Identification of a Domain within the Human T-Cell Leukemia Virus Type 2 Envelope Required for Syncytium Induction and Replication

Abstract

In vitro infection by human T-cell leukemia virus type 1 and 2 (HTLV-1 and HTLV-2) can result in syncytium formation, facilitating viral entry. Using cell lines that were susceptible to HTLV-2-mediated syncytium formation but were nonfusogenic with HTLV-1, we constructed chimeric envelopes between HTLV-1 and -2 and assayed for the ability to induce syncytia in BJAB cells and HeLa cells. We have identified a fusion domain composed of the first 64 amino acids at the amino terminus of the HTLV-2 transmembrane protein, p21, the retention of which was required for syncytium induction. Construction of replication-competent HTLV genomic clones allowed us to correlate the ability of HTLV-2 to induce syncytia with the ability to replicate in BJAB cells. Differences in the ability to induce syncytia were not due to differences in the levels of total or cell membrane-associated envelope or in the formation of multimers. Therefore, we have localized a fusion domain within the amino terminus of the transmembrane protein of HTLV-2 envelope that is necessary for syncytium induction and viral replication.

Human T-cell leukemia virus types 1 and 2 (HTLV-1 and HTLV-2) are type C retroviruses that have been associated with a variety of human malignancies. HTLV-1 is the etiological agent of adult T-cell leukemia as well as a degenerative neurological disorder, HTLV-1-associated myelopathy/tropical spastic paraparesis (28, 40, 58, 60, 83). Recent reports have also implicated HTLV-1 infection with arthropathy (42, 65), polymyosis (23, 37), and uveitis (48, 49, 51). HTLV-2 has been associated with a rare form of atypical hairy cell leukemia (62, 63, 68) as well as some cases of neuropathy (33, 39). It is estimated that between 10 million and 20 million individuals worldwide are infected with HTLV, with an overall risk of 5% of disease progression in infected individuals (14). HTLV is endemic in southern Japan, the Caribbean Basin, and Central and South America. In the United States, recent reports have identified a high proportion of HTLV, especially HTLV-2, infection in intravenous-drug abusers (44, 61, 64).

Cell-to-cell contact is considered critical for the in vivo and in vitro transmission of HTLV-1 and HTLV-2, as infection by cell-free HTLV virus is inefficient in vitro and in vivo. By analogy with other enveloped viruses, HTLV infection of susceptible cells is likely mediated by the envelope glycoprotein. Antibodies against HTLV envelope are protective against infection in vivo (71, 80), and multiple epitopes that elicit neutralizing antibodies have been identified throughout the protein (31, 34, 56). Initially synthesized as a precursor protein, gp61, HTLV envelope is subsequently modified by glycosylation and cleaved into two subunits, gp46 and p21. The external surface glycoprotein, gp46, is anchored to the cell surface by noncovalent association with the transmembrane envelope glycoprotein, p21. Interaction of envelope with the as yet unidentified cellular receptor leads to cell-to-cell fusion and can result in syncytium formation.

We were interested in identifying the molecular determinants of HTLV involved in syncytium formation and viral entry. Our laboratory has several cell lines that are permissive to HTLV-2- but not HTLV-1-mediated cell fusion. Therefore, we constructed recombinants between the HTLV-1 and -2 envelope genes and assayed for the loss of syncytium induction in BJAB cells and HeLa cells. Loss of a 64-amino-acid (aa) domain located at the amino terminus of the HTLV-2 transmembrane protein, p21, correlated with a loss in the ability of the envelope chimera to induce cell fusion. When the chimeric envelopes were expressed in the context of replication-competent genomic clones, there was a good correlation between syncytium induction and the ability to replicate in permissive cells. Present within the identified fusion domain is a hydrophobic region and a heptad repeat resembling a leucine zipper. We examined the contribution of the fusion domain to the structural integrity of the HTLV-2 envelope by using a vaccinia virus expression system. None of the recombinants affected the synthesis, transport, or oligomer formation of the HTLV glycoprotein complex.

MATERIALS AND METHODS

Cells.

BJAB cells were grown in RPMI 1640 medium with 10% fetal bovine serum (FBS) (Gemini, Calabasas, Calif.). HeLa cells were grown in Dulbecco modified Eagle medium containing 10% calf serum. 729ph6neo and SLB1, which express infectious HTLV-2 and HTLV-1, respectively, were grown in Iscove’s medium with 20% FBS.

Oligonucleotide-directed mutagenesis and DNA manipulation.

The HTLV-2 envelope sequences from nucleotide (nt) 5123 to 7392 (72) from BCHTLV (30) and the HTLV-1 envelope sequences from nt 5126 to 7480 (69) from Env1A (kindly provided by D. Slamon, University of California, Los Angeles) were subcloned between the SphI and MluI restriction sites of CDM7 (kindly provided by D. Camerini, University of Virginia), replacing a 2.6-kb fragment and creating CDM7-II and CDM7-I, respectively. Following preparation of single-stranded DNA using M13 K07 phage, mutagenesis was performed according to the protocol for the T7-GEN in vitro mutagenesis kit (United States Biochemical). The nucleotide positions of the HTLV-1 sequences in clones NH, NK, KH, KM, MH, and NKM are indicated in parentheses. Clones NH (nt 5179 to 6680), NK (nt 5179 to 6118), KH (nt 6118 to 6680), KM (nt 6118 to 6290), MH (nt 6290 to 6680), and NKM (nt 5179 to 6118 and 6290 to 6680) were constructed by introducing compatible restriction sites within HTLV-1 and -2 envelope. A novel HindIII site was introduced at the ends of both genes. KpnI and MstI sites, already present within the HTLV-1 envelope, were introduced into the same locations within the HTLV-2 envelope. All clones were confirmed by DNA sequencing. Oligonucleotides containing base pair substitutions that created the unique restriction sites (underlined) were as follows: for HTLV-2, 5′ CTG CTA TTG GTA CCG CAC GGC GGC G 3′ (KpnI site 6105), 5′ GCT GCA AAG CTT GCA GGT CTA 3′ (HindIII site 6650), and 5′ CGT CTA TTT TGC GCA GCA TAC TGT GC 3′ (MstI site 6292); for HTLV-1, 5′ GGC TGG GGA AGC TTG AGG CGA TGA G 3′ (HindIII site 6682).

Constructs BM and KB were made by using overlapping oligonucleotides that consisted of nt 6118 to 6200 (sense) and 6195 to 6310 (antisense) of HTLV-2 and HTLV-1 envelope and nt 6116 to 6195 (sense) and 6184 to 6300 (antisense) of HTLV-1 and HTLV-2 envelope, respectively. The oligonucleotides were annealed by heating in extension buffer (40 mM Tris [pH 7.4], 20 mM MgCl₂, 50 mM NaCl) at 65°C for 5 min and allowing the reaction to cool to 25°C. Extension occurred at 25°C for 10 min by the addition of 1 μl Sequenase (United States Biochemical). The resulting 200-bp fragment, containing a unique BamHI restriction site at nt 6200, was subsequently cloned into CDM7-II between the newly created KpnI and MstI restriction sites.

The genomic clones were obtained by inserting the chimeric envelopes from CDM7 between the SphI and MluI restriction sites of H6H11. H6H11 contains the HTLV-2 genomic sequences from λH6 (8) inserted between the HindIII site of pBR322.

The vaccinia virus-driven expression vectors containing the HTLV envelopes were obtained by PCR amplification of the envelope sequences from CDM7 into the pTM-3 expression vector (kindly provided by B. Moss, National Institute of Allergy and Infectious Diseases). The HTLV-2 envelope was amplified by using forward primer (5′ GCG GAA TTC TTT TCT TCC TAC TTT TAT TC 3′) and reverse primer (5′ GAA TCG AGT TAG GGC TGG 3′). The HTLV-1 envelope was amplified by using forward primer (5′ GGC GAA TTC TTC TCG CCA CTT TGA TTT 3′) and reverse primer (5′ CGC AGA TCT TAT CGG CGG GAG CGG GAT CC 3′). The resulting 1.7-kb PCR fragments were digested with EcoRI and PstI and inserted downstream of the bacteriophage T7 promoter of pTM-3.

p24 assay.

Supernatants from transfected BJAB cells were collected and clarified by centrifugation at 3,000 rpm for 5 min. Analysis for the presence of HTLV p24 antigen was performed by enzyme-linked immunosorbent assay (ELISA), using Coulter kit 6604252 according to the manufacturer’s procedure.

Protein analysis.

Cells were lysed on ice for 10 min in OGL lysis buffer (100 mM Tris [pH 8.0], 100 mM NaCl, 1 mM CaCl₂, 250 mM octyl-glucosidase). For detection of denatured proteins, samples were heated at 100°C for 5 min in 1× loading buffer (100 mM Tris [pH 6.8], 20% glycerol, 0.02% bromophenol blue) with the addition of 4% sodium dodecyl sulfate (SDS) and 5% β-mercaptoethanol, and separated on SDS–10% polyacrylamide gels. For detection of native proteins, samples were diluted in 1× loading buffer and separated on 4 to 20% gradient acrylamide gels in the presence of 0.01% SDS in the running buffer (79). Western analysis was performed with anti-HTLV-1 gp46 antibody (clone 65/6c2.2.34; Cellular Products Inc., Buffalo, N.Y.) and developed by using the Amersham enhanced chemiluminescence assay.

Transfections and blue syncytium assay.

Transient transfection assays were performed by resuspending 5 × 10⁶ BJAB cells in electroporation medium (RPMI 1640 with 20% FBS) and electroporating with 15 μg of DNA at 230 V and 960 μF.

The blue syncytium assay was performed as previously described (45), with the following modifications. HeLa cells were infected with wild-type vaccinia virus (WR) or VTF7-3 (multiplicity of infection [MOI] = 1.0) for 2 h at 37°C. Cells were subsequently transfected with 15 μl of Lipofectin (Gibco BRL) and 10 μg of DNA/10⁵ cells in serum-free Dulbecco modified Eagle medium as specified by the manufacturer. Cells infected with VTF7-3 were transfected with vaccinia virus-driven HTLV envelope constructs. Cells infected with WR were transfected with pEM-ClacZβgAn (45). At 16 h posttransfection, cells expressing the HTLV envelope clones were harvested by treatment with phosphate-buffered saline-EDTA and washed twice with medium. Then 10⁵ cells from each of the two sources, WR infected and VTF7-3 infected, were mixed in the presence of actinomycin D (1 μg/ml) to prevent vaccinia virus spread. After a 7-h incubation at 37°C, cultures were fixed and stained in situ for 1 h at 37°C with solution containing 4 mM potassium ferrocyanide, 2 mM MgCl₂, and 0.5 mg of 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-Gal) per ml. Giant blue syncytia were counted microscopically.

Flow cytometry.

Cells were stained with 1 μg of anti-HTLV-1 gp46 antibody clone (65/6c2.2.34; Cellular Products) diluted in 100 μl of FACS (fluorescence-activated cell sorting) buffer (phosphate-buffered saline with 2% FBS and 0.01% sodium azide) at 4°C for 20 min. After one wash with FACS buffer, cells were incubated at 4°C, with fluorescein isothiocyanate-conjugated sheep anti-mouse antibody (diluted 1/200; Cappel) for an additional 20 min. Cells were resuspended in FACS buffer, and data were acquired on a FACScan (Becton Dickinson) and analyzed with the Lysis 2 software program.

RESULTS

The amino terminus of the HTLV-2 transmembrane protein contains sequences required for syncytium formation in BJAB cells.

HTLV-1 and -2 can induce syncytium formation upon infection of a variety of cell types, including T- and non-T-cell lines (35, 47, 80). We had previously observed that an Epstein-Barr virus-negative human B-cell line (32), BJAB, formed numerous multinucleated cells upon cocultivation with 729ph6neo, an HTLV-2-producing cell line. In contrast, cocultivation with HTLV-1-producing cells, such as SLB1 cells, did not result in visible syncytia. We took advantage of the differential ability of HTLV-1 and -2 to form syncytia in BJAB cells to define the regions of the HTLV-2 genome involved in syncytium formation.

Previous experiments have mapped the fusion phenotype to the HTLV envelope protein, as expression of envelope alone was sufficient to induce syncytia (18, 45, 57). HTLV-1 and -2 show 75% amino acid homology between the envelope proteins, allowing construction of recombinants which should maintain the functionality of the protein. The chimeric envelope genes were subsequently substituted into H6H11, an infectious HTLV-2 genomic clone (Fig. 1). These full-length clones containing HTLV-1 envelope sequences allowed us to study not only the syncytium-inducing phenotype but also the contributions of the envelope domains in viral replication. The genomic clones were transfected into BJAB cells and, at 3 days posttransfection, analyzed microscopically for syncytium formation. At this time, the ability of the clones to cause fusion is due to the transient expression of viral proteins from the transfected DNA. Syncytia were defined as giant cells greater in diameter than three single cells.

FIG. 1.

The amino terminus of HTLV-2 p21 is necessary but not sufficient for syncytium formation in BJAB cells. BJAB cells (5 × 10⁵) were transfected with genomic constructs containing chimeric envelopes (shown) as described in Materials and Methods. At day 3 posttransfection, cells were analyzed for syncytia by microscopic analysis. + and − indicate that greater than or less than 10% of the cell population were undergoing cell fusion, respectively. HTLV-2 sequences are depicted by open boxes; HTLV-1 sequences are depicted by dark boxes.

Transfection with the genomic HTLV-2 clone, H6H11, produced syncytia in 10 to 30% of the cell population. In contrast, a genomic clone in which the entire HTLV-2 envelope had been substituted by the HTLV-1 envelope (NH) did not cause any visible cell fusion. These results are in agreement with our previous observation that HTLV-2- but not HTLV-1-producing cells can induce syncytia in BJAB cells. These results also confirmed the role of the envelope gene in syncytium formation.

The HTLV envelope is composed of the external surface glycoprotein, gp46, and the transmembrane anchoring protein, p21. We studied the contributions of these domains in syncytium induction by substituting these individual domains from HTLV-1 into the HTLV-2 envelope. A clone (NK) that had the gp46 domain substituted by HTLV-1 envelope sequences but retained the p21 transmembrane domain of HTLV-2 was capable of inducing syncytia in BJAB cells. The reciprocal clone (KH), containing a replacement of the p21 domain with HTLV-1 sequences, lost the fusogenic phenotype. These results appear to map the syncytium induction phenotype in BJAB cells to the HTLV-2 p21 transmembrane protein.

Additional chimeras were constructed to further define the regions within p21 necessary for cell fusion. Replacement of the carboxyl portion of HTLV-2 p21 with HTLV-1 sequences (MH) did not affect syncytium formation. However, substitution of the amino terminus of HTLV-2 p21 with HTLV-1 sequences (KM) resulted in loss of cell fusion. Therefore, this 64-aa region within the HTLV-2 p21 transmembrane protein appeared to be necessary for syncytium induction.

We next substituted either the amino or carboxyl portion of these 64 aa from HTLV-2 p21 with HTLV-1 envelope sequences. These smaller substitutions within the amino terminus of HTLV-2 p21 revealed that a region from HTLV-2 envelope comprised of aa 330 to 365 was required for fusion. Substitution of these sequences with the corresponding HTLV-1 sequences (BM) correlated with a loss of syncytium induction. This conclusion is further supported by the retention of these sequences from HTLV-2 p21 in all clones that were able to cause syncytia in BJAB cells (NK, MH, and KB). Therefore, we have defined a region essential for HTLV-2 fusion in BJAB cells to a 34-aa domain located at the amino terminus of the HTLV-2 envelope transmembrane protein.

Loss of this 34-aa domain from HTLV-2 envelope negatively affected syncytium induction, as seen by the inability to fuse BJAB cells. We next tested whether the sequences from the amino terminus of HTLV-2 p21 were sufficient for syncytium induction in the context of a recombinant where all other envelope sequences were derived from HTLV-1. A clone containing the first 64 aa of HTLV-2 p21 in the context of HTLV-1 envelope (NKM) did not induce syncytia in BJAB cells. A substitution of the smaller 34-aa domain from HTLV-2 also did not result in syncytium induction by the HTLV-1 envelope (data not shown). Therefore, the fusion domain identified in HTLV-2 was not sufficient on its own to confer the syncytium induction phenotype upon the HTLV-1 envelope.

The amino terminus of HTLV-2 p21 is required for HTLV-2 replication in BJAB cells.

Cell fusion is believed to be the major route of HTLV viral spread, as cell-free transmission is highly inefficient both in vivo and in vitro (6, 46, 52, 59). Therefore, we examined the ability of the recombinants to replicate in BJAB cells and correlated it with their ability to cause syncytia. The replicative potential of the chimeras in BJAB cells was determined by ELISA for the production and secretion of viral core antigen, p24, over several weeks (Fig. 2). All of the clones were p24 positive on day 3, indicating that all of the chimeras were capable of viral protein production from the transfected DNA (data not shown). As expected, the genomic HTLV-2 clone, H6H11, replicated to high levels over this time period, as shown by an increase in the production of p24 in the supernatant. The increased p24 production correlated with increased cell fusion with syncytia in over 80% of the cells during this time period. All clones that were capable of causing cell fusion (NK, MH, and KB) were also able to replicate in BJAB cells, producing p24 at levels similar to those of H6H11 (Fig. 2A). Similarly, clones that did not induce syncytia (KH, KM, BM, and NKM) did not replicate, producing an initial burst of p24 due to the transfected DNA that was not sustained over time (Fig. 2B). These results support the premise that HTLV-2 spread is dependent on the cell fusion activity, as only the clones capable of syncytium induction were also able to replicate in BJAB cells. As the replication-competent clones (NK, MH, and KB) also retained the amino terminus of the HTLV-2 p21 domain, we conclude that the same 34-aa sequences within the transmembrane protein p21 confer both the syncytium induction phenotype and the replicative capability in BJAB cells.

FIG. 2.

HTLV-2 replication in BJAB cells requires sequences at the amino terminus of p21. BJAB cells (5 × 10⁵) were transfected with genomic constructs containing chimeric envelopes as described in Materials and Methods. Every 3 days, the medium was changed by allowing the cells to settle and replacing half of the medium. At 3, 7, 14, and 21 days posttransfection, supernatants were collected and cell debris was removed by centrifugation at 3,000 rpm for 5 min as described in Materials and Methods. The amount of secreted HTLV p24 was quantitated by ELISA as described in Materials and Methods. (A) Clones that induced syncytia in BJAB cells (Fig. 1) which included full-length HTLV-2 envelope (II), as well as clones NK, MH, and KB. (B) Clones that did not induce syncytia in BJAB cells (KH, KM, BM, and NKM). (C) Results from the full-length HTLV-2 envelope and the full-length HTLV-1 envelope (NH). The data are representative of three independent experiments. OD₅₅₀, optical density at 550 nm.

Interestingly, the clone containing the entire HTLV-1 envelope sequences (NH) was able to replicate to the same degree as the wild-type HTLV-2 (Fig. 2C), albeit at a slightly lower rate and without detectable cell fusion. Cocultivation of BJAB cells with lethally irradiated SLB1 cells, the original source of the HTLV-1 envelope clone, did not result in productively infected BJAB cells, as evidenced by the eventual decline in p24 production over time (data not shown). Therefore, there may be interactions occurring between the HTLV-1 envelope and other HTLV-2 viral components that allow replication in BJAB cells.

Fusion of HeLa cells by HTLV-2 envelope alone requires the amino terminus of p21.

The ability of the construct containing the HTLV-1 envelope to replicate in BJAB cells suggested that interactions of envelope with other viral proteins may be occurring. Therefore, we investigated the contributions to syncytium formation of the HTLV envelope protein alone in the absence of other viral components. We next constructed vaccinia virus/T7 promoter-driven vectors that expressed only the HTLV chimeric envelopes. HTLV envelope expressed by the vaccinia virus/T7 polymerase expression system has been reported to be produced in a properly processed and folded form (3).

We also used a modification of a sensitive blue syncytium assay to quantitate the amount of cell fusion caused by the envelope chimeras (45). Briefly, cells were transfected with a construct containing the β-galactosidase gene under the control of the T7 RNA polymerase promoter. A second set of cells were transfected with the HTLV envelope construct, as well as infected with vaccinia virus expressing the T7 RNA polymerase (VTF7-3). Upon mixing of the two populations, cell fusion would allow the T7 RNA polymerase expressed from VTF7-3 to activate the β-galactosidase gene, resulting in a blue syncytia.

In these experiments, HeLa cells were used, as they produced the highest level of expression of HTLV envelope (data not shown). In addition, syncytia were easier to score in HeLa cells due to the adherent nature of the cells. HeLa cells, similar to BJAB cells, also showed a differential ability to be fused by HTLV-1 and -2 envelope (Fig. 3). HTLV-2 envelope gave rise to 780 blue syncytia, compared to the 9 syncytia produced by the HTLV-1 envelope (NH). Clones that were capable of inducing syncytia in BJAB cells (NK, MH, and KB) also resulted in significant amounts of blue syncytia. The majority of the clones that were unable to cause fusion in BJAB cells (KH, KM, and NKM) were also unable to cause fusion in HeLa cells. In general, the blue syncytium assay in HeLa cells, using vaccinia virus-expressed envelope, appeared to confirm the fusion results of the genomic clones in BJAB cells, with one exception.

FIG. 3.

The amino terminus of HTLV-2 p21 is required for syncytium formation in HeLa cells. HeLa cells (10⁶) were infected with wild-type vaccinia virus (WR) or VTF7-3 (MOI = 1.0) for 2 h at 37°C. HeLa cells that were infected with WR were transfected with pEM-ClacZβgAn, a plasmid in which the Escherichia coli lacZ gene has been linked to the T7 promoter. Cells that were infected with VTF7-3 were transfected as described in Materials and Methods with vaccinia virus constructs expressing chimeric HTLV envelopes. At 16 h posttransfection, 10⁵ cells from each of the two sources were mixed and scored for blue syncytia as described in Materials and Methods. These data represent the average of two duplicate wells and are representative of two independent experiments. □, HTLV-2; ▪, HTLV-1.

Although loss of the first 64 aa at the amino terminus of HTLV-2 p21 (KM) resulted in loss of syncytia in both HeLa cells and BJAB cells, cell fusion could be induced in HeLa cells by clones that retained either the amino or carboxyl portion of this region (BM or KB, respectively). In contrast, in BJAB cells, clone KB was fusogenic whereas clone BM did not form visible syncytia. Therefore, syncytium induction in HeLa cells required two discrete adjacent domains within p21 that included the 34-aa fusion domain identified in BJAB cells. It is possible that clone BM caused a low level of fusion in BJAB cells that was undetectable microscopically and caused a detectable amount of fusion in HeLa cells due to the greater sensitivity of the assay. Alternatively, there may be differences in the interaction of regions of envelope with cellular components from these two cell lines. There is also the possibility that BJAB cells and HeLa cells differ in their cell surface requirements for fusion, akin to the different coreceptors in the case of human immunodeficiency virus type 1 (HIV-1) cell fusion.

The syncytium-inducing phenotype of the chimeric envelopes is not due to differences in the levels of intracellular envelope.

The ability of viral envelope proteins to induce syncytia is influenced by several factors, including intracellular expression levels of the envelope, density of the fusion protein on the cell surface, and correct protein configuration. We were unable to examine the envelope expressed by transfection of nonproductive genomic clones, as the levels of envelope expression were below detection by both radioimmunoprecipitation and flow cytometry (data not shown). We therefore used the high levels of HTLV envelope produced by the vaccinia virus system to analyze expression of the recombinants.

We first examined the intracellular levels of envelope in the vaccinia virus-infected cells, using a monoclonal antibody that recognizes gp46 from HTLV-1 and HTLV-2 (Fig. 4). By Western analysis on SDS-PAGE, approximately equal amounts of precursor gp61 envelope were detected from all of the chimeric envelope clones. The majority of the envelope protein existed as uncleaved precursor, similar to findings for infected cells (data not shown). These results indicate that the difference in the ability of the clones to cause fusion in HeLa cells was not attributable to differential expression levels of the chimeric envelopes.

FIG. 4.

The chimeric HTLV envelopes produce equivalent amounts of intracellular gp61. HeLa cells (10⁵) were infected with VTF7-3 (MOI = 1) for 2 h at 37°C and transfected with vaccinia virus constructs expressing chimeric HTLV envelopes as described in Materials and Methods. Cells were harvested 16 h posttransfection. Total cell lysates (10 μg) were separated on SDS–10% polyacrylamide gels and analyzed for the presence of HTLV envelope by Western blot analysis as described in Materials and Methods. Sizes (in kilodaltons) of the molecular weight standards are indicated on the left; gp61 is indicated by the arrow on the right.

The chimeric envelopes are expressed at similar levels on the cell surface of infected cells.

We next compared the levels of expression of the recombinant envelopes on the cell surface by flow cytometric analysis (Fig. 5). Compared to uninfected cells, both wild-type HTLV-1 and -2 envelopes were efficiently expressed by the vaccinia virus expression system. Similar levels of HTLV envelope were detected on the surface of cells infected with vaccinia virus expressing the various envelope chimeras regardless of their ability to induce syncytia. The anti-gp46 antibody was specific for the HTLV envelope, as we detected a significant level of HTLV envelope on the cell surface of the transfected cells over an isotype control antibody (data not shown). The expression by all the recombinant envelope constructs of comparable levels of gp46 on the cell surface suggested that the loss of syncytium induction was not due to a defect in the transport of gp46 to the membrane.

FIG. 5.

The chimeric HTLV envelopes are expressed efficiently on the cell surface of infected cells. HeLa cells (10⁵) were infected with VTF7-3 for 2 h at 37°C and transfected with vaccinia virus constructs expressing chimeric HTLV envelopes as described in Materials and Methods. Cells were harvested 16 h posttransfection, and 5 × 10³ cells were analyzed by flow cytometry for expression of envelope on the cell surface as described in Materials and Methods. The histograms indicate relative cell number (y axis) as a function of relative amount of gp46 on the cell surface (x axis). Mock-infected cells are represented by the white histogram area, and cells transfected with chimeric HTLV envelopes are represented by the dark histogram area.

The HTLV envelope chimeras are capable of forming multimers.

Oligomerization of envelope proteins is necessary for their proper function and viral infectivity. HTLV-1 envelope has been previously reported to be capable of forming multimers (55). The 64-aa domain in HTLV-2 envelope that we have identified as important in cell fusion and viral replication in BJAB cells contains a region that bears similarities to a leucine zipper motif (Fig. 6A) (7). The different syncytium-inducing phenotypes may therefore be due to effects on the multimerization of the envelope protein. We examined the oligomerization potential of the chimeric envelopes by Western analysis of native protein complexes separated on nondenaturing protein gels (Fig. 6B). The wild-type HTLV-1 migrated at approximately 125 kDa, confirming that the HTLV-1 envelope exists as a multimer. Similarly, the HTLV-2 envelope also migrated at 125 kDa, indicating that the HTLV-2 envelope may also function as a multimer. Analysis of the chimeric envelopes revealed that all of the recombinant envelopes also migrated as multimers. As an additional control, we first denatured the HTLV protein complexes at 100°C and then separated them on the same nondenaturing protein gels as specified above. These proteins migrated at 61 kDa, the expected size for the monomeric form of the HTLV envelope (Fig. 6C). There was no correlation between the ability of the chimeric envelopes to induce syncytia and their ability to form oligomers. Therefore, the inability of some of the chimeric envelopes to induce syncytia was not due to a failure to multimerize.

FIG. 6.

The chimeric HTLV envelopes are similar in the ability to form multimers. HeLa cells (10⁶) were infected with VTF7-3 for 2 h at 37°C and transfected with vaccinia virus constructs expressing chimeric HTLV envelopes as described in Materials and Methods. Cells were harvested 16 h posttransfection and lysed in 500 μl of OGL buffer. (A) Sequence comparison of the amino terminus of HTLV-2 and HTLV-1 p21 proteins. The first 64 aa of HTLV-2 and HTLV-1 p21 are represented. The nonconservative amino acid changes between HTLV-2 and HTLV-1 are boxed. (B) Western analysis of native envelope proteins. Total-cell lysates (20 μl) were separated on a 5 to 20% gradient nondenaturing acrylamide gel with 0.01% SDS in the running buffer as described in Materials and Methods. HTLV envelope was detected by Western blot analysis as described in Materials and Methods. Sizes (in kilodaltons) of the molecular weight standards are indicated on the left; the multimeric forms of HTLV envelope are indicated by the arrow on the right. (C) Western analysis of denatured envelope proteins. Total-cell lysates (20 μl) were heated to 100°C for 5 min and separated on a 5 to 20% gradient nondenaturing acrylamide gel with 0.01% SDS in the running buffer as described in Materials and Methods. HTLV envelope was detected by Western blot analysis as described in Materials and Methods. Sizes of the molecular weight standards are indicated on the left; gp61 is indicated by the arrow on the right.

DISCUSSION

The amino terminus of HTLV-2 p21 encodes a fusion domain required for HTLV-2 syncytium induction and replication.

We have analyzed the requirements for syncytium induction by the HTLV-2 envelope protein. Our results have identified a 64-aa domain located within the amino terminus of the transmembrane protein as necessary for HTLV-2-mediated cell fusion of BJAB cells and HeLa cells. This region was not sufficient by itself to confer the syncytium-inducing phenotype, as the presence of this domain within the HTLV-1 envelope did not permit fusion. The ability of HTLV-2 genomic clones to replicate in permissive cells correlated with the ability to induce syncytia. We were unable to attribute the loss of the syncytium-inducing phenotype to total cell levels, levels of membrane associated, or oligomerization of the envelope protein.

In this study, we used two cell lines, BJAB and HeLa, that were permissive for HTLV-2- but not HTLV-1-mediated cell fusion. However, other groups have previously reported on the ability of HTLV-1 to induce syncytia in HeLa cells (16, 18). Agadjanyan et al. have also reported that several subclones of a BJAB cell line were fusogenic upon cocultivation with both HTLV-2 and HTLV-1 (1). Our inability to observe fusion with HTLV-1 envelope in BJAB cells or HeLa cells may be a property of the cell lines upon propagation. Distinct properties may arise in cell lines upon long-term culture, as illustrated by the isolation of subclones of BJAB cells that have lost the ability to form syncytia with HTLV-2 (1). Alternatively, the HTLV-1 envelope used in our studies may have cell tropisms distinct from those of other HTLV-1 envelope clones. Sequence analysis between the Env1A envelope from SLB1 cells and the envelope derived from a replication-competent HTLV-1 clone (41) revealed no amino acid differences within the putative fusion domain (data not shown). However, there are a total of seven amino acid substitutions elsewhere in the envelope protein which may account for the differences observed.

Previous work in our lab has suggested that the transmembrane protein of HTLV-2 is required for cell fusion, as expression of gp46 alone was not sufficient to induce syncytia (45). In addition, large substitutions within the HTLV-1 transmembrane protein by murine leukemia virus envelope sequences abolished HTLV-1-mediated cell fusion (17, 18). These studies localized the fusion domain within the extracellular portion of the transmembrane protein. We have further defined the fusion domain in HTLV-2 to the amino terminus of p21. It is likely that the corresponding region in HTLV-1 envelope serves a similar role since linker insertion mutations within the hydrophobic stretch in the N-terminal part of HTLV-1 p21 resulted in loss of syncytium induction (57).

Potential mechanisms of viral entry mediated by the HTLV-2 fusion domain.

Fusion domains have been localized to the transmembrane protein, and specifically to the amino terminus, of other retroviral envelopes, including HIV-1 and simian immunodeficiency virus (25, 81). These domains are highly hydrophobic with a predicted α-helical structure. It has been hypothesized that these domains form sided helixes with most of the bulky hydrophobic residues on one side of the helix (15). Located adjacent to these hydrophobic sequences are heptad repeat sequences with nonpolar residues at the first and fourth positions, similar to a leucine zipper motif (7). This region has been postulated to play a role in stabilizing the oligomeric form of these molecules. However, mutations in the HIV-1 leucine zipper, while abolishing syncytium formation with CD4⁺ cells and impairing infectivity, did not affect the ability of the envelope protein to form oligomers (10, 22). Mutagenesis of the HIV-1 envelope point to a critical role of the leucine zipper motif in HIV-1 membrane fusion and virus entry, likely at a post-CD4 binding step (82).

The HTLV-2 fusion domain that we have identified contains a hydrophobic domain adjacent to a region with homology to a leucine zipper motif. In BJAB cells, the syncytium-inducing phenotype appeared to map to the leucine zipper region. However, substitution of the HTLV-2 leucine zipper domain did not ablate fusion in HeLa cells, indicating that in some cell types this domain alone may not be necessary for fusion. The mechanism by which the HTLV-2 fusion domain contributes to syncytium induction remains to be clarified. There was no apparent correlation between the ability of the HTLV envelope chimeras to induce syncytia and the intracellular or cell surface expression, nor was there an apparent effect of the recombinants on the formation of multimers, similar to the HIV-1 results. One possible hypothesis is that the fusion domain in HTLV-2 p21 is required for receptor-mediated conformational changes. It is known that conformational changes occur upon HIV-1 envelope binding to CD4. Post-CD4 binding events included enhanced antibody binding and cleavage by an exogenous proteinase of the V3 loop, a major neutralizing determinant in gp120 (13, 66), as well as exposure of previously cryptic epitopes (67, 77). One indicator of overall changes in the HIV-1 envelope protein is the observed enhanced shedding of gp120 upon CD4 binding (4, 76, 78). Introduction of prolines into highly conserved leucine or isoleucine residues of the HIV-1 leucine zipper affected secretion of gp120 (9), suggesting the involvement of this region in modifying the envelope tertiary structure. Other viral fusion proteins similarly undergo conformational changes in order to acquire their fusion potential (29, 81). Therefore, the loss of syncytium induction and infectivity by our chimeric envelope constructs may be due to an effect on the ability of the envelope to undergo the appropriate conformational changes required for HTLV viral entry.

A second hypothesis for the role of the fusion domain in HTLV-2 in syncytium induction may be at the level of receptor utilization. Precedence for this possibility exists for HIV-1, where syncytium formation can be mediated by sequences which include the V3 loop in gp120 (13, 26, 54). The V3 loop has also been implicated in the cell tropism observed for various strains of HIV-1. Changes in the V3 loop can influence the efficiency of entry of different HIV-1 strains into different cell types, such as T-cell lines, macrophages, and microglial cells (11, 36, 38, 50, 70, 73, 74). Although the primary receptor for HIV-1 binding to cells is CD4, recent reports have identified secondary receptors, CCR-5 and CXCR-4, that modulate viral entry. CCR-5 and CXCR-4 are both members of the chemokine receptor family and mediate infection at an early stage by macrophage-tropic and T-cell-tropic viruses, respectively (2, 5, 19–21, 24, 43). The ability of HIV-1 to utilize CCR-5 or CXCR-4 has been mapped to the V3 loop (12, 53), previously implicated in syncytium formation and viral entry. Using a highly sensitive syncytium assay, we have described cell lines that display differential abilities to be fused by HTLV-1 and -2. Yet, based on receptor interference studies, HTLV-1 and -2 have been postulated to share a common receptor on the tk-1 region of human chromosome 17 (27, 75). Therefore, it is conceivable that, analogous to HIV-1, the fusion capability of HTLV-1 and -2 in BJAB and HeLa cells is mediated by interaction of the fusion domain with accessory molecules that determine cellular tropism.