We describe mutants of human immunodeficiency virus type-1 (HIV-1) strain NL4-3, which are lacking the thirteenth, fifteenth, or seventeenth sites for N-linked glycosylation (g13, g15, g17) of the envelope protein gp120. All three sites are located within the hypervariable V3 loop region of gp120. Those mutants lacking carbohydrates g15 or combinations of g15/g17 showed markedly higher infectivity for GHOST cells (human osteosarcoma cells) expressing CXCR4 (GHOST-X4), compared to the fully glycosylated NL4-3 wild type virus. In addition, these mutants could also infect cells which exhibits low background expression of CXCR4, corresponding to <10% of that observed for GHOST-X4 cells. In addition to the enhanced infectivity observed, mutants lacking g15 and g17 showed increased resistance to inhibition by SDF-1, the natural ligand of CXCR4. Thus, loss of the oligosaccharides g15 and g17 in the V3 region of gp120 markedly influences CXCR4-specific infection.

The present study was designed to determine the antibody specificity for the human immunodeficiency virus type 1 (HIV-1) V3 domains of infectious and noninfectious virions present in the serum of AIDS patients. To accomplish this, HIV-1 was isolated in the presence of autologous antibodies from the serum samples of six AIDS patients in HIV-1-negative donor peripheral blood mononuclear cells by short-term cultivation. The isolated virus, defined as the infectious cell-free virus (iCFV), was characterized by sequence analysis of the proviral DNA coding for the third hypervariable (V3) region of the external glycoprotein gpl20. This was carried out by amplifying and cloning the V3 region. In all six cases studied, 20 randomly selected V3 clones derived from the proviral DNA of the iCFV, 20 clones from patient cell-free virus, and 20 clones from cell-integrated virus were sequenced to study the distribution and frequency of the intrapatient virus population. The number of major virus variants in the six patients ranged from three to nine. The various V3 sequences found in the AIDS patients showed the typical amino acid pattern of the syncytium-inducing and non-syncytium-inducing viral phenotypes characteristic for the late stage of infection. However, only one patient-specific iCFV variant was detected within the 20 V3 clones analyzed per virus isolation. For the six patients a total of 34 V3-loop variants, either iCFV or non-iCFV, was observed. All 34 V3-loop sequences were expressed as glutathione-S-transferase fusion proteins (V3-GST). The autologous antibody response to the V3-GST fusion proteins was studied by Western immunoblot analysis. A strong antibody response to almost all non-iCFV V3-GST proteins was found in the sera of the six patients. In contrast, the autologous antibody response to the six iCFV V3 loops was undetectable (in four patients) or very faint (in two patients) compared with that to the non-iCFV V3 loops. Five of the six iCFV loops showed positively charged amino acids at positions strongly associated with the syncytium-inducing phenotype. These findings suggest that our in vitro isolation system selects for virions which are not recognized by V3-specific antibodies and are infectious both in vitro and in vivo.

No abstract

SUMMARYThis study was performed to analyse correlates of viral escape in AIDS patients. Peripheral blood mononuclear cells (PBMC) from HIV ¹ donors were inoculated with AIDS patients' serum to detect neutralization-resistant cell-free virus. Infectious virus was detected by polymerase chain reaction (PCR) and analysed by sequencing the V3 region. The escaped virus species was compared with all V3 virus variants found in the patients' PBMC and plasma. In one patient escaped virus was also compared with variants found in CD4 þ T cells isolated by FACS from blood, spleen and lymph node. The frequency of the virus variants was determined by cloning and sequence analysis of 20 V3 clones for each PCR amplification. To monitor anti-V3 antibodies by ELISA, each V3 sequence was expressed as fusion with glutathione S-transferase (GST-V3). In our AIDS patients, a V3-directed antibody response against the infectious virus V3 loop was not detectable. In contrast, virus variants unable to infect the donor PBMC in vitro were well recognized by homologous V3-directed antibody. After an interval of 1 year the frequency of these variants clearly decreased, while at the same time the escaped variants grew out and finally represented the predominant viral species both in plasma and PBMC. The infectious variants lacking V3 antibody response were also predominant in CD4 þ T cells in spleen and lymph node. Our data indicate that the escape of virus variants is closely related to the lack of V3-directed antibody.

Purpose of the study: Determination of HIV-1 coreceptor tropism is a major prerequisite before starting treatment with a CCR5-antagonist. While most of the patients currently under treatment with maraviroc are probably infected with HIV-1 subtype B viruses, recently published data show differences in the distribution of coreceptor tropism in different HIV-1 subtypes. Methods: In a Germany-wide project within the HIV-GRADE society, V3-loop sequences of 2466 isolates were analysed with geno2pheno for coreceptor tropism using a FPR cut-off of 10%. HIV-1 subtype was determined by using the COMET HIV subtyping tool. Sequences consisted of at least the V3 loop fragment. The ratio of CCR5 vs CXCR4 tropic viruses was calculated for each subtype. A normalized mean for all analyzed subtypes was calculated to extrapolate the overall ratio of coreceptor usage distribution. From this the expected distribution in the particular subtype was calculated and compared to the observed one. Statistical analysis was performed using the chi2 test. Summary of Results: Most samples were classified as HIV-1 subtype B (79%, n=1952). Other subtypes present in at least 23 samples were A1 (9.5%, n=234), C (4.8%, n=118), CRF01_AE (2.2%, n=55), G (1.6%, n=39), D (1.1%, n=27), F (0.9%, n=23). The calculated normalized mean distribution over all subtypes was 71% CCR5- vs. 29% CXCR4-tropic viruses. No significant difference compared to the mean distribution could be observed for HIV-1 subtypes B (71/29%), C (76/24%) and F (70/30%). Higher rates of CXCR4 tropic virus were detected in subtypes D (52/48%, p=0.01) and CRF01_AE (49/51%, p=0.001), while in HIV-1 subtypes A1 (22/78%, p=0.02) and G (13/87%, p=0.02), a higher rate of CCR5-tropic virus was observed. Conclusions: Our analysis shows a different distribution of CCR5 and CXCR4 tropic virus in some subtypes. In contrast to other publications, we could not observe a statistically significant difference in subtype C compared to the overall mean distribution, while we could confirm a higher rate of CXCR4-tropic virus in subtype D, as previously described. Without further data on treatment success of patients with non-B subtypes under treatment with maraviroc, it remains unclear if subtype-specific differences in the distribution of tropism are biased by differences in clinical variables before test or if there is a bias in the tropism interpretation system. In the latter case, individual interpretation cut-offs for different subtypes may be necessary