By David N. Leff

No one has yet come up with a better metaphor for the way a protein binds to its receptor — typically, an antibody to its antigen — than the way a key unlocks its lock.

Keys have specific high points and low points, called bits, that mesh with matching contours inside the lock. As a rule, one key fits one door, not more. But a master key, a.k.a. a skeleton or passkey, has some of its bits filed off, so it opens a number of doors.

Of course, Nature laughs at locksmiths. Viruses got there first.

Take HIV, the AIDS virus. Its "key" binds preferentially to receptors on two types of cells in the human immune system. These are, specifically, T lymphocytes and monocyte/macrophages. At the other extreme, the keys on mouse leukemia viruses (MLV) practice "one shape fits all." They invade any and all cells in the murine body.

And there's the rub, as far as human gene therapy is concerned.

"One of the gene-delivery vectors that's been very popular among people who are trying to develop gene therapy," observed structural biologist Peter Kim, "is the MLV. It's very well characterized, and has the potential to combat disease by inserting normal genes into the tissues of patients with severe genetic disorders. So people in the gene therapy field have chosen MLV as the major retrovirus they're trying to use."

The rub lies in MLV's lack of specificity.

Kim, a member of the Whitehead Institute of Biomedical Research, at Massachusetts Institute of Technology, in Cambridge, explained: "In order for gene therapy to be maximally effective, one needs to be able to target specific cells in a specific way, and not just come into a human being with a virus that infects every cell in the body."

To confer this single-lock specificity on the MLV vector, he and his colleagues have mapped the three-dimensional molecular structure of its skeleton key, so that gene therapists can direct it to desired target cells only.

Today's Science, dated Sept. 12, 1997, reports their feat under the title: "Structure of a murine leukemia virus receptor-binding glycoprotein at 2.0 Angstrom resolution." The article images the crystal structure of a piece of the protein that studs the outside of the virus's envelope — the protein that recognizes and binds to receptors on the surface of a mammalian cell.

"This is the first time we've been able to get a high-resolution look at the structure of any retrovirus receptor-binding domain," Kim told BioWorld Today. He and five of his six co-authors are investigators of the Howard Hughes Medical Institute, which funded the project.

3-D Road Map To Precision DNA Transfer

"What's interesting about the structure," Kim observed, "is that the parts of the protein that we know are important for binding to the receptor, all mapped to one place on the molecule up at the top of its L-shaped conformation. The parts that are conserved form the central stalk.

"So what we hope," he continued, "is that gene therapists who are trying to design better retroviral vectors with more specificity will be able to take this structure and use it as a guide to where to make changes, such as where to insert hormones."

By way of example, he cited "inserting things like erythropoietin (EPO) into the receptor-binding domain, so that now the virus will be targeted to certain types of blood cells that contain the EPO receptor on them. Let's say you wanted to target that receptor," he continued. "You could take our structure, and you would have a good candidate as to where to put the hormone to make a chimeric protein. That way, hopefully, you would direct the virus to just those cells so they would be more specifically infected."

Picturing "the scenarios that one could envision," Kim mentioned gene therapy to treat cancer. "There," he pointed out, "one would want to target malignant tumors, and not healthy cells. In muscular dystrophy," he went on, "one might take aim at a selected group of muscle cells, but not every muscle in the body. Another common one that people talk about," he added, " is trying to target stem cells in the hematopoietic system. One could go down the list."

The Whitehead Institute has deposited the coordinates of this new structure at the Brookhaven Laboratory National Data Bank, Kim said, adding, "which is available on the web. So people in both academia and industry can use the 3-D information to try to design their favorite vectors to do what they want them to do."

However, the structure per se is not the object of a patent application. "I think," the structural biologist explained, "that as people develop things that have to do with cell specificity, then that will serve as the basis for patents."

Structure Meets Function At Harvard

Gene therapist and molecular geneticist Richard Mulligan, a professor of genetics at Harvard University Medical School, in Boston, collaborates with Whitehead's Kim and his associates.

Commenting on their new structural map, Mulligan told BioWorld Today, "One of our first targets with this approach is we're trying to specifically transduce the tumor vasculature, which consists of replicating endothelial cells. We're interested in trying to generate viruses that will specifically release things that either clot the blood vessels, kill the tumor, and essentially inhibit angiogenesis."

Mulligan continued: "Now we have a road map that's far superior to how to go about getting targeted infection at a high efficiency. So Peter Kim and I are trying to make use of what he knows about structural information, with what we know about the vector part of things, to try to make much more precise gene insertions and substitutions to achieve targeted infection at high efficiency.

"This is not novel," he pointed out, "in the sense that we will for the first time be able to get a targeted infection. That has been demonstrated before, but it's never been demonstrated in a way that was sufficiently efficient to be practically useful.

"In the past," he continued, "previous to having this structure, people were making a rather futile, brute-force effort to target retroviruses to specific cell types. That is, without having the structural information, they were generally lopping off big chunks of the envelope and trying to paste to what was left of it single-chain antibody sequences, or ligands or other protein-coding sequences that would hopefully give them specificity of infection." *