Matrix and envelope coevolution revealed in a patient monitored since primary infection with human immunodeficiency virus type 1

Abstract

Lentiviruses, including human immunodeficiency virus type 1 (HIV-1), typically encode envelope glycoproteins (Env) with long cytoplasmic tails (CTs). The strong conservation of CT length in primary isolates of HIV-1 suggests that this factor plays a key role in viral replication and persistence in infected patients. However, we report here the emergence and dominance of a primary HIV-1 variant carrying a natural 20-amino-acid truncation of the CT in vivo. We demonstrated that this truncation was deleterious for viral replication in cell culture. We then identified a compensatory amino acid substitution in the matrix protein that reversed the negative effects of CT truncation. The loss or rescue of infectivity depended on the level of Env incorporation into virus particles. Interestingly, we found that a virus mutant with defective Env incorporation was able to spread by cell-to-cell transfer. The effects on viral infectivity of compensation between the CT and the matrix protein have been suggested by in vitro studies based on T-cell laboratory-adapted virus mutants, but we provide here the first demonstration of the natural occurrence of similar mechanisms in an infected patient. Our findings provide insight into the potential of HIV-1 to evolve in vivo and its ability to overcome major structural alterations.

FIG. 1.

Schematic representation of the proviral constructs used. The upper diagram shows the HIV-1 gag and env genes with the corresponding proteins. The parental viral proteins are indicated by boxes of different colors: black, NL(AD8); gray, early clones of HIV-1 from patient 153; white, late clones of HIV-1 from patient 153. The arrows indicate the stop codons shortening the gp41 CTs by 20 amino acids (clones L1 and L2). (A) Schematic representation of pNL(AD8)-NX proviral constructs harboring the sequences encoding the gp41 CTs derived from HIV-1 present at early (clone E) and late stages (clones L1 and L2) in patient 153. (B) Schematic representation of pNL(AD8)-NX proviral constructs carrying various combinations of sequences encoding the gp41 CTs (clones E, L1, and L2) and matrix proteins (early clone ME, late clone ML) derived from HIV-1 present at early or late stages in patient 153. (C) Schematic representation of pNL(AD8)-NX proviral constructs harboring the sequence encoding the truncated gp41 CT (clone L1) associated with the sequence encoding the early matrix protein (clone ME) with various combinations of the three amino acid substitutions (K25R, V34I, or I45L) found in all HIV-1 variants present at the late stage in patient 153.

FIG. 2.

Amino acid sequences of gp41 CTs and matrix proteins from patient 153. (A) Sequence alignments for gp41 CTs are shown for an early viral variant with a full-length sequence (clone E), for two late viral variants harboring a 20-amino-acid truncation (clones L1 and L2), and for the NL(AD8)-NX-WT virus. The Env sequences upstream from the NheI restriction site and downstream from the XbaI restriction site are derived from NL(AD8). The highlighted domains include the C-terminal region of the membrane-spanning domain (msd) and the amphipathic alpha-helical domains LLP-1, LLP-2, and LLP-3. Trafficking motifs (Y₇₁₂SPL, Y₈₀₂W₈₀₃, and L₈₅₅L₈₅₆) are boxed. Amino acid identity (.), stop codons (*), and substitutions are indicated. (B) Sequence alignments for matrix proteins are shown for an early viral variant (clone ME) and for a late viral variant (clone ML). The three amino acid substitutions specific to late matrix proteins are highlighted by gray shading. The matrix sequences upstream from the AatII restriction site (localized in the 5′ noncoding region of the viral genome) and downstream from the BsrGI restriction site (localized at the junction between the matrix protein p17 and the capsid protein p24) are derived from NL(AD8). Amino acid identity (.) and substitutions are indicated. Amino acid numbers correspond to the HXB2 sequence.

FIG. 3.

Replication kinetics of NL(AD8)-NX viruses carrying gp41 CTs and matrix proteins derived from the HIV-1 present at early or late stages in patient 153. Virus stocks, obtained by transfecting 293T cells with the indicated molecular clones, were normalized as a function of p24 capsid protein concentration and used to infect MT4.R5 cells (A, B, and C), PBMC (D, left), or MDMs (D, right), as described in Materials and Methods. Variations in p24 concentrations were monitored in the culture supernatant over time. Each experiment was performed at least twice in duplicate, with similar results obtained in terms of both the hierarchy and the extent of replication. Errors bars indicate the standard deviation for duplicate infections.

FIG. 4.

Singe-round infectivity of NL(AD8)-NX viruses carrying gp41 CTs and matrix proteins derived from HIV-1 present at the early or late stages in patient 153. Virus stocks of the indicated molecular clones, obtained by transfecting 293T (A) and HeLa cells (B) or by infecting MT4.R5 cells (C), PBMC (D), or MDMs (E), were used to infect TZM-bl cells, as described in Materials and Methods. Eight 10-fold serial dilutions of each virus stock were tested in 96-well plates. After 48 h, infection of the target cells was assessed by staining with X-Gal. The infectivity of each virus is expressed as the number of infected cells per ng of p24. The data are expressed as means ± the standard deviations.

FIG. 5.

Env incorporation of NL(AD8)-NX viruses carrying gp41 CTs and matrix proteins derived from HIV-1 present at the early or late stages in patient 153. Virus stocks of the indicated molecular clones were obtained by transfecting 293T (A) and HeLa cells (B) or by infecting MT4.R5 cells (C), PBMC (D), or MDMs (E), as described in Materials and Methods. Env incorporation into virions was assessed, using viruses purified by centrifugation through a 20% sucrose cushion. Viral pellets were lysed in TNE buffer containing 1% Triton X-100 and gp120 was quantified by ELISA. The results shown were obtained from a representative experiment performed in duplicate. Errors bars indicate standard deviations. (F) Western blot analysis of Env incorporation into virions produced by HeLa cells. The upper panel shows the blot probed with anti-gp120 and anti-p24 antibodies. The positions of the Env glycoprotein gp120, the Gag precursor Pr55, the Gag processing intermediate Pr41 and the processed p24 capsid protein are indicated. The blot was stripped and reprobed with an anti-gp41 antibody to detect the Env glycoprotein gp41 (lower panel; see Materials and Methods).

FIG. 6.

Analysis of Gag and Env intracellular distributions by confocal microscopy. HeLa cells transfected with the indicated molecular clones were fixed and processed for immunofluorescence analysis with 2G12 anti-Env MAb and a mouse MAb specific for the matrix p17 protein. Before fixation, all of the cells were treated with 50 μg of cycloheximide/ml for 3 h. Protein distribution was assessed by confocal microscopy and the Pearson channel correlation coefficient (R) (“1” indicating perfect colocalization, “0” indicating no correlation) was calculated for each sample, as described in Materials and Methods. Scale bar, 10 μm.

FIG. 7.

Cell-to-cell HIV transfer. Cell-to-cell viral transfer was measured by a flow cytometry-based assay. Primary CD4⁺ T lymphocytes (A and B) or MT4R5 cells (C and D) were infected with VSV-G-pseudotyped HIV-1 particles carrying different combinations of matrix and Env. The cells were then cocultured with CFSE-labeled target cells, such that targets could be distinguished from donors. The appearance of Gag⁺ target cells was followed over time (A and C). The efficiency of transmission (B and D) was measured by calculating the area under the curve in three independent experiments, with values for the NL(AD8) clone being defined as 100%. Means and standard deviations are shown.