A 19-NUCLEOTIDE SEQUENCE UPSTREAM OF THE 5' MAJOR SPLICE DONOR IS PART OF THE DIMERIZATION DOMAIN OF HUMAN-IMMUNODEFICIENCY-VIRUS-1 GENOMICRNA

The genome of all retroviruses, including human immunodeficiency virus type 1 (HIV-1), consists of two identical RNAs noncovalently Linked n ear their 5' end. Dimerization of genomic RNA is thought to modulate s everal steps in the retroviral life cycle, such as recombination, tran slation, and encapsidation. We report the results of experiments desig ned to identify the 5' and 3' boundaries of the dimerization domain of the HIV-1 genome: (1) An HIV-1 RNA starting at nucleotide 252 or at o ther downstream positions (four tested) does not dimerize despite the inclusion of the whole of a previously proposed dimerization domain (n ucleotides 295-401); (2) an RNA starting between nucleotides 242 and 2 49 (five positions tested) dimerizes to a variable extent depending on the starting position; (3) an RNA starting at nucleotide 233 or at ot her upstream positions (five tested) is fully or >80% dimeric; (4) an RNA starting at nucleotide 1 but lacking the 233-251 or the 242-251 re gion is, respectively, fully monomeric or about 50% monomeric; (5) the 343-401 region contains two strings of G's (GGGGG367 and GGG384) that had been postulated to promote genome dimerization through the format ion of,guanine quartets. We have deleted the 379-401, 358-401, and 343 -401 regions from otherwise dimeric RNAs without changing their abilit y to dimerize. We reach three conclusions: (1) a dimerization signal e xists upstream of the major 5' splice donor (nucleotide 290); (2) the previously proposed downstream dimerization domain is insufficient to promote dimerization and has a 3' half that is not necessary to obtain fully dimeric RNAs; (3) the 5' boundary of the HIV-1 dimerization dom ain is located somewhere between nucleotides 233 and 242, and the 3' b oundary is located no farther than at nucleotide 342, making it possib le that the 5' and 3' boundaries of the HIV-1 dimerization domain are both located within the leader sequence. We speculate that the 248-270 or 233-285 region forms a hairpin that is the core dimerization domai n of HIV-1 RNA.