775 G H AN A A N D IV ORY COAS T are situated between the two major initial foci of the African HIV-1 and HIV-2 epidemics. It is in such countries that dual HIV-1/HIV-2 serological reactivity (dual seropositivity) is most common. The proportion of dual seropositivity has increased in West Africa countries, where HIV-2 was initially the main cause of the epidemic. The diagnosis of dual HIV-1/HIV-2 infections may be complicated, as HIV-2 DNA cannot always be detected and the levels of HIV-2 DNA in dual seropositives may decrease with progression to AIDS. Also, even low levels of HIV-2 antibodies can imply dual infection. 4 The lack of HIV-2 DNA and/or low HIV2 antibody level may thus be related not only to the diagnostic methods used, but possibly to suppression of HIV-2 by HIV-1. HIV-2 infection may protect against HIV-1 superinfection, which could have impact on the development of protective vaccines. It has been suggested that the HIV-2 protection from subsequent HIV-1 infection is mediated by b -chemokines and CD8 1 cells, and is restricted to HIV-1 strains using the CCR5 receptor. Furthermore, conflicting reports on cross-neutralization of HIV-1 isolates by HIV-2 sera exist. Since the V3 loops of the viruses are important for chemokine receptor usage and neutralization, a critical look at this region is of interest for understanding the interaction between HIV-1 and HIV-2 in vivo. We therefore determined the V3 sequences of HIV-1 strains obtained from epidemiologicall y matched patients seropositive for both types or HIV-1 only. Peripheral blood mononuclear cells (PBMCs) were obtained from patients who were HIV seropositive by World Health Organization (WHO) criteria (Table 1), attending semi rural or urban AIDS clinics in the southern part of Ghana during 1996. Ethical permission was obtained from the Ministry of Health (Accra, Ghana), the University of Ghana Medical School (Accra, Ghana), and the Huddinge Hospital (Huddinge/ Stockholm, Sweden). Lymphocyte subset determination was performed by a flow cytometric method and the patients classfied as asymptomatic or symptomatic on the basis of the WHO expanded clinical definition of AIDS. Two commercial kits were used for analysis of anti-HIV reactivity. In brief, a dot immunobinding assay (Target HIV1-HIV2; V-Tech, Pomona, CA), which is based on homologous synthetic peptides in the HIV-1 and HIV2 transmembrane proteins, was used for screening. Confirmatory tests with a line immunoassay including recombinant proteins and peptides for HIV-1, and a peptide for HIV-2 (Innolia; NV Innogenetics, Antwerp, Belgium), were done according to the manufacturer instructions. Samples that reacted with HIV1 gp41 (and other bands) with a rating of 1 1 or above were considered HIV-1 seropositive, and those having additional reactivities of at least 11 with the HIV-2 gp36 band were classified as dual seropositive. Two million PBMCs were lysed and DNA was extracted by phenol±chloroform and ethanol precipitation. To amplify an env fragment encoding the V3 loop, the following outer primers were used: JA 167 [5 9 TAT C(C/T)T TTG AGC CAA TTC C(C/T)A TAC A] and JA 170 [5 9 GTG ATG TAT T(A/G)C A(A/G)T AGA AAA ATT C], with conditions previously described. For the nested step, primers V3-3 (59 biotin-AATGGCAGTCTAGCAGAA) and V3-4-M13 (5 9 M13-CTGGGTCCCCTCCTGAGG) were used. Direct sequencing was performed using magnetic beads with covalently coupled streptavidin (Dynabeads, M280-streptivid in; Dynal A.S., Oslo, Norway), an AutoRead sequencing kit (Pharmacia Biotech, Uppsala, Sweden), and an automated laser fluorescence sequencer (ALFexpress; Pharmacia Biotech). Sequences approximately 252 or 249 nucleotides long were manually aligned. The phylogenetic analysis was done with the TREEDON package and 1000 bootstrap resamplings. Distance calculations were derived by neighbor joining, using the Jin and Nei method ( g distribu-