The HIV-1 Tat protein has a versatile role in activating viral transcription

Abstract

It is generally acknowledged that the Tat protein has a pivotal role in HIV-1 replication because it stimulates transcription from the viral long terminal repeat (LTR) promoter by binding to the TAR hairpin in the nascent RNA transcript. However, a multitude of additional Tat functions have been suggested. The importance of these functions is difficult to assess in replication studies with Tat-mutated HIV-1 variants because of the dominant negative effect on viral gene expression. We therefore used an HIV-1 construct that does not depend on the Tat-TAR interaction for transcription to reevaluate whether or not Tat has a second essential function in HIV-1 replication. This HIV-rtTA variant uses the incorporated Tet-On gene expression system for activation of transcription and replicates efficiently upon complete TAR deletion. Here we demonstrated that Tat inactivation does nevertheless severely inhibit replication. Upon long-term culturing, the Tat-minus HIV-rtTA variant acquired mutations in the U3 region that improved promoter activity and reestablished replication. We showed that in the absence of a functional TAR, Tat remains important for viral transcription via Sp1 sequence elements in the U3 promoter region. Substitution of these U3 sequences with nonrelated promoter elements created a virus that replicates efficiently without Tat in SupT1 T cells. These results indicate that Tat has a versatile role in transcription via TAR and U3 elements. The results also imply that Tat has no other essential function in viral replication in cultured T cells.

Fig. 1.

Mutations in Tat. (A) Schematic of the HIV-rtTA proviral DNA genome, with the LTR region subdivided in the U3, R, and U5 domains. The Tat-TAR axis of transcription regulation was inactivated by multiple nucleotide substitutions in the bulge and loop sequences of TAR (TAR^m; crossed boxes). Transcription and replication of the virus were made dox dependent by the introduction of tetO elements in the U3 promoter region and replacement of the Nef gene by the rtTA gene. We constructed different HIV-rtTA variants with either a wild-type or a mutated Tat gene. (B) Mutations in Tat. The sequence of the first Tat-coding exon from splice acceptor 3 (SA3) to the splice donor site (SD) is indicated with the translated amino acid sequence. The mAUG, stop, fs, and Y26A mutations in the Tat gene are boxed in gray. These mutations do not affect the Vpr and Rev amino acid sequences (the Vpr TAG stop codon and the Rev AUG start codon are underlined), splice acceptor (SA) sites, or known splice enhancer (ESE) and silencer (ESS) motifs. (C) Activity of Tat mutants. The HIV-rtTA plasmids and the control plasmid pBluescript (−) were cotransfected with a Tat-responsive HIV-1 LTR promoter/luciferase reporter construct into 293T cells. Luciferase production was measured after the cells were cultured with 1 μg/ml of dox for 48 h. Average values obtained in multiple experiments are shown, with the wild-type level set at 100% and the error bars indicating standard deviations (n = 8 for HIV-rtTA plasmids; n = 4 for pBluescript). Statistical analyses (two-sided unpaired sample t tests) demonstrated that all mutated constructs and the pBluescript control differed significantly from the wild-type construct (P < 0.01). Furthermore, the wt, Y26A, and mAUG constructs differed significantly from the pBluescript control, while the stop and fs mutants did not. (D) Replication of HIV-rtTA variants. SupT1 T cells were transfected with the different HIV-rtTA plasmids and cultured with 1 μg/ml of dox. CA-p24 levels in the culture supernatants were measured by ELISA. No virus replication was observed in the absence of dox (data not shown). The replication curves shown are representative of data obtained in four experiments.

Fig. 2.

Evolution of HIV-rtTA-Tat^stop. (A) Comparison of the U3-R regions from HIV-1 and HIV-rtTA. The HIV-rtTA variants have two tetO elements inserted between the NF-κB and Sp1 binding sites and several nucleotide substitutions in TAR (TAR^m). (B) Mutations observed in the tetO region of the HIV-rtTA-Tat^stop mutant upon culturing with dox for up to 272 days (culture I) or 283 days (culture II). Cellular proviral DNA was isolated at 108, 206, and 272 days (culture I) and at 55, 130, and 283 days (culture II), and the LTR region was subsequently PCR amplified and either directly sequenced (population [pop] sequence; indicated on the left) or cloned into the TA-cloning vector followed by sequencing of three to six TA clones for each sample (with the frequency at which each sequence was observed indicated on the left). The nucleotide sequences of the tetO region between the NF-κB binding sites and the SpI binding sites are shown for HIV-rtTA-Tat^stop (upper line) and the evolved viruses. The tetO core sequence is boxed in gray. The number of tetO elements present in each sequence is indicated on the right, with asterisks indicating the sequences that were recloned into HIV-rtTA-Tat^stop (Fig. 3). Nucleotide substitutions are indicated with lowercase characters. Δ, nucleotide deletion.

Fig. 3.

Multiplication of tetO elements rescues replication of HIV-rtTA-Tat^stop. (A) The HIV-rtTA-Tat^stop mutant has two tetO sites in the LTR promoter region. The 4- and 6-tetO configurations that were observed upon long-term culturing (sequences indicated with asterisks in Fig. 2) were recloned into the 5′ and 3′ LTRs of HIV-rtTA-Tat^stop. (B to D) Replication of the HIV-rtTA-Tat^stop variants with different tetO configurations was assayed in SupT1 cells (B), PBMCs (C), and 174 × CEM cells (D). Cells were transfected with the different HIV-rtTA plasmids (B and D) or infected with 293T-produced virus (C) and cultured with 1 μg/ml of dox. Replication was monitored by CA-p24 ELISA on culture supernatant samples. For comparison, we included HIV-rtTA-Tat^wt and HIV-rtTA-Tat^Y26A. The replication curves shown are representative of data obtained in four (B), three (C), or two (D) experiments. (E) LTR-tetO-promoter/luciferase reporter constructs with two, four, or six tetO sites were cotransfected with an rtTA-expressing plasmid into 293T cells. Cells were cultured without or with 1 μg/ml of dox for 2 days, after which luciferase production was quantified to measure promoter activity. Average values obtained in eight experiments are shown, with error bars indicating standard deviations. The average activity of the 2-tetO construct with dox was set at 100%. Statistical analyses (two-sided unpaired sample t tests) demonstrated that the promoter activity of the 2-tetO, 4-tetO, and 6-tetO constructs with dox differed significantly (2 tetO versus 4 tetO, 2 tetO versus 6 tetO, and 4 tetO versus 6 tetO; P < 0.01).

Fig. 4.

Tat activates HIV-rtTA gene expression. 293T cells were transfected with the HIV-rtTA constructs and 0 to 50 ng of pTat. After the cells were cultured with dox for 48 h, the CA-p24 levels in the culture supernatants were measured by ELISA. Average values obtained in multiple experiments are shown (n = 6 to 9), with error bars indicating standard deviations. A two-sided unpaired-sample t test was used to identify statistically significant differences (P < 0.01) between the wt and Tat mutants at 0 ng of pTat (^a, asterisk indicates a statistically significant difference between the wt and mutant) and between the HIV-rtTA construct at 0 ng of pTat and the same construct at 0.5, 5, or 50 ng of pTat (^b, asterisk indicates a statistically significant difference between the constructs without and with pTat). When no asterisk is shown, the difference was not statistically significant (P > 0.01).

Fig. 5.

Tat activates gene expression via U3 elements. HIV-rtTA-Tat^wt (wt) and HIV-rtTA-Tat^stop (stop) constructs with different 5′-LTR-promoter configurations were cotransfected with 0 to 50 ng of pTat into 293T cells. CA-p24 levels in the culture supernatants were measured after the cells were cultured with dox for 48 h. Average values obtained in multiple experiments are shown, with error bars indicating standard deviations. (A) HIV-rtTA promoter configuration with two tetO sites placed between the NF-κB and Sp1 binding sites (n = 12 to 16). (B) ΔNF constructs in which the NF-κB sites and upstream U3 region were deleted (n = 3). (C) In ΔNF ΔTAR constructs, the TAR hairpin was replaced with the unrelated ER3 hairpin (n = 6) (22). (D) tetO-CMV constructs in which the 5′ U3 and TAR regions were replaced by a promoter sequence consisting of three tetO elements coupled to a 30-bp minimal TATA box region from the CMV IE promoter and a hairpin that is based on the +1/+16 region of the CMV promoter (n = 12 to 16). (E) In the tetO-CMV-Sp1 constructs, the HIV-1 Sp1 sites were cloned between the tetO and TATA box regions of tetO-CMV (n = 3). All constructs have the tetO-CMV configuration in the 3′ LTR. A two-sided unpaired sample t test was used to identify statistically significant differences (P < 0.01) between the wt and corresponding stop constructs at 0 ng of pTat (^a, asterisk indicates a statistically significant difference between the wt and stop constructs) and between the construct at 0 ng of pTat and the same construct at 0.5, 5, or 50 ng of pTat (^b, asterisk indicates a statistically significant difference between the constructs without and with pTat). When no asterisk is shown, the difference was not statistically significant (P > 0.01). CA-p24 production of all HIV-rtTA constructs was very low (>10-fold reduced) in the absence of dox (data not shown).

Fig. 6.

The tetO-CMV promoter. (A) In the tetO-CMV HIV-rtTA variant, the U3-TAR promoter region was replaced with a promoter consisting of three tetO elements coupled to the 30-bp minimal TATA box region from the CMV IE promoter and a hairpin sequence that is based on the +1/+16 sequence of the CMV transcript. This promoter was developed in another HIV-rtTA study (to be published). The EcoRV and HindIII restriction enzyme sites in the U3 and R regions, respectively, are indicated. HIV-1-derived nucleotides are shown in uppercase and non-HIV nucleotides in lowercase. Reintroduction of the three Sp1 sites from the HIV-1 LAI strain between the tetO and TATA elements resulted in the tetO-CMV-Sp1 variant. (B) Secondary structure of the new 5′-hairpin element that replaces TAR, as predicted with MFold RNA-folding software (81).

Fig. 7.

The tetO-CMV virus replicates efficiently without Tat. Replication of HIV-rtTA variants with the tetO-CMV or tetO-CMV-Sp1 promoter configuration in both the 5′ and 3′ LTRs and with the Tat^wt or Tat^stop gene (wt and stop, respectively) was tested in SupT1 cells (A) and PBMCs (B) cultured with 1 μg/ml of dox. The wild-type HIV-1 LAI strain (cultured without dox) and HIV-rtTA-Tat^wt (cultured with 1 μg/ml of dox) were included for comparison. SupT1 cells (A) were transfected with 5 μg of the proviral constructs. PBMCs (B) were infected with equal amounts of 293T-produced virus (5 ng of CA-p24). The CA-p24 levels in the culture supernatants were measured by ELISA. The replication curves shown are representative of data obtained in three (A) or two (B) experiments.