HIV-1 Gag mutants demonstrate more relaxed RNA binding specificities. (A) S. cerevisiae L40-coat, stably expressing the LexA-MS2 coat fusion protein, was transformed with different combinations of plasmids encoding the indicated RNA hybrids and fusion proteins. Colonies from the selected transformants were replicated on nitrocellulose filters and assayed for β-galactosidase activity (indicated by the appearance of a blue spot). To avoid color changes due to differences in density, we spotted pools of yeast transformants consisting of many colonies and then assayed the pools. The RNA binding specificities of the randomly mutated Gag proteins were compared to that of the wild-type (wt) protein. The negative control included in this assay was a fusion protein consisting of the residues from the matrix to the capsid, excluding the NC domain (MA-CA). The red color of the yeast transformed with the plasmid expressing the IRE RNA was due to the absence of the ADE2 nutritional marker from the plasmid (57), which was irrelevant to this screen. (B) Quantitative results for β-galactosidase activities in the yeast three-hybrid system, as measured in a liquid assay. The β-galactosidase units for wild-type (wt) Gag and two of the mutants (Mut1 and -2) were calculated as described previously (5). β-Galactosidase activities following interactions with HIV-1Ψ (white bars), HaMSVΨ (black bars), and IRE (gray bars) are presented as means (n = 3). (C) RNA filter-binding assay. DIG-labeled, in vitro-transcribed HaMSVΨ RNA was incubated with a recombinant wild-type HIV-1 Gag protein (wt), an NC deletion Gag mutant (deltaNC), or Mut1. Samples were spotted in triplicate on a nitrocellulose membrane, and bound RNAs were specifically detected with an anti-DIG antibody (α-DIG-POD; 1:1,000) by standard Western blot techniques.

Zinc fingers with site-directed lysine substitutions demonstrate broader RNA binding affinities. (A) Zinc finger mutations. The locations of the lysine substitutions were decided following an alignment of the two zinc fingers. These substitutions are demonstrated above the sequence, and the mutation numbering and random mutations are demonstrated below the sequence. (B) The lysine substitution mutants (indicated at the top of the panel) were tested in the yeast three-hybrid assay, and their binding to HIVΨ, HaMSVΨ, or IRE RNA was determined. The activation of lacZ was determined by measurement of the β-galactosidase activity (as described in the legend for Fig. 2A).

HIV-1 NC mutants have different transduction efficiencies. 293T cells were cotransfected with pHR′-CMV-GFP, pMD.G, and DNAs encoding the listed NC mutations in the context of pHIVgptSVPA. Two days following transfection, the cells were analyzed by fluorescence-activated cell sorting to quantify GFP expression. The culture medium from the cells was collected and used to infect naïve cells in the presence of 8 μg/ml of Polybrene (hexadimethrine bromide). After 48 h, the infected cells were analyzed by fluorescence-activated cell sorting to detect GFP. The data are presented as infection indexes normalized to transfection efficiencies, with the samples standardized against the wild type as described before (48). The data shown represent a minimum of three independent experiments (means ± standard deviations).

Schematic representation of the yeast three-hybrid screen utilized to select for HIV-1 Gag mutants with enhanced binding to HaMSVΨ RNA. Binding of the RNA hybrid (consisting of the MS2 coat binding sites and the viral Ψ sequence [5]) to the LexA-MS2 coat protein and to the Gal4AD-HIV Gag protein leads to the formation of a complex which activates the transcription of the lacZ gene. In this system, wild-type HIV-1 Gag interacts strongly with HIV-1Ψ but not with HaMSVΨ, resulting in minimal activation of the lacZ gene in the latter case (top panel). Plasmid libraries expressing random Gag mutants were screened with this system to identify mutants with enhanced binding activities for HaMSVΨ resulting in increased β-galactosidase expression (curved arrow, bottom panel).

Schematic representation of HIV-1 Gag and zinc finger sequences demonstrating the positions of random mutations. The random mutants that were generated and selected for enhanced binding to HaMSVΨ all had mutations in the zinc finger motifs of the NC, either as single point mutations or double mutations. For two of the mutants (M46K and M46V), a zinc finger mutation was accompanied by an additional mutation in the carboxy region of the CA domain.

Expression of and VLP production with HIV-1 Gag mutants. The figure shows the results of a Western blot analysis of Gag proteins in cell lysates (top panels) and virion pellets (bottom panels). 293T cells were transfected with pHIVgptSVPA encoding wild-type Gag or the listed NC mutations. Two days following transfection, the cells were harvested and lysed, the medium was collected, and the virions were purified from it by centrifugation through a 25% sucrose cushion. The different panels represent different experiments and thus show the wild-type activity for that specific experiment. The proteins were electrophoresed in a 10% acrylamide gel and transferred to a polyvinylidene difluoride membrane for Western analysis with an anti-CA antibody.

Increased RNA encapsidation in virions of the N17K mutant. (A) For quantification of the vector and helper RNAs packaged in virions of the NC N17K mutant, 293T cells were transfected with pHR′-CMV-GFP, pMD.G, and pHIVgptSVPA expressing the N17K mutant or wild-type Gag. At 2 days posttransfection, RNAs were extracted from virions purified through a 25% sucrose cushion and analyzed by slot blot hybridization with either a GFP (top panel) or a GPT (bottom panel) riboprobe to detect vector and helper RNAs, respectively. To compare the encapsidation of foreign RNAs by wild-type and N17K virions, we transfected 293T cells with pQCXIP-gfp-C1 and pHIVgptSVPA expressing the N17K mutant or wild-type Gag. At 2 days posttransfection, total RNAs were extracted from the cells and from virions normalized by exogenous RT. The RNAs were reverse transcribed and the indicated dilutions of cDNA were amplified by PCR with primers specific to GFP. This semiquantitative analysis was performed with virion (B) and cellular (C) samples. Undiluted samples of virion and cellular RNA prior to reverse transcription (no RT) were used to verify the absence of contaminating plasmid DNA (D). PCR samples were electrophoresed in a 2% agarose gel.