|Home | About | Journals | Submit | Contact Us | Français|
We and others have recently uncovered the existence of human T-cell lymphotropic virus type 3 (HTLV-3), the third member of the HTLV family. We have now sequenced the full-length HTLV-3Pyl43 provirus. As expected, HTLV-3Pyl43 contains open reading frames corresponding to the gag, pol, env, tax, and rex genes. Interestingly, its long terminal repeat (LTR) includes only two Tax-responsive elements, as is the case for type 3 simian T-cell lymphotropic viruses (STLV-3). Phylogenetic analyses reveal that HTLV-3Pyl43 is closely related to central African STLV-3. Unexpectedly, the proximal pX region of HTLV-3Pyl43 lacks 366 bp compared to its STLV-3 counterpart. Because of this deletion, the previously described RorfII sequence is lacking. At the amino acid level, Tax3Pyl43 displays strong similarities with HTLV-1 Tax, including the sequence of a PDZ class I binding motif. In transient-transfection assays, Tax3Pyl43 activates the transcriptions from HTLV-3, HTLV-1, and HTLV-2 LTRs. Mutational analysis indicates that two functional domains (M22 and M47) important for transactivation through the CREB/ATF or NF-κB pathway are similar but not identical in Tax1 and Tax3Pyl43. We also show that Tax3Pyl43 transactivates the human interleukin-8 and Bcl-XL promoters through the induction of the NF-κB pathway. On the other hand, Tax3Pyl43 represses the transcriptional activity of the p53 tumor suppressor protein as well as the c-Myb promoter. Altogether, these results demonstrate that although HTLV-3 and HTLV-1 have only 60% identity, Tax3Pyl43 is functionally closely related to the transforming protein Tax1 and suggest that HTLV-3, like HTLV-1, might be pathogenic in vivo.
Four types of primate T-cell lymphotropic viruses (PTLVs) have been discovered. While three of them, i.e., PTLV-1, PTLV-2, and PTLV-3, include human T-cell lymphotropic viruses (HTLV-1, HTLV-2, and HTLV-3) and simian T-cell lymphotropic viruses (STLV-1, STLV-2, and STLV-3), the fourth type (HTLV-4) has so far been found to consist only of a unique human isolate (4, 52).
STLV-3 (formerly named PTLV-L) was isolated shortly before STLV-2 (14). The STLV-3 lineage is composed of at least two subtypes that correspond more or less to the geographical sources of the isolates (East Africa and west-central Africa) (7, 14, 24, 25, 27, 46, 47, 49, 51). Although STLV-3 is much more widespread than STLV-2 in African monkeys, STLV-3 infection has not been linked to any pathology so far, yet neither longitudinal follow-up nor clinical studies have been performed. Sequence comparisons of STLV-3 full-length proviruses pointed out that these strains are highly divergent from HTLV-1, HTLV-2, and STLV-2 (40%, 38%, and 38% divergence, respectively) (7, 14, 24, 25, 27, 46, 47, 49, 51). The overall STLV-3 genomic organization is similar to that of HTLV-1 and HTLV-2; the gag, pro, pol, env, tax, and rex genes are present (24, 25, 46, 47, 51). Sequence analyses also revealed that STLV-3 long terminal repeat (LTR) sequences possess only two 21-bp repeats (or Tax-responsive elements [TREs]), while HTLV-1 and HTLV-2 LTRs retain three of these sequences (24, 25, 46). The impact of the lack of a TRE on viral transcription is not known. Interestingly, apart from Tax and Rex, the STLV-3 prototype (STLV-3PH969) pX region was reported to contain only one additional open reading frame (ORF) whose corresponding mRNA can be amplified by reverse transcriptase (RT)-PCR (47). This mRNA could be translated into a putative 84-amino-acid-long protein designated RorfII, which shares some similarities with the HTLV-1 p12 protein (47). Of note, the RorfII mRNA was not detected in the two other STLV-3 strains (PPAF3 and CTO604) that have been analyzed lately by RT-PCR. In fact, sequence analysis of these isolates revealed either a mutation in the splice acceptor sequence and/or a stop codon in the RorfII sequence (24, 25). At this point, it is therefore not clear whether, apart from Tax and Rex, additional ORFs are commonly present in the STLV-3 pX sequence.
We and others have uncovered the existence of HTLV-3, a third member of the HTLV family, in two Cameroonian individuals (4, 52). We therefore ought to obtain the complete sequence of HTLV-3Pyl43 in order to determine its genomic organization and to characterize its Tax protein.
Full-length genome comparisons and phylogenetic analyses allow us to demonstrate that the HTLV-3 genome organization is similar to that of HTLV-1 and HTLV-2 and is related to some central African STLV-3 strains. Like STLV-3, HTLV-3 LTRs contain only two Tax-responsive elements. Unexpectedly, a 366-bp sequence is lacking in the HTLV-3 proximal pX region. This sequence corresponds in part to the RorfII sequence that was previously described for STLV-3PH969 and also to the 3′ end of a putative antisense transcript. We have lately demonstrated that HTLV-3 Tax protein is expressed in vivo (6). Our results now reveal that Tax3Pyl43 is functionally more closely related to the transforming HTLV-1 Tax protein than to HTLV-2 Tax, suggesting that HTLV-3, like HTLV-1, might be pathogenic in vivo.
The HTLV-3 sample originated from a 62-year-old Bakola pygmy living in a remote settlement in the Ocean Department of southern Cameroon (4).
HeLa and 293T cell lines were grown in Dulbecco's modified Eagle's medium supplemented with fetal bovine serum (10%) and antibiotics (penicillin, 100 U/ml; streptomycin, 100 μg/ml). Jurkat cells were grown in RPMI 1640 medium supplemented with 10% fetal bovine serum, antibiotics, and l-glutamine. All cell lines were maintained at 37°C in 5% CO2.
High-molecular-weight DNA was extracted from the peripheral blood buffy coat of the Pyl43-infected individual by using a QIAmp blood mini kit (QIAGEN). The reaction mixture contained 1 μg of DNA, 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 2 mM MgCl2 (ABgene), a 200 μM concentration of the deoxynucleoside triphosphates (Roche), 10 pmol of each primer, and 1.25 U of Thermo-start DNA polymerase (Abgene). Table S1 in the supplemental material shows the primer sequences and the corresponding annealing temperatures. The nested-PCR products were purified on a 1.5% agarose gel using a QIAquick gel extraction kit (QIAGEN) and cloned into the pCR2.1 vector (TA cloning kit; Invitrogen). For each PCR fragment, 6 to 10 clones were sequenced using a BigDye Terminator cycle kit and an ABI 3100 automated sequencer (Applied Biosystem).
Alignment of multiple nucleotide sequences with 6,812 bp corresponding to the concatenated gag-pol-env-tax genes of HTLV-3Pyl43 together with PTLV sequences that are available in GenBank was performed with the DAMBE program as previously described (53). The general time-reversible model was used to create a tree based on the neighbor-joining method. The parameters (base frequencies, substitution ratios, proportions of invariant sites, and gamma distributions [alpha shape]) iteratively estimated in this analysis were then used to construct the neighbor-joining trees. With the same parameters, the maximum likelihood trees were constructed using a heuristic research mode with the nearest-neighbor interchange algorithm and then with the tree bisection and reconnection algorithm. The bootstrap values were evaluated with 1,000 replicates (44, 53). To estimate the separation between HTLV-3Pyl43 and STLV-3CTO604 strains, we tested a log-normal, relaxed molecular-clock hypothesis for PTLV on the previous alignment using the general time-reversible I+G substitution model implemented with BEAST software (free) from the University of Oxford (10). The clock was calibrated by estimating a divergence for PTLV-1 strains of (50 ± 10) × 103 years. The molecular clock was found to be appropriate, as the effective sample size for each taxon tested was high. The evolutionary rate was around 1.4 × 10−6 nucleotide substitutions per site per year.
To amplify and clone the HTLV-3 LTR, we used a previously described PCR technique (23) with three partially overlapping HTLV-3Pyl43 clones corresponding to nucleotides 1 to 137 (part of clone 18), 44 to 213 (clone 19), and 111 to 446 (part of the clone 1). The primers used were LTR-pyl43-Mlu1 sense (CCCCCCACGCGTTGTCAGTGATGATGAGCCTCG), Tax8511 antisense (GGGTAGGGAGAGACGTCAGAGCC), LTR8268 sense (GGAGGACAAATAGCTGAATCATCCG), LTR8411 antisense (CAAATTTTTAGGGTTATCGTCAGAGCC), LTR111 sense (CCAAGGCTCTGACGTCTCTCCCTAC), and LTR-pyl43-Bgl2 antisense (AAAAAAAGATCTGGAGTGATGGCCTAGCTCGAC). The 446-bp-long LTR fragment encompassing the U3 sequence (i.e., with the TREs) and part of the R sequence was then digested and inserted into the pGL2 basic plasmid as previously described (6). The nucleotide sequence of the construct was determined as described above.
Because human HTLV-3 viral RNA is not available, the tax3 cDNA sequence from STLV-3CTO602, which is highly related to Pyl43, was used as a matrix for in vitro site-directed mutagenesis. The simian Tax3 sequence was therefore converted into the human Tax3 sequence (S322→N322, L333→P333, and D335→G335). The nucleotide sequences of both the human pSG5M-Tax3 and the human green fluorescent protein (GFP)-Tax3 constructs were determined as described above. The pSG5M-Tax3 and GFP-Tax3 constructs for the M22 (A130M131→A130S131) and the M47 (L319L320→R319S320) domains were also constructed using a QuikChange site-directed mutagenesis kit (Stratagene) as previously described (23).
Twenty-four hours after transfection, cells were washed twice with phosphate-buffered saline (PBS), lysed (Tris-HCl [pH 7.4], 50 mM; NaCl, 120 mM; EDTA, 5 mM; NP-40, 0.5%; Na3VO4, 0.2 mM; dithiothreitol, 1 mM; phenylmethylsulfonyl fluoride, 1 mM) in the presence of protease inhibitors (Complete; Boehringer), and incubated on ice (5). Cell debris was pelleted by centrifugation. The protein concentration was determined by the Bradford method (Bio-Rad). Samples were loaded into 10% Tris-glycine gels (Invitrogen), subjected to electrophoresis, and transferred onto a polyvinylidene fluoride membrane (Immobilon-P; Millipore). Membranes were blocked in a 5% PBS-milk solution and incubated with anti-GFP (1:1,000; catalog no. 8371; Clontech) antibodies. The next day, the membranes were washed and incubated with anti-mouse horseradish peroxidase-conjugated secondary antibodies (Amersham Biosciences) and developed using a SuperSignal West Pico chemiluminescent-substrate kit (Pierce). To control for the amount of protein loaded per well, membranes were stripped with the Reblot Plus kit (Chemicon International) and reprobed with a specific anti-β-tubulin antibody (1:1,000; catalog no. sc9104; Santa Cruz Biotechnology).
For microscopic analyses, HeLa cells were seeded on an eight-well glass slide at a concentration of 3.5 × 104 cells/well and transfected the next day with 0.3 μg of the different GFP plasmids using Effectene reagent (QIAGEN), as described previously (23). Twenty-four hours after transfection, the cells were washed with PBS, fixed with 4% paraformaldehyde (Sigma), and washed with PBS. Nucleic acids were stained with DAPI (4′,6′-diamidino-2-phenylindole)-containing mounting medium (Vectashield; Vector). Cells were visualized with a Zeiss Axioplan 2 imaging microscope at a magnification of ×40 using a Zeiss Axiocam HRc (color) camera and Zeiss Apotome software as described previously (5).
Jurkat cells (3 × 106) were transfected with HTLV-3, HTLV-1, HTLV-2-LTR-luciferase (luc), NF-κB-luc, Pr-Bcl-XL, MRE-luc, and interleukin-8 reporter (IL-8536)-luc plasmids, together with pSG5M-Tax3, Tax3M22, and Tax3M47 plasmids, using Superfect reagent (QIAGEN) as previously described (26, 32, 33). HeLa cells were transfected with the PG13py-luc plasmid and the p53-expressing plasmid, as previously described (26). The amount of DNA transfected was always equalized by the addition of a control pSG5M empty vector. All the transfections were carried out in the presence of a phRG-TK vector in order to normalize the results for the transfection efficiencies. Reporter activities were assayed 24 hours posttransfection using a dual-luciferase reporter assay system (Promega) with a Berthold LB9500C luminometer, as described previously (23).
Twenty-four hours after transfection of Jurkat cells, total RNA was extracted with an RNeasy mini kit (QIAGEN) and treated with a DNase RNase-free DNA set (QIAGEN) to avoid any carryover of the Tax-containing plasmid. Total RNA (0.5 μg) was then used as a matrix for RT-PCR according to instructions accompanying the OneStep RT-PCR kit (QIAGEN). PCR was performed using the Tax 602TaxEcoRI and 602TaxBamHI primers, which allow the amplification of a 1,050-bp PCR product, which was then resolved on a 1% agarose gel as previously described (6).
293T cells were transfected using Polyfect reagent (QIAGEN) with 4 μg of pGFP, pGFP-Tax1, pGFP-Tax2, or Pyl43 pGFP-Tax3 and pSGF-cl.15, a PDZ domain-containing protein. Twenty-four hours later, cells were washed twice with cold PBS and lysed in CHRIS buffer (50 mM Tris-HCl, 0.5% NP-40, 200 mM NaCl, 0.1 mM EDTA, 10% glycerol) in the presence of protease inhibitors for 30 min. Cell extracts (350 μg) were precleared with immunoglobulin G anti-mouse M280 Dynabeads (Dynal) for 1 h. Fifteen microliters of beads was mixed with anti-FLAG monoclonal antibody (Sigma) and incubated for 1 h. The complex of beads and anti-FLAG antibody was then added to the precleared cell extracts and incubated for 1 h. After the beads were washed three times with CHRIS buffer (same as described above except with 300 mM NaCl), they were run on 10% Tris-glycine gels (Invitrogen) and electro-blotted onto a polyvinylidene fluoride membrane (Immobilon-P). Western blot analysis was performed to detect GFP fusion proteins using anti-GFP antibody (BD).
Nucleotide sequence accession number. The HTLV-3Pyl43 sequence has been deposited in GenBank under accession no. DQ462191.
A series of 19 successive nested-PCR amplifications allowed us to amplify and to clone the complete genome of HTLV-3Pyl43 (8,553 bp) (GenBank accession number DQ462191) from the high-molecular-weight DNA extracted from HTLV-3Pyl43-infected uncultured blood cells (Fig. (Fig.1A).1A). In all cases, nested PCR was necessary to amplify the HTLV-3 viral sequences. This might reflect a low proviral load. The overall genomic organization of HTLV-3 is similar to that of HTLV-1 and HTLV-2, including the presence of the gag, pro, pol, env, tax, and rex genes (Fig. (Fig.1B).1B). There is, however, one major difference between the HTLV-3 and HTLV-1 and HTLV-2 sequences: the HTLV-3 LTR is 695 bp long and possesses only two 21-bp repeat sequences, while the HTLV-1 and HTLV-2 LTRs have three of these TREs (data not shown). This is due to a deletion of the TATA-distal 21-bp repeat in the U3 region, a characteristic that has previously been reported for STLV-3 strains (47).
The HTLV-1 pX region encodes at least five proteins (Tax, Rex, p12, p13, and p30) whose functions are now well characterized (for a review, see reference 31). Apart from Tax and Rex, the existence of a doubly spliced viral mRNA that would be generated by alternative splicing was reported for the STLV-3PH969 prototypical strain (14). This transcript is predicted to encode an 84-amino-acid-long, highly hydrophobic and leucine-rich protein designated RorfII. However, the RorfII transcript was not detected when investigators analyzed two other STLV-3 strain transcripts by RT-PCR. This was due either to mutations in the RorfII splice acceptor sequence, to a stop codon in the RorfII coding sequence, or to both. Surprisingly, HTLV-3Pyl43 RorfII could not be amplified when primers positioned within this sequence were used (data not shown). Two independent sets of PCR primers that were predicted to amplify a larger proviral sequence, encompassing the RorfII putative ORF, however, allowed amplification of HTLV-3 DNA fragments that were shorter than the expected size (Fig. (Fig.1C,1C, compare lane 1 versus lane 3 and lane 4 versus lane 6; see Table S1 in the supplemental material for primers and sequences). Sequence analysis subsequently revealed a 366-bp deletion in the HTLV-3 proximal pX region, leading to the absence of the putative RorfII DNA sequence (Fig. (Fig.1D).1D). Of importance, this deletion does not alter Env, Tax, or Rex ORFs. Whether this deletion is also present in the other known HTLV-3 strain (HTLV-32026ND) remains to be determined.
Recent studies have demonstrated that the antisense HTLV-1 strand encodes a protein, HTLV-1 b-ZIP factor (HBZ), whose roles might be to support the proliferation of adult T-cell leukemia cells (2, 12, 37) and to enhance infectivity and persistence in inoculated animals (54). Despite the 366-bp deletion, a putative ORF corresponding to an “antisense encoded protein” (AEP) is still present in the HTLV-3 provirus (data not shown). However, this protein, if it exists, would be shorter and most likely have impaired functions.
We then compared the HTLV-3Pyl43 genome with those of all available STLV-3 as well as with HTLV-1 and HTLV-2 prototypes. To this end, sequence comparisons of the different Gag-, Pol-, Env-, and Tax-encoding genes and LTR sequences were performed (Table (Table1).1). Comparisons revealed that the HTLV-3Pyl43 sequence was highly related to that of the published STLV-3CTO604 isolate (99% identity for Gag, Pol, Env, and Tax sequences and 98% identity for the LTR sequences). The percentages of similarity between the genome of HTLV-3Pyl43 and those of the other STLV-3 viruses were roughly equivalent for Gag, Pol, Env, and Tax sequences (85 to 92% similarity) and ranged from 86% to 88% for the LTR sequences. Finally, HTLV-3Pyl43 is as divergent from HTLV-1 as it is from HTLV-2 (Table (Table1).1). In a comparison of full-length sequences, HTLV-3Pyl43 displays 99.12% homology with the STLV-3CTO604 strain, if the 366-bp deletion is excluded from the analysis or 95.06% homology if the deletion is included in the sequence comparison. On the other hand, HTLV-3Pyl43 has 87%, 88%, and 86% similarity with the East African STLV-3PH969, the central African STLV-3CTONG409, and the West African STLV-3PPAF3 isolates, respectively.
Phylogenetic analyses based on the concatenated gag-pol-env-tax genes clearly demonstrate that HTLV-3Pyl43 belongs to the PTLV-3 west-central African group (see Fig. Fig.22 for neighbor-joining analysis; data not shown for the maximum likelihood analysis). Finally, we performed a molecular-clock analysis with the free program BEAST, v1.3, using the same gag-pol-env-tax alignment. The calculated dates for the emergence of the different HTLV-1 types or for the separation between the different STLV-3 strains were in the range of those previously published (24, 50; data not shown). Our analysis allowed us to estimate that the separation between HTLV-3Pyl43 and STLV-3CTO604, its closest homolog, occurred within the last 3,800 ± 70 years (mean 95% confidence interval, 1,000 to 7,000 years). This is not without precedent for the PTLV, as a previous study demonstrated that the spread of HTLV-1 subtype F in humans also occurred within the last 3,000 years (50).
We have lately demonstrated that Tax3Pyl43 is expressed in vivo and elicits a humoral response in HTLV-3Pyl43-infected individuals (6). Sequence comparisons reveal that the Tax3Pyl43 protein displays 26% and 30% divergence at the amino acid level from HTLV-1 and HLTV-2 Tax, respectively (Table (Table1).1). Nevertheless, most of the previously reported Tax1 domains are conserved in the Tax3Pyl43 sequence (Fig. (Fig.3).3). Of interest, the Tax3Pyl43 putative KID-like domain (amino acids 81 to 95), which is critical for binding to CBP/p300 (15), is similar to that of Tax1. There are, however, two amino acid changes, one of which is on lysine 85, which is important for the ability of Tax1 to bind CBP/p300 (15). The 2 amino acids corresponding to the “M22 domain” (43, 44) are also different in the Tax1 and Tax3Pyl43 proteins (T130L131→A130M131). There is also some variation in the putative Tax3Pyl43 CR2 binding sequence (amino acids 312 to 319). This domain has been shown to be important for the ability of Tax1 to bind CBP/p300 (38). In contrast, the Tax3Pyl43 “M47 domain” is identical to that of Tax1. Finally, Tax3Pyl43 contains a PDZ binding motif at its C terminus. Such a domain has repeatedly been shown to be required for the ability of Tax1 to transform rat fibroblasts or to induce the proliferation of human peripheral blood mononuclear cells (11, 54).
Tax1 is a shuttling protein that is found predominantly in the cell nucleus, where it forms speckled structures (39). Tax1 contains a noncanonical nuclear localization signal (NLS) and a nuclear export signal. By contrast, HTLV-2 Tax (Tax2), governed by a minimal domain encompassing amino acids 90 to 100, is located mainly in the cytoplasm of the HTLV-2-immortalized or transformed T cells (5, 23). Since the localization of Tax in the nucleus is critical for its ability to transactivate the viral LTR (13), we used a previously described GFP-based technique to determine the intracellular localization of the Tax3Pyl43 protein (23). In transiently transfected HeLa cells, GFP-Tax3Pyl43 has a strong nuclear localization that is similar to that of Tax1 and simian Tax3 but different from that of Tax2 (Fig. (Fig.4A,4A, compare panels b, d, and g with panel c). Nevertheless, some cytoplasmic speckles that are reminiscent of Tax2 are also visible in Tax3Pyl43-expressing cells. These structures are absent from Tax1-expressing HeLa cells. Based on the known domains of Tax1 (43), we constructed two Tax3Pyl43 mutants, i.e., Tax3M22 (A130M131→A130→S131) and Tax3M47 (L319L320→R319S320). The localization of both mutants is similar to that of Tax3, but both mutants have fewer cytoplasmic speckles (Fig. (Fig.4A,4A, compare panel d with panels e and f). Western blot analysis demonstrated that all GFP fusion proteins were expressed to similar levels (Fig. (Fig.4B,4B, upper panel). As a control for protein loading, the membrane was stripped and probed with an anti-β-tubulin antibody (Fig. (Fig.4B,4B, lower panel).
As stated above, sequence analysis of the C terminus of Tax3Pyl43 revealed that the protein contains a class I PDZ binding motif, “SSV-COOH” (Fig. (Fig.5A).5A). We therefore determined whether Tax3Pyl43 could bind clone 15, a protein that contains such a sequence (Fig. (Fig.5B,5B, upper panel). 293T cells were transfected with clone 15 and either GFP or GFP-Tax constructs. As a positive control, the Tax1 protein, which binds strongly to clone 15, was used. We also used Tax2 as a negative control, which, as previously shown (6), does not bind to clone 15. As anticipated from the sequence comparison, Tax3Pyl43 binds to clone 15, demonstrating that this protein contains a PDZ binding motif. We also used a series of Tax3Pyl43 PDZ mutants. As expected for type 1 PDZ binding motif, mutations at either position 348 or 350 of Tax3Pyl43 disrupted its ability to bind to clone 15, while mutations at position 349 had no effect (Fig. (Fig.5B).5B). Finally, an assessment of GFP, GFP-Tax1, GFP-Tax2, and GFP-Tax3 input levels revealed that identical amounts of proteins had been used for the immunoprecipitations (Fig. (Fig.5C5C).
Tax1 protein activates both the CREB/ATF and the NF-κB signaling pathways. These events are responsible to some extent for the cell immortalization/transformation that is observed in vivo and in vitro (for a review, see reference 19). In order to study Tax3-mediated transactivation, we performed a series of transient-transfection assays with different HTLV LTR constructs. We first determined whether Tax3Pyl43 could mediate transcription from the HTLV-3 LTR in lymphocytes. This cell population is likely to be the natural target of HTLV-3 infection in vivo, as it is for HTLV-1 and HTLV-2 (22). As stated above, the HTLV-3 LTR sequence lacks the distal TRE. In HTLV-1, these three TREs, of which the middle TRE is the most critical, are functionally nonequivalent for the basal transcription (3). Nevertheless, the presence of only two TREs is sufficient to induce a high level of LTR activation in the presence of Tax1. As expected, despite the lack of one TRE, Tax3Pyl43 efficiently transactivated the HTLV-3 promoter in Jurkat cells (Fig. (Fig.6A).6A). As anticipated, Tax3Pyl43 M22 was able to activate the HTLV-3 LTR, while Tax3Pyl43 M47 activation of the CREB/ATF pathway was impaired (Fig. (Fig.6A).6A). Because HTLV-3 was discovered in a human population in whom HTLV-1 and HTLV-2 are endemic (data not shown), we also investigated whether Tax3Pyl43 could transactivate the heterologous human retroviral promoters (LTR1 and LTR2, respectively). As shown in Fig. 6B and C, Tax3Pyl43 efficiently activates transcription from both LTRs in transiently transfected lymphocytes.
Since Tax1, Tax3Pyl43, and Tax3Pyl43 M47 activate NF-κB as well, it was surprising that Tax3Pyl43 M22 transcriptional activity was only partially impaired (Fig. (Fig.6D).6D). The latter result suggests that the Tax3 sequence may contain another domain that can, to some extent, compensate for the M22 mutation.
Tax1- and Tax2-specific antibodies do not allow detection of Tax3 (data not shown), and a Tax3-specific antibody is not yet available. Therefore, Tax3 expression was monitored by RT-PCR. A Tax3-specific signal was present in Tax3Pyl43-transfected Jurkat cells (Fig. (Fig.6E,6E, lanes 4, 6, and 8). The band was absent from mock-transfected cells (Fig. (Fig.6E,6E, lane 2) or when the reverse transcriptase was not added to the PCR mix (Fig. (Fig.6E,6E, lanes 1, 3, 5, and 7).
Next, we investigated whether Tax3Pyl43 represses or activates some cellular promoters whose regulation was previously demonstrated to be altered upon Tax1 expression through the activation of the NF-κB or the CREB/ATF pathway. We first tested whether Tax3Pyl43 or the different Tax3Pyl43 mutants inhibit p53 transcriptional activity (Fig. (Fig.7A).7A). As with Tax1 (26, 29, 35), Tax3Pyl43 and Tax3Pyl43 M22 repressed p53 activity in HeLa cells, suggesting that, in these cells, the activation of NF-κB is not needed for inhibiting p53 functions. We performed a series of Western blot analyses with cellular extracts obtained from the transfected HeLa cells. Consistent with the ability of Tax3Pyl43 and Tax3Pyl43 M22 to inhibit p53 activity better than Tax3Pyl43 M47 does, the p53 protein was slightly more efficiently stabilized in these cells (Fig. (Fig.7A,7A, lower panel). This observation was also reported when p53 functions were repressed with HTLV-1 Tax. A control Western blot using an antibody specific for β-tubulin demonstrated that equivalent amounts of protein were analyzed (Fig. (Fig.7A,7A, lower panel).
We have previously demonstrated that the activation of the NF-κB pathway by Tax1 protein leads to transcriptional inactivation of c-Myb (32). c-Myb is essential for a controlled balance between cell growth and differentiation, and aberrant c-Myb activity has been reported for numerous human cancers. To further determine whether Tax3Pyl43 is functionally related to Tax1, we performed another series of transfections using a c-Myb promoter construct (MRE). Jurkat cells express c-Myb and therefore do not require the addition of ectopically expressed c-Myb. In T cells, Tax3Pyl43 and Tax3Pyl43 M47 but not Tax3Pyl43 M22 transrepress the MRE construct (Fig. (Fig.7B).7B). These results are similar to those obtained with Tax1 (32).
We wanted to determine whether Tax3Pyl43 transactivates two cellular promoters that were previously reported to be upregulated by Tax1 through the activation of the NF-κB pathway (28, 33). In the first series of experiments, a Bcl-XL promoter construct was used. Unexpectedly, the Tax3M22 mutant increased Bcl-XL promoter reporter activity, but no significant activation was observed with the Tax3M47 mutant (Fig. (Fig.7C).7C). These results are different from those obtained with Tax1. Given the fact that Tax3Pyl43 M22 partially activates the NF-κB pathway (Fig. (Fig.7D),7D), we could not determine whether or not NF-κB is involved in Bcl-XL transactivation. The experiment was therefore repeated in the presence of the dominant negative IκBαS32/36A mutant. Strikingly, the presence of this mutant prevented the effect of Tax3Pyl43 and Tax3Pyl43 M22 (Fig. (Fig.7C).7C). This result confirms that NF-κB is involved in Bcl-XL transactivation but also suggests that Tax3Pyl43 M47, even if it activates this pathway, has another defect that prevents it from transactivating the Bcl-XL promoter.
Finally, we tested an interleukin-8 promoter construct that is also known to be upregulated by Tax1, through the activation of the NF-κB pathway (28). As with the Bcl-XL reporter, both Tax3Pyl43 and Tax3Pyl43 M22 transactivated the IL-8 reporter, while the Tax3M47 mutant was impaired. The IκBαS32/36A mutant also prevented the effect of Tax3Pyl43 (Fig. (Fig.7D).7D). As a control, the experiment was repeated with Tax1M22 and Tax1M47. As previously reported (45) and in contrast to Tax3Pyl43 M47, Tax1M47 transactivated the Bcl-XL promoter while Tax1M22 did not (data not shown). Altogether, these results suggest that, in vitro, the activation of the NF-κB pathway by Tax3 is necessary but is not sufficient to transactivate the Bcl-XL and the IL-8 promoters, suggesting that Tax1 and Tax3, although functionally highly related in vitro, are not entirely equivalent.
Emerging viral pathogens have either invaded a new host species or expanded into new geographic populations of host species. As represented by the human immunodeficiency virus epidemic or the Spanish influenza outbreak in the early 1920s, viruses can be highly transmissible and virulent. These episodes appear to be unpredictable. A viral pathogen may be benign while residing within a reservoir species, yet on entering a new host, it increases in virulence (41).
Until recently, the HTLV family comprised only two members: HTLV-1, known as the etiological agent of adult T-cell leukemia/lymphoma, which is one of the worst leukemia viruses; and HTLV-2, which although very similar in sequence to HTLV-1, is barely leukemogenic. We and others have recently discovered a third member of the HTLV group (4, 52), and our next challenge is to determine whether HTLV-3 is more similar to HTLV-1 than to HTLV-2 in terms of pathogenicity. Of note, studies of the first STLV-3-infected animals were reported almost a decade ago (14), but those animals as well as the STLV-3-infected monkeys described more recently (24, 25, 27, 48, 51) were never studied longitudinally. For this reason, it is not possible to determine whether, as is the case for some STLV-1-infected animals, animals can suffer from adult T-cell leukemia/lymphoma-like or other diseases (9). We have now sequenced the HTLV-3 provirus, which is very similar to a central African STLV-3 strain (STLV-3CTO604). Furthermore, HTLV-3Pyl43 shares some sequence features with all STLV-3 strains (24, 25, 46). As an example, the viral promoter contains only two Tax-responsive elements. Together with the phylogenetic analyses, these results confirm that, as is the case for HTLV-1 (21, 41), HTLV-3 arose in humans from interspecies transmissions, which, according to our molecular-clock analysis, is estimated to have occurred within the last 3,800 years. Of note, this estimate is not without precedent, as HTLV-1 subtypes B, D, and F were previously estimated to have arisen in humans between 3,000 and 13,600 years ago (50).
Unexpectedly, however, the proximal pX region of HTLV-3, in which the RorfII sequence is removed, is shorter than that of its simian counterpart. For this reason, it would be of interest to compare the proximal pX region of HTLV-3Pyl43 with that of the strain reported by Wolfe et al. (52). This would allow us to determine whether this feature is common to all HTLV-3 or whether it is specific to HTLV-3Pyl43.
The STLV-3 minus strand is predicted to contain an ORF that encodes a 221-amino-acid-long protein (data not shown). Similar to HTLV-1 HBZ, this protein would possess several NLS sequences. The NLS sequences that are present in HTLV-1 HBZ are critical for its localization and, likely, for its functions (18). On the other hand, the putative HTLV-3 antisense protein would be only 103 amino acids long due to the 366-bp deletion in the proximal pX region. This protein would lack the NLS domains and would probably be functionally impaired. Because an HTLV-3-infected cell line is not available for analysis, the existence and the putative role of such a protein cannot be investigated at this point. The lack of a protein in a PTLV genome is not without precedent. Various studies have shown that defective proviruses comprise 25 to 40% of all HTLV-1 genomes present in lymphocytes from infected individuals or in cells infected with HTLV-1 in vitro (8, 17, 20, 34). These viruses retain different parts of the genomes, i.e., both LTRs, the amino terminus of gag, and the portions of pX encoding the Tax, Rex, p12, and p30 proteins (16, 40). Whether or not these defective viruses are the only genomes that are present in these individuals is very difficult to determine.
Full-length genomes that lack the expression of some viral proteins were also reported. As an example, apart from HTLV-1 subtype A, which has always been used for molecular studies, the other HTLV-1 genomes that are available for analysis (subtypes B and C) lack the p12 ATG codon. Therefore, it is very unlikely that the p12 protein is present in individuals infected with HTLV-1 subtypes B and C (A. Vandamme and S. Van Dooren, personal communication).
Because Tax is one of the key players in HTLV-1 pathogenicity, we examined HTLV-3 Tax properties. Sequence analysis first revealed that, as with Tax1 but not Tax2, Tax3Pyl43 contains a PDZ binding motif. Such a domain, which is absent from Tax2, is critical for the ability of the viral protein to induce T-cell proliferation and to persist in vivo (54). This is one argument that strongly suggests that HTLV-3 possesses some of the HTLV-1 properties. We show here that Tax3Pyl43 is capable of inducing the CREB/ATF and the NF-κB pathways. We also provide evidence that Tax3Pyl43 transactivates both the HTLV-1 and the HTLV-2 promoters. This is of importance as HTLV-3Pyl43 was discovered in a population in whom both HTLV-1 and HTLV-2 are endemic. Interestingly, central African monkeys that were coinfected by STLV-1 and STLV-3 were described earlier (7). Whether coinfection would accelerate the occurrence of a PTLV-associated disease remains to be determined. So far, both cases of HTLV-3 infection were reported in individuals who were not coinfected with other types of HTLV.
Unexpectedly, we have also found that, even if Tax3Pyl43, like Tax1, ultimately transrepresses some cellular gene products (p53 or c-Myb) and transactivates others (IL-8 and Bcl-XL), the mechanisms that are involved might be slightly different. In particular, we have demonstrated here that the activation of NF-κB by Tax3Pyl43 is not sufficient for transactivating the interleukin-8 and the Bcl-XL promoters.
For these reasons, the construction of both an HTLV-3 and an STLV-3 molecular clone is now needed to investigate viral expression in the context of a provirus. This will allow us to determine in particular whether the 366-bp deletion has an impact on the HTLV-3 viral life cycle. Indeed, earlier analyses have allowed us to conclude that not only Tax but other pX-encoded proteins (p12, p13, p30, and HBZ) are involved in viral pathogenesis. Due to the deletion in the putative HTLV-3 AEP sequence, we anticipate that HTLV-3 is likely to encode a nonfunctional protein compared to its STLV-3 counterpart. Based on HTLV-1 data (2) and the possibility that STLV-3 AEP is functionally related to HBZ, it is expected that a loss of the HTLV-3 AEP functions would result in a lower proviral load in vivo.
We thank K. E. Boulukos for the generous gift of the Pr-Bcl-XL plasmid, L. Boxer for the MRE-luc plasmid, K. Matsushima for the IL-8536-luc plasmid, W. Greene for the pCMV4-HA-IκBS32/36Α plasmid, P. Jalinot for the pSGF-cl.15 plasmid, A.-M. Vandamme for the generous gift of STLV-3PH969 DNA and for sharing unpublished results, S. Van Dooren for help with the phylogeny, and the Plate Forme d'Imagerie Dynamique laboratory.
This work was supported by a grant from l'Association de Recherche sur le Cancer (ARC grant no. 4781) to R.M. and fellowships from le Ministère de la Recherche to S.A.C. and from la Ligue Contre le Cancer to S.C. R.M. is supported by INSERM.