Amino Acid Changes at Positions 173 and 187 in the Human T-Cell Leukemia Virus Type 1 Surface Glycoprotein Induce Specific Neutralizing Antibodies

Abstract

The nucleotide sequence of human T-cell leukemia virus type 1 (HTLV-1) is highly conserved, most strains sharing at least 95% sequence identity. This sequence conservation is also found in the viral env gene, which codes for the two envelope glycoproteins that play a major role in the induction of a protective immune response against the virus. However, recent reports have indicated that some variations in env sequences may induce incomplete cross-reactivity between HTLV-1 strains. To identify the amino acid changes that might be involved in the antigenicity of neutralizable epitopes, we constructed expression vectors coding for the envelope glycoproteins of two HTLV-1 isolates (2060 and 2072) which induced human antibodies with different neutralization patterns. The amino acid sequences of the envelope glycoproteins differed at four positions. Vectors coding for chimeric or point-mutated envelope proteins were derived from 2060 and 2072 HTLV-1 env genes. Syncytium formation induced by the wild-type or mutated envelope proteins was inhibited by human sera with different neutralizing specificities. We thus identified two amino acid changes, I173→V and A187→T, that play an important role in the antigenicity of neutralizable epitopes located in this region of the surface envelope glycoprotein.

Human T-cell leukemia virus type 1 (HTLV-1) is the etiologic agent of adult T-cell leukemia and a chronic neurological disease, tropical spastic paraparesis or HTLV-1-associated myelopathy (14, 19, 29, 31, 39). The virus infects 10 to 20 million persons worldwide, 4% of whom will develop one of these diseases. In common with that of other retroviruses, the entry of HTLV-1 into the target cell is mediated by the viral envelope glycoproteins. These are two noncovalently linked subunits, a 46-kDa surface glycoprotein (SU) which is responsible for attachment of the virus to a cell surface receptor and a 21-kDa transmembrane glycoprotein (TM) which fuses the viral envelope to the target cell membrane, allowing penetration of the viral core into the cytoplasm. Several regions involved in viral entry have been identified on the HTLV-1 envelope glycoproteins by the use of neutralizing antibodies or peptides that inhibit fusion (1, 2, 10, 17, 30, 38) and by functional analysis (7, 8, 32).

HTLV-1 is distributed worldwide but exhibits relatively little sequence variation. HTLV-1 strains from Japan, Africa, the West Indies, and the Americas and belonging to the cosmopolitan clade have at least 95% sequence similarity. More distantly related strains displaying 8% nucleotide sequence variation have been found in remote populations from the Solomon Islands, Papua New Guinea, and Australia (15). HTLV-1 infection has been successfully transmitted to rats, rabbits, and monkeys in the laboratory (25, 27, 37). This infection can be prevented by passive immunization with immunoglobulins purified from HTLV-1-infected patients (21, 26, 33) or by vaccination with various versions of HTLV-1 envelope proteins (3, 12, 18, 22, 27, 36). These observations suggest that genetically engineered HTLV-1 envelope proteins or synthetic peptide-based subunits could be used in a vaccine against HTLV-1. However, protective humoral and cellular immune responses elicited by vaccine components could be foiled by the existence of different antigenic forms of HTLV-1 proteins. In this respect, incomplete cross-reactivity between some cosmopolitan and Melanesian strains of HTLV-1 has been reported (4). More recently, we showed that sera from some patients infected with cosmopolitan HTLV-1 strains with only a few amino acid changes in their envelope glycoproteins displayed different neutralization patterns (5). These patterns could be classified into three categories that fit well with groups of viruses each harboring the same residues in the major immunodominant and neutralizable domain (amino acids [aa] 175 to 199) of SU. Since within each group, different amino acids could be substituted at other positions, the residues involved in the observed differences have yet to be identified.

To identify the amino acid changes involved in the antigenic specificity of neutralizable epitopes, we constructed expression vectors coding for the envelope proteins of two HTLV-1 isolates (2060 and 2072) which induced human antibodies with different neutralization patterns. The serum of the patient infected with virus 2060 completely neutralized cosmopolitan HTLV-1 of the three groups mentioned above, whereas the serum of the virus 2072-infected patient had a higher neutralization potential against the autologous virus than against cosmopolitan viruses of the other two groups. The amino acid sequences of the envelope glycoproteins of viruses 2060 and 2072 differed at four positions located in surface gp46. Vectors coding for chimeric or point-mutated envelope proteins were derived from 2060 and 2072 HTLV-1 env genes. Their ability to induced syncytium formation after transfection in COS-LTR_HIV-LacZ cells was assessed, as was the inhibition of syncytium formation by sera from HTLV-1-infected patients.

MATERIALS AND METHODS

Sera.

Human sera used for syncytium inhibition were provided by J. C. Vernant (La Meynard Hospital, Fort-de-France, Martinique), J. F. Moreau and J. L. Sarthou (Institut Pasteur de Guyane, Cayenne, Cayenne, French Guiana), S. Sainte-Foie and C. Hajjar (Centre Hospitalier Intercommunal de Basse-Terre/Sainte-Claude, Basse-Terre, Guadeloupe), and M. C. Georges-Courbot (CIRMF, Franceville, Gabon). All sera were heated for 30 min at 56°C before use. The presence of HTLV-1 antibodies in these sera was assessed with a commercially available Western blot diagnostic kit (Diagnostic Biotechnology 2.3).

Cells.

HTLV-1-infected cell lines (infected with isolates 2060 and 2072) (28) were maintained in RPMI 1640 medium (Whittaker) supplemented with 10% heat-inactivated fetal calf serum (Whittaker), 100 U of penicillin per ml, 100 μg of streptomycin per ml, and 200 U of interleukin 2 (Chiron) per ml. COS-LTR_HIV-LacZ cells were grown in Dulbecco modified Eagle medium supplemented with the serum and antibiotics listed above and with 150 μg of hygromycin B per ml. XC-tat cells were maintained in RPMI 1640 medium supplemented with serum, penicillin, and streptomycin as described above plus 400 μg of Geneticin per ml (9).

Construction of env expression vectors.

The sequences of an HTLV-1 provirus cloned from 2060 HTLV-1-infected cells were used for construction of the expression vectors. The proviral clone, p4.39 (28), contains a full-length provirus in a 15.5-kbp EcoRI fragment. Sequence determination of the env gene in this clone revealed a stop codon in the gp21 coding sequence instead of a tryptophan (position 427). The proviral sequences were first excised by use of the BstEII enzyme and then subcloned in plasmid pBR327 to eliminate most of the cellular flanking sequences. The stop codon in the gp21 coding sequence of clone p4.39 was eliminated by exchanging the StuI fragment with a fragment from the MT2 provirus, which codes for the same amino acid sequence as the nonmutated 2060 provirus (5). A large deletion (2,750 bp) in the gag and pol genes was created by HindIII digestion, and all sequences of the 5′ long terminal repeat (LTR) were replaced by the cytomegalovirus promoter as an NsiI-NheI fragment from pBK-CMV (Stratagene). The resulting plasmid, pB1D, was used as an env expression vector after transfection in COS cells.

The env gene of the 2072 provirus was obtained by PCR amplification of the proviral DNA integrated in lymphocyte DNA from the patient infected with isolate 2072. Two primer pairs were used: 5′ ACCATGGCTAAGTTTCTCG 3′ and 5′ GGAGACAAGCTTGACCGC 3′ for amplification of all gp46 coding sequences and 5′ GTCGACGCTCCAGGATATGACC 3′ and 5′ GGAGGATTTGATGGGAGA 3′ for amplification of DNA sequences coding for the carboxyl-terminal half of gp46 and gp21.

Site-directed mutagenesis of the 2060 HTLV-1 env gene was performed by use of the Sculptor in vitro mutagenesis system (Amersham). The presence of mutations was verified by DNA sequencing.

Western blot analysis of envelope glycoproteins.

After transfection, the cellular proteins were solubilized in 20 mM Tris (pH 7.9)–0.15 M NaCl–1 mM Na₂HPO₄–1 mM phenylmethylsulfonyl fluoride–0.5% sodium deoxycholate–0.5% Nonidet P-40. The glycoproteins were affinity purified on lentil lectin-Sepharose (Pharmacia), fractionated on 12.5% polyacrylamide gels as described by Laemmli (22a), and transferred to nitrocellulose membranes (Pall Gelman Sciences). The nitrocellulose sheets were incubated overnight at room temperature with 2% nonfat dry milk and 0.5% bovine serum albumin in phosphate-buffered saline–20 mM Tris (pH 7.4)–0.05% Tween 20. Next, the membranes were incubated in the same buffer containing 20 μg of monoclonal antibody MF2 per ml, specific for a peptide of HTLV-1 gp46 (24), or immunoglobulin G purified from an HTLV-1-infected patient for 2 h at 37°C; the membranes were then washed five times with phosphate-buffered saline containing 0.1% Tween 20. Next, the membranes were incubated with peroxidase-labelled goat anti-mouse or anti-human immunoglobulin G F(ab′)₂ fragments (Immunotech) for 1 h at room temperature. After five washes, the bound antibodies were revealed with a Super Signal chemiluminescence kit (Pierce).

Transfection procedure and envelope fusion assay.

A β-galactosidase assay was used for the quantitative evaluation of syncytium formation. Briefly, 15 ng of plasmid DNA was transfected into COS-LTR_HIV-LacZ cells (40,000 cells per well in 24-well microplates) as described by Cullen (6). At 48 h posttransfection, 80,000 XC-tat indicator cells were added to each well and incubation was continued for 24 h. Each experimental point was determined in triplicate. β-Galactosidase activity was revealed by in situ staining with 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-Gal) or by a colorimetric method with chlorophenol red-β-d-galactopyranoside (CPRG) (Roche Diagnostics) as an enzyme substrate as previously described (1).

Assay of syncytium inhibition by human sera.

At 48 h posttransfection with env expression vectors, COS-LTR_HIV-LacZ cells were incubated for 1 h at 37°C with human sera at various serial log₃ dilutions (in 0.5 ml of RPMI 1640, starting at dilution 1:50). XC-tat cells were then added in 0.5 ml, and fusion assays were performed as described above. The syncytium inhibition titer (ID₅₀) was the reciprocal of the serum dilution resulting in a 50% reduction in the number of blue cells or in the absorbance at 620 nm of the positive control (coculture in the absence of human serum) after subtraction of background values obtained with untransfected COS-LTR_HIV-LacZ cells cocultivated with XC-tat cells.

Statistical analyses.

For serum neutralization pattern analysis, each experiment was repeated three to five times. The significance of differences between the patterns was determined with paired two-tailed Student’s t tests for all experiments. Confirmation was obtained with a two-way analysis of variance. Differences were considered statistically significant when the P value was <0.05. To assess the statistical significance of the fusogenic properties of the envelope constructs, the 95% confidence interval (95% CI) centered on the mean was determined after calculation of the standard deviation.

RESULTS

Chimeric and mutant envelope proteins derived from 2060 and 2072 virus env genes.

In a previous report (5), we showed that the amino acid sequences of the envelope glycoproteins of viruses 2060 and 2072 differed at five positions: four located in surface gp46 and one located in transmembrane gp21. A new determination of the complete env gene sequences of viruses 2060 and 2072 confirmed the differences previously found in gp46 of both viruses, but the amino acid change at position 375 in gp21 was not retrieved. Consequently, both viruses had identical amino acid sequences for gp21.

Figure 1 shows the alignment of amino acid changes observed in the surface envelope glycoproteins of viruses 2060 and 2072. The sequence of the ATK-HTLV-1 envelope protein was included in this alignment for comparison (35). The surface envelope protein sequences of these three viruses differed at eight positions, and only four amino acid changes were observed when the 2060 and 2072 viral envelope glycoproteins were compared (positions 39, 93, 173, and 187).

FIG. 1.

Alignment of amino acid (aa) changes observed in the envelope glycoprotein sequences of different HTLV-I proviruses. Only positions in which changes were observed in at least one of the sequences are listed. Amino acid sequences were deduced from published nucleotide sequences for ATK (35) and 2056, 2060, 2072, and 2085 (5). ATL, adult T-cell leukemia; TSP/HAM, tropical spastic paraparesis or HTLV-1-associated myelopathy; AS, asymptomatic carrier.

To identify amino acid changes that might be involved in the reactivity of neutralizable epitopes harbored by the proteins of the two viruses under study, chimeric and mutant envelope constructs were derived from the virus 2060 env gene inserted downstream from the cytomegalovirus promoter (pB1D vector). Two chimeric vectors were constructed: pB1D/SU72, containing all coding sequences of virus 2072 gp46, and pB1D/V173T187, containing the SalI-BamHI fragment of the virus 2072 env gene, which codes for the carboxy terminal half of gp46 (Fig. 2). Finally, single amino acid changes specific for virus 2072 gp46 were introduced into the virus 2060 envelope glycoprotein sequence. The names of the resulting env expression vectors refer directly to the mutated positions (Fig. 2). Western blot analysis of glycoproteins of COS cells transfected with the seven vectors by use of monoclonal antibody MF2 indicated that all vectors synthesized equivalent amounts of gp61 precursor (Fig. 3, lanes 2 to 8). Under these experimental conditions, monoclonal antibody MF2 recognized only the precursor. A protein corresponding to gp46 was visualized with IgG purified from an HTLV-1-infected patient (Fig. 3, lanes 10 to 13), indicating that maturation of the precursor occurred in COS cells.

FIG. 2.

Schematic representation of the seven HTLV-1 env expression vectors. CMV, cytomegalovirus; LTR, long terminal repeat; pX, X region of HTLV-1.

FIG. 3.

Western blot analysis with monoclonal antibody MF2 (lanes 1 to 8) or human IgG (lanes 9 to 13) of glycoproteins of transfected COS cells. Five hundred micrograms of total proteins of transfected COS cells was affinity purified on lentil lectin-Sepharose. The eluted glycoproteins were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to nitrocellulose membranes, and HTLV-1 glycoproteins were revealed with specific antibodies. Lentil lectin-Sepharose-purified proteins were from untransfected COS cells (lanes 1 and 9), COS cells transfected with pB1D (lanes 2 and 10), COS cells transfected with pB1D/I173V-A187T (lane 3), COS cells transfected with pB1D/SU72 (lane 4), COS cells transfected with pB1D/E39K (lane 5), COS cells transfected with pB1D/N93Y (lanes 6 and 11), COS cells transfected with pB1D/I173V (lanes 7 and 12), and COS cells transfected with pB1D/A187T (lanes 8 and 13).

Fusogenic properties of recombinant HTLV-1 glycoproteins.

The fusogenic properties of the 2060 (pB1D) and chimeric or point-mutated envelope glycoproteins were monitored after transfection into COS-LTR_HIV-LacZ cells and coculturing with XC-tat cells. In situ staining with X-Gal showed that the envelope glycoproteins expressed by all the vectors induced extensive formation of large syncytia. A quantitative evaluation of this phenomenon by the colorimetric method with CPRG confirmed that the seven envelope glycoproteins exhibited equivalent fusogenic activities. The results of five independent experiments are listed in Table 1. The absorbances obtained for the different envelope glycoproteins ranged from 1.021 to 2.213. This range was also observed for the same envelope glycoproteins in independent transfection series. Nevertheless, every mean fell within the 95% CI of every other one. This result was thought to reflect variations in cell culture conditions and/or transfection efficiencies. Consequently, in subsequent experiments, all syncytium inhibition assays performed with transfected COS-LTR_HIV-LacZ cells were repeated three to five times.

TABLE 1.

Fusogenic properties of parental, chimeric, and point-mutated envelopes^a

Transfected vector	β-Galactosidase activity (absorbance)
	In expt:					Mean	SD	95% CI
	1	2	3	4	5
pB1D	1.757	—	2.213	1.394	1.242	1.652	0.432	0.96–2.34
pB1D/SU72	2.018	—	1.135	2.109	—	1.754	0.538	0.42–3.09
pB1D/I173V-A187T	1.111	1.297	1.233	—	1.114	1.189	0.092	1.04–1.34
pB1D/E39K	1.365	—	2.164	1.021	1.292	1.461	0.492	0.68–2.24
pB1D/N93Y	1.783	—	2.190	1.503	—	1.825	0.345	0.97–2.68
pB1D/I173V	1.065	—	1.034	—	—	1.049	0.022	0.85–1.25
pB1D/A187T	1.254	2.417	1.233	1.217	1.954	1.615	0.546	0.94–2.29

^aenv expression vectors were transfected into COS-LTR_HIV-LacZ cells in 24-well plates (3 wells for each DNA). At 48 h posttransfection, XC-tat cells were added, and culturing was continued for 24 h. Fusion was monitored by colorimetric determination of β-galactosidase activity. —, not done. 

Inhibition of syncytium formation induced by recombinant HTLV-1 glycoproteins by 2060 and 2085 sera.

The 2085 serum was from a patient infected by an HTLV-1 strain with a gp46 identical to that of virus 2060. We showed that both 2060 and 2085 sera inhibited with the same efficiency syncytium formation induced by viruses 2060 and 2072 produced by lymphoid cell lines originating from the respective patients. These two sera were assayed for inhibition of syncytia induced by the HTLV-1 glycoproteins synthesized from the seven vectors described above. For comparison, inhibition of syncytium formation induced by viruses 2060 and 2072 was tested in parallel with the 2085 serum. Figures 4 and 5 show the levels of inhibition plotted against the reciprocal of the serum dilution. Consistent with our previous results, the 2085 serum inhibited with the same efficiency syncytium formation induced by viruses 2060 and 2072. The maximum levels of inhibition (99%) and the ID₅₀s (4,600 and 5,700 respectively) were similar to those found previously (Fig. 4). The values were slightly lower when we tested the inhibition of syncytia induced by 2060 and 2072 envelope glycoproteins produced by COS-LTR_HIV-LacZ cells transfected with pB1D or pB1D/SU72 (maximum levels of inhibition, 77 and 72%, respectively). This result may have been due to differences in protein maturation and/or presentation at the cell surface of HTLV-1-infected lymphoid cell lines and of transiently transfected COS-LTR_HIV-LacZ cells. The inhibition curves obtained for syncytia induced by all chimeric and point-mutated envelope glycoproteins were similar with both 2060 and 2085 sera (Fig. 4 and 5). The maximum levels of inhibition ranged from 71 to 77% with the 2085 serum and from 60 to 79% with the 2060 serum; the ID₅₀s differed by only one dilution (1/300 or 1/900). Statistical analysis of the inhibition patterns confirmed that both sera inhibited syncytia induced by the seven recombinant envelope glycoproteins with the same efficiency (P, 0.105 to 0.930).

FIG. 4.

Inhibition curves for syncytia induced by 2060 (solid circles)- and 2072 (open circles)-infected cells (broken lines) or COS-LTR_HIV-LacZ cells transfected with HTLV-1 env expression vectors (solid lines) with the 2085 serum. Each point corresponds to the mean of two to five values obtained in independent experiments. Expression vectors were pB1D (solid circles), pB1D/I173V-A187T (open triangles), pB1D/N93Y (solid triangles), pB1D/A187T (open diamonds), pB1D/SU72 (solid squares), pB1D/E39K (open squares), and pB1D/I173V (solid diamonds).

FIG. 5.

Inhibition curves for syncytia induced by COS-LTR_HIV-LacZ cells transfected with HTLV-1 env expression vectors with the 2060 serum. Each point corresponds to the mean of two to five values obtained in independent experiments. Expression vectors were those listed in the legend to Fig. 4.

These results demonstrated that the neutralizing activity of 2060 and 2085 sera was not affected by amino acid differences between the 2060 and 2072 envelope glycoproteins, whether these proteins were presented on the surface of either chronically infected lymphoid cells or transiently transfected COS-LTR_HIV-LacZ cells.

Inhibition of syncytium formation induced by recombinant HTLV-1 glycoproteins by 2072 serum.

The serum from the patient infected with virus 2072 displayed different neutralization patterns based on the HTLV-1 strain produced by the infected cell lines used in the syncytium and reporter gene inhibition assays. The 2072 serum neutralized the autologous virus (i.e., 2072) with a higher efficiency than the heterologous 2060 virus (5). To determine if amino acid changes in envelope glycoproteins affected their recognition by the 2072 serum, inhibition experiments were performed with syncytia induced by 2060 and 2072 envelope glycoproteins and by the five chimeric and point-mutated envelope glycoproteins described above. Control experiments were also performed with syncytia induced by viruses 2060 and 2072 produced by lymphoid cell lines. Figure 6 shows the inhibition curves obtained in each case. The data obtained on syncytia induced by viruses from lymphoid cell lines confirmed our previous findings, indicating that the 2072 serum showed a maximum level of inhibition and a higher ID₅₀ against virus 2072 than against virus 2060. The inhibition curves obtained when transiently transfected COS-LTR_HIV-LacZ cells were used as syncytium inducers could be classified into two groups.

FIG. 6.

Inhibition curves for syncytia induced by 2060 (solid circles)- and 2072 (open circles)-infected cells (broken lines) or COS-LTR_HIV-LacZ cells transfected with HTLV-1 env expression vectors (solid lines) with the 2072 serum. Each point corresponds to the mean of two to five values obtained in independent experiments. Expression vectors were those listed in the legend to Fig. 4.

The first group included the inhibition curves obtained with COS-LTR_HIV-LacZ cells transfected with the pB1D/SU72 and pB1D/I173V-A187T vectors. Relatively high levels of inhibition of syncytium formation were obtained in both cases (47 and 66% maximum inhibition at a 1/100 dilution, respectively). The second group included the inhibition curves obtained with COS-LTR_HIV-LacZ cells transfected with the pB1D, pB1D/E39K, pB1D/N93Y, pB1D/I173V, and pB1D/A187T vectors. In this group, the maximum levels of inhibition were very low or null (maximum inhibition at a 1/100 dilution, 0.7 to 26.3%).

Detailed analysis of the results obtained for the first group (Fig. 6 and Table 2) showed that the higher inhibitory activities of the 2072 serum were observed for syncytia induced by the chimeric proteins with the entire 2072 gp46 sequence (pB1D/SU72) and with only 2 amino acid changes in the 2072 gp46 sequence (pB1D/I173V-A187T) (47 and 66%, respectively). Statistical analysis confirmed that syncytium inhibition by the 2072 serum for cells transfected with pB1D/SU72 and pB1D/I173V-A187T was significantly different from that for cells transfected with pB1D (expressing 2060 gp46) (P, 0.012 and 0.002, respectively). The inhibition curves obtained for syncytia induced by point-mutated envelope glycoproteins did not differ from that obtained with the pB1D-encoded glycoprotein, (P, 0.152 to 0.735), indicating that none of the envelope glycoproteins displayed the antigenic structure specifically recognized by the 2072 serum.

TABLE 2.

Inhibition of syncytium formation induced by parental, chimeric, and point-mutated envelope proteins by 2072 serum

Envelope protein source	Inhibitiona at the following reciprocal of the serum dilution:
	100	300	900	2,700	8,100	24,300
2060 virus	38.15	30.76	16.33	9.69	13.00	0
2072 virus	86.27	79.81	54.26	44.00	31.33	15.67
pB1D	23.61	12.59	13.06	3.93	5.00	0
pB1D/SU72	44.98	47.23	20.67	11.97	0	0
pB1D/I173V-A187T	65.70	53.73	22.33	11.33	12.50	0
pB1D/E39K	26.30	11.55	3.33	1.49	0	0
pB1D/N93Y	10.00	24.05	11.67	0	0	0
pB1D/I173V	7.92	12.03	11.59	0	0	0
pB1D/A187T	0.70	15.50	17.90	1.60	0	0

^aReported as the percent reduction in the number of blue cells or in the absorbance at 620 nm of the positive control (coculture in the absence of 2072 serum) after subtraction of the background value obtained with untransfected COS-LTR_HIV-LacZ cells cocultivated with XC-tat cells. 

Taken together, these results indicated that specific recognition of envelope glycoproteins by neutralizing antibodies in the 2072 serum required a valine at position 173 and a threonine at position 187.

Inhibition of syncytium formation induced by recombinant HTLV-1 glycoproteins by 2056 serum.

In a previous study (5), we found that the 2056 serum showed a maximum level of inhibition and a higher ID₅₀ for virus 2072 than for virus 2060. However, the observed differences were not statistically significant. This result may have been due to the relatively low antibody titer of the 2056 serum compared to that of the 2072 serum or to differences in gp46 amino acid sequences at two positions (108 and 173) (Fig. 1). The inhibitory activity of this serum was reevaluated by use of syncytium inhibition assays performed with COS-LTR_HIV-LacZ cells transfected with different pB1D-derived vectors. The results presented in Table 3 and Fig. 7 showed that the maximum level of inhibition was slightly higher for syncytia induced by the pB1D/SU72 vector (expressing gp46 of virus 2072) (58%) than for syncytia induced by all other chimeric or point-mutated envelope glycoproteins and the 2060 envelope glycoprotein (33 to 44%). However, the observed differences were not statistically significant (P value for pB1D compared to pB1D/SU72, 0.39). These observations could be explained in light of those obtained with the 2072 serum indicating that V173 and T187 are required for the specific recognition of envelope glycoproteins by this serum. As shown in Fig. 1, virus 2056 has an isoleucine at position 173 and a threonine at position 187.

TABLE 3.

Inhibition of syncytium formation induced by parental, chimeric, and point-mutated envelope proteins by 2056 serum

Envelope protein source	Inhibitiona at the following reciprocal of the serum dilution:
	100	300	900	2,700	8,100
pB1D	43.25	25.70	16.62	5.90	6.83
pB1D/SU72	57.68	35.44	26.38	14.48	0
pB1D/I173V-A187T	40.87	28.99	14.86	5.14	0
pB1D/E39K	38.75	25.65	7.98	0	0
pB1D/N93Y	33.53	22.44	11.32	5.00	8.00
pB1D/I173V	37.24	13.78	10.90	8.94	0
pB1D/A187T	44.17	37.33	8.17	4.17	0

^aReported as the percent reduction in the number of blue cells or in the absorbance at 620 nm of the positive control (coculture in the absence of 2056 serum) after subtraction of the background value obtained with untransfected COS-LTR_HIV-LacZ cells cocultivated with XC-tat cells. 

FIG. 7.

Inhibition curves for syncytia induced by COS-LTR_HIV-LacZ cells transfected with HTLV-1 env expression vectors with the 2056 serum. Each point corresponds to the mean of two to five values obtained in independent experiments. Expression vectors were those listed in the legend to Fig. 4.

DISCUSSION

In a previous study (5), we showed that sera from some patients infected with cosmopolitan HTLV-1 strains with few amino acid changes in their envelope glycoproteins displayed different neutralization patterns. These patterns could be classified into three categories that fit well with groups of viruses each harboring the same residues in the major immunodominant and neutralizable domain (aa 175 to 199) of SU (5). The sera from the first group completely neutralized, with equivalent titers, viruses in the different groups, whereas some sera from the other two groups partially neutralized viruses in the different groups. Since within each group, different amino acids could be found at other positions, the residues involved in the observed differences could not be determined.

To identify the amino acid changes responsible for the antigenic specificity of neutralizable epitopes, we constructed an expression vector expressing the envelope glycoproteins of virus 2060. We derived chimeric or point-mutated vectors expressing envelope glycoproteins with the amino acid substitutions found in virus 2072, which induced some specific neutralizing antibodies in the infected patient. We found that two amino acids (positions 173 and 187) were important for correct recognition of the envelope glycoproteins by neutralizing antibodies present in the 2072 serum. Indeed, substitutions I173→V and A187→T in 2060 gp46 restored recognition by neutralizing antibodies in the 2072 serum. The involvement of these two amino acids in the induction of specific antibodies was corroborated by our observation that the serum of a patient infected with a virus (2056) harboring a threonine at position 187 but an isoleucine at position 173 has a phenotype intermediate between those of the 2060 and 2072 sera. The existence in 2072 SU of V173 and T187 instead of I173 and A187, as in 2060 SU, could alter the structure of the envelope glycoprotein, thereby affecting the immunogenicity of the proteins in infected patients. Indeed, aa 173 is located close to a conserved GYDP motif (aa 168 to 172). This motif, probably a β turn, is proposed to be a hinge in gp46 (13) and could be involved in an SU-TM association; a Y170→S mutation results in SU secretion into culture medium (7). The aa 187 is located in a major immunodominant and neutralizable region (aa 175 to 210) of gp46. Synthetic peptides with sequences in this region were recognized by 75 to 100% of HTLV-1-positive patient sera (20). It has been shown that an A187→T mutation in a gp46 peptide affects the binding of monoclonal antibody 2.30g (34). In contrast, a linear neutralizable epitope including aa 173 has not so far been identified. A peptide encompassing aa 172 to 194 of gp46 and recognized by 56% of HTLV-1-positive human sera has been shown to inhibit significantly the fusion induced by HTLV-1 (10). The two amino acids (V173 and T187) involved in specific recognition by the 2072 serum are located in this peptide. Amino acid changes at these positions could readily affect the structure and antigenicity of a neutralizable epitope located in this gp46 domain involved in cell fusion or of a conformational epitope involving two domains of the molecule.

As the 2060 and 2085 sera completely inhibited syncytium induction by envelope glycoproteins regardless of the amino acids at positions 173 and 187, it could be argued that the differences observed were not due to variations in env sequences but were due to differences in immune responses between individuals. This suggestion seems unlikely, as it is now established that variations in the amino acid sequence of the major immunodominant domain in gp46 modify recognition by human sera (11, 23). This finding was confirmed by the isolation of monoclonal antibodies specific for the amino acid sequence of the injected gp46 peptide (24, 34). Definite proof that I173→V and A187→T mutations affected the specificity of neutralizing antibodies will depend on the immunization of animals with recombinant glycoproteins which induce specific neutralizing antibodies, as the virus does in humans.

The nucleotide sequences of HTLV-1 proviruses originating from different areas in the world are highly conserved (1 to 5% variation for viruses found in Japan, the Caribbean Basin, the Americas, or Africa and 8% variation for viruses infecting some Melanesian populations). However, the results reported here showed that the influence of amino acid changes in the envelope glycoproteins, albeit rare, on neutralizing and cytotoxic activities induced in humans is worthy of investigation. In particular, the effects of amino acid changes in domains of gp46 known to contain neutralizable epitopes need to be determined, as such information has implications for the development of an effective vaccine against a wide range of HTLV-1 strains. The experimental approach described in this report could also help identify conformational epitopes involved in viral neutralization (16).