HIV - 1 Virus Like Particles Produced by Stably Transfected 1 Drosophila S 2 cells - a Desirable Vaccine Component

22 The development of a successful vaccine against HIV-1 likely requires 23 immunogens that elicit both broadly neutralizing antibodies against envelope 24 spikes and T cell responses that recognize multiple viral proteins. HIV-1 virus like 25 particles (VLP), because they display authentic envelope spikes on the particle 26 surface, may be developed into such immunogens. However, in one way or the 27 other current systems for HIV-1 VLP production have many limitations. To 28 overcome these, in the present study we developed a novel strategy to produce 29 HIV-1 VLP using stably transfected Drosophila S2 cells. We co-transfected S2 30 cells with plasmids encoding HIV-1 envelope, gag and rev proteins and a 31 selection marker. After stably transfected S2 clones were established, HIV-1 VLP 32 and their immunogenicity in mice were carefully evaluated. Here, we report that 33 HIV-1 envelope proteins are properly cleaved, glycosylated and incorporated into 34 VLP with gag. The amount of VLP released into culture supernatants is 35 comparable to those produced by insect cells infected with recombinant 36 baculoviruses. Moreover, cryo-EM tomography revealed average 17 spikes per 37 purified VLP and antigenic epitopes on the spikes were recognized by broadly 38 neutralizing antibodies 2G12, b12, VRC01 and 4E10, but not by PG16. Finally, 39 mice primed with DNA and boosted with VLP in the presence of CpG exhibited 40 anti-envelope antibody responses including ELISA-binding, neutralizing, ADCC 41 and ADCVI as well as envelope and gag-specific CD8 T cell responses. Thus, we 42 conclude that HIV-1 VLP produced by the S2 expression system has many 43 desirable features to be developed into a vaccine component against HIV-1. 44 45 on A uust 0, 2017 by gest http/jvi.asm .rg/ D ow nladed fom Introduction 46 Developing a safe and effective vaccine to control HIV-1 pandemic is a major 47 global health priority. The encouraging results from a recent phase III study 48 (RV144) of a combination vaccine regimen conducted in Thailand have created 49 optimism that a preventive vaccine can be developed, although the efficacy of 50 that regimen was judged to be marginal, short-lived and not sufficient to be useful 51 at a population level (40). Thus, optimal vaccine may require a component that 52 elicits broadly neutralizing antibodies that are capable of binding to the envelope 53 spikes on the virion surface as well as memory T cells that recognize multiple T 54 cell epitopes on viral proteins (31). 55 HIV-1 virus like particles (VLP), because they display authentic envelope 56 spikes on the particle surface, may be developed into such a vaccine component 57 to elicit both neutralizing antibody and memory T cell responses (11, 57, 58). 58 Indeed, immunization of HIV-1 VLP has been shown to generate promising 59 immune responses in animals. For example, Hammonds et al. demonstrated that 60 in a guinea pig model the breadth of neutralizing antibody response elicited with 61 HIV-1 VLP produced by stably transfected 293T cells was enhanced as 62 compared with subunit protein of the same HIV-1 isolate (16). Buonaguro et al. 63 showed that systemic and mucosal cross-subtype neutralizing antibody 64 responses were elicited in mice with HIV-1 VLP produced by insect cells infected 65 with recombinant baculoviruses (RB) (5). McBurney et al. showed that HIV-1 VLP 66 produced by transfected COS cells elicited broader cell-mediated peripheral and 67 mucosal immune responses than polyvalent and monovalent envelope vaccines 68 on A uust 0, 2017 by gest http/jvi.asm .rg/ D ow nladed fom (30). However, in macaque challenge models definitive proof of protection has 69 not been clearly demonstrated. Immunization with SIV/HIV VLP elicited 70 anamnestic response to HIV-1 gp120, which correlated with accelerated 71 clearance of SHIV (34); immunization with single cycle SIV elicited broad SIV72 specific T cell responses and significantly reduced viral loads after intravenous 73 SIV challenge (22); repeated vaccination with VSV-G-pseudotyped SIV VLP 74 significantly reduced peak viremia after mucosal SIV challenge, but persistent 75 suppression of viral load was not achieved (25); vaccination with chemically 76 inactivated SIV particles elicited both SIV envelope-specific binding and 77 neutralizing antibody responses and significantly reduced viral loads after 78 intravenous homologous SIV challenge, but failed to resist subsequent 79 heterologous SIV challenge (26); whereas immune responses elicited by VLP 80 alone or by heterologous poxvirus-VLP prime-boost did not protect macaques 81 from SHIV or SIV challenge (33, 50). 82 Although HIV-1 VLP as immunogens has shown great promising, in one way 83 or the other the production of HIV-1 VLP by current systems have many 84 limitations. For example, yeast (42) or mammalian 293T (16), COS (30) and Vero 85 (36) cells transiently co-transfected with DNA plasmids encoding HIV-1 envelope 86 and gag proteins can produce enough HIV-1 VLP for small animal studies, but 87 not enough for large animals and humans. Because of this, attempts have been 88 made to establish stable mammalian cell transfectants for HIV-1 VLP production, 89 in which genes encoding HIV-1 gag and envelope were driven by a tetracycline90 inducible promoter (16, 17). Although this approach yields more HIV-1 VLP, the 91 on A uust 0, 2017 by gest http/jvi.asm .rg/ D ow nladed fom amount is still too low to be practical for large animals and humans. 92 Using insect cells infected with RB yields much higher amount of HIV-1 VLP 93 (37, 41). The HIV-1 VLP produced by infected insect cells elicits both humoral 94 and cellular immune responses (5). However, there are three major drawbacks in 95 using insect cells infected with RB for HIV-1 VLP production. First, culture 96 supernatants contain both HIV-1 VLP and RB. Routine purification procedures 97 such as gradient centrifugation or ultrafiltration can not strictly discriminate HIV-1 98 VLP from RB. As a result, significant amount of live baculovirus is present in the 99 HIV-1 VLP-containing fractions (17, 41). Contamination of HIV-1 VLP 100 preparations with live baculovirus is important with regard to immunogenicity and 101 regulatory issues. Because of this, as an alternative to the baculovirus system, 102 piggyback transposition that expressed HIV-1 gag protein was developed to 103 create transgenic insect cell lines for continuous HIV-1 gag VLP production (27). 104 Second, HIV-1 gp160 precursor envelope proteins are not cleaved in insect cells 105 infected with RB (17), although the effect of uncleaved gp160 on immunogenicity 106 has yet to be investigated. Third, since insect cells infected with RB are short107 lived, new infection is required each time when new batch of HIV-1 VLP is going 108 to be produced. Because of this, the amount and the quality of HIV-1 VLP could 109 vary among different batches. 110 Therefore, to overcome the limitations associated with current HIV-1 VLP 111 production systems, in the present study we developed a novel strategy to 112 produce HIV-1 VLP using stably transfected Drosophila melanogaster Schneider 113 2 (S2) cells. S2 cells have been widely used to produce ectropic proteins, 114 on A uust 0, 2017 by gest http/jvi.asm .rg/ D ow nladed fom including HIV-1 envelope proteins (8, 21). HIV-1 envelope precursor gp160 has 115 been shown to be properly cleaved in S2 cells during its biogenesis (4, 21). We 116 hypothesized that S2 cells, due to their high efficiency in expressing ectropic 117 proteins and proper cleavage of HIV-1 envelope precursor, can be used to 118 produce HIV-1 VLP. Furthermore, because the system uses only plasmids, HIV-1 119 VLP produced by this system will have no recombinant virus contamination. 120 Finally, stably transfected S2 cells can grow to high viable cell density, resulting in 121 high amount of HIV-1 VLP production in culture supernatants. Because of these, 122 stably transfected S2 cells may overcome many limitations and drawbacks in 123 HIV-1 VLP production by the current systems. 124 To test this hypothesis, we co-transfected S2 cells with plasmids encoding an 125 envelope glycoprotein (consensus B or C), a rev-independent gag (Pr55) protein 126 and a rev protein along with a pCoBlast selection marker. For the comparison, 127 we also co-transfected S2 cells with plasmids encoding a rev-independent gag 128 and a pCoBlast marker or with plasmids encoding gp120 (consensus B or C), a 129 rev protein and a pCoBlast marker. After the stable cell transfectants were 130 generated, limiting dilution assay was performed to establish stably transfected 131 S2 clones. After stably transfected clones were established, expression, process 132 and glycosylation of envelope proteins, production, morphology and antigenicity 133 of HIV-1 VLP, and the immunogenicity of HIV-1 VLP in mice were carefully 134 evaluated. 135 on A uust 0, 2017 by gest http/jvi.asm .rg/ D ow nladed fom Materials and Methods 136 Cell lines 137 Drosophila S2 cells were provided by Dr. Vincent Deubel at the Institut 138 Pasteur of Shanghai and cultured in complete Express FIVE SFM medium [i.e. 139 Express FIVE SFM medium supplemented with 10% FBS, 100 U/ml penicillin, 140 100 U/ml streptomycin and 100 mg/L L-glutamine] at 28°C without CO2. Every 3 141 or 4 days the cells were split at the density of 1 × 10 cells per ml. 142 The human embryonic kidney cell line 293T was purchased from Invitrogen 143 Life Technologies and maintained in complete DMEM medium [i.e. high glucose 144 DMEM supplemented with 10% FBS, 2 mM L-glutamine, 1 mM sodium pyruvate, 145 penicillin (100 U/ml), streptomycin (100 μg/ml)] plus G418 (500 μg/ml) (Invitrogen 146 Life Technologies). TZM-bl cells were obtained from the NIH AIDS Research and 147 Reference Reagent Program (ARRRP,Germantown, MD) (9, 38). Human CD4 148 T cell line CEMss-CCR5 was generated as described before (54). These cells 149 were maintained in complete DMEM. 150 Chronically HIV-1-infected CEMss-CCR5 cells were prepared by incubating 151 CEMss-CCR5 cells with a HIV-1 strain AD8 at one multiplicity of infection (MOI) 152 overnight. After the infection, cells were maintained in complete DMEM medium. 153 Periodically, HIV-1 replication and cell surface expression of HIV-1 envelope 154 protein were measured (see below). 155 Antibodies and poled HIV-1 patient sera 156 Human anti-gp120 or gp41 antibodies b12, 2G12, PG16 and 4E10 and 157 mouse anti-gag p24 antibody (183-H12-5C) were obtained from ARRRP. Pooled 158 on A uust 0, 2017 by gest http/jvi.asm .rg/ D ow nladed fom HIV-1 patient plasmas were provided by Dr. Ping Zhong at the Shanghai 159 Municipal Center for Disease Control and Prevention, Shanghai, China. Mouse 160 anti-gp120 antibody (cat. #4301) and sheep anti-gp120 C5 capture antibody (cat. 161 #D7324) were purchased from Advanced Bioscience Laboratory INC (Silver 162 Spring, MD) and Aalto BioReagents (Dublin, Ireland), respectively. Human anti163 gp120 antibody VRC01 and human anti-H5 HA antibody 100F4 were produced 164 and purified in our laboratory by stably transfected S2 clones as described before 165 (53). APC-conjugated anti-CD8, PerCP-conjugated anti-CD4, FICT-conjugated 166 anti-IFNγ, PE-Cy7-conjugated anti-TNFα, purified anti-CD49d and anti-CD28 167 antibodies were purchased from BD Bioscience (Mountain View, CA). 168 Gene constructs 169 cDNA sequence encoding HIV-1 rev-independent gag p55 (M1-10) (44) was 170 generated by recursive PCR (39) and cloned into a TA vector (Invitrogen) and 171 sequenced as described before (59). The correct insert was inserted into EcoR I 172 and Not I sites of a constitutive S2 expression vector pAc5.1/V5-His (Invitrogen). 173 The resulting construct was designated as pAC-Gag(Pr55) (Fig. 1A). 174 cDNA encoding consensus B and C HIV-1 gp160 envelope proteins without a 175 signal peptide were PCR-amplified using HIV-1 consensus B and C envelope 176 genes (23, 24) (generously provided by Dr. B. H. Hahn at the University of 177 Alabama) as templates and inserted into the EcoR I and Xho I sites of an 178 inducible S2 expression vector pMT/BiP/V5-His (Invitrogen). The resulting 179 constructs were designated as pMT/BiP-gp160 (consensus B and C), 180 respectively (Fig. 1A). 181 on A uust 0, 2017 by gest http/jvi.asm .rg/ D ow nladed fom cDNA sequence encoding HIV-1 rev was amplified using a pZeoSV/rev as a 182 template and inserted into the EcoR I and Not I sites of an inducible expression 183 vector pMT/V5-His (Invitrogen). The resulting construct was designated as pMT184 Rev (Fig. 1A). 185 cDNA encoding consensus B and C HIV-1 gp120 envelope proteins without a 186 signal peptide were inserted into the BamH I and Xho I sites of an inducible S2 187 expression vector pMT/BiP/V5-His. The resulting constructs were designated as 188 pMT/BiP-gp120/V5-His (consensus B and C), respectively. 189 For the mammalian expression, cDNA encoding consensus B and C HIV-1 190 gp120 envelope proteins with a signal peptide or HIV-1 gag (Pr55) were inserted 191 into mammalian expression vector pCMV/R (3). The resulting plasmids were 192 designated as CMV/R-gp120 (consensus B and C) and CMV/R-Gag (Pr55), 193 respectively. Plasmid was produced at the concentration of 1 mg/ml in sterile 194 water and found to be predominantly supercoiled by gel electrophoresis with 195 endotoxin levels below 1.5 EU/mg DNA. 196 Generation of stable Drosophila S2 cell clones 197 To generate stable S2 transfectants to produce HIV-1 VLP, S2 cells were co198 transfected with 9.5 μg pAc-Gag(Pr55), 4.75 μg pMT/Bip-gp160 (consensus B or 199 C), 4.75 μg pMT-Rev along with 1 μg selection vector pCoBlast which contains 200 blasticidin resistance gene (Invitrogen) using a calcium phosphate precipitation 201 method (47). After overnight incubation, cells were washed once with PBS and 202 cultured in complete Express FIVE SFM medium. Seventy two h later blasticidin 203 (Invitrogen) at the final concentration 25 μg/ml was added and stable S2 204 on A uust 0, 2017 by gest http/jvi.asm .rg/ D ow nladed fom transfectants were generated in 2 to 3 weeks. 205 To generate stable S2 clones, limiting dilution assay was performed as 206 described before (54). Briefly, 1.5 ×10 S2 transfectants were mixed with 1×10 207 parental S2 cells (feeder layer cells) in 10ml Drosophila Schneider 2 medium 208 supplemented with 10% FBS, 100 U/ml penicillin, 100 U/ml streptomycin and 100 209 mg/L L-glutamine. 100 μl cell mixture was added into each well in 96-well plates 210 and incubated at 28°C for 2 days. Blasticidin at the final concentration 25 μg/ml 211 was then added into the culture medium. After 3 weeks, single clones were 212 isolated for further expansion. The stably transfected S2 clones were induced by 213 5 μM of CdCl2 and the expression of HIV-1 envelope and gag proteins in culture 214 supernatants were analyzed by western blot analysis (see below). The S2 clones 215 that produced the highest HIV-1 VLP (consensus B and C) were selected for HIV216 1 VLP production (see below). 217 To generate stably transfected S2 cells expressing soluble gp120, S2 cells 218 were co-transfected with 9.5 μg pMT/BiP-gp120/V5-His (consensus B or C), 9.5 219 μg pMT-Rev along with 1 μg selection vector pCoBlast and selected with 220 blasticidin as described above. After stable transfectants were established, 221 limiting dilution assay was performed as described above. The S2 clones that 222 produced the highest gp120 were selected for gp120 production (see below). 223 Production of HIV-1 VLP and soluble gp120 by S2 clones 224 Wave bioreactor 20/50EHT systems with a WAVEPOD Process Control Unit 225 (GE Healthcare) were used to grow S2 clones in fed batch culture for HIV-1 VLP 226 and soluble gp120 production as described before (53). Briefly, 6 x 10 cells of 227 on A uust 0, 2017 by gest http/jvi.asm .rg/ D ow nladed fom the S2 clones in 300 ml of complete Express FIVE SFM medium were 228 inoculated into a 2-L cellbag and cultured at 28°C without CO2. At 5, 7 and 8 229 days after the initial culture in the Wave bioreactor 300, 200 and 200 ml of fresh 230 complete Express FIVE SFM medium were added, respectively. At 8 days after 231 the initial culture CdCl2 at the final concentration of 5 μM was added into culture 232 medium to induce the expression of HIV-1 envelope proteins. Three days after 233 the induction, culture supernatants were harvested, clarified by centrifugation at 234 6,000 × g for 30 min at 4°C and filtered through a 0.45 μm filter. Filtered 235 supernatants were concentrated 5-fold using QuixStand Benchtop System 236 incorporated with a 50,000 NMWC Hollow Fiber Cartridge (Model UFP-50-C237 4MA). HIV-1 VLP in concentrated supernatant was pelleted by ultracentrifuge 238 through 20% sucrose cushion, resuspended in PBS, and stored in aliquots at 239 80C until further use (see below). During 11 day’s fed batch culture, small 240 aliquot of cell suspension was collected every 24 h. The number and viability of 241 cells were counted by trypan blue exclusion assay. The amount of HIV-1 VLP, as 242 measured by HIV-1 gp120 and gag p55, was determined (see below). Soluble 243 gp120 in concentrated supernatants were purified by Ni-NTA column according 244 to manufacturer’s instruction (Invitrogen). The amount and the purity of proteins 245 was quantified by BCA assay (Pierce) and determined by 10% SDS/PAGE 246 followed by Coomassie blue staining. 247 Western blot analysis 248 To test HIV-1 envelope and gag protein expression, S2 clones were induced 249 with 5μM of CdCl2 for 3 days. Culture supernatants were harvested, precipitated 250 on A uust 0, 2017 by gest http/jvi.asm .rg/ D ow nladed fom by TCA and dissolved in PBS. Samples were separated through 10% SDS/PAGE 251 and transferred onto PVDF membranes (Millipore). The membranes were 252 blocked in a solution of Tris-buffered saline containing 5% nonfat dry milk and 253 0.05% Tween 20 and subsequently probed with an anti-gag p24 antibody and 254 with indicated antibodies specific for HIV-1 gp120 and gp41. Antigens were 255 visualized with an AP-conjugated anti-mouse IgG antibody according to 256 manufacturer’s instruction (Promega). 257 To test the incorporation of HIV-1 envelope and gag proteins into VLP, the 258 VLP-containing supernatants were harvested, clarified by centrifugation at 6,000 259 × g for 30 min at 4°C and filtered through a 0.45 μm filter, loaded onto 20% 260 sucrose cushion and ultra-centrifuged at 20,000 rpm for 2.5 h at 4°C in a SW28 261 rotor (Beckman Coulter, Fullerton, CA). The pellets were resuspended and 262 fractionated through a 25 – 65 % sucrose density gradient at 25,000 rpm for 16 h 263 at 4°C in a SW41 rotor (Beckman Coulter). Twelve fractions (1 ml each) were 264 collected from top to bottom of the gradient and the density of each fraction was 265 measured. The samples were then TCA precipitated, separated by 10% SDS266 PAGE and transferred onto PVDF membranes. The presence of HIV-1 gag and 267 envelope proteins in unfractionated and fractionated samples were detected as 268 described above. 269 To test mammalian expression of plasmids CMV/R-gp120 (Consensus B and 270 C) and CMV/R-Gag(Pr55), 293 T cells were transiently transfected with plasmids 271 encoding soluble consensus B and C HIV-1 gp120 envelope proteins or 272 gag(Pr55). Two days after transfection, culture supernatants were harvested, 273 on A uust 0, 2017 by gest http/jvi.asm .rg/ D ow nladed fom pelleted by ultra-centrifugation, separated by 10% SDS-PAGE and transferred 274 onto PVDF membranes. The presence of HIV-1 gp120 or gag(Pr55) proteins was 275 detected by indicated antibodies against gp120 or gag p24. 276 To quantify the amount of gag p55 in supernatants produced by S2 clones in 277 fed batch culture, western blot and densitometry analyses were performed as 278 reported by Lynch et al. (27). Briefly, culture supernatants were collected daily 279 during 11 day’s fed batch culture (see above) and loaded onto 4-12% Bis-Tis gel 280 (Invitrogen) along with a serially 2-fold diluted p24 standards (starting at 40 ng) 281 (Aalto BioReagents) and transferred onto PVDF membranes. The membranes 282 were blocked in a solution of TBST containing 5% nonfat dry milk and 283 subsequently probed with an anti-gag p24 antibody. Antigens were visualized 284 with an HRP-conjugated anti-mouse IgG antibody (MultiSciences) and EZ-ECL 285 substrate (Thermo). Gag concentrations were estimated by comparing calculated 286 densities of the p55 bands in experimental samples and p24 standards using 287 Quantity One software (Bio-Rad). Consistent with what was previously reported 288 (27), we also found that HIV-1 VLP concentrations determined by calculating 289 densities of the p55 band on western blots is a more reliable indicator of actual 290 VLP concentration in the sample than using p24 ELISA quantification (Song et al. 291 data not shown). 292 ELISA 293 To estimate the amount of HIV-1 envelope incorporation into HIV-1 VLP, 96294 well EIA/RIA flat bottom plate (Costar) were coated overnight with 1μg/ml anti295 HIV-1 gp120 C5 capture antibody. The plates were blocked with PBS containing 296 on A uust 0, 2017 by gest http/jvi.asm .rg/ D ow nladed fom 5% BSA for 1 h at 37 °C. The VLP containing culture supernatants, above 297 concentrated VLP-containing samples, or a serial diluted purified gp120 (serve 298 as standards) in assay diluent (10% BSA, 0.5% Triton X-100 in PBS) were added 299 onto anti-gp120 antibody-coated 96 well plates and incubated at 37°C for 2 h. 300 The plates were then washed 5 times with PBST buffer (0.05% Tween 20 in 301 PBS). Anti-gp120 antibody (Cat. #4301) at 1:2,000 dilution was then added and 302 incubated for another 1 h. Horseradish peroxidase (HRP)-conjugated goat anti303 mouse IgG (Chemicon) at 1:5,000 dilution was added. Colorimetric analysis was 304 performed using the TMB Substrate Kit (Pierce) and absorbance was read at 450 305 nm by a spectrophotometer (BioTek Instruments, Winooski, VT, USA). 306 To determine the antigenicity of HIV-1 VLP, a similar sandwich ELISA was 307 performed with 10-fold diluted human anti-gp120 and gp41 antibodies (b12, 308 VRC01, 2G12, PG16 and 4E10) and control antibody (100F4) followed by HRP309 conjugated goat anti-human Ig(H+L) (IgM+IgG+IgA, SBA) at 1:2,000 dilution as a 310 secondary antibody. Colorimetric analysis was performed as described above. 311 To test total IgG, IgG1 and IgG2a antibody responses against consensus B 312 and consensus C gp120, 96-well EIA/RIA flat bottom plates were coated with 100 313 ng/well of the respective consensus B or consensus C gp120 proteins at 4°C 314 overnight. Plates were blocked with PBS containing 5% BSA at 37°C for 1 h and 315 incubated with serial diluted mouse serum samples or pooled human plasmas at 316 37°C for 2 h. Plates were washed 5 times with PBST buffer and incubated with 317 HRP-conjugated anti-mouse IgG (1:5,000) or anti-human IgG (1:2,000), anti318 mouse IgG1 (1:500) or anti-mouse IgG2a (1:500) antibodies, respectively, at 319 on A uust 0, 2017 by gest http/jvi.asm .rg/ D ow nladed fom 37°C for 30 min. The wells were washed 5 times with PBST buffer. Colorimetric 320 analysis was performed as described above. End-point titers were determined as 321 the last reciprocal serial serum or plasma dilution at which the absorption at 450 322 nm was greater than the mean plus 2 SD of the background signals generated by 323 pre-immune serum samples. 324 Characterization of glycosylation of HIV-1 envelope proteins produced by 325 S2 clones 326 The glycosylation pattern of HIV-1 gp120 produced by stably transfected S2 327 cells was measured using DIG Glycan Differentiation Kit according to the 328 manufacturer’s instruction (cat. #11210238001, Roche). Briefly, purified HIV-1 329 gp120 produced by stably transfected S2 cells along with internal controls 330 supplied with the Kit were separated on 10% SDS/PAGE and transferred onto 331 PVDF membranes. The membranes were blocked in 20 ml blocking solution 332 supplied by the Kit for 30 min. After washed twice with TBS, membranes were 333 incubated with digoxigenin-labeled lectins: GNA (galanthus nivalis agglutinin), 334 SNA (sambucus nigra agglutinin), DSA (datura stramonium agglutinin), MAA 335 (maackia amurensis agglutinin) and PNA (peanut agglutinin), respectively, for 1 h. 336 After washed twice with TBS glycosylation pattern was visualized with an AP337 conjugated anti-digoxigenin antibody according to manufacturer’s instruction. 338 To deglycosylate envelop protein produced by S2 cells, 1μg of soluble 339 gp120, HIV-1 VLPor wild type HIV-1-containing culture supernatants equivalent 340 to 1μg of total proteins were denatured by heating at 100°C for 10 min in the 341 presence of 0.5% SDS and 40 mM DTT. After cooling, 1μl PNGase F or Endo-H 342 on A uust 0, 2017 by gest http/jvi.asm .rg/ D ow nladed fom (New England Biolabs, Ipswich, MA) was added and the mixture was incubated 343 in 1 × G7 or G5 reaction buffers (New England Biolabs) at 37°C for 1 h. 344 Subsequently, samples were subjected to protein gel electrophoresis and 345 Western blot analyses detected by anti-HIV-1 gp120 antibody (#4301) as 346 described above. 347 Cryo-EM tomography 348 To purify HIV-1 VLP for cryo-EM study, culture supernatants produced by 349 stably transfected S2 clones were harvested and clarified by low speed 350 centrifuge at 6,000 × g for 30 min at 4°C, passed through the 0.45 μm filter 351 (FISHER), loaded on to 20% sucrose cushion and ultra-centrifuged at 25,000 352 rpm for 2 h using SW28 rotor. The pellets were resuspended with PBS and 353 loaded onto 25%-65% linear sucrose gradient and ultra-centrifuged at 25,000 354 rpm for 16 h (SW41 rotor). VLP-containing fractions were pooled and pelleted by 355 ultracentrifugation at 25,000 rpm for 2 h (SW41 rotor). Pellets were resuspended 356 with PBS and loaded onto 30% and 45% non-linear sucrose gradient and ultra357 centrifuged at 110,000 × g for 3 h (MLS-50 rotor). Two fuzzy bands, one close to 358 the top of sucrose gradient (upper band) and the other at the sucrose interface 359 (lower band), were collected, diluted in PBS and passed through 0.2 μm low 360 protein binding and non-pyrogenic syringe filter (cat. #PN4612, PALL). The 361 samples were then pelleted by ultracentrifuge at 110,000 × g for 2 h (MLS-50 362 rotor), resuspended in 20 μl of PBS and stored at -80°C. 363 To observe VLP by cryo-EM tomography, 3.5 μl of upper and lower band 364 samples were loaded on a 300 mesh R2/1 Quantifoil holey grid (Quantifoil Micro 365 on A uust 0, 2017 by gest http/jvi.asm .rg/ D ow nladed fom Tools, GmbH, Jena, Germany) for 1 minute, blotted for 4s with a filter paper, and 366 rapidly vitrified into liquid ethane using an FEI vitrobot Mark IV. The grid was 367 transferred to a Gatan 626 cryoholder (Gatan, Pleasanton, CA) and examined 368 under low dose conditions on an FEI Tecnai 20 electron microscope operated at 369 200 kV. Micrographs were recorded on a Gatan Ultrascan 1,000 charge-coupled 370 device camera in an FEI Tecnai F20 electron microscope operated at 200 kV 371 with a magnification of 38,000 × and a defocus of ~2.5 μm. Cryo-EM tomography 372 was performed at an FEI 300 kV Titan Krios cryo-EM equipped with a Gatan 373 UltraScan4000 (model 895) 4K x 4K pixels CCD. Single axis Tilt series were 374 collected at 47,000x magnification using 2K x 2K CCD mode around a single axis 375 at a 2° increment between -62° and +60° at a pixel size of 0.38 nm. The defocus 376 was set at -3 to -6 μm and the cumulative dose was about 72 e/Å. 377 Immunization and sampling 378 All animal protocols were approved by the Institutional Animal Care and Use 379 Committee at the Institut Pasteur of Shanghai (Approval No. A2011014). Female 380 BALB/c mice (Mus musculus) at the age of 6 to 8 weeks were purchased from 381 Shanghai Institutes of Biological Sciences Animal Center and housed in micro382 isolator cages ventilated under negative pressure with HEPA-filtered air and a 383 12/12-h light dark cycle. For the immunization, mice were randomly divided into 2 384 groups. Group one was intramuscularly (i.m.) injected of both hind legs with total 385 200 μl PBS (pH 7.4). Group two was i.m. primed twice with 150 μg of three 386 plasmids encoding consensus B and C HIV-1 gp120 proteins and rev387 independent HIV-1 gag (50μg each) and then subcutaneously (s.c.) boosted 388 on A uust 0, 2017 by gest http/jvi.asm .rg/ D ow nladed fom twice with a mixture of HIV-1 VLP (consensus B and C) equivalent to 5 μg gp120 389 plus 5 μg phosphorothioate CpG oligodeoxynucleotides (CpG-ODN 1826 5′-TCC 390 ATG ACG TTC CTG ACG TT-3′) (18). HIV-1 VLP used for boost contained a 391 mixture of Gag-alone particles vs. Env-Gag particles, which was likely in a 1 to 1 392 ratio, because the optical density of two fussy bands (upper band containing 393 Gag-alone particles vs. lower band containing Env-Gag particles) collected after 394 the ultracentrifugation in 30% and 45% non-linear sucrose gradient were vary 395 similar (see above). The DNA prime was done on day 0 and 28 and the HIV-1 396 VLP/CpG boost was done on day 56 and 84. Seven days before prime and 10 397 days after boost, blood samples were collected by retro-orbital plexus puncture. 398 Serum samples were collected, heat inactivated at 56°C and stored in aliquots at 399 -20°C. Ten days after the boost spleens were harvested and splenocytes were 400 isolated by Ficoll-Hiquac gradient centrifugation, depleted of erythrocytes by 401 treatment with NH4Cl (0.1 M, pH7.4). Cells were then washed with PBS and 402 resuspended in complete RPMI 1640 medium [i.e. RPMI 1640 supplemented 403 with 10% inactivated FBS, 100 U/ml penicillin, 100 U/ml streptomycin and 100 404 mg/L L-glutamine] for intracellular cytokine staining (see below). 405 HIV-1 and 10A1 pseudotypes and a single-cycle infectivity assay 406 HIV-1 and 10A1 pseudotypes were generated and a single-cycle infectivity 407 assay was performed as described before (32, 54). 408 FACS analysis 409 To study cell surface expression of HIV-1 envelope, chronically HIV-1 410 infected CEMss-CCR5 cells (see above) were incubated with the post-immune 411 on A uust 0, 2017 by gest http/jvi.asm .rg/ D ow nladed fom sera, sera from PBS control mice or pooled HIV-1 infected patient sera for 45 min 412 on ice. Cells were then washed twice with FACS buffer (PBS containing 1% BSA 413 and 0.02% NaN3) and stained with PE-conjugated anti-mouse or anti-human IgG 414 antibody (Sigma) for another 45 min on ice. Cells were then washed twice with 415 FACS buffer and fixed with 1% formaldehyde in 0.5 ml of FACS buffer. FACS 416 analysis was performed on a BD LSRII flow cytometer (BD Biosciences). 417 ADCC assay 418 The rapid fluorometric ADCC assay was performed as previously described 419 (14, 45). Briefly, 5,000 HIV-1-infected CEMss-CCR5 target cells were doubly 420 labeled with 5 μM PKH-26 (Sigma-Aldrich) and 0.5 μM CFSE (Molecular Probes). 421 Labeled target cells were resuspended in RPMI 1640 medium containing 10% 422 FBS and incubated with preand post-immune sera from PBS control and 423 immunized mice, naïve mice sera (negative control) or pooled HIV-1 patient sera 424 (positive control) at the final dilution of 1:50 in a 96-well micro-plate for 30 min at 425 room temperature. Splenocytes (effector cells) from naïve mice were added at a 426 50:1 effector/target cell (E:T) ratio. The plates were centrifuged for 5 min at 400 × 427 g to promote cell-to-cell interactions and then incubated for 4 h at 37 °C in 5% 428 CO2. Cells were washed twice with PBS, fixed in 3.7% paraformaldehyde–PBS 429 (v/v) for flow cytometry. Fifty thousand non-gated events in duplicate wells were 430 acquired within 18 h by using BD LSRII flow cytometer. Data were analyzed using 431 FlowJo (Tree Star Inc., USA). The percent ADCC killing was determined by back432 gating on the PKH-26 population of target cells that lost the CFSE viability dye 433 and substracted the nonspecific effect by pre-immune sera. Non-stained and 434 on A uust 0, 2017 by gest http/jvi.asm .rg/ D ow nladed fom single-stained targets were included in every experiment to compensate for 435 single-stained CFSE and PKH-26 emission. 436 ADCVI assay 437 Target and effector cells used for the ADCVI assay were the same HIV-1438 infected CEMss-CCR5 cells and naïve splenocytes used for ADCC (see above). 439 Target cells (5,000) were washed to remove cell-free virus. Effector cells (naïve 440 splenocytes) were added to target cells at an E:T ratio of 20:1. Preand post441 immune sera from PBS control and immunized mice, naïve mice sera (negative 442 control) or pooled HIV-1 infected patient sera (positive control) were added to 443 target and effector cells to achieve a final dilution of 1:50. Control wells lacking 444 sera but containing effector cells and viral replication control wells lacking both 445 sera and effector cells were included. Two days later, supernatants were 446 collected, and gag p24 was measured by ELISA (Zeptometrix). The percent 447 inhibition due to ADCVI was calculated relative to pooled pre-immune mice sera 448 as follows: percent inhibition = 100[1 – (p24post)/(p24pre)], where (p24post) and 449 (p24pre) are concentrations of p24 in supernatant fluid from wells containing a 450 source of postor pre-immune sera, respectively. Individual serum samples were 451 assayed in triplicate and in two separate assays. In general, values from the two 452 independent assays were in close agreement with each other. 453 Intracellular cytokine staining 454 2×10 splenocytes from immunized and control mice were seeded into 455 complete RPMI 1640 and stimulated with two mixtures of peptides (2μg/ml each 456 peptide) as previously described (30), one mixture containing peptides derived 457 on A uust 0, 2017 by gest http/jvi.asm .rg/ D ow nladed fom from HIV-1 envelope protein and the other from gag protein, as well as 5 μg of 458 anti-mouse CD28 and 5μg of anti-mouse CD49d for 6 h at 37°C and 5% CO2. 459 Golgi plug (BD Pharmingen) was added during the final 4 h of incubation. Cells 460 were then stained with anti-mouse CD16/32 (Fc block) antibody followed by 461 surface staining with PerCP-conjugated anti-CD4 and APC-conjugated anti-CD8 462 antibodies. Cells were then fixed, permeabilized with cytofix cytoperm (BD 463 Pharmingen) and stained with FITC-conjugated anti-IFNγ and PE-Cy7464 conjugated anti-TNFα. 1 × 10 cells per sample were acquired on a BD LSRII 465 flow cytometer and FACS data were analyzed using FlowJo software. 466 Statistics 467 Analyses were performed with GraphPad Prism v 5.0. Unpaired t test was 468 used to compare two data sets. Correlation between the ADCC and ADCVI was 469 assessed by the exact Spearman rank correlation method. All P values are two470 tailed. 471 472 on A uust 0, 2017 by gest http/jvi.asm .rg/ D ow nladed fom Results 473 HIV-1 envelope and gag proteins were properly produced, processed and 474 incorporated into VLP 475 To generate stably transfected S2 clones that produce high amount of HIV-1 476 VLP, we co-transfected Drosophila S2 cells with plasmids pAC-Gag(Pr55), 477 pMT/BiP-gp160 (consensus B or C), pMT-Rev and pCoBlast, in which genes 478 encoding HIV-1 envelope and rev proteins were driven by an inducible pMT 479 promoter; while gene encoding HIV-1 gag(Pr55) was driven by a constitutive pAC 480 promoter (Fig. 1A). After the stable transfectants were established, limiting 481 dilution assay was performed to generate stably transfected S2 clones. Fig. 1B 482 shows that in supernatants from S2 clone VB2, but not from parental S2 cells, 483 HIV-1 gag p55 was readily detected with or without induction; whereas without 484 induction a faint gp120 was detected, while after induction the expression of 485 gp120 increased dramatically. Fig. 1C shows that after induction both gp41 and 486 gp160 of HIV-1 envelope proteins were detected by antibody 4E10, but only 487 gp120 was detected by antibody VRC01, indicating that although precursor HIV-1 488 envelope gp160 were properly cleaved into gp120 and gp41, because the high 489 level of expression of HIV-1 gp160 in stable S2 transfectants, only over 60% of 490 gp160 was cleaved by cellular endopeptidase(s) during their biogenesis through 491 Golgi apparatus. Similar studies on the induction of envelope protein and 492 cleavage of gp160 into gp120 and gp41 were also observed in S2 clone VC1 493 (that expressed consensus C HIV-1 envelope protein) (Yang et al. data not 494 shown). 495 on A uust 0, 2017 by gest http/jvi.asm .rg/ D ow nladed fom To test the incorporation of envelope and gag proteins into HIV-1 VLP, S2 496 clone VB2 was induced with CdCl2 and supernatants were harvested, 497 concentrated, and then fractionated through a 25-65% sucrose gradient. The 498 envelope and gag proteins in each fraction were detected with a mixture of two 499 mouse antibodies against gp120 and gag p24. As a control, supernatants from 500 S2 cells stably transfected with pAC-Gag(Pr55) alone were also harvested, 501 concentrated and fractionated through the same sucrose gradient. As shown in 502 Fig. 1D gag proteins derived from S2 cells transfected with pAC-Gag(Pr55) alone 503 were peaked in the fractions with buoyant density between 1.10 and 1.14; 504 whereas both gp160 and gp120 and gag proteins derived from S2 clone co505 transfected with pAC-Gag(Pr55) and pMT/Bip-gp160 were peaked in the 506 fractions with buoyant density between 1.15 and 1.19 (Fig. 1E), indicating likely 507 both uncleaved gp160 and cleaved gp120/gp41 were incorporated into the VLP 508 and the difference in 160 kD gp160 band shown in Fig. 1E, but not in Fig. 1C was 509 because two different anti-HIV-1 gp120 antibodies (VRC01 and #4301) were 510 used to detect HIV-1 glycoproteins. In western blot antibody VRC01 only reacts 511 with gp120; whereas antibody #4301 reacts with both gp120 and gp160. Similar 512 study was performed in the stable S2 clone VC1 with similar results (Yang et al. 513 data not shown). Thus, taken together, these results clearly indicate that in the 514 S2 clones HIV-1 envelope proteins are properly induced, cleaved and both 515 envelope and gag are incorporated into HIV-1 VLP. 516 Glycosylation of envelope proteins produced by S2 clones 517 To test glycosylation of envelope proteins on HIV-1 VLP produced by S2 518 on A uust 0, 2017 by gest http/jvi.asm .rg/ D ow nladed fom clone VB2, we first performed deglycosylation assays with N-endoglycosidase 519 PNGase F and Endo H. For a comparison, glycosylation of soluble gp120 520 produced by S2 clone B5 (that expressed consensus B gp120) and wild type 521 HIV-1 (Bru-3) were also tested. PNGase F removes all types of N-linked 522 oligosaccharides from glycoproteins, whereas Endo H cleaves the chitobiose 523 core of high-mannose and hybrid oligosaccharides from N-linked glycoproteins. 524 Fig. 2A shows that without PNGase F or Endo H treatment, 120 KD band was 525 detected by anti-gp120 antibody in soluble gp120 as well as in HIV-1 VLP; 526 whereas with PNGase F treatment, gp120 bands of soluble gp120 and HIV-1 527 VLP were reduced to 70 KD band; whereas with Endo H treatment gp120 bands 528 of soluble gp120 and HIV-1 VLP were reduced to about 74 KD and 85 KD bands, 529 respectively. Similar, but not identical, deglycosylation pattern was seen in 530 envelope proteins of wild-type HIV-1 virions with PNGase F or Endo H treatment 531 (Fig. 2A). Thus, these results not only revealed that soluble gp120 and envelope 532 protein on HIV-1 VLP produced by S2 clones are N-link-glycosylated, but also 533 highlighted some difference in N-linked glycosylation between soluble gp120 and 534 envelope proteins on HIV-1 VLP even though these envelope glycoproteins were 535 both produced by clones derived from the same S2 cells. 536 To further test glycosylation patterns, soluble gp120 produced by S2 clone 537 B5 were probed with lectins GNA, SNA, MAA, PNA and DSA that recognize 538 different sugar moieties (see Materials and Methods). As shown in Fig. 2B large 539 amount of gp120 produced by the S2 clone was detected by GNA and small 540 amount was detected by PNA and DSA, but not by SNA and MAA, indicating that 541 on A uust 0, 2017 by gest http/jvi.asm .rg/ D ow nladed fom likely, on a given HIV-1 envelope glycoprotein produced by S2 clones there are 542 more high-mannose glycans and less hybrid and complex forms. 543 Quantification of HIV-1 VLP produced by S2 clones 544 To test the cell growth and HIV-1 VLP production by S2 clones, S2 clones 545 were grown in fed batch culture using the Wave bioreactor for a total of 11 days. 546 Eight days after the initial fed batch culture CdCl2 was added to induce HIV-1 547 envelope expression (see Materials and Methods). Fig. 3A shows the growth of a 548 representative S2 clone VB2 in fed batch culture. Starting with 300 ml of seed cell 549 suspension, 300, 200 and 200 ml of fresh culture medium were added on day 5, 550 7 and 8, respectively, and cell density grew from 2 x 10 cells per ml on day 0 to 551 2.2 x 10 cells per ml on day 11. By so doing, the total viable cell numbers grew 552 about 37 folds from 6 x 10 cells on day 0 to 2.2 x 10 cells on day 11 (Fig. 3B). 553 Fig. 3C shows that while the amount of gag p55 in the supernatants grew steadily 554 and reached 9.5 μg per ml on day 11; the amount of gp120 in the supernatants 555 before the CdCl2 induction were very low, but after the induction the amount of 556 gp120 increased dramatically and reached 4.2 μg per ml on day 9 and 7.5 μg per 557 ml on day 11. Similar cell growth and HIV-1 VLP production were also observed 558 in another S2 clone VC1 (Song et al. data not shown). 559 Morphology and spike number on HIV-1 VLP produced by S2 clones 560 To assess the morphology of the HIV-1 VLP produced by S2 clones, HIV-1 561 VLP in culture supernatants were concentrated and purified (see Materials and 562 Methods) and the morphology and spikes on the surface of purified HIV-1 VLP 563 were analyzed by cryo-EM and electron tomography. Fig. 4A and 4B show the 564 on A uust 0, 2017 by gest http/jvi.asm .rg/ D ow nladed fom particles obtained from upper and lower bands of representative HIV-1 VLP 565 samples derived from S2 clone VB2 (see Materials and Methods for the detail). 566 Particles obtained from both upper and lower bands were intact and round. The 567 diameters of particles ranged from 96 nm to 185 nm with an average diameter of 568 125.7 ±23.2 nm (Table 1), which is slightly larger but comparable with those of 569 native SIV and HIV-1 virions we measured in the similar cryoEM tomograms (109 570 ± 14 and 110 ± 8 nm for SIV and HIV-1 virus respectively) (62). It is also 571 consistent with the measurements of SIV and HIV-1 virions in the negative stain 572 electron tomograms (61). At this time, we do not really know what accounts for 573 the heterogeneity in the size of the VLPs. But it may be due to the lack of HIV-1 574 protease in VLPs to process immature p55 Gag. Interestingly, particles collected 575 from the upper band did not exhibit any spikes on the HIV-1 VLP surface (Fig. 576 4A); whereas on the surface of particles collected from lower band, envelope 577 spikes were readily observed (Fig. 4B). Electron tomography with selected 12 578 particles collected from the lower band showed average 17 ± 2 spikes per 579 particle (ranging from 13 to 20 spikes) (Fig. 4C and Table 1). Furthermore, like 580 those observed on the surface of SIV or HIV-1 virions (62), no discrete periodic 581 distances between spikes on the surface of HIV-1 VLP were observed (Fig. 4D). 582 Antigenicity of HIV-1 envelope glycoproteins produced by S2 clones 583 To test antigenicity of HIV-1 envelope glycoproteins produced by stably 584 transfected S2 clones, we performed three sets of experiments with a panel of 585 broadly neutralizing antibodies 2G12, b12, VRC01, PG16 and 4E10 (6, 7, 52, 55, 586 60, 63)as well as control antibody 100F4 antibody (53). The antigenicity results of 587 on A uust 0, 2017 by gest http/jvi.asm .rg/ D ow nladed fom HIV-1 envelope glycoproteins were summarized in Table 2. 588 In the first experiment we compared antigenicity of HIV-1 envelope 589 glycoproteins on cell surface between human kidney cell line 293 T transiently 590 transfected with CMV/R-gp160 (consensus B) and S2 clone VB2 that expressed 591 the same gp160 (consensus B). Supplementary Fig. 1 shows that on the surface 592 of transduced 293 T cells HIV-1 envelope glycoproteins were recognized by all 593 broadly neutralizing antibodies 2G12, b12, VRC01, PG16 and 4E10, but not by 594 control antibody 100F4; whereas on the surface of S2 clone HIV-1 envelope 595 glycoproteins were recognized by antibodies 2G12, b12, VRC01and 4E10, but 596 not by PG16 and control antibody 100F4. 597 In the second experiment, we measured antigenicity of HIV-1 VLP by 598 western blot analysis. Supplementary Fig. 2 shows that HIV-1 VLP was only 599 recognized by broadly neutralizing antibodies VRC01, 2G12 and 4E10, but not by 600 b12, PG16 and control antibody 100F4. 601 In the third experiment, we measured antigenicity of HIV-1 VLP by sandwich 602 ELISA using serially diluted antibodies (see Materials and Methods). 603 Supplementary Fig. 3 shows that HIV-1 VLP was only bound by broadly 604 neutralizing antibodies 2G12, b12, VRC01 in a dose-dependent manner, but not 605 by PG16 and 4E10 nor by control antibody 100F4. Thus, taken together we 606 conclude that except for antigenic epitope PG16, all other broadly neutralizing 607 antigenic epitopes 2G12, b12, VRC01 and 4E10 tested are preserved on spikes 608 of HIV-1 VLP produced by S2 clones. 609 ELISA-binding antibody responses elicited by DNA-VLP prime-boost 610 on A uust 0, 2017 by gest http/jvi.asm .rg/ D ow nladed fom In our initial studies on immunogenicity of HIV-1 VLP, we found that HIV-1 611 VLP produced by S2 clones was immunogenic even without a supplement of any 612 adjuvant. However, priming with DNA and boosting with HIV-1 VLP enhanced 613 immunoginicity (Yang et al. data not shown), which was similar to what was 614 reported by Buonaguro et al.(5). In addition, we demonstrated that heterologous 615 DNA-VLP prime-boost elicited superior neutralizing antibody response and 616 protection in mice against HPAI H5N1 virus than homologous DNA-DNA or VLP617 VLP vaccination (10). Therefore, in the present study, immunogenicity of HIV618 VLP produced by stably transfected S2 clones was evaluated in the heterologous 619 DNA-VLP prime-boost setting. Specifically, BALB/c mice were i.m. primed twice 620 with three plasmids CMV/R-gp120 (consensus B and C) and CMV/R-Gag (Pr55) 621 and s.c. boosted twice with a mixture of HIV-1 VLP expressing consensus B and 622 C envelope proteins in the presence of CpG. BALB/c mice injected with PBS 623 were used as controls. Serum samples were collected before and after 624 immunization and splenocytes were harvested after sacrifice. No animals showed 625 any signs of toxicity, and all remained healthy up to the end of the immunization 626 protocol. 627 To evaluate HIV-1 envelope-specific serum antibody titers, ELISA was 628 performed on microwell plates coated with consensus B and C gp120, 629 respectively. For the comparison, pooled HIV-1 patient plasmas were used. The 630 plasmas were pooled by 8 plasma samples selected from over 800 samples of 631 HIV-1 B’C recombinant virus-infected individuals in China. In our unpublished 632 study these 8 individual plasmas broadly neutralized various subtypes of HIV-1 633 on A uust 0, 2017 by gest http/jvi.asm .rg/ D ow nladed fom strains (Yang et al. data not shown). Fig. 5A and B show that both mouse 634 immune sera and pooled human plasmas exhibited very high end-point titers with 635 average titers of 1.77 x 10 in mouse immune sera and 2.37 x 10 in human 636 plasmas against consensus B gp120 and average titers of 5.43 x 10 in mouse 637 immune sera and 3.67 x 10 in human plasmas against consensus C gp120. Fig. 638 5C and D show that high titers of specific serum IgG1 against consensus B or C 639 gp120 were elicited in the immunized group. At 1 to 10 dilutions, OD values 640 ranging from 0.7 to 1.7 were detected against consensus B gp120 and from 0.8 641 to 1.6 against consensus C in the immunized group. Compared to OD values 642 detected in PBS control group the difference is very significant (P < 0.0001 for 643 consensus B gp120; P < 0.0001 for consensus C gp120). Fig. 5E and F show 644 that high titers of specific serum IgG2a against consensus B or C gp120 were 645 elicited in the immunized group. At 1:10 dilutions, OD values ranging from 0.1 to 646 0.8 were detected against consensus B gp120 and from 0.1 to 0.7 against 647 consensus C in the immunized group. Compared to OD values detected in PBS 648 control group the difference is statistically significant (P = 0.0371 for consensus B 649 gp120; P = 0.0111 for consensus C gp120). Thus we conclude that heterologous 650 DNA-VLP prime-boost elicits very high titers of ELISA-binding total IgG, IgG1 and 651 IgG2a responses against both consensus B and C envelope proteins. 652 Neutralizing antibody responses elicited by DNA-VLP prime-boost 653 To test neutralization activity of the immune sera, HIV-1 pseudotype-based 654 neutralization (PN) assay was performed against a panel of HIV-1 pseudotypes 655 as well as 10A1 control. The retroviral envelope 10A1 recognizes either Ram-1 or 656 on A uust 0, 2017 by gest http/jvi.asm .rg/ D ow nladed fom Glvr-1 as a receptor for cell entry (32) and was used here as negative control. 657 The panel of HIV-1 pseudotypes consists of HIV-1 envelopes derived from clade 658 A (Q168), clade B (consensus B, AD8, HXBc2 and Yu2), clade B’(CNE11), clade 659 C (consensus C) and CRF01_AE (CNE3). Fig. 6A shows neutralization activity of 660 immune sera at 1:50 dilutions against the pseudotype panel. As expected, 661 compared with pseudotypes alone, immune sera from all 6 immunized mice 662 exhibited no neutralization activity against control 10A1 pseudotype. Compared 663 with neutralization activity against control 10A1 pseudotype, there was no 664 significant neutralization activity of the immune sera against AD8 pseudotype. In 665 contrast, the immune sera from 6 immunized mice exhibited statistically 666 significant neutralization activity against Q168, consensus B, HxBc2, Yu2, 667 consensus C, CNE3 and CNE11 pseudotypes. Among them, on average there 668 were 60% and 40% neutralization activity against autologous consensus B and C 669 pseudotypes respectively, and 51, 42, 41, 36 and 21% against heterologous 670 CNE11, HXBc2, CNE3, Yu2 and Q168 pseudotypes, respectively. 671 To further test neutralization activity, the immune sera of individual 672 immunized mice were titrated against HIV-1 pseudotypes expressing consensus 673 B, consensus C, CNE11 and CNE3 envelope proteins (Fig. 6B). On average at 674 1:20 dilution 70% neutralization activity against consensus B, 54% neutralization 675 activity against consensus C, 60% neutralization activity against CNE11 and 55% 676 neutralization activity against CNE3 were detected and at 1:60 dilution 677 neutralization activity dropped to 50% neutralization activity against consensus B, 678 20% neutralization activity against consensus C, 40% neutralization activity 679 on A uust 0, 2017 by gest http/jvi.asm .rg/ D ow nladed fom against CNE11 and 35% neutralization activity against CNE3. The pre-immune 680 sera from immunized mice did not show any neutralization activity. With the same 681 pooled human plasmas, we also titrated their neutralizing antibody activity 682 against HIV-1 pseudotypes expressing consensus B, consensus C, CNE3 and 683 CNE11. As shown in Supplementary Fig. 4, these highly selected, pooled human 684 plasmas exhibited comparable to or even higher neutralizing antibody activity 685 than mouse immune sera shown in Fig. 6 B. Thus, we conclude that DNA-VLP 686 prime-boost elicits neutralizing antibody responses against a number of 687 autologous and heterologous envelope proteins. But their potency and breadth 688 still need to be improved. 689 ADCC and ADCVI responses elicited by DNA-VLP prime-boost 690 To test whether immune sera elicited with DNA-VLP prime-boost could 691 mediate antibody-dependent cell-mediated viral inhibition (ADCVI), we first 692 infected CEMss-CCR5 cells with HIV-1 AD8 strain and 15 days later measured 693 virus replication by HIV-1 gag p24 ELISA and stained cell surface expression of 694 HIV-1 envelope protein using pooled HIV-1 patient sera followed by FACS. We 695 found that HIV-1 replicated well in infected CEMss-CCR5 cells (Yang et al. data 696 not shown) and HIV-1 envelope proteins were expressed on the surface of 697 infected cells (the right panel of Fig. 7A). We then stained cell surface expression 698 of HIV-1 envelope protein using sera from PBS control mice and immune sera 699 elicited with DNA-VLP prime-boost. The left and middle panels of Fig. 7A show 700 that immune sera elicited with DNA-VLP prime-boost mice, but not from PBS 701 control, recognized HIV-1 envelope proteins on the surface of infected cells. 702 on A uust 0, 2017 by gest http/jvi.asm .rg/ D ow nladed fom We next investigated whether immune sera could mediate ADCC using AD8703 infected cells as target cells and naïve mouse splenocytes as effector cells. Fig. 704 7B shows that at a serum dilution of 1:50, ADCC activity ranging from 5 to 12% 705 was observed in immune sera from all 6 immunized mice; whereas no ADCC 706 activity was observed in sera from all 6 PBS control mice. The difference was 707 also highly statistically significant (P = 0.0002). The experiment was repeated 708 twice with similar results. 709 ADCVI of immune sera was also investigated using the same AD8-infected 710 cells and naïve mouse splenocytes as target and effector cells, respectively. Fig. 711 7C shows that at a serum dilution of 1:50, ADCVI activity ranging from 15% to 712 40% was observed in immune sera from all 6 immunized mice; whereas no 713 ADCVI activity was observed in sera from all 6 PBS control mice. The difference 714 was highly statistically significant (P = 0.006). The experiment was repeated 715 twice with similar results. Fig. 7D shows that there was a positive correlate 716 between % of killing by ADCC and % of inhibition by ADCVI (r = 0.9161) when 717 sera from all immunized and PBS control mice were analyzed. 718 CD8 T cell responses elicited by DNA-VLP prime-boost 719 To test whether DNA-VLP prime-boost could elicit HIV-1 envelope and gag720 specific T cell responses, splenocytes from DNA-VLP prime-boost and PBS 721 control mice were assayed against envelope or gag peptide mixtures using 722 intracellular cytokine staining for IFNγ and TNFα. Fig. 8A shows that compared to 723 CD8 T cells from PBS control mice, CD8 T cells from DNA-VLP prime-boost mice 724 exhibited statistically significant peptide-specific responses against envelope 725 on A uust 0, 2017 by gest http/jvi.asm .rg/ D ow nladed fom peptides. On average, 0.35% of CD8 T cells secreted IFNγ and 0.39% of CD8 T 726 cells secreted TNFα. In contrast, no CD8 T cell response against envelope 727 peptides was observed in splenocytes from all 6 PBS control mice. The 728 differences were highly statistically significant (P < 0.0001 for both IFNγ and 729 TNFα ). Fig. 8B shows that compared to CD8 T cells from PBS control mice, CD8 730 T cells from DNA-VLP prime-boost mice exhibited statistically significant peptide731 specific responses against gag peptides. On average, 0.06% of CD8 T cells 732 secreted IFNγ and 0.07% of CD8 T cells secreted TNFα. In contrast, no CD8 T 733 cell response against gag peptides was observed in splenocytes from all 6 PBS 734 control mice. The differences were highly statistically significant (P = 0.009 for 735 IFNγ, P = 0.0226 for TNFα). Thus, we conclude that the DNA-VLP prime-boost 736 elicits both HIV-1 envelope and gag-specific CD8 T cell responses. 737 738 on A uust 0, 2017 by gest http/jvi.asm .rg/ D ow nladed fom Discussion 739 In the present study, we developed stably transfected Drosophila S2 cell 740 clones to produce HIV-1 VLP and demonstrated that in one liter of culture 741 supernatant the amount of HIV-1 VLP produced by S2 cell clones is equivalent to 742 7.5 mg of gp120 and 9.5 mg of gag p55 (Fig. 3C), which is comparable to that 743 produced by insect cells infected with RB (5, 41) and much higher than those 744 produced by mammalian cells (16, 17, 30, 36, 42). The high HIV-1 VLP 745 production is mainly due to the fact that stably transfected S2 clones could grow 746 into much higher viable cell density than mammalian cells. In our present study, 747 stably transfected S2 clones in the fed batch culture using Wave bioreactor grew 748 to 2.2 x 10 cells per ml without significantly compromising cell viability (Fig. 3A 749 and 3B). In our study with other S2 clones, we found that using Wave bioreactor 750 stably transfected S2 clones in perfusion culture could grow to 1 x 10 cells per 751 ml without compromising cell viability (Wang et al. Mol. Biotech. In press). 752 Therefore, there is still large room to improve HIV-1 VLP production by our stably 753 transfected S2 clones. In addition, unlike HIV-1 VLP produced by insect cells 754 infected with RB, envelope proteins in HIV-1 VLP produced by stably transfected 755 S2 clones were properly cleaved into gp120 and gp41 (Fig. 1C). Moreover, unlike 756 HIV-1 VLP produced by insect cells infected with RB, culture supernatants 757 produced by stably transfected S2 clones are not contaminated with recombinant 758 viruses. Thus, compared to current expression systems, stably transfected S2 759 clones truly have several distinct advantages for HIV-1 VLP production. 760 In the present study, we also demonstrated that mice primed with plasmids 761 on A uust 0, 2017 by gest http/jvi.asm .rg/ D ow nladed fom expressing HIV-1 gp120 and gag and boosted with HIV-1 VLP produced by stably 762 transfected S2 clones in the presence of CpG exhibited both anti-envelope 763 antibody responses (see Fig. 5, 6 and 7) and envelope and gag-specific CD8 T 764 cell responses (see Fig. 8). Anti-HIV-1 envelope antibody responses include 765 ELISA-binding, neutralizing, ADCC and ADCVI. Percentage of killing by ADCC 766 and percentage of inhibition by ADCVI are positively correlated (Fig. 7D). The 767 protective role of envelope and gag-specific CD8 T cells in the control of SIV 768 replication has been inferred by several studies in macaques (29, 43, 48). The 769 protective role of neutralizing antibodies has also been demonstrated in SHIV770 infected macaque models through passive immunization (1, 28, 49). However the 771 in vivo mechanism of protection through passive immunization of antibody b12 772 was in part, FcγR mediated (19). Indeed, vaccine studies in macaques (2, 15, 20, 773 56) and analyses of sera from HIV-1 infected individuals (13, 46) have shown an 774 inverse correlation between the level of ADCC or ADCVI and virus load. Thus, 775 although exact correlates of protection against HIV-1 is still not fully established, 776 taking into account the correlates of immune protection against other viral 777 pathogens as well as understanding the immune components that control HIV-1 778 in vivo, the consensus view now is that a highly effective vaccine will need to 779 elicit coordinated antibody and T cell responses (31). Thus, the demonstration 780 that DNA-VLP prime-boost could elicit both antibody and T cell responses in mice 781 and that antibody responses include ELISA binding, neutralizing, ADCC and 782 ADCVI in the present study makes HIV-1 VLP produced by stably transfected S2 783 clones a promising vaccine component. 784 on A uust 0, 2017 by gest http/jvi.asm .rg/ D ow nladed fom Although our present study focused on DNA prime and VLP/CpG boost to 785 evaluate immunogenicity of HIV-1 VLP, in our preliminary study, we compared 786 immunogenicity between DNA prime and VLP/CpG boost versus DNA prime and 787 soluble gp120/CpG boost in mice. We found that although DNA-VLP/CpG elicited 788 higher, but not statistically significant, ELISA-binding antibody and CD8 T cell 789 responses than DNA-gp120/CpG, neutralizing antibody responses elicited by 790 DNA-VLP/CpG vaccination are significantly higher than those elicited by DNA791 gp120/CpG (Yang et al. data not shown). We also compared immunogenicity 792 between DNA-DNA, VLP/CpG-VLP/CpG and DNA-VLP/CpG in mice. We found 793 that DNA-VLP/CpG elicited significantly higher ELISA binding and neutralizing 794 antibody responses than those elicited by DNA-DNA and VLP/CpG-VLP/CpG 795 vaccination (Yang et al. data not shown). 796 In the present study we also show that envelope proteins in HIV-1 VLP 797 produced by stably transfected S2 clones are recognized by several broadly 798 neutralizing antibodies 2G12, b12, VRC01 and 4E10, but not by PG16 (see Table 799 2 and supplementary Fig. 1, 2 and 3). We found that while antibody 2G12 could 800 bind to envelope proteins on both transfected 293 T cells and S2 cells; antibody 801 PG16 could only bind to envelope proteins on the surface of transfected 293 T 802 cells, but not of transfected S2 cells (Table 2 and Supplementary Fig. 1). 803 Antibody 2G12 recognizes terminal dimannose (Manα1,2Man) moieties on 804 oligomannose glycans (7) and antibody PG16 recognizes a conformational 805 epitope that is dependent on glycosylation at specific variable loop N-linked sites 806 (12, 52). The difference in PG16 recognition of envelope proteins expressed on 807 on A uust 0, 2017 by gest http/jvi.asm .rg/ D ow nladed fom the surface of transfected 293 T cells and S2 cells, therefore, is likely due to the 808 heterogeneity in glycosylation. Recently, six new human broadly neutralizing 809 antibodies PGT 125, 126, 127, 128, 130 and 131 that bind specifically to the 810 Man8/9 glycans on gp120 were reported (35, 51). Among them, PGT 128, the 811 broadest of these antibodies, neutralizes over 70% of globally circulating viruses 812 and is, on average, an order of magnitude more potent than PG16 and VRC01 813 (51). PGT128 was found to bind to two conserved glycans and a short β-strand 814 segment of the gp120 V3 loop (35), therefore it will be interesting to test whether 815 these new antibodies can recognize glycans on envelope proteins produced by 816 stably transfected S2 clones, which was found to be mainly glycosylated in a high 817 mannose form with a few in complex and hybrid (Fig. 2A and B).. 818 Finally, careful analysis of morphology of HIV-1 VLP produced by stably 819 transfected S2 cell clones also points out an inherited shortcoming in the vector 820 system we used in the present study. We found that because in the vector 821 system the gag gene was under a constitutive promoter and the envelope gene 822 was under an inducible promoter, there are two kinds of HIV-1 VLP in 823 supernatants. One kind of HIV-1 VLP produced before cells were induced by 824 CdCl2 are mainly present in the upper band of non-linear sucrose gradient and do 825 not have spikes on their surface (Fig. 3C and 4A) and have lower buoyant 826 density between 1.10 and 1.14 (Fig. 1D); while the other kind produced after cells 827 were induced by CdCl2 are present only in the lower band of non-linear sucrose 828 gradient and do have spikes on their surface (Fig. 3C and 4B) and have higher 829 buoyant density between 1.15 and 1.18 (Fig. 1E). Importantly, using cryo-EM 830 on A uust 0, 2017 by gest http/jvi.asm .rg/ D ow nladed fom and electron tomography method, we were able to analyze spikes on the surface 831 of a dozen of representative spike-containing HIV-1 VLP and found that average 832 17 spikes per particle are present on particle surface after the purification (Table 833 1), which is similar to what was previously reported in wild type HIV-1 particles 834 (62). Furthermore, like those observed on the surface of SIV or HIV-1 virions (62) 835 we also did not find discrete periodic distances between spikes on the surface of 836 HIV-1 VLP (Fig. 4D) as one may have expected if spike placement reflected an 837 underlying geometry of matrix and capsid gag structure. Furthermore, since we 838 did observe that uncleaved gp160 as well as cleaved gp120/gp41 co-migrated 839 with HIV-1 Gag in sucrose gradient (Fig. 1E), likely the spikes on the surface of 840 HIV-1 VLP observed in Fig. 4C represent both uncleaved gp160 as well as 841 cleaved gp120/gp41, which can not be distinguished by the current 842 reconstruction with cryoEM tomography. Therefore, further refinement of vector 843 system to express gag and envelope proteins under both constitutive or both 844 inducible promoters, as well as production, induction and purification protocols 845 are needed to maximize the yield of envelope spike containing HIV-1 VLP. 846 847 on A uust 0, 2017 by gest http/jvi.asm .rg/ D ow nladed fom Acknowledgment 848 849 The authors wish to thank Dr. L. Naldini at the University Torino Medical 850 School, Torino, Italy for lentiviral transfer vector, Dr. B.H. Hahn at the University of 851 Alabama for DNA plasmid encoding consensus B and C HIV-1 envelope protein, 852 Dr. V. Deubel at the Institut Pasteur of Shanghai, Chinese Academy of Sciences 853 for Drosophila S2 cells, and Dr. P. Zhong at the Shanghai Municipal Center for 854 Disease Control and Prevention, Shanghai, China for pooled HIV-1 patient 855 plasmas. Cell line TZM-bl, human and mouse monoclonal antibodies b12, 4E10, 856 2G12 and 183-H12-5C and HIV-1 molecular clones pBru-3 and pBru-Yu2 as well 857 as expression vectors pADA, pAD8, pQ168ENVa2 and pNL4-3.luc.R-Ewere 858 obtained through the AIDS Research and Reference Reagent Program, Division 859 of the AIDS, National Institute of Allergy and Infectious Diseases, National 860 Institutes of Health, Germantown, MD. These reagents were originally developed 861 and contributed by Drs. J. Kappes, X. Wu, N. Landau, R. Risser, J. Overbaugh, E. 862 Freed, T. Ndung’u, A. Adachi, M.A. Martin, I.R. Chen, Dennis Burton, John 863 Mascola, G.W. Shaw and B.H. Hahn. This work was supported by research 864 grants from the Chinese National Science Foundation (#31170871), Chinese 865 Science and Technology Ministry 973 Program Project (2006CB504308), 866 National Science and Technology Major Project (2012ZX10001-008), the 867 Shanghai Pasteur Foundation (#SPHRF2007001) and French Energy Company 868 Areva to Paul Zhou and by research grants from the Chinese National Science 869 Foundation (U0832604), Chinese Science and Technology Ministry 973 Program 870 Project (2010CB912404) to Ping Zhu. 871 872onAuust0,2017bygesthttp/jvi.asm.rg/Downladedfom References8731. Baba, T. W., V. Liska, R. Hofmann-Lehmann, J. Vlasak, W. Xu, S. Ayehunie, L. A. Cavacini,874M. R. Posner, H. Katinger, G. Stiegler, B. J. Bernacky, T. A. Rizvi, R. Schmidt, L. R. Hill, M.875E. Keeling, Y. Lu, J. E. Wright, T. C. Chou, and R. M. Ruprecht. 2000. Human neutralizing876monoclonal antibodies of the IgG1 subtype protect against mucosal simian-human877immunodeficiency virus infection. Nat Med 6:200-6.8782. Banks, N. D., N. Kinsey, J. Clements, and J. E. Hildreth. 2002. Sustained antibody-dependent879cell-mediated cytotoxicity (ADCC) in SIV-infected macaques correlates with delayed progression880to AIDS. AIDS Res Hum Retroviruses 18:1197-205.8813. Barouch, D. H., Z. Y. Yang, W. P. Kong, B. Korioth-Schmitz, S. M. Sumida, D. M. Truitt, M.882G. Kishko, J. C. Arthur, A. Miura, J. R. Mascola, N. L. Letvin, and G. J. Nabel. 2005. A883human T-cell leukemia virus type 1 regulatory element enhances the immunogenicity of human884immunodeficiency virus type 1 DNA vaccines in mice and nonhuman primates. J Virol 79:8828-88534.8864. Brighty, D. W., and M. Rosenberg. 1994. A cis-acting repressive sequence that overlaps the887Rev-responsive element of human immunodeficiency virus type 1 regulates nuclear retention of888env mRNAs independently of known splice signals. Proc Natl Acad Sci U S A 91:8314-8.8895. Buonaguro, L., M. L. Visciano, M. L. Tornesello, M. Tagliamonte, B. Biryahwaho, and F. M.890Buonaguro. 2005. Induction of systemic and mucosal cross-clade neutralizing antibodies in891BALB/c mice immunized with human immunodeficiency virus type 1 clade A virus-like particles892administered by different routes of inoculation. J Virol 79:7059-67.8936. Burton, D. R., J. Pyati, R. Koduri, S. J. Sharp, G. B. Thornton, P. W. Parren, L. S. Sawyer,894R. M. Hendry, N. Dunlop, P. L. Nara, and et al. 1994. Efficient neutralization of primary895isolates of HIV-1 by a recombinant human monoclonal antibody. Science 266:1024-7.8967. Calarese, D. A., C. N. Scanlan, M. B. Zwick, S. Deechongkit, Y. Mimura, R. Kunert, P. Zhu,897M. R. Wormald, R. L. Stanfield, K. H. Roux, J. W. Kelly, P. M. Rudd, R. A. Dwek, H.898Katinger, D. R. Burton, and I. A. Wilson. 2003. Antibody domain exchange is an899immunological solution to carbohydrate cluster recognition. Science 300:2065-71.9008. Culp, J. S., H. Johansen, B. Hellmig, J. Beck, T. J. Matthews, A. Delers, and M. Rosenberg.9011991. Regulated expression allows high level production and secretion of HIV-1 gp120 envelope902glycoprotein in Drosophila Schneider cells. Biotechnology (N Y) 9:173-7.9039. Derdeyn, C. A., J. M. Decker, J. N. Sfakianos, X. Wu, W. A. O'Brien, L. Ratner, J. C.904Kappes, G. M. Shaw, and E. Hunter. 2000. Sensitivity of human immunodeficiency virus type 1905to the fusion inhibitor T-20 is modulated by coreceptor specificity defined by the V3 loop of906gp120. J Virol 74:8358-67.90710. Ding, H., C. Tsai, R. A. Gutierrez, F. Zhou, P. Buchy, V. Deubel, and P. Zhou. 2011. Superior908neutralizing antibody response and protection in mice vaccinated with heterologous DNA prime909and virus like particle boost against HPAI H5N1 virus. PLoS ONE 6:e16563.91011. Doan, L. X., M. Li, C. Chen, and Q. Yao. 2005. Virus-like particles as HIV-1 vaccines. Rev911Med Virol 15:75-88.91212. Doores, K. J., and D. R. Burton. 2010. Variable loop glycan dependency of the broad and potent913HIV-1-neutralizing antibodies PG9 and PG16. J Virol 84:10510-21.91413. Forthal, D. N., G. Landucci, and E. S. Daar. 2001. Antibody from patients with acute human915immunodeficiency virus (HIV) infection inhibits primary strains of HIV type 1 in the presence of916natural-killer effector cells. J Virol 75:6953-61.91714. Gomez-Roman, V. R., R. H. Florese, L. J. Patterson, B. Peng, D. Venzon, K. Aldrich, and M.918Robert-Guroff. 2006. A simplified method for the rapid fluorometric assessment of antibody-919dependent cell-mediated cytotoxicity. J Immunol Methods 308:53-67.92015. Gomez-Roman, V. R., L. J. Patterson, D. Venzon, D. Liewehr, K. Aldrich, R. Florese, and M.921Robert-Guroff. 2005. Vaccine-elicited antibodies mediate antibody-dependent cellular922cytotoxicity correlated with significantly reduced acute viremia in rhesus macaques challenged923with SIVmac251. J Immunol 174:2185-9.92416. Hammonds, J., X. Chen, T. Fouts, A. DeVico, D. Montefiori, and P. Spearman. 2005.925Induction of neutralizing antibodies against human immunodeficiency virus type 1 primary926onAuust0,2017bygesthttp/jvi.asm.rg/Downladedfom isolates by Gag-Env pseudovirion immunization. J Virol 79:14804-14.92717. Hammonds, J., X. Chen, X. Zhang, F. Lee, and P. Spearman. 2007. Advances in methods for928the production, purification, and characterization of HIV-1 Gag-Env pseudovirion vaccines.929Vaccine 25:8036-48.93018. Hartmann, G., and A. M. Krieg. 2000. Mechanism and function of a newly identified CpG DNA931motif in human primary B cells. J Immunol 164:944-53.93219. Hessell, A. J., L. Hangartner, M. Hunter, C. E. Havenith, F. J. Beurskens, J. M. Bakker, C.933M. Lanigan, G. Landucci, D. N. Forthal, P. W. Parren, P. A. Marx, and D. R. Burton. 2007.934Fc receptor but not complement binding is important in antibody protection against HIV. Nature935449:101-4.93620. Hidajat, R., P. Xiao, Q. Zhou, D. Venzon, L. E. Summers, V. S. Kalyanaraman, D. C.937Montefiori, and M. Robert-Guroff. 2009. Correlation of vaccine-elicited systemic and mucosal938nonneutralizing antibody activities with reduced acute viremia following intrarectal simian939immunodeficiency virus SIVmac251 challenge of rhesus macaques. J Virol 83:791-801.94021. Ivey-Hoyle, M., and M. Rosenberg. 1990. Rev-dependent expression of human941immunodeficiency virus type 1 gp160 in Drosophila melanogaster cells. Mol Cell Biol 10:6152-9.94222. Jia, B., S. K. Ng, M. Q. DeGottardi, M. Piatak, E. Yuste, A. Carville, K. G. Mansfield, W. Li,943B. A. Richardson, J. D. Lifson, and D. T. Evans. 2009. Immunization with single-cycle SIV944significantly reduces viral loads after an intravenous challenge with SIV(mac)239. PLoS Pathog9455:e1000272.94623. Kothe, D. L., J. M. Decker, Y. Li, Z. Weng, F. Bibollet-Ruche, K. P. Zammit, M. G. Salazar,947Y. Chen, J. F. Salazar-Gonzalez, Z. Moldoveanu, J. Mestecky, F. Gao, B. F. Haynes, G. M.948Shaw, M. Muldoon, B. T. Korber, and B. H. Hahn. 2007. Antigenicity and immunogenicity of949HIV-1 consensus subtype B envelope glycoproteins. Virology 360:218-34.95024. Kothe, D. L., Y. Li, J. M. Decker, F. Bibollet-Ruche, K. P. Zammit, M. G. Salazar, Y. Chen,951Z. Weng, E. A. Weaver, F. Gao, B. F. Haynes, G. M. Shaw, B. T. Korber, and B. H. Hahn.9522006. Ancestral and consensus envelope immunogens for HIV-1 subtype C. Virology 352:438-49.95325. Kuate, S., C. Stahl-Hennig, H. Stoiber, G. Nchinda, A. Floto, M. Franz, U. Sauermann, S.954Bredl, L. Deml, R. Ignatius, S. Norley, P. Racz, K. Tenner-Racz, R. M. Steinman, R. Wagner,955and K. Uberla. 2006. Immunogenicity and efficacy of immunodeficiency virus-like particles956pseudotyped with the G protein of vesicular stomatitis virus. Virology 351:133-44.95726. Lifson, J. D., J. L. Rossio, M. Piatak, Jr., J. Bess, Jr., E. Chertova, D. K. Schneider, V. J.958Coalter, B. Poore, R. F. Kiser, R. J. Imming, A. J. Scarzello, L. E. Henderson, W. G. Alvord,959V. M. Hirsch, R. E. Benveniste, and L. O. Arthur. 2004. Evaluation of the safety,960immunogenicity, and protective efficacy of whole inactivated simian immunodeficiency virus961(SIV) vaccines with conformationally and functionally intact envelope glycoproteins. AIDS Res962Hum Retroviruses 20:772-87.96327. Lynch, A. G., F. Tanzer, M. J. Fraser, E. G. Shephard, A. L. Williamson, and E. P. Rybicki.9642010. Use of the piggyBac transposon to create HIV-1 gag transgenic insect cell lines for965continuous VLP production. BMC Biotechnol 10:30.96628. Mascola, J. R., M. G. Lewis, G. Stiegler, D. Harris, T. C. VanCott, D. Hayes, M. K. Louder,967C. R. Brown, C. V. Sapan, S. S. Frankel, Y. Lu, M. L. Robb, H. Katinger, and D. L. Birx.9681999. Protection of Macaques against pathogenic simian/human immunodeficiency virus 89.6PD969by passive transfer of neutralizing antibodies. J Virol 73:4009-18.97029. Matano, T., R. Shibata, C. Siemon, M. Connors, H. C. Lane, and M. A. Martin. 1998.971Administration of an anti-CD8 monoclonal antibody interferes with the clearance of chimeric972simian/human immunodeficiency virus during primary infections of rhesus macaques. J Virol97372:164-9.97430. McBurney, S. P., and T. M. Ross. 2009. Human immunodeficiency virus-like particles with975consensus envelopes elicited broader cell-mediated peripheral and mucosal immune responses976than polyvalent and monovalent Env vaccines. Vaccine.97731. McElrath, M. J., and B. F. Haynes. 2010. Induction of immunity to human immunodeficiency978virus type-1 by vaccination. Immunity 33:542-54.97932. Miller, A. D., and F. Chen. 1996. Retrovirus packaging cells based on 10A1 murine leukemia980virus for production of vectors that use multiple receptors for cell entry. J Virol 70:5564-71.98133. Montefiori, D. C., J. T. Safrit, S. L. Lydy, A. P. Barry, M. Bilska, H. T. Vo, M. Klein, J.982onAuust0,2017bygesthttp/jvi.asm.rg/Downladedfom Tartaglia, H. L. Robinson, and B. Rovinski. 2001. Induction of neutralizing antibodies and gag-983specific cellular immune responses to an R5 primary isolate of human immunodeficiency virus984type 1 in rhesus macaques. J Virol 75:5879-90.98534. Notka, F., C. Stahl-Hennig, U. Dittmer, H. Wolf, and R. Wagner. 1999. Accelerated clearance986of SHIV in rhesus monkeys by virus-like particle vaccines is dependent on induction of987neutralizing antibodies. Vaccine 18:291-301.98835. Pejchal, R., K. J. Doores, L. M. Walker, R. Khayat, P. S. Huang, S. K. Wang, R. L. Stanfield,989J. P. Julien, A. Ramos, M. Crispin, R. Depetris, U. Katpally, A. Marozsan, A. Cupo, S.990Maloveste, Y. Liu, R. McBride, Y. Ito, R. W. Sanders, C. Ogohara, J. C. Paulson, T. Feizi, C.991N. Scanlan, C. H. Wong, J. P. Moore, W. C. Olson, A. B. Ward, P. Poignard, W. R. Schief, D.992R. Burton, and I. A. Wilson. 2011. A potent and broad neutralizing antibody recognizes and993penetrates the HIV glycan shield. Science 334:1097-103.99436. Persson, R. H., S. X. Cao, G. Cates, F. L. Yao, M. H. Klein, and B. Rovinski. 1998.995Modifications of HIV-1 retrovirus-like particles to enhance safety and immunogenicity.996Biologicals 26:255-65.99737. Pillay, S., E. G. Shephard, A. E. Meyers, A. L. Williamson, and E. P. Rybicki. 2010. HIV-1998sub-type C chimaeric VLPs boost cellular immune responses in mice. J Immune Based Ther999Vaccines 8:7.100038. Platt, E. J., K. Wehrly, S. E. Kuhmann, B. Chesebro, and D. Kabat. 1998. Effects of CCR51001and CD4 cell surface concentrations on infections by macrophagetropic isolates of human1002immunodeficiency virus type 1. J Virol 72:2855-64.100339. Prodromou, C., and L. H. Pearl. 1992. Recursive PCR: a novel technique for total gene1004synthesis. Protein Eng 5:827-9.100540. Rerks-Ngarm, S., P. Pitisuttithum, S. Nitayaphan, J. Kaewkungwal, J. Chiu, R. Paris, N.1006Premsri, C. Namwat, M. de Souza, E. Adams, M. Benenson, S. Gurunathan, J. Tartaglia, J.1007G. McNeil, D. P. Francis, D. Stablein, D. L. Birx, S. Chunsuttiwat, C. Khamboonruang, P.1008Thongcharoen, M. L. Robb, N. L. Michael, P. Kunasol, and J. H. Kim. 2009. Vaccination with1009ALVAC and AIDSVAX to Prevent HIV-1 Infection in Thailand. N Engl J Med.101041. Roldao, A., M. C. Mellado, L. R. Castilho, M. J. Carrondo, and P. M. Alves. 2010. Virus-like1011particles in vaccine development. Expert Rev Vaccines 9:1149-76.101242. Sakuragi, S., T. Goto, K. Sano, and Y. Morikawa. 2002. HIV type 1 Gag virus-like particle1013budding from spheroplasts of Saccharomyces cerevisiae. Proc Natl Acad Sci U S A 99:7956-61.101443. Schmitz, J. E., R. P. Johnson, H. M. McClure, K. H. Manson, M. S. Wyand, M. J. Kuroda, M.1015A. Lifton, R. S. Khunkhun, K. J. McEvers, J. Gillis, M. Piatak, J. D. Lifson, G. Grosschupff,1016P. Racz, K. Tenner-Racz, E. P. Rieber, K. Kuus-Reichel, R. S. Gelman, N. L. Letvin, D. C.1017Montefiori, R. M. Ruprecht, R. C. Desrosiers, and K. A. Reimann. 2005. Effect of CD8+1018lymphocyte depletion on virus containment after simian immunodeficiency virus SIVmac2511019challenge of live attenuated SIVmac239delta3-vaccinated rhesus macaques. J Virol 79:8131-41.102044. Schneider, R., M. Campbell, G. Nasioulas, B. K. Felber, and G. N. Pavlakis. 1997.1021Inactivation of the human immunodeficiency virus type 1 inhibitory elements allows Rev-1022independent expression of Gag and Gag/protease and particle formation. J Virol 71:4892-903.102345. Sheehy, M. E., A. B. McDermott, S. N. Furlan, P. Klenerman, and D. F. Nixon. 2001. A novel1024technique for the fluorometric assessment of T lymphocyte antigen specific lysis. J Immunol1025Methods 249:99-110.102646. Stratov, I., A. Chung, and S. J. Kent. 2008. Robust NK cell-mediated human immunodeficiency1027virus (HIV)-specific antibody-dependent responses in HIV-infected subjects. J Virol 82:5450-9.102847. Tsai, C., C. Caillet, H. Hu, F. Zhou, H. Ding, G. Zhang, B. Zhou, S. Wang, S. Lu, P. Buchy,1029V. Deubel, F. R. Vogel, and P. Zhou. 2009. Measurement of neutralizing antibody responses1030against H5N1 clades in immunized mice and ferrets using pseudotypes expressing influenza1031hemagglutinin and neuraminidase. Vaccine 27:6777-90.103248. Vaccari, M., J. Mattapallil, K. Song, W. P. Tsai, A. Hryniewicz, D. Venzon, M. Zanetti, K. A.1033Reimann, M. Roederer, and G. Franchini. 2008. Reduced protection from simian1034immunodeficiency virus SIVmac251 infection afforded by memory CD8+ T cells induced by1035vaccination during CD4+ T-cell deficiency. J Virol 82:9629-38.103649. Veazey, R. S., R. J. Shattock, M. Pope, J. C. Kirijan, J. Jones, Q. Hu, T. Ketas, P. A. Marx, P.1037J. Klasse, D. R. Burton, and J. P. Moore. 2003. Prevention of virus transmission to macaque1038onAuust0,2017bygesthttp/jvi.asm.rg/Downladedfom monkeys by a vaginally applied monoclonal antibody to HIV-1 gp120. Nat Med 9:343-6.103950. Wagner, R., V. J. Teeuwsen, L. Deml, F. Notka, A. G. Haaksma, S. S. Jhagjhoorsingh, H.1040Niphuis, H. Wolf, and J. L. Heeney. 1998. Cytotoxic T cells and neutralizing antibodies induced1041in rhesus monkeys by virus-like particle HIV vaccines in the absence of protection from SHIV1042infection. Virology 245:65-74.104351. Walker, L. M., M. Huber, K. J. Doores, E. Falkowska, R. Pejchal, J. P. Julien, S. K. Wang, A.1044Ramos, P. Y. Chan-Hui, M. Moyle, J. L. Mitcham, P. W. Hammond, O. A. Olsen, P. Phung,1045S. Fling, C. H. Wong, S. Phogat, T. Wrin, M. D. Simek, P. G. Principal Investigators, W. C.1046Koff, I. A. Wilson, D. R. Burton, and P. Poignard. 2011. Broad neutralization coverage of HIV1047by multiple highly potent antibodies. Nature.104852. Walker, L. M., S. K. Phogat, P. Y. Chan-Hui, D. Wagner, P. Phung, J. L. Goss, T. Wrin, M.1049D. Simek, S. Fling, J. L. Mitcham, J. K. Lehrman, F. H. Priddy, O. A. Olsen, S. M. Frey, P.1050W. Hammond, G. Miiro, J. Serwanga, A. Pozniak, D. McPhee, O. Manigart, L.1051Mwananyanda, E. Karita, A. Inwoley, W. Jaoko, J. Dehovitz, L. G. Bekker, P. Pitisuttithum,1052R. Paris, S. Allen, S. Kaminsky, T. Zamb, M. Moyle, W. C. Koff, P. Poignard, and D. R.1053Burton. 2009. Broad and Potent Neutralizing Antibodies from an African Donor Reveal a New1054HIV-1 Vaccine Target. Science.105553. Wang, L., H. Hu, J. Yang, F. Wang, C. Kaisermayer, and P. Zhou. 2011. High Yield of1056Human Monoclonal Antibody Produced by Stably Transfected Drosophila Schneider 2 Cells in1057Perfusion Culture Using Wave Bioreactor. Mol Biotechnol.105854. Wen, M., R. Arora, H. Wang, L. Liu, J. T. Kimata, and P. Zhou. 2010. GPI-anchored single1059chain Fv--an effective way to capture transiently-exposed neutralization epitopes on HIV-11060envelope spike. Retrovirology 7:79.106155. Wu, X., Z. Y. Yang, Y. Li, C. M. Hogerkorp, W. R. Schief, M. S. Seaman, T. Zhou, S. D.1062Schmidt, L. Wu, L. Xu, N. S. Longo, K. McKee, S. O'Dell, M. K. Louder, D. L. Wycuff, Y.1063Feng, M. Nason, N. Doria-Rose, M. Connors, P. D. Kwong, M. Roederer, R. T. Wyatt, G. J.1064Nabel, and J. R. Mascola. 2010. Rational Design of Envelope Identifies Broadly Neutralizing1065Human Monoclonal Antibodies to HIV-1. Science.106656. Xiao, P., J. Zhao, L. J. Patterson, E. Brocca-Cofano, D. Venzon, P. A. Kozlowski, R. Hidajat,1067T. Demberg, and M. Robert-Guroff. 2010. Multiple vaccine-elicited nonneutralizing1068antienvelope antibody activities contribute to protective efficacy by reducing both acute and1069chronic viremia following simian/human immunodeficiency virus SHIV89.6P challenge in rhesus1070macaques. J Virol 84:7161-73.107157. Yao, Q., Z. Bu, A. Vzorov, C. Yang, and R. W. Compans. 2003. Virus-like particle and DNA-1072based candidate AIDS vaccines. Vaccine 21:638-43.107358. Young, K. R., S. P. McBurney, L. U. Karkhanis, and T. M. Ross. 2006. Virus-like particles:1074designing an effective AIDS vaccine. Methods 40:98-117.107559. Zhou, P., S. Goldstein, K. Devadas, D. Tewari, and A. L. Notkins. 1998. Cells transfected with1076a non-neutralizing antibody gene are resistant to HIV infection: targeting the endoplasmic1077reticulum and trans-Golgi network. J Immunol 160:1489-96.107860. Zhou, T., I. Georgiev, X. Wu, Z. Y. Yang, K. Dai, A. Finzi, Y. Do Kwon, J. Scheid, W. Shi, L.1079Xu, Y. Yang, J. Zhu, M. C. Nussenzweig, J. Sodroski, L. Shapiro, G. J. Nabel, J. R. Mascola,1080and P. D. Kwong. 2010. Structural Basis for Broad and Potent Neutralization of HIV-1 by1081Antibody VRC01. Science.108261. Zhu, P., E. Chertova, J. Bess, Jr., J. D. Lifson, L. O. Arthur, J. Liu, K. A. Taylor, and K. H.1083Roux. 2003. Electron tomography analysis of envelope glycoprotein trimers on HIV and simian1084immunodeficiency virus virions. Proc Natl Acad Sci U S A 100:15812-7.108562. Zhu, P., J. Liu, J. Bess, Jr., E. Chertova, J. D. Lifson, H. Grise, G. A. Ofek, K. A. Taylor, and1086K. H. Roux. 2006. Distribution and three-dimensional structure of AIDS virus envelope spikes.1087Nature 441:847-52.108863. Zwick, M. B., A. F. Labrijn, M. Wang, C. Spenlehauer, E. O. Saphire, J. M. Binley, J. P.1089Moore, G. Stiegler, H. Katinger, D. R. Burton, and P. W. Parren. 2001. Broadly neutralizing1090antibodies targeted to the membrane-proximal external region of human immunodeficiency virus1091type 1 glycoprotein gp41. J Virol 75:10892-905.109210931094onAuust0,2017bygesthttp/jvi.asm.rg/Downladedfom Table 1. Measurement of purified HIV-1 VLP by cryo-EM and tomography.1095 1096HIV-1 VLP diameter (nm)Spikes (number)1128.380182142.229193134.887184123.321175101.999146114.596137128.440208124.15517996.2702010111.7781411186.6332012115.47918Mean±SD 125.681±23.24317±210971098onAuust0,2017bygesthttp/jvi.asm.rg/Downladedfom Table 2. Summary of antigenicity of HIV-1 VLP measured by FACS, Western1099Blot and ELISA.11001101VLP(S2)VLP(S2)gp160(S2)gp160(293T)ELISA WB FACS FACS 2G12++ +++ +++ VRC01++ +++ ++ b12++++ ++ PG16---++ 4E10-+++ 100F4----1102onAuust0,2017bygesthttp/jvi.asm.rg/Downladedfom Figure Legends1103Fig. 1 Expression, cleavage and incorporation of HIV-1 envelope and gag1104proteins produced by S2 clones. A. Schematic diagrams of plasmids encoding1105 gp160 (consensus B or C), rev, gag p55 proteins and pCoBlast selection marker.1106pMT/Bip stands for sequences encoding metallothionien promoter and Bip signal1107peptide; pMT stands for sequence encoding metallothionein promoter and pAC1108 stands for sequence encoding Ac5 promoter. Poly A: poly adenolytion signal. B.1109Western blot analysis of HIV-1 gag and envelope protein expression with or1110without the induction of 5 μM CdCl2. C. Western blot analysis of HIV-1 envelope1111 protein expression by anti-gp120 and gp41 antibodies, respective. D. Western1112blot analysis of HIV-1 gag protein in sucrose gradient-unfractionated and -1113 fractionated samples from HIV-1 gag alone transfected S2 cells. E. Western blot1114analysis of HIV-1 envelope and gag proteins in sucrose gradient-unfractionated1115and -fractionated samples from stably transfected S2 clone expressing both HIV-11161 envelope and gag proteins.1117Fig. 2 Glycosylation of HIV-1 gp120 or HIV-1 VLP produced by S2 clones. A.1118 Deglycosylation of HIV-1 gp120 protein, HIV-1 VLP and wild type HIV-1 (Bru-3)1119with the treatment of PNGase F and endo H. B. Probing glycosylation pattern of1120 HIV-1 gp120 with a panel of lectins. GNA stands for galanthus nivalis agglutinin;1121 SNA stands for sambucus nigra agglutinin; DSA stands for datura stramonium1122agglutinin; MAA stands for maackia amurensis agglutinin; and PNA stands for1123 peanut agglutinin. I.C. stands for an internal control.1124Fig. 3 Growth and HIV-1 VLP production by S2 clones in Wave bioreactor. A.1125onAuust0,2017bygesthttp/jvi.asm.rg/Downladedfom The viable cell density and the culture volume of S2 clone VB2 during 11 day’s1126fed batch culture; B. The cell number and the culture viability of S2 clone VB21127during 11 day’s fed batch culture; C. gag p55 and gp120 in supernatants1128 produced by S2 clone VB2 during 11 day’s fed batch culture. Arrow stands for1129the day in which CdCl2 inducer was added.1130Fig. 4 Representative cryoEM tomographic images of HIV-1 VLP produced by S21131clone VB2. A. Representative computationally derived transverse sections1132electron micrographs of HIV-1 VLP taken from upper band of non-linear sucrose1133gradient revealed using cryo-EM. Scale bars = 100 nm. B. Representative1134 computationally derived transverse sections electron micrographs of HIV-1 VLP1135taken from lower band of non-linear sucrose gradient revealed using cryo-EM.1136 Scale bars = 100 nm. C. Z-stack Computationally derived transverse Z-stack1137tomographic images of a selected HIV-1 VLP taken from lower band of non-linear1138sucrose gradient. Red arrows indicate the location of spikes. D. A surface-1139rendered model of a selected HIV-1 VLP produced by stably transfected S2 clone1140VB2 with highlighted presumptive ENV spikes (white) derived from electron1141 tomogram.1142Fig. 5 ELISA-binding antibody responses elicited with heterologous DNA-VLP1143prime-boost. The left panels show the end-point titers of total IgG (panel A), IgG11144 (panel C) and IgG2a (panel E) responses specifically against consensus B gp1201145and the right panels show the end-point titers of total IgG (panel B), IgG1 (panel1146 D) and IgG2a (panel F) responses specifically against consensus C gp120. In1147each panel, data represent the mean ± SD from 6 mice immunized with DNA-1148onAuust0,2017bygesthttp/jvi.asm.rg/Downladedfom VLP prime-boost and 6 mice injected with PBS control. Pooled HIV-1 patient1149plasmas were combined 8 plasma samples selected from over 800 samples of1150HIV-1 B’C recombinant virus-infected individuals in China, which broadly1151 neutralized various subtypes of HIV-1 strains.1152Fig. 6 Neutralization activity in individual immune serum samples elicited with1153heterologous DNA-VLP prime-boost. A. Neutralization activity of immune serum1154 samples at 1:50 dilution against a panel of HIV-1 pseudotypes and 10A1 control1155in TZM-bl cells. Percentage (%) of neutralization was calculated by [the amount1156of relative luciferase activity (RLA) in pseudotype alone – the amount of RLA in1157 pseudotype plus immune serum sample]/the amount of RLA in pseudotype alone.1158Statistic analysis was performed between mean ± SD of a given HIV-11159 pseudotype and mean ± SD of 10A1 control. The dash line shows 50% of1160neutralization; bars stands for means of % neutralization to a given HIV-11161pseudotype; and stands for P < 0.001. B. Titration of neutralization activity of1162immune serum samples against HIV-1 pseudotypes consensus B and C, CNE111163and CNE3.1164Fig. 7 ADCC and ADCVI responses of individual immune serum samples elicited1165with heterologous DNA-VLP prime-boost. A. Cell surface expression of HIV-11166 envelope proteins in HIV-1 AD8-infected CEMss-CCR5 cells detected by pooled1167 sera from PBS control and immunized mice as well as pooled HIV-1 patient sera.1168B. HIV-1 AD8-infected CEMss-CCR5 target cells were doubly labeled with PKH-1169 26 and CFSE. The labeled target cells were incubated with splenocytes (effector1170cells) from a naïve BALB/c mouse (E:T = 50:1) and with 1:50 diluted individual1171onAuust0,2017bygesthttp/jvi.asm.rg/Downladedfom sera samples from PBS control and immunized mice. Eighteen h later, ADCC1172was determined as described in Materials and Methods. C. HIV-1 AD8-infected1173CEMss-CCR5 target cells were incubated with splenocytes (effector cells) from1174 naïve BALB/c mice (E:T = 20:1) and with 1:50 diluted individual sera samples1175from PBS control and immunized mice. Two days later, gag p24 in culture1176supernatants was determined by ELISA, and virus inhibition was determined as1177 described in Materials and Methods. D. Linear regression analysis between %1178killing by ADCC and % inhibition by ADCVI among all immunized and control1179mice.1180Fig. 8 HIV-1 envelope and gag-specific CD8 T cell responses elicited with1181heterologous DNA-VLP prime-boost. Intracellular cytokine (IFN-γ and TNF-α)1182 staining was performed to analyze the CD8 T cell responses against a panel of1183HIV-1 envelop-specific peptides (A) and a panel of HIV-1 gag-specific peptides1184(B). The percentages of activated CD8 T cells that produce IFN-γ and TNF-α are1185shown. Splenocytes from mice (n=6 per group) immunized with DNA-VLP prime-1186boost or mice (n=6 per group) from PBS control were assessed and immune1187 responses were measured 10 days after the final boost.1188onAuust0,2017bygesthttp/jvi.asm.rg/Downladedfom