrase protein. We have made this totalsequence much more soluble than the partial region we had before,so it looks like a very good candidate for crystallization."

Cragie explained the expected added benefit: "At some point it'sgoing to be important to identify inhibitors that block the action ofthe enzyme. By having the full structure, there's going to be apotentially wide range of targets that we might be able to thinkabout."

The protein that Goff's group has cloned, which works hand-in-glove with integrase, is not a likely target for inhibition, theColumbia researcher said. "We are coming to appreciate that thisprotein is part of a complex that has the job in the cell of actuallyunfolding chromatin, which is the tightly wound, packed, orcondensed DNA in the chromosome." As this complex "is requiredfor the expression of many many genes in the cell, to inhibit itsactivity would probably be a pretty bad thing to do."

Instead, Goff surmised, "one might like to inhibit the protein'sbinding to integrase. That's a reasonable goal," he said, "whichinvolves identification of the protein's surface to which the virusbinds." He went on, "We know a bit about this binding site, andwe're expecting to know a great deal more."

In their Science paper, Goff and his co-authors report having clonedthe gene for the chromatin-opening protein, which they call INi1(integrase interactor 1). Its long open reading frame predicts aprotein of 44,131 daltons containing 385 amino acids. Initialexperiments in yeast and in vitro suggest that its first 107 aminoacids are not required for binding.

First identified in mutant yeast cells, then in fruitflies and laboratorymammals, the integrase-binding gene now cloned in humans "is veryancient," Goff pointed out. This genetic conservation suggests thatits protein has a fundamental role in DNA unfolding across a broadeukaryotic spectrum.

"From the point of view of the virus," Goff explained, "we showedin the Science paper that this protein binds to integrase. Andfunctionally that it can stimulate the enzyme, in vitro, to better carryout its enzymatic activities of joining the viral DNA to the hostDNA."

The Columbia group is now conducting two experiments to gauge"the virus uses or needs for the protein." In one, they've set up celllines that greatly overexpress the protein "We're about to challengethose with virus, to find out whether they will stimulate the virus togrow better, or whether the overexpression will mess them up, andactually prevent normal infection."

Their second effort is to determine the target positions in the humangenome where viral insertion takes place.

Craigie observes that "There's no way of knowing until you canmake gene knockouts, but if the interactions [between integrase andthe protein complex are not at one specific site], "the virus willprobably integrate elsewhere instead." n

-- David N. Leff Science Editor

(c) 1997 American Health Consultants. All rights reserved.