Structure of AIDS virus binding site unraveled The atomic structure of a fragment of the cell-surface glycoprotein CD4, which plays a critical role in the pro cess by which the human immuno deficiency virus infects immune sys tem cells, has been determined by two independent research groups. The structure sheds light on the lock-and-key binding between CD4 and HIV and on CD4's normal func tion in the immune response. Per haps most important, it provides re searchers with a template for devel oping drugs that might act to block HIV's infectivity. CD4 belongs to the "immunoglo bulin superfamily" of molecules, most of whose members function in biochemical recognition processes. The most familiar members of this superfamily are antibodies. The primary amino acid sequence of CD4 indicates it consists of a large extracellular segment, a transmem brane span, and a short cytoplasmic tail. The extracellular segment has four domains, designated D l , D2, D3, and D4, whose sequences resem ble those of immunoglobulins. Soluble CD4, which consists of the extracellular portion of the molecule, has been under investigation as a po tential agent to combat HIV, the vi rus that causes AIDS. The first step in HIV infection of a cell is the highly specific binding of a viral glycopro tein called gpl20 to CD4. Biomedical researchers speculated that soluble CD4 might coat HIV particles and block them from infecting the Τ lym phocytes that carry the CD4 protein, but so far clinical results have been disappointing in that AIDS patients have shown no improvement. The research on CD4 structure, re ported in Nature [348, 411 and 419, (1990)], was carried out by Stephen C. Harrison, of Howard Hughes Medi cal Institute and Harvard University, and coworkers at Harvard, Harvard Medical School, and Upjohn; and Wayne A. Hendrickson, of Howard Hughes Medical Institute and Co lumbia University, and coworkers at Columbia, the University of Califor nia (San Diego), and SmithKline Beecham Pharmaceuticals. Efforts to determine the structure
Harrison: unusual "spined" molecule of soluble CD4, say both Harrison and Hendrickson, have been ham pered because, thus far, crystals of the protein containing all four do mains diffract x-rays rather poorly. Both researchers, while continuing efforts to produce suitable crystals of soluble CD4, turned to a fragment of the protein containing the Dl and D2 domains, for which excellent
crystals have been obtained. Al though all four domains appear to be involved in the normal function of CD4, gpl20's interaction with CD4 is limited to the Dl domain. Both groups found that the Dl and D2 domains possess structures that are quite similar to immunoglobu lins, each domain possessing what is known as a ^-barrel structure. Never theless, Hendrickson says, the struc tures for CD4 that have been predict ed on the basis of immunoglobulin structure "don't capture the essence of CD4." The structure reveals that several amino acid residues that had been identified as important in bind ing gpl20 exist on an exposed ridge about 25-Â long on the surface of CD4. The data suggest that the binding site on gpl20 is a groove that is at least 25-Â long and 12-Â wide. Another striking feature of the molecule is the linkage between Dl and D2, Harrison says. Rather than being a flexible polypeptide strand— as might have been expected from antibody structure—the final βstrand of Dl continues directly into the first β-strand of D2, acting as "spine" to produce a single, rodlike molecule about 65 Â in length. Rudy Baum
Direct methane-to-formaldehyde route found Formaldehyde, a valuable petro chemical intermediate, may be the next commodity chemical to be pro duced by direct oxidation of meth ane. A research group at the Univer sity of Liverpool in the U.K. has been able to produce appreciable yields of the chemical by simply varying the operating conditions of a reactor that normally would pro duce ethylene and ethane. Richard W. Joyner of Liverpool's chemistry department notes that most of the effort for the catalytic oxidation of methane is currently aimed at producing ethylene and/or ethane. But these C 2 hydrocarbons have few uses beyond polymeriza tion. This is not the case for such Cx compounds as methanol and formal dehyde, and there is growing inter est in their production via direct ox idation of methane. Joyner's group, which includes Justin S. J. Hargreaves and Graham
J. Hutchings, has studied the partial oxidation of methane at a constant temperature of 1023 Κ using magne sia catalysts [Nature, 348, 428 (1990)]. At this low temperature, the reac tion selectivity is more sensitive to oxygen conversion, which turns out to be critical in controlling product distribution. The usual procedure is to use ex tremely severe conditions to pro duce ethylene and ethane with oxy gen conversions of up to 90%. In the Liverpool procedure, much milder conditions are used, and oxygen conversion is kept to 10% or less. Be low 10% conversion, the character of the reacting system changes com pletely. Most significantly, formal dehyde becomes the major product when reactor flow rates are high. The changes in selectivity are re versible and depend only on the flow rate through the reactor. At oxygen conversions below December 3, 1990 C&EN 7
