De novo design of peptide immunogens that mimic the coiled coil region of human T-cell leukemia virus type-1 glycoprotein 21 transmembrane subunit for induction of native protein reactive neutralizing antibodies.

Peptide vaccines able to induce high affinity and protective neutralizing antibodies must rely in part on the design of antigenic epitopes that mimic the three-dimensional structure of the corresponding region in the native protein. We describe the design, structural characterization, immunogenicity, and neutralizing potential of antibodies elicited by conformational peptides derived from the human T-cell leukemia virus type 1 (HTLV-1) gp21 envelope glycoprotein spanning residues 347-374. We used a novel template design and a unique synthetic approach to construct two peptides (WCCR2T and CCR2T) that would each assemble into a triple helical coiled coil conformation mimicking the gp21 crystal structure. The peptide B-cell epitopes were grafted onto the epsilon side chains of three lysyl residues on a template backbone construct consisting of the sequence acetyl-XGKGKGKGCONH2 (where X represents the tetanus toxoid promiscuous T cell epitope (TT) sequence 580-599). Leucine substitutions were introduced at the a and d positions of the CCR2T sequence to maximize helical character and stability as shown by circular dichroism and guanidinium hydrochloride studies. Serum from an HTLV-1-infected patient was able to recognize the selected epitopes by enzyme-linked immunosorbent assay (ELISA). Mice immunized with the wild-type sequence (WCCR2T) and the mutant sequence (CCR2T) elicited high antibody titers that were capable of recognizing the native protein as shown by flow cytometry and whole virus ELISA. Sera and purified antibodies from immunized mice were able to reduce the formation of syncytia induced by the envelope glycoprotein of HTLV-1, suggesting that antibodies directed against the coiled coil region of gp21 are capable of disrupting cell-cell fusion. Our results indicate that these peptides represent potential candidates for use in a peptide vaccine against HTLV-1.

Peptide vaccines able to induce high affinity and protective neutralizing antibodies must rely in part on the design of antigenic epitopes that mimic the threedimensional structure of the corresponding region in the native protein. We describe the design, structural characterization, immunogenicity, and neutralizing potential of antibodies elicited by conformational peptides derived from the human T-cell leukemia virus type 1 (HTLV-1) gp21 envelope glycoprotein spanning residues 347-374. We used a novel template design and a unique synthetic approach to construct two peptides (WCCR2T and CCR2T) that would each assemble into a triple helical coiled coil conformation mimicking the gp21 crystal structure. The peptide B-cell epitopes were grafted onto the ⑀ side chains of three lysyl residues on a template backbone construct consisting of the sequence acetyl-XGKGKGKGCONH2 (where X represents the tetanus toxoid promiscuous T cell epitope (TT) sequence 580 -599). Leucine substitutions were introduced at the a and d positions of the CCR2T sequence to maximize helical character and stability as shown by circular dichroism and guanidinium hydrochloride studies. Serum from an HTLV-1-infected patient was able to recognize the selected epitopes by enzyme-linked immunosorbent assay (ELISA). Mice immunized with the wild-type sequence (WCCR2T) and the mutant sequence (CCR2T) elicited high antibody titers that were capable of recognizing the native protein as shown by flow cytometry and whole virus ELISA. Sera and purified antibodies from immunized mice were able to reduce the formation of syncytia induced by the envelope glycoprotein of HTLV-1, suggesting that antibodies directed against the coiled coil region of gp21 are capable of disrupting cell-cell fusion. Our results indicate that these peptides represent potential candidates for use in a peptide vaccine against HTLV-1.
The efficacy of potential peptide vaccines for bacterial, parasitic, and viral diseases as well as cancer depends partly on the induction of broadly protective, high affinity neutralizing antibody responses. Such responses rely on the design of epitopes able to mimic the three-dimensional structure of the respective B cell epitope in the native protein as well as the activation of the helper T cell component of the immune response. Over the last several years, our laboratory has conceptualized and implemented several approaches to engineer conformational epitopes in tandem with a chimeric strategy to deliver potential vaccine candidates. The design involves selecting surface-oriented antigenic epitopes that display certain secondary structural attributes (␣-helix, ␤-turn, and loop) that are then linked to a "promiscuous" T-helper cell epitope via a four-residue turn sequence (GPSL). These chimeric peptides allow independent folding of the B-cell and T-cell epitope and are very effective in overcoming major histocompatibility complex restriction in various strains of mice and in eliciting high titered and high affinity antibody responses (1)(2)(3)(4). Additionally, the incorporation of a promiscuous T-helper epitope was effective in generating the required "help" for the induction of antigen specific antibodies without the problems associated with coupling peptides to carrier proteins. Furthermore, to construct epitopes in various defined orientations and conformations, a novel single-matrix multicomponent template strategy was developed previously (3,5,6). In this approach, a core ␤-sheet template consisting of alternating Gly/Lys is synthesized such that the ⑀ side chains of the lysine residues allow for the growth of individual epitopes in varying permutations. Model peptides incorporating different combinations of B-cell and T-cell epitopes from either the lactate dehydrogenase C 4 (LDH-C 4 ) antigen or the human T-cell leukemia virus type 1 (HTLV-1) 1 envelope surface-exposed subunit synthesized in a template format resulted in enhanced immunogenicity in several inbred and outbred strains of mice and rabbits eliciting antibodies reactive with the native protein (4,(7)(8)(9).
HTLV-1 is a retrovirus that is associated with various diseases including adult T-cell leukemia/lymphoma(10) and trop-ical spastic paraparesis/HTLV-1-associated myelopathy (11). It has also been associated with a number of inflammatory diseases, including pediatric infectious dermatitis (12), uveitis (13), and some cases of arthropathy (14) and polymyositis (15). HTLV-1 transmission occurs mainly via fusion between infected Env-expressing cells and receptor-bearing cells because infection by cell-free HTLV-1 virus is inefficient in vitro and in vivo. The HTLV-1 envelope proteins expressed on the surface of virus-infected cells and on viral particles are the first to be recognized by the host in the course of a natural immune response (16,17). Several studies have focused on linear and conformational immunodominant regions of the surface-exposed glycoprotein subunit to elicit neutralizing antibody responses against HTLV-1 for the purpose of vaccine development and diagnostic screening (18 -26). However, there has been far less focus on the gp21 transmembrane (TM) subunit. Recent data has shown that the transmembrane gp21 domain plays a critical role in the postbinding steps during infection that is required for the viral core to be delivered into the target cell cytoplasm (27). Peptides derived from the gp21 region encompassing amino acids 361-430 showed specific reactivity to sera from HTLV-1-infected individuals (28). Likewise, later studies with overlapping synthetic peptides revealed the importance of amino acids 400 -429 in inhibiting the formation of syncytia between infected cells (29).
There is little structural homology between the surface-exposed subunits of various retroviral envelope proteins. However, the TM subunits of retroviruses display remarkable homology with conserved spatial conformations (30). In general, they can be divided into an N-terminal hydrophobic fusion peptide, an adjacent leucine zipper-like motif that is capable of self-assembly into a coiled coil, and a disulfide-bonded region followed by a C-terminal region that contains ␣-helical segments. The transmembrane and cytoplasmic tail regions are located at the C terminus. The recently solved crystal structure of the central segment of the HTLV-1 transmembrane subunit as a chimera with maltose-binding protein shows that the HTLV-1 TM does not deviate from the conserved spatial organization and general organization (31).
The gp46 subunit is believed to be responsible for the recognition of the cellular receptor. After receptor recognition, conformational changes in the transmembrane region exposes the fusion peptide, which brings the viral and cellular membranes into close proximity, allowing fusion to occur. Fusion is required for the introduction of the viral core into the cytoplasm of target cells. Although there is limited information regarding the precise mechanism of envelope-mediated fusion and the role of the TM glycoproteins, their critical role in the fusion process is suggested by their similarity in sequence and structure (32). Mutational studies have defined certain regions such as the coiled coil segment and the disulfide-bonded region of chain reversal C-terminal to the coiled coil segment as being critically involved during the later stages of the fusion process after receptor binding (27,33).
Based on the above observations, we hypothesized that a rationally designed vaccine consisting of B-cell epitopes derived from these important regions of the TM subunit should elicit antibodies that could interfere with the fusion process by binding to the relevant region on the envelope glycoprotein. We focused on a B-cell epitope derived from the central coiled coil region for our studies that is critical for fusion. This report describes the structural characterization and immunogenicity studies of peptides derived from the central region of the TM subunit (residues 347-374) that required a new template design and a different synthetic approach to synthesize a peptide that would assemble into a triple helical coiled coil conforma-tion mimicking the gp21 solved crystal structure. The rationale for this design was to constrict the three strands at one end and also to bring them in close proximity to promote interactions between the hydrophobic residues at the a and d positions to form a trimeric coiled coil. Leucine substitutions at the a and d positions of the heptad repeat were also made to maximize hydrophobic interactions between these residues that would potentiate the triple helical coiled-coil formation. Circular dichroism (CD) measurements of the mutated coiled coil peptide revealed a stable ␣-helical structure that was concentrationindependent. The wild type peptide, on the other hand, was less helical under similar conditions. Serum of a HTLV-1-infected patient was able to recognize the selected epitopes by ELISA. Both peptides were highly immunogenic in outbred mice. Antibodies against both peptides were able to cross-react with the native envelope protein as determined by ELISA with the recombinant gp21 protein. Both peptides were shown to induce high antibody titers, which were capable of recognizing the native protein as shown by Flow cytometry and whole virus ELISA. Furthermore, antibodies to the leucine-substituted peptide bound with higher affinity to the native whole viral protein, indicating that the peptide folds in the native conformation. Competition experiments using various peptides as inhibitors revealed that antibodies against the leucine-substituted template peptide were highly specific to the conformation of the immunogen. Sera and purified antibodies from mice immunized with either peptide immunogen were able to reduce the formation of syncytia induced by the envelope glycoprotein of HTLV-1. These results indicate that these two peptides represent potential candidates for future multivalent vaccine studies.

EXPERIMENTAL PROCEDURES
Peptide Synthesis and Purification-Template peptides were synthesized by solid phase peptide synthesis after Fmoc chemistry with benzotriazole-1-yl-oxy-tris-(dimethylamino)-phosphonium hexafluorophosphate/N-hydroxybenzotriazole as a coupling reagent on a fully automated peptide synthesizer (Model 9600 Peptide Synthesizer Milli-Gen/Biosearch) as described previously (9) with modifications. The template peptide backbone (GKGKGKG) ( Fig. 1) with the Lys side chains protected (ivDde) was assembled on Rink-Amide-CLEAR Resin (substitution 0.41 mmol of amino groups/g (Peptides International, Louisville, KY) using Fmoc-Lys(ivDde) (Bachem, Torrance, CA)). The N terminus of the template peptide was acetylated using acetylimidazole (Aldrich, Milwaukee, WI) in N,N-dimethylformamide. The Lys side chain deprotection (ivDde) was achieved using 2% hydrazine hydrate in N,N-dimethylformamide (3 and 10 min, positive Kaiser test). The gp21 (residues 347-374) peptides were assembled on the template using the peptide synthesis protocol described above and acetylated at N terminus. The peptides were cleaved from support using Reagent R. The crude peptides were purified by semipreparative reversed phase HPLC using a C-4 column (Vydac, Hesperia, CA). Analytical HPLC was performed using a Vydac column and a linear gradient of 90% acetonitrile in water containing 0.1% trifluoroacetic acid. The identities of the peptides were confirmed by electrospray ionization time of flight mass spectrometry.
Circular Dichroism Measurements-Circular dichroism measurements were performed on an AVIV model 62A DS CD instrument (Lakewood, NJ). All spectral measurements were obtained at 25°C under continuous nitrogen purging of the sample chamber using a quartz cuvette of 0.1-cm path length. Spectral measurements of the peptides were obtained over a range of concentrations (25-100 M) in water or in 50% trifluoroethanol (TFE). Water or 50% TFE blanks were subtracted from the CD spectra. Molar ellipticity values were calculated using the formula [] M, ϭ ( ϫ 100 ϫ M r )/(n ϫ c ϫ ), where is the recorded ellipticity (deg), M r is the molecular weight of the peptide, n is the number of amino acid residues in peptide, c is the peptide concentration (mg/ml), and is the path length of the cuvette (34). Helicity of the peptides was calculated according to the Chen equation (35), with reference to the mean residue ellipticity of polylysine for 100% ␣-helix () 222 nm ϭ Ϫ33,000.
Guanidinium Hydrochloride Denaturation-The structural stability of the synthesized coiled coil template peptides was determined by chemical denaturation experiments using guanidinium hydrochloride (GnHCl) as described previously (36). Peptide concentration was maintained at 50 M, and the concentration of GnHCl was increased from 0 to 13 M in water. The ellipticity of the peptide was then measured at 222 nm, similar to that described above. Ellipticity was then normalized to the fraction folded by using the equation: f folded ϭ ( observed Ϫ unfolded )/ ( folded Ϫ unfolded ) (37).
Animal Immunizations-Six-eight-week-old female ICR outbred mice were purchased from The Jackson Laboratory (Bar Harbor, ME). Groups of 8 -9 mice were immunized separately with 100 g of peptide dissolved in water with 100 g of a muramyl dipeptide adjuvant (Nacetylglucosamine-3 yl-acetyl-L-alanyl-D-isoglutamine nor-MDP) and emulsified (50:50) in Squalene/Arlacel A oil (4:1) as described elsewhere (38). Booster immunizations were administered at 3, 6, and 9 weeks. Mice were bled retro-orbitally every week for antibody titer determination. Sera collected were complement-inactivated by heating at 56°C for 30 min. High titrated sera were purified on a protein A/G-agarose column (Pierce), and eluted antibodies were concentrated and exchanged in PBS using 100-kDa cutoff centrifuge filter units (Millipore). Antibody concentrations were determined with the Coomassie Plus protein assay reagent kit (Pierce).
ELISA-Antibody titers were determined using ELISA as previously described (8). Titers were defined as the highest dilution of sera with an absorbance greater than 0.2 after subtracting the background. All data represent the average of duplicate samples.
Direct ELISA with HTLV-1-infected Patient Serum-Plates were coated with peptide immunogens and B-cell epitopes as described for direct ELISA. HTLV-1-infected asymptomatic patient serum was used as a primary antibody beginning with a 1:50 dilution. After a 2-h incubation, a 1:500 dilution of horseradish peroxidase-conjugated goat anti-human secondary antibody was added (Pierce). Substrate development was preformed as described for the direct ELISA protocol.
Competitive ELISA-Plates were coated with peptide/antigen and blocked as described for the direct ELISA. Primary antibody was added as 50 l of a constant dilution (50% maximal binding as determined by ELISA) with 50 l of a serially diluted competitive inhibitor, with concentrations ranging from 0 to 60 M. Positive inhibitors were peptide/antigen for which the antibodies were specific. Inhibitor and antiserum were incubated for 2 h. Horseradish peroxidase-linked goat anti-mouse IgG addition and color development were performed as described for the direct ELISA protocol.
Whole Virus ELISA-Reactivity of peptide antisera to native HTLV-1 whole viral lysate was determined using whole virus ELISA. A Vironostika HTLV-I/II Microelisa assay (Organon Teknika, Durham, NC) was performed according to manufacturer's instructions, modified for mouse serum. Antibody dilutions used were 1/100 for immune and preimmune negative control serum. A 1/2000 dilution of horseradish peroxidase-linked goat anti-mouse IgG (0.8 mg/ml) secondary antibody was used. Plates were read at 450 nm, and the data are represented as absorbance units.
Cell Lines-HTLV-1-infected T-cell line ACH (39) was grown in RPMI 1640 with 10% fetal calf serum and 1% penicillin/streptomycin. For the ACH cells, 5% human interleukin 2 was also added to the culture medium. HTLV-1-infected MT-2 cells were cultivated in the same medium described above but without interleukin-2 and supplemented with 2 mM glutamine. HeLa-Tat cells containing the human immunodeficiency virus Tat transactivator gene (a kind gift from C. Pique) previously described in (40) were cultivated in Dulbecco's modified Eagle's medium containing 10% fetal calf serum, 1% penicillin/ streptomycin. CosZ28 cells stably transfected with the ␤-galactosidase gene under the control of the human immunodeficiency virus long terminal repeat promoter (a kind gift from C. Pique) and previously described by Dragic et al. (40) were cultivated in same medium as HeLa-Tat cells, complemented with 300 g/ml hygromycin.
Flow Cytometry-Binding of peptide antibodies to HTLV-1 ACHinfected cells was tested by flow cytometry as adopted from Hudziak et al. (41). 0.5 ϫ 10 6 cells were incubated with a 1 ⁄5 dilution of the peptide antibodies. The same dilution of preimmune serum was used as a negative control. HTLV-1-infected rabbit serum was used as a positive control. The cells were stained with primary antibodies for 2 h at 4°C in 100 l of PBS, 2% fetal calf serum. The cells were then washed twice with PBS and stained by incubation with fluorescein isothiocyanatelabeled F(ab) 2 fragment goat anti-mouse IgG secondary antibody (1:50 dilution) for 45 min at 4°C in 100 l of PBS, 2% fetal calf serum, then fixed in PBS, 2% paraformaldehyde and analyzed by coulter ELITE flow cytometer (Coulter, Hialeah, FL). A total of 10,000 gated events were collected, and final processing was performed. Debris, cell clusters, and dead cells were gated out by light scatter assessment before single parameter histograms were drawn.
Syncytia Inhibition Assay-A ␤-galactosidase assay was used for the quantitative evaluation of syncytia formation. A three-cell line-based protocol modified from Kazanji et al. (42), Blanchard et al. (43), and Delamarre et al. (44) was used. Briefly, in 24-well plates 5 ϫ 10 4 MT-2 cells were incubated in 300 l of MT-2 culture medium (containing a 1:50 dilution of donor herd horse serum (Invitrogen)) with 4-fold serial dilutions of mouse antisera and HTLV-1 asymptomatic patient sera beginning with 1:50 or three dilutions of purified antibody solution (250, 60, 1 ng/ml) for 1 h at 37°C and 5% CO 2 . After incubation, 5 ϫ 10 4 CosZ28 and 5 ϫ 10 4 HeLa-Tat cells were added to each well in 600 l of HeLa-Tat culture medium, bringing the final volume to 0.9 ml/well. All sera and purified antibody dilutions were preformed in triplicate. Plates were then incubated for 20 h at 37°C and 5% CO 2 . Cells were washed once with PBS, and ␤-galactosidase production was detected using the Galactolight® kit (Tropix) as described by Delamarre et al. (44). 90 l of lysis buffer, as provided in the kit, was used to harvest the cells, and 20 l of lysate per well was added to 96-well chemiluminescent detection plates (Costar) in triplicate and incubated for 30 min with substrate (as per the manufacturer's instructions). Plates were read using a Lmax® luminometer (Molecular Devices). After the injection of 100 l/well of accelerator, plates were read for 10 s after a 2-s delay. The % of syncytia inhibition was defined as (luminescence of positive control wells without experimental sera Ϫ luminescence of experimental wells with sera)/ (luminescence of positive control wells without sera) ϫ 100. Positive control wells contained MT-2, CosZ28, and HeLa-Tat cells in the absence of experimental antibodies. The syncytia inhibition (ID 50 ) titer was defined as the reciprocal of the serum dilution, resulting in a 50% or greater reduction of syncytia formation as compared with the positive control wells. Sera from HTLV-1-seropositive individuals were used as positive controls for the syncytia inhibition assay. Sera from HTLV-1 patients were titered using the gelatin-particle agglutination assay (Serodia HTLV-1 kit, Fujirebio Inc., Tokyo, Japan).

RESULTS
Engineering of Template Peptides to Mimic Coiled Coil Conformation-It has been previously shown that the minimum peptide length for stable formation of a coiled coil was 29 FIG. 1. Template design for synthesis of the coiled coil region of gp21 envelope subunit. Amino acids 347-374 corresponding to the coiled-coil region of HTLV-1 gp21 was synthesized on a template peptide backbone of three Gly-Lys repeats. A promiscuous T-helper epitope (TT) derived from tetanus toxoid (sequence 580 -599) was synthesized colinear with the template. The peptide was acetylated at all N termini and amidated at the C terminus to limit charge repulsions.
residues (45), which corresponds to ϳ4 heptads with 7 residues/ heptad. Coiled coils are usually made up of heptad (abcdefg) repeats that are stabilized by hydrophobic interactions between the a and d positions in the different strands (46,47). The HTLV-1 coiled coil segment consists of amino acids 340 -392. We selected the region between amino acids 347 and 374, which spans the central region of the coiled coil and has four potential heptad repeats. To increase the proximity of the three peptide strands such that they interact with one another to form a coiled coil, the peptides were synthesized simultaneously through the ⑀-NH 2 side chains of 3 lysyl residues separated by intervening glycine residues as depicted in Fig. 1. The three strands were each acetylated at the N terminus to limit charge repulsions between each strand. A promiscuous T-helper epitope derived from the tetanus toxoid protein (residues 580 -599) was also synthesized colinear with the template backbone sequence at the N terminus. First, we synthesized the wild type sequence 347-374 designated WCCR2T. Additionally, another template construct designated CCR2T was designed with five substitutions at the a and d positions (V349L, I353L, I360L, N363L, and I370L). The rationale for these substitutions was to increase the hydrophobic interactions that should further stabilize the coiled coil conformation. We also synthesized the two peptides as single strands that corresponded only to the B-cell epitopes (residues 347-374) of WCCR2T and CCR2T that were acetylated and amidated to stabilize the helix dipole. The peptides were purified by reverse phase HPLC, and the identities of all peptides were confirmed by matrix-assisted laser desorption ionization time-of-flight or electrospray ionization mass spectrometry. All peptides synthesized along with their molecular weights and sequences are shown in Table I.   TABLE I Chimeric template peptides synthesized from central region of gp21 envelope subunit Leucine substitutions introduced into the wild type sequence are shown in bold. Lysine residues used for extension of the gp21 B-cell epitope (amino acids 347-374) are indicated by an asterisk CCR2E and WCCR2E are the B-cell epitopes synthesized separately as free peptides. All peptides were acetylated (CH 3 CO) at the N terminus and amidated (CONH 2 ) at the C terminus.  Fig. 3. High helical content (ϳ55%) that was stable over a wide range of concen-trations was observed in water with the CCR2T peptide. When the measurements were taken in 50% TFE there was a marginal increase in the helicity of the peptide. This implies that the peptide attained its maximum helical potential in water and that TFE did not have any major affect on inducing a helical conformation. Furthermore, the concentration-independent helical content implied that there was no oligomerization or aggregation of the peptide at higher concentrations. The WCCR2T peptide, on the other hand, showed a much lower helicity in water as compared with the CCR2T (ϳ13%), which increased significantly to 55% when TFE was added. TFE stabilizes ␣-helical structure (49); hence, the wild type template peptide did not fold completely in aqueous solution. The individual B-cell epitopes synthesized as free peptides were also tested for their helical potential. CCR2E and WCCR2E showed similar levels of helicity of (ϳ12 and 10%, respectively) in water. However, upon the addition of TFE the helicity increased to about 49% for WCCR2E and 55% for CCR2E (Fig. 3).
Here again there was no increase in helicity as the concentration was increased, suggesting that there was no oligomerization.
Additionally, the stability of the CCR2T and WCCR2T peptides in water was tested by chemical denaturation experiments using increasing concentrations of GnHCl. As observed in Fig. 4, the CCR2T peptide was highly resistant to chemical denaturation, and significant unfolding was observed only at higher concentrations of GnHCl (Ն6 M). In contrast, WCCR2T revealed a steady unfolding of the secondary structure with increasing concentrations of GnHCl. This implied that the CCR2T template peptide had adopted a more stable ␣-helical secondary structure that was probably due to the increased hydrophobic interactions with the leucine-substituted residues.

Recognition of Epitope Sequence by HTLV-1-infected Patient
Serum-An ELISA test was performed to evaluate whether the selected peptide epitope is recognized by naturally occurring anti-HTLV-1 antibodies present in infected patients. As shown in Table II, high titers of antibodies could be detected against both WCCR2T and CCR2T. ELISA titers of 50 -100 were obtained when the plates were coated with immunogen, and titers of 50 -800 were obtained when only the B-cell epitopes was used on the plates. These results confirm that the epitopes chosen are indeed antigenic and that the peptides can mimic the natural protein.
Immunogenicity Studies of Template Peptides-The two template peptides were synthesized to include a promiscuous Thelper epitope TT from tetanus toxoid (amino acids 580 -599) (6). This epitope was chosen since it was devoid of lysine residues that may interfere with the synthesis of the three strands of the gp21 B-cell epitope. WCCR2T and CCR2T were used to immunize groups of female outbred mice. Both template peptides were highly immunogenic, and high-titered antibodies were obtained (Fig. 5). Both CCR2T and WCCR2T elicited high antibody responses after the first booster immunization, and subsequent booster injections did not increase the titers. The WCCR2T peptide was slightly more immunogenic, and higher titers in the range of 32,000 -64,000 (reciprocal of highest sera dilution with absorbance units Ն0.2) were obtained 2 weeks after the first booster. Overall, WCCR2T titers were slightly higher than CCR2T titers, although both peptides elicited hightitered antibodies in outbred mice.
Cross-reactivity of Peptide Antibodies to Native Protein-For peptide antibodies to be effective in neutralizing virus infection, it is imperative that the antibodies cross-react with the native protein from which the epitope was derived. To test the cross-reactivity of template peptide antibodies to the native gp21 protein we utilized three separate assays, 1) flow cytometry using HTLV-1-infected cells, 2) direct ELISA using recombinant maltose-binding protein (MBP)-gp21 chimera (50), and 3) whole virus ELISA using plates coated with HTLV-1 viral lysate.
ACH cells are primary human peripheral blood cells that were immortalized using an infectious molecular clone of HTLV-1 (51). These cells express the envelope protein of HTLV-1 and were used to test the binding of CCR2T and WCCR2T mouse antisera. As depicted in Fig. 6, upper panels, we observed that both of the antisera bound the surface of ACH cells, indicating that both antibodies were capable of recognizing the native protein. Preimmune serum from outbred mice was used as a negative control. These results were further confirmed using direct ELISA with recombinant MBP-gp21 protein chimera. The antibodies against CCR2T and WCCR2T were able to specifically bind the gp21 protein (Fig. 6, lower  panel).
Virus neutralization has been defined as the "loss of infectivity" that occurs as a result of anti-viral antibodies binding to free virus particles (52,53). The ability of peptide-induced TABLE II Direct ELISA with HTLV-I-infected patient serum ELISA plates were coated with WCCR2T and CCR2T immunogens and B-cell epitopes and screened for their cross-reactivity with HTLV-1-infected asymptomatic patient serum by direct ELISA. A 1:500 dilution of goat anti-human IgG (HϩL) conjugated with horseradish peroxidase was used as a secondary antibody.

Treatment
Immunogen titer a B-cell epitope titer a WCCR2T 100 50 CCR2T 50 800 a Titer is defined as the reciprocal of the lowest dilution with an absorbance values for clarity: Ͼ0.2 after subtracting backround.

FIG. 5. Immune responses to CCR2T and WCCR2T in ICR outbred mice.
Antibody titers against the corresponding immunogen in the sera of each individual mouse were determined by direct ELISA. A 1:500 dilution of rabbit anti-mouse IgG (HϩL) conjugated with horseradish peroxidase was used as secondary antibody. Arrows represent booster immunizations. antibodies to bind native protein on viral particles was tested in a whole virus ELISA system using microtiter plates coated with an HTLV-1 viral lysate and various dilutions of CCR2T and WCCR2T antisera. As indicated in Fig. 7, both CCR2T and WCCR2T antisera were able to recognize whole viral lysate. A significant difference between the levels of binding of CCR2T antisera and WCCR2T antisera was observed. The CCR2T antisera showed a much higher affinity as compared with the binding observed for the WCCR2T antisera. The preimmune mouse serum did not show any binding under similar conditions. These results indicate that the CCR2T antisera had higher levels of cross-reactive antibodies against the native protein than WCCR2T.
Specificity of Peptide Antibodies-The relative affinities or specificities of the two template peptide antisera for the native gp21 protein, the corresponding template peptide immunogen, and the B-cell epitope for each template by itself were assayed using competitive inhibition ELISA. Immune CCR2T antisera (1/16,000 dilution) could be effectively inhibited from binding to immobilized immunogen CCR2T by the CCR2T inhibitor in solution. In the same analysis 16-fold higher concentrations of the B-cell epitope CCR2E was required to completely abrogate binding of CCR2T antisera to the immobilized immunogen (Fig. 8A). These results indicate that the CCR2T antiserum was highly specific to the immunizing template peptide or to its specific conformation. In contrast, on WCCR2T template peptide-coated plates, both WCCR2T template peptide and WCCR2E B-cell epitope peptide in solution were effective in inhibiting binding of WCCR2T antisera (1/16000 dilution), the WCCR2T template peptide being only slightly more effective than WCCR2E (Fig. 8B). FIG. 7. Reactivity of template peptide antisera against native protein in a whole virus ELISA using plates coated with a HTLV-1/2 viral lysate. ELISA was performed as described under "Experimental Procedures" with a starting dilution of 1/100 of the antisera. Pooled mouse antisera against both CCR2T and WCCR2T were used. Preimmune mouse serum did not show any reactivity against whole viral lysate under similar conditions. FIG. 6. Relative binding of CCR2T and WCCR2T antisera to gp21 protein on the surface of HTLV-1 infectious molecular clone immortalized ACH primary cells. Upper panels, indirect immunofluorescence staining was used to determine the relative binding. Gray histograms represent staining with preimmune serum, and solid line histograms represent sera from an HTLV-1infected individual, used as a positive control. CCR2T and WCCR2T antisera (dashed line) were tested at a 1:5 dilution. Lower panel, cross-reactivity of peptide antibodies to native gp21 protein. Reactivity of CCR2T and WCCR2T mouse antisera was tested against recombinant MBP-gp21 protein chimera in a direct ELISA. Preimmune mouse serum did not show any reactivity against gp21 protein or whole viral lysate.
FIG. 8. Competitive inhibition ELISA curves for CCR2T and WCCR2T antisera using corresponding immunogen-coated plates. Using CCR2T-coated plates and a 1/16,000 dilution of CCR2T antisera (A) or WCCR2Tcoated plates and a 1/16,000 of dilution WCCR2T antisera (B), the inhibitory capacity of increasing concentrations of various peptide inhibitors as indicated was tested. The binding of both antisera was most efficiently inhibited by its corresponding immunogen. The insets in both panels show the serum dilution curve for each antibody on the same microtiter plates. C, competitive inhibition ELISA curves for CCR2T and WCCR2T antisera using recombinant gp21 protein-coated plates. The inhibitory capacity of increasing concentrations of gp21 protein inhibitor on the binding of CCR2T and WCCR2T antisera (1/20 dilution) to immobilized gp21 protein on the microtiter plate was tested.
We also tested the relative affinity of the fraction of gp21 protein cross-reactive antibodies in CCR2T antisera and WCCR2T antisera. Both antisera (1/20 dilution) were tested for their binding to recombinant gp21 protein immobilized on the plate in the presence of increasing concentrations of gp21 protein inhibitor in solution. As seen in Fig. 8C, CCR2T antisera and WCCR2T antisera binding to immobilized protein could be inhibited by gp21 protein in solution with similar affinity (Fig. 8C).
Reduction in Syncytia Formation by Peptide-induced Antibodies-To test the ability of our peptide-induced antibodies to reduce cell fusion mediated by HTLV-1-infected MT-2 cells, we used antisera and purified antibodies from the terminal bleeds of WCCR2T (titers 32,000 and 64,000, respectively)-and CCR2T (titers 16,000 and 32,000, respectively)-immunized mice. Sera from two HTLV-1-infected patients (patient 1 and 7) were used as positive control. Titers for these patients were determined to be 16,000 and 64,000, respectively, using the gelatin-particle agglutination assay. As shown in Fig. 9, both WCCR2T and CCR2T antibodies were very similar in their ability to reduce the formation of syncytia. Using crude sera, mice immunized with WCCR2T were found to have an ID 50 titer of 50 (57.05% inhibition), whereas CCR2T-immunized mice possessed an ID 50 titer of 50 (55.62% inhibition). The ID 50 titer of HTLV-1-infected patient 1 was 200 (64.31% inhibition), whereas patient 7 was 800 (64.41% inhibition) (Fig. 9). In studies with purified antibodies (Fig. 10) using antibody concentrations of 250, 60, and 1 ng/well, WCCR2T inhibited syncytia better than CCR2T at all three dilutions (43.71, 36.69, and 30.30% respectively), although CCR2T better retained its inhibiting activity over the range of dilutions (27.83, 25.61, and 22.96%, respectively). Nonspecific inhibition from preimmune serum was quickly diluted out over the course of dilutions (18.04, 12.52, and 5.79%, respectively).

DISCUSSION
Peptides that mimic the native structure in an antigenic protein are more likely to elicit antibodies that have high affinity for the native protein, which may have neutralizing potential. Likewise, many of the neutralizing antibodies against the envelope glycoprotein of HTLV-1 that have been identified during natural infection have been specific for conformational epitopes, i.e. those that arise due to specific folding of the protein or those antibodies that recognize specific conformations adopted by immunogenic contiguous segments of the protein. Hence, in recent years there has been a great thrust in the design of peptides that are able to adopt a defined conformation in solution (54 -56). The central region of the HTLV-1 TM subunit has been implicated to be very important in the fusion process between the viral or infected cell membrane and the receptor-expressing target cells (32). This region forms a leucine zipper-like motif that undergoes conformational changes to adopt a parallel trimeric coiled coil structure that brings the fusion peptide in close proximity with the target cell membrane, leading to fusion. Indeed, the crystal structure of the gp21 subunit in the fusion-activated state shows the central region to be a coiled coil. Furthermore, most of the point mutations within this region result in loss of infectivity (33). From a vaccine standpoint, we hypothesized that antibodies specific to these transitional structures would be able to interfere with the fusion process and, hence, arrest viral infection. The goal of the present study, therefore, was to design peptides that mimic the native coiled coil conformation of the gp21 central region and to generate antibodies that bind with high affinity to the native protein. The HTLV-1 coiled coil region is naturally stabilized by various internal interactions. This includes among other interactions an N-terminal hydrophobic cap consisting of the Met-338 side chain that makes extensive contact with the first helical residue Leu-340 and a hydrophobic core structure formed by Leu-384, Leu-385, Phe-386, and Ala-395 at the C terminus. The N-terminal cap has been implicated in stabilizing the fusion-activated coiled coil state similar to the HA2-capping structure of influenza virus (33). ␣-Helical segments as isolated peptides are generally devoid of any structure and assume random coil-like conformations in aqueous medium unless constrained by disulfide bridges and appro- FIG. 9. Mouse antisera generated against CCR2T and WCCR2T peptide immunogens were tested for their ability to inhibit syncytia formation mediated by the HTLV-1 Env protein. 4-Fold serial dilutions of antisera were used beginning with 1:50 to inhibit the fusion of MT-2, CosZ28, and HeLa-Tat cells. Inhibition of syncytia was measured by the reduction in ␤-galactosidase production detected by chemiluminescence. Sera from two HTLV-1-infected patients were used as positive controls, and mouse presera was used as a negative control. FIG. 10. Protein A/G-purified mouse antibodies generated against CCR2T and WCCR2T peptide immunogens were tested for their ability to inhibit syncytia formation mediated by the HTLV-1 Env protein. Three concentrations of purified antibody (250, 60, and 1 ng/well) were used to inhibit the fusion of MT-2, CosZ28, and HeLa-Tat cells. Inhibition of syncytia was measured by the reduction in ␤-galactosidase production detected by chemiluminescence. Mouse preimmune antibodies were used as a negative control. priate cyclizations (46). To compensate for the lack of additional interactions with adjacent residues when designing the coiled coil peptide segment as an isolated species, a template strategy was applied that has been successfully used in the past by our laboratory to synthesize peptides that have defined structure. This strategy involves a single-matrix multicomponent synthesis that constrains the individual peptide strands at one end on a ␤-sheet template and serves to bring each strand in close proximity to promote hydrophobic interactions to fold into a coiled coil. Similar strategies such as the use of rigid organic macromolecules have been applied for the synthesis of model coiled coils from other proteins referred to as template-assembled synthetic proteins (36,57,58). Additionally, based on studies of proteins that form helical bundles, leucine was found to be the most frequent occupant at positions a and d of the heptad repeat (59 -61). Hence, we synthesized a similar template peptide where all the residues at the a and d positions were mutated to leucine (CCR2T) as the latter residue displays side chain orientation that is conducive to ideal packing. The interdigitation of the hydrophobic core allows the display of hydrophilic residues that are important for antibody recognition. Circular dichroism experiments revealed that the peptide with leucine substitutions had a high helical content in aqueous solution, and the addition of helix-inducing TFE solvent did not increase the helicity significantly. Furthermore, the CCR2T peptide was highly resistant to chemical denaturation, and the peptide remained almost completely folded in 6 M GnHCl. This was in sharp contrast to the wild type peptide WCCR2T synthesized on a similar template whose structure showed steady disruption at much lower concentrations of Gn-HCl. Likewise the individual B-cell epitope CCR2E also displayed very low helical content in water. These results taken together suggest that the leucine substitutions combined with the template design for synthesis was essential for the formation of a defined ␣-helical structure.
To further test whether our design principles for an helical coiled-coil structure was correct, we used the whole virus ELISA assay to test the binding potential of CCR2T antisera versus the WCCR2T wild type antisera. Our results show that the CCR2T antibody had a higher binding affinity to the HTLV-1 viral lysate. These results indicated that the CCR2T template peptide was able to better mimic the native protein conformation. More importantly, antibodies from an HTLV-1infected patient were able to recognize the triple helical coiledcoil constructs. Competitive inhibition experiments further confirmed the specificity of CCR2T antisera to a defined secondary structure. In addition, peptide competition analysis showed the CCR2T template peptide was a better inhibitor of CCR2T antisera than the corresponding B-cell epitope, which lacked a defined secondary structure in aqueous solution. In contrast, WCCR2T antisera could be efficiently inhibited by both WCCR2T template peptide and the corresponding B-cell epitope. Similar analysis using recombinant gp21 protein, however, showed CCR2T and WCCR2T antisera to have similar affinity. Furthermore, in direct ELISA very low levels of reactivity were observed with both antisera (Fig. 6, lower panel). One possible explanation is that the gp21 protein used in the ELISA and competition studies was a chimera of MBP and gp21 (amino acids 338 -425), with the gp21 segment forming a very small part of the entire chimera. This could lead to steric hindrances in binding analysis when the protein is immobilized onto ELISA plates. Hence, the binding and titers we observed may not be an accurate representation of actual affinity for the native protein. This could also explain the results obtained in the competition experiments using the gp21 protein as inhibitor (Fig. 8C). However, in the whole virus ELISA experiments we tested the binding of CCR2T and WCCR2T antisera against a disrupted virus, which is more relevant. This is the first study showing that antibodies directed against the coiled coil region of the HTLV-1 gp21 envelope glycoprotein can lead to a reduction in cell fusion. In syncytia inhibition assays with crude sera, WCCR2T and CCR2T inhibition values were almost identical. HTLV-1-infected patient values for inhibition were somewhat higher than the immunized mice values. Although it should be noted that the difference in inhibition between patient 1 and patient 7 is almost as large as the difference between patient 1 and the immunized mice (Fig. 9). Indeed, the capacity of sera-containing antibodies against the entire proteins of HTLV-1 virus should have a greater ability to neutralize syncytia than antibodies directed against a single epitope. We also conducted syncytia inhibition assays with purified antibodies to eliminate the effects of extraneous proteins in our assay donated by the sera. In these assays we observed similar trends in the efficacy of WCCR2T and CCR2T antibodies because the nonspecific effects of preimmune antibodies were lost much more quickly than the reduction of inhibition observed in the immunized mice antibodies. However, when purified antibodies were used instead of crude sera, WCCR2T and CCR2T showed more disparity in inhibition because WCCR2T inhibited syncytia better that CCR2T. This may be related to the difference in the titers of WCCR2T-and CCR2T-purified antibodies (Fig. 10). Furthermore, the antibodies generated in this study retain their neutralizing activity at low concentrations (Fig. 10), suggesting that a large quantity of antibody is not necessary to induce the disruptive effects mediated by the coiled-coil peptide induced antisera. This effect is desirable, as a vaccine containing these constructs may require fewer boosts over the course of the life of an individual.
These new data confirm a previous report based on mutagenesis studies (27), showing the direct implication of this TM region in the cell to cell fusion process. It is worth noting that an antibody raised against the linear epitope of the same region does not form a helical trimeric coiled coil and does not inhibit the HTLV-1 Env-mediated cell fusion (29). These results demonstrate the validity of our template design, which allows the assembly of peptides able to adopt a trimeric coiled-coil conformation, thereby mimicking the structure of the native protein.
Previous studies show that antibodies directed against the coiled-coil region of other retroviruses such as human immunodeficiency virus and bovine leukemia virus can lead to virus neutralization (62,63).
In conclusion, we utilized a novel strategy to synthesize peptides that would structurally resemble the fusion-activated coiled-coil conformation of the gp21 subunit to elicit antibodies that would bind with high affinity to the native protein. Although the antisera was able to recognize native protein, further modifications aimed at stabilizing the structure such as the introduction of disulfide bridges to covalently link the different strands on the template can be attempted that may serve to increase helix-stabilizing interactions. Indeed, a recent report by Lu and Hodges (64) describes the use of a template strategy with a disulfide bridge to elicit antibodies that are specific to an ␣-helical structure. Such studies are not only important from a vaccine perspective to elicit neutralizing antibodies against transient structures of the viral fusion process but also form important tools to probe the structure-function relationship of the envelope glycoprotein of HTLV-1.
To further develop a human vaccine for HTLV-1, we have recently initiated and begun studies in a non-human primate model (Saimiri sciureus). In these ongoing studies, we have immunized monkeys with the gp21 coiled-coil peptides as well as in combination with Env gp46 peptides and Tax cytotoxic T lymphocytes epitopes. These animals will be challenged with an HTLV-1-transformed monkey cell line to determine the protective efficacy of these potential vaccine candidates. Such studies are important in the pursuit of a multivalent peptide vaccine for HTLV-1 as well as for other pathogens that employ a trimeric coiled coil for host cell infection.