VIROLOGY 229, 269–278 (1997) ARTICLE NO. VY978445

HIV-1 Glycoproteins from Two Series of Primary Isolates:

View metadata, citation and similar papersReplication at core.ac.uk Phenotype and Immunogenicity brought to you by CORE provided by Elsevier - Publisher Connector FARAH MUSTAFA,*,1,2 JOAN F. L. RICHMOND,*,1 ROBERTO FERNANDEZ-LARSSON,†,1 SHAN LU, ROBERT FREDRIKSSON,‡ EVA MARIA FENYO¨ ,‡ MARYELLEN O’CONNELL,* ERIC JOHNSON, JIAYU WENG,* JOSEPH C. SANTORO,* and HARRIET L. ROBINSON*,3

*Department of Pathology, University of Massachusetts Medical Center, Worcester, Massachusetts 01655; †Children’s Research Institute, Center for Virology, Immunology, and Infectious Diseases, Washington, DC 20010-2970; and ‡Microbiology and Tumorbiology Center, Karolinska Institute, Stockholm, Sweden Received October 4, 1996; returned to author for revision November 18, 1996; accepted January 7, 1997

Seven envelope regions from two series of patient isolates have been molecularly cloned and analyzed for replication phenotypes and immunogenicity. Growth potential was analyzed for env sequences substituted into an HIV-1-NL4-3 backbone (NL4-3/env recombinants). Immunogenicity studies were conducted on secreted monomeric (gp120) and oligomeric (gp140) forms of the Envs using Env-expressing plasmid DNAs for immunizations. The env regions of the patient isolates conferred a spectrum of replication kinetics and cytotropisms on the NL4-3/env recombinants. Both patient series included non- syncytium-inducing with no ability to grow on T-cell lines, and highly syncytium inducing viruses which grew well on T-cell lines. These differences in growth potential did not correlate with the ability of the DNA-expressed Envs to raise antibody in rabbits. Rather, the relative immunogenicity of the Envs was patient and form specific. The Envs from patient 5 raised higher titers of antibody than the Envs from patient 6. For each primary Env, the gp120 form of the Env raised higher titers of antibody than the gp140 form. Thus, structural features of Env that affect replication do not necessarily affect the ability to raise antibody. ᭧ 1997 Academic Press

INTRODUCTION isolates from the symptomatic phase of the disease often have ‘‘rapid/high’’ replication kinetics on both PBMCs and A central problem in the development of a vaccine for T-cell lines. These viruses frequently induce syncytia. HIV-1 has been the identification of immunogens capable To date no studies have been done on the relative of raising neutralizing antibody for the envelope glycopro- immunogenicity of ‘‘slow/low’’ and ‘‘rapid/high’’ Envs. In teins (Env) of primary isolates. The Envs of primary iso- general, poor replicative function has correlated with lates have marked differences in amino acid sequence slow entry into cells and relative resistance to antibody (for compendium, see Myers et al., 1994), replication phe- binding (Fernandez-Larsson et al., 1992; Sullivan et al., notype (for review, see Fenyo et al., 1989), and the ability 1995). This suggests that epitopes important to the entry- to be recognized by antibody (for reviews, see Moore and Ho, 1995; Fenyo et al., 1996). In this manuscript, we mediated steps of binding and fusion may be masked in have undertaken an initial analysis of the effect of the ‘‘slow/low’’ viruses. Indeed, targets for neutralizing anti- replication phenotype of an Env on the ability of that Env body, such as the V3 loop and the CD4-binding domain, to elicit antibody. can be largely hidden from cognate antibodies in ‘‘slow/ Patient isolates from different phases of HIV-1 infection low’’ Envs (see Bou-Habib et al., 1994; Sullivan et al., exhibit stage-specific replication characteristics (Asjo et 1995; Stamatos and Cheng-Mayer, 1995). This raises the al., 1986; Cheng-Mayer et al., 1988; Fenyo et al., 1988; possibility that a greater exposure of entry epitopes in Tersmette et al., 1988; 1989). Viruses isolated from the ‘‘rapid/high’’ viruses would cause ‘‘rapid/high’’ Envs to be asymptomatic phase of the disease frequently exhibit better candidates for raising neutralizing antibody than ‘‘slow/low,’’ non-syncytium-inducing (NSI) replication in ‘‘slow/low’’ Envs. peripheral blood mononuclear cells (PBMCs). In general, The ‘‘rapid/high’’ and ‘‘slow/low’’ Envs used in this study these viruses do not replicate on cell lines. By contrast, were molecularly cloned from serial isolates recovered from two HIV-1-positive Swedish homosexual men (pa- tient 5 and patient 6) (von Gegerfelt et al., 1991). In both 1 These authors contributed equally to this work. 2 series, the growth characteristics of the isolates changed Present address: Department of Microbiology, The University of from ‘‘slow/low’’ to ‘‘rapid/high’’ as disease progressed Texas at Austin, Austin, TX 78712. 3 To whom correspondence and reprint requests should be ad- from lymphadenopathy to ARC/AIDS. Replication pheno- dressed. types of the Envs were assessed by substituting env

0042-6822/97 $25.00 269 Copyright ᭧ 1997 by Academic Press All rights of reproduction in any form reserved.

AID VY 8445 / 6a2a$$$181 02-04-97 03:43:30 vira AP: Virology 270 MUSTAFA ET AL. sequences in a constant HIV-1 backbone (HIV-1-NL4-3). tions, vectors, and V3 sequence for the clones used to Immunogenicity was examined using DNA-based immu- construct the NL4-3/env recombinants and vaccine vec- nization, a technique with the potential to raise neutraliz- tors are summarized in Fig. 2. ing antibody for both human and simian immunodefi- ciency viruses (Wang et al., 1993; Lu et al., 1995, 1996; Substitution of serial env fragments into NL4-3 Okuda et al., 1995). Patient env fragments from the pcRII or pSVIII vectors Our results reveal the serial Envs conferring a spec- were substituted for the homologous NL4-3 env fragment trum of growth characteristics on the otherwise constant by three-fragment ligation of 12.1-kb EcoRI–BamHI and NL4-3 backbone. They also demonstrate that the replica- 0.6-kb EcoRI–KpnI fragments of pNL4-3 and the 2.1-kb tion phenotype of an Env is not the primary determinant KpnI–BamHI patient env fragments (See Figs. 1A–1C). of the ability of an Env to raise antibody. Recombinant clones were identified by restriction endo- nuclease digestion and V3 sequence analysis. Due to MATERIALS AND METHODS problems with instability, the pNL4-3/env recombinants were grown in Stable II E. coli (Life Sciences Inc. Gaith- Plasmids ersburg, MD) in 20-ml vol using 50-ml tubes. pNL4-3 (Adachi et al., 1986) was obtained from the NIAID AIDS Repository, (Rockville, MD); pSVIIIenv Construction of gp120 and gp140 vaccine plasmids (Helseth et al., 1990) from Dr. J. Sodroski (Harvard Medi- Vaccine plasmids expressing gp120 and gp140 forms of cal School, Boston, MA); pJW4303, from Dr. J. I. Mullins the Envs were constructed by cloning env fragments in (University of Washington, Seattle, WA); pWR61602 (a hu- frame with a synthetic tissue plasminogen activator-(tPA)- man growth hormone (hGH) expressing plasmid) from leader sequence in pJW4303 (Figs. 1D and 1E). pJW4303 Dr. J. Haynes (Auragen Inc., Madison, WI); and pCRII from is an eukaryotic expression vector that uses enhancer and Invitrogen Corp. (San Diego, CA). Plasmids were grown promoter elements, including intron A from the cytomegalo- in the HB101 or Stable II (Life Technologies, Gaithers- immediate-early promoter, and polyadenylation se- burg, MD) strains of Escherichia coli and purified using quences from the bovine growth hormone (for details, see Qiagen columns (Chatsworth, CA). Lu et al., 1996). pJW4303 supports Env expression in the absence of Rev (Chapman et al., 1991). A 5؅ sense primer -Molecular cloning of env sequences from serial (JApcr501 5؅-GTCGCTCCGCTAGCTTGTGGGTCACAGTC patient isolates TATTATGGGTACC-3؅) used an NheI site (underlined) to Serial patient isolates that had undergone limited cul- clone env fragments in frame with the tPA-sequence in ture on PBMCs were obtained for molecular cloning (von pJW4303. For construction of gp120, a 3؅ antisense primer -Gegerfelt et al., 1991). The env regions were molecularly (JApcr504: 5؅-GGTCGGATCCttaCTGCACCACTCTTCTCTT cloned from infected peripheral blood mononuclear cell TGCC-3؅) introduced a stop codon (in lowercase) at the (PBMC) DNA using a nested PCR approach (Fig. 1). First, boundary of the surface (SU) and transmembrane (TM) a 3283-bp fragment containing the entire env gene and subunits of Env followed by a BamHI site (underlined) for flanking sequences was PCR amplified using sense oli- cloning into the BamHI site in pJW4303. The 3؅ antisense -gonucleotide 226, 5؅-TAATAAGAATTCTGCAACAACTGC- primer used to construct gp140 forms of env (JApcr502:5؅ -TGTTTAT-3؅ (nt 5737–5768 in NL4-3) and antisense oli- CGACGGATCCttaTGTTTATGTCAAACCAATTCCAC-3؅) in -gonucleotide 231, 5؅-CTTAAAGGTACCTGAGGTGTG- troduced a stop codon immediately prior to the transmem .ACTGGAAA-3؅ (nt 9016–8988 in NL4-3) (Fig. 1A). After brane domain of TM gel purification, this fragment was further amplified using Heteroduplex tracking assay (HTA) -sense oligonucleotide 227, 5؅-CACAGTCTATTATGG GGTACCTGTGTGGAA-3؅ (nt 6328–6358) and antisense A nested PCR approach was used to amplify a 941- -oligonucleotide 212, 5؅-GCTAAGGATCCGTTCACTAAT- bp internal V1–V3 region of the env sequences pre CGAATG G-3؅ (nt 8475–8448) to generate a 2150-bp env sent in the infected PBMC DNAs used for cloning. The fragment encompassing the KpnI (nt 6343) and BamHI first round of amplification was performed using the sites (nt 8465) of NL4-3 (Fig. 1B). The PCR-amplified frag- GeneAmp kit (Perkin–Elmer, Foster City, CA) and pri- (ments were cloned directly into the pCRII TA-cloning mers MNA (5؅-GCGAAAGAGCAGAAGACAGTGGC-3؅ -vector or digested with KpnI and BamHI and substituted and MN13 (5؅-CAGCTCGTCTCATTCTTTCCC-3؅) to pro for homologous sequences in pSVIIIenv. V3 loop se- duce a 2.6-kb fragment encompassing the entire env quences of each cloned fragment were obtained using gene. Hot-start PCR reactions used 30 cycles (94Њ for 60 the Sequenase Version 2.0 DNA sequencing kit (United sec, 60Њ for 90 sec, and 72Њ for 3 min) followed by a 10- States Biochemical, Cleveland, OH). These signature se- min final extension step at 72Њ. Amplification products quences were used to verify the identity of env se- were desalted on Sepharose G-50 spin columns (Sigma quences in subsequent cloning procedures. Designa- Biochemicals, St. Louis, MO) in TEN buffer (40 mM Tris–

AID VY 8445 / 6a2a$$$182 02-04-97 03:43:30 vira AP: Virology REPLICATION PHENOTYPES AND IMMUNOGENICITY OF HIV-1 ENVS 271

HCl, pH 7.5, 1 mM EDTA, 150 mM NaCl) and further Western blot analyses amplified in a second round of PCR using nested primers Protein blots were conducted on lysates of transiently -5؅-ACCCCAACCCACAAGAAG-3؅) and 218 (5؅-CCA) 209 transfected COS cells as previously described (Lu et al., .AGTACTATTAAACAG-3؅) to produce a 941-bp product 1996). Proteins were visualized by the Western-Light The same conditions were used for the second round of Chemiluminescent Detection System (Tropix, Bedford, PCR except that the hot-start was omitted and a 55Њ MA). HIV-Ig (1:3000 dilution) was used for the primary annealing temperature was used. Radiolabeled probes antibody. were made by amplifying molecularly cloned env genes with primers 209 and 218 in the presence of 0.1 mlof10 Antigen capture ELISA Ci/ml[32P]dCTP (New England Nuclear, Dupont, Boston, MA). Annealing conditions and the mobility of hetero- HIV-1 antigen was captured overnight at 4Њ on Easy duplexes were as described by Delwart et al. (1994a, Wash microtiter plates (Corning, Corning, NY) coated 1994b). with a 1:5000 dilution of HIV-Ig in phosphate-buffered saline (PBS). A standard curve was constructed in dupli- Recovery of viral stocks and growth studies cate using a known quantity of Env protein serially diluted in whey buffer (PBS, 4% whey powder (Davisco, Le Sueur, Viral stocks from recombinant NL4-3/env plasmids MN), 0.05% Tween-20). This standard Env was either a were generated by transfection of Cos cells using a baculovirus-produced IIIb-gp120 (Intracel, Cambridge, DEAE-dextran procedure (Gibbs et al., 1994). Transfec- MA) or a virus (vCB-14)-produced BH-8-gp140 tion efficiencies were controlled by adding a hGH-ex- (Earl et al., 1994). Captured antigen was detected using pressing plasmid as an internal control and testing for biotinylated HIV-Ig followed by streptavidin-linked horse- levels of hGH expression at 72 hr using an antigen cap- radish peroxidase (Vector Laboratories Inc. Burlingame, ture ELISA (Boehringer-Mannheim, Indianapolis, IN). CA) and TMB substrate. O.D. was measured at 450 nm. Fresh PHA-stimulated PBMCs were added to transfected All incubations were at 23Њ for 60 min. COS cells at 36 hr posttransfection and the culture me- dium was changed to RPMI 1640, 10% fetal calf serum, Env-specific antibody and 3% IL-2 (Kabat et al., 1994). Two days later, PBMCs were removed and cultured separately. PBMC cultures Titers of Env-specific IgG were determined by ELISA were regularly fed and assayed for virus replication using (Robinson et al., 1990) using vaccinia-produced gp140 indirect immunofluorescence analyses and a HIV-Ig (Ab- as the solid-phase antigen. Briefly, microtiter plates were bott, Abbott Park IL)-based antigen capture ELISA (Fer- coated with concanavalin A followed by Env antigen at nandez-Larsson et al., 1992). Cytotropism studies were approximately 1 mg of Env per milliliter. Sample sera carried out by cocultivating 1:1 mixtures of infected were assayed in duplicate at serial threefold dilutions PBMCs with T-cell lines (MT-2, H9, A3.01, and CEM1174 ranging from 1:100–1:8100. A standard curve was con- obtained from the NIAID AIDS Repository, Rockville, MD) structed using pooled sera from rabbits immunized with in RPMI 1640 supplemented with 10% fetal calf serum a plasmid expressing HXB-2 gp120. This curve was as- and antibiotics, and scoring virus-expressing cells using signed values that correlated with the concentration of indirect immunofluorescence over a 2-week period. Syn- specific antibody in the pooled sera. The concentration of cytium induction was assayed by cocultivation with MT- specific antibody in the standard was determined using 2 cells as described by Koot et al. (1992). purified rabbit IgG as a standard. Bound Env-specific IgG was detected using biotinylated goat anti-rabbit antisera In vitro expression of vaccine plasmids followed by streptavidin-horse radish peroxidase and TMB substrate (all from Vector Laboratories Inc. Burlin- Expression of plasmids was tested in transiently trans- game, CA). The sensitivity of the assay was {0.005 mg/ fected COS cells. Transfections used 1 to 3 mg of a test ml. Rabbit sera had backgrounds in this assay ranging DNA, 100 ng of the hGH-expressing plasmid, 10 mlof from 0.01 to 0.05 mg/ml. lipofectAMINE (Life Technologies, Gaithersburg, MD) and 5 105 subconfluent COS cells as described by Ç 1 DNA immunizations Reiter et al. (1995). At 48 to 72 hr posttransfection, super- nates were harvested and lysates were prepared by de- DNA immunizations were delivered to freshly shaved taching cells from plates in TEN buffer, collecting the abdominal skin of anesthetized female NZW rabbits with cells by centrifugation, and resuspending the pellet in Accel electric discharge or helium pulse gene guns (Aur- 10 mM Tris–HCl, pH 7.5, 1% Triton X-100, and 1 mM agen Inc., Middleton). DNA was precipitated onto 0.95 phenylmethylsulfonylfluoride. Lysates were clarified by or 1- to 3-mm gold beads using spermidine in the pres- microcentrifugation at 12,000 rpm for 60 min at 4Њ. Anti- ence of Ca2/ (Eisenbraun et al., 1993). Over the 2-year gen capture ELISAs were used to determine the concen- course of these experiments various conditions and trations of Env (see below) and hGH. schedules were used for immunizations. Immunizations

AID VY 8445 / 6a2a$$$182 02-04-97 03:43:30 vira AP: Virology FIG. 1. Cloning and expression of patient Envs. A–C depict positions of primers and restriction endonuclease sites used in the PCR amplification and subcloning of primary env fragments. D and E depict the structure and junctions of gp120 and gp140 env inserts in the pJW4303 vaccine expression vector. For more detail, see Materials and Methods. with the electric discharge gun included 6 shots of 0.8 ras (Figs. 1A–1C). Chimeras were constructed from mg of gold with 7.2 mg of DNA, and 3 shots of 0.8 mg patient 5 isolates taken at times A, B, and C (designated of gold with 2 mg of DNA. The amount of DNA per bead NL4-3/5A, NL4-3/5B, and NL4-3/5C); and from patient 6 was reduced as studies revealed that levels of §0.02 isolates taken at times A, B, C, and D (designated NL4- mg of DNA per milligram of gold did not increase the 3/6A, NL4-3/6B, NL4-3/6C, and NL4-3/6D) (von Gegerfelt efficiency of immunization (Pertmer et al., 1995). The et al., 1991). Figure 2 summarizes the cloning vectors switch from the electric discharge gun to the helium and signature V3 loop sequences for each of these envs. pulse gun was made when the helium gun became Hetroduplex tracking analyses revealed that each of available. Immunizations with the helium pulse gun in- the molecularly cloned envs represented the predomi- cluded 6, 10, 36, and 40 shots of 0.5 mg of gold loaded nant env in the infected culture from which it had been with 1.25 mg of DNA. The large number of shots with the cloned (Fig. 3 and data not shown). This was shown by helium pulse gun were used for the final DNA boosts in the presence of limited or undetectable heteroduplexes all groups. for envs hybridized with sequences from the culture from which they had been cloned (Fig. 3, lanes 1, 5, 9, 12). RESULTS The 6B env clone had V3 loop sequence with features of both the 5 and 6 series (Fig. 2). Hetroduplex tracking Cloning and construction of NL4-3/env recombinants analysis of this env confirmed that the 6B env belonged for two series of patient isolates to the patient 6 series (Fig. 3, lanes 10–16). This was Env regions from two series of Swedish patient iso- evidenced by heteroduplexes formed with DNA from pa- lates were PCR amplified from DNA isolated from in- tient 6 cultures having greater complementarity (migrat- fected PBMC and substituted into a replication-compe- ing more rapidly) than hetroduplexes with DNA from pa- tent HIV-1-NL4-3 backbone to produce NL4-3/env chime- tient 5 cultures.

FIG. 2. Designations, cloning vectors, and signature sequence data for the molecularly cloned Envs. The cysteines bounding the V3 loop are boxed.

AID VY 8445 / 6a2a$$$182 02-04-97 03:43:30 vira AP: Virology REPLICATION PHENOTYPES AND IMMUNOGENICITY OF HIV-1 ENVS 273

env recombinants (Table 1). NL4-3/5B and NL4-3/5C ex- hibited the most restricted growth, barely replicating on PBMCs. NL4-3/6A had an intermediate replication phe- notype, replicating well on PBMCs but poorly on T-cell lines. The remaining recombinants had high replication kinetics, yet distinctive tropisms for the four tested T-cell lines. The NL4-3 parent exhibited the best replication kinetics, replicating well on PBMCs as well as all tested T-cell lines. The NL4-3/Env recombinants also differed in their abil- ities to induce syncytia in MT-2 cells (Table 1). The four recombinants with high replication kinetics and broad cytotropisms were syncytium-inducing (SI). The two re- combinants with low replication kinetics, as well as the one with an intermediate replication phenotype (NL4-3/ 6A), did not induce syncytia.

Expression and immunogenicity of the gp120 and FIG. 3. Use of heteroduplex tracking analyses to establish the rela- gp140 vaccine plasmids tionship of the molecularly cloned Envs to the envelope sequences A representative group of Envs were selected for im- present in infected PBMCs. The PBMC DNAs and Env clones used in the analysis are indicated at the top of the gel. The positions of single- munogenicity studies. These were subcloned into the stranded DNA and homoduplexes are on the right side. The hetero- pJW4303 vaccine vector in gp120 and gp140 forms (see duplexes run between the single-stranded DNA and the homoduplexes. Figs. 1D and 1E). Western blot analyses revealed that each of these vaccine subclones expressed the antici- pated form of Env in transiently transfected COS cells Effects of env sequences from the patient isolates on (Fig. 4 and data not shown). The gp140 form underwent the growth characteristics of pNL4-3 partial proteolytic cleavage at the SU/TM boundary, as The cloned env sequences conferred distinct replica- evidenced by a gp140 form with slightly more rapid mo- tion kinetics on the NL4-3/env recombinant viruses (Ta- bility than full-length Env (gp160) and a gp120 form (the ble 1). Replication studies were initiated by cocultivating SU subunit of gp160). COS cells transiently transfected with the recombinant The total levels of expression of the vaccine plasmids viruses with PBMCs. Within each patient series, trans- in transiently transfected COS cells were fairly similar fected COS cells showed less than a 2.5-fold variation (Table 2). With one exception, both forms of all Envs were in the levels of the expression of the transfected virus or expressed at 12 to 62 ng per transfection. This level of a control cotransfected hGH plasmid (Table 1 and data expression was similar to that of the gp120 and gp140 not shown). In contrast, the cocultivations revealed forms of the T-cell-line-adapted HXB-2 Env. The one ex- marked differences in the replication phenotypes of the ception, the gp120 form of 5B, was expressed at much

TABLE 1

Cytotropism and Syncytium Inducing Activity of NL4-3/Env Recombinants

Transfection Cytotropism: Syncytium DNA efficiencya Expressionb PBMC MT-2 H9 A3.01 CEMx174 inductionc pNL4-3/5A 18 10% (3.7) ///// //// // // / SI pNL4-3/5B 31 15% (4.9) { 0{00 (NSI) pNL4-3/5C 15 11% (2.6) { 0{00 (NSI) pNL4-3/6A 10 7% (2.4) ///{{00 NSI pNL4-3/6B 6 8% (1.9) ///// ///// //// // ///// SI pNL4-3/6C 10 15% (3.6) ///// ///// /// // // SI pNL4-3/6D 14 9% (2.3) ///// ///// // / // SI pNL4-3 parent 17 13% (4.2) ///// ///// ///// ///// ///// SI

a Nanogram of human growth hormone per milliliter of culture supernate at 72 hr posttransfection. b Percentage of transfected COS cells scoring for virus proteins using an indirect immunofluorescence assay (relative units of HIV-1 proteins in COS cell lysates). The assays were performed at 72 hr posttransfection. c SI, syncytium induction; NSI, non-syncytium-inducing; (NSI) syncytium not observed; however, poor replication of the virus could have precluded the formation of syncytia. Syncytia were scored according to Koot et al. (1992).

AID VY 8445 / 6a2a$$$183 02-04-97 03:43:30 vira AP: Virology 274 MUSTAFA ET AL.

FIG. 4. Example of Western blot analyses of pJW4303-expressed gp120 and gp140. The Envs under analysis are indicated at the tops of the gels and sizes of the marker HXB-2 gp160 Env on the sides. All envs were expressed in transiently transfected COS cells. lower levels of 2 ng per transfection. In these transfec- tions, the cotransfected hGH plasmid was expressed at similar levels to those observed in transfections with the other vaccine plasmids (118 to 282 ng per transfection). This suggested that the low level of expression of the 5B gp120 was specific to sequences in this construct. FIG. 5. Analysis of the ability of the gp120 and gp140 forms of the The extent of secretion of the gp120 and gp140 pro- patient 5 Envs to raise antibody in NZW rabbits. Gene gun DNA immuni- teins was assessed in COS cells by using an antigen zations are indicated as dotted lines. Symbols used for rabbits follow: capture ELISA to quantitate Env in cell lysates and cul- 5A gp120, R41(᭺), R42(ᮀ); 5B gp120, R28(᭺), R29(ᮀ), R30(᭝); 5B gp120, ture medium. These analyses revealed that the gp120 R43(᭺), R44(ᮀ), R45(᭝); 5B gp140, R61(᭺), R62(ᮀ), R88(᭝). forms of all of the Envs were efficiently secreted. Be- tween 92 and 95% of the total gp120 was found in the ciently secreted. Only 18 to 60% of these proteins were culture medium (Table 2). The gp140s were less effi- present in culture medium (Table 2). Despite the fairly similar levels of expression in trans- fected COS cells, the DNA-expressed Envs differed in TABLE 2 their abilities to raise antibody in rabbits (Table 2, Figs. Expression and Immunogenicity of Patient Envs 5 and 6). Multiple DNA immunizations of all of the Envs were required to elicit antibody (Figs. 5 and 6). In in- Expression, ng stances where antibody was raised, titers generally (percentage secreted)a Anti-Env IgG, mg/mlb peaked at 2 weeks postboost and declined by 8 weeks Env gp120 gp140 gp120 gp140 postboost. Both peaks and troughs of antibody titers rose with sequential immunizations. As the number of immuni- HXB-2 12 (92%) 15 (50%) 6.3 3.3 zations increased, there tended to be less than a 10-fold 5.7 1.6 difference between peaks and troughs. 5A 62 (93%) 37 (60%) 8.6 0.6 0.2 0.5 A number of gene gun delivery conditions were used 0.2 over the course of the experiment (see Materials and 5B 2 (94%) 12 (27%) 4.6 0.3 Methods). These differences in conditions did not corre- 4.3 õ late with immunogenicity or antibody titers (see rabbits 2.8 õ with identical spacings of DNA deliveries, Figs. 5 and 6). 6A 42 (93%) 43 (30%) 1.0 0.3 õõThus, the multiple boosts appeared to have evened out õõpotential differences in efficiency of the different immuni- 6C 16 (94%) 24 (18%) õõzation conditions. õõAs a whole, the primary gp120s raised higher titers of 6D 44 (95%) 25 (27%) õõantibody than the primary gp140s (Table 2). Within the õõ õõgp120s, the patient 5 gp120s were more immunogenic than the patient 6 gp120s (Table 2). The patient 5 gp120s a Numbers are nanograms of Env per transfection. Secreted protein raised antibody in each of five immunized rabbits is represented as the percentage of the total expressed protein present whereas the patient 6 gp120s raised antibody in only in the culture medium. The transfection efficiency was monitored using one of eight immunized rabbits. Peak anti-Env titers for human growth hormone which ranged from 114 to 282 ng per transfec- tion. the patient 5 gp120s ranged from 0.2 to 8.6 mg of specific b Numbers are peak titers for individual rabbits, listed from the high- IgG per milliliter, whereas the one seroconverter for pa- est to the lowest responder for each group. tient 6 gp120s had a peak titer of 1 mg of specific IgG

AID VY 8445 / 6a2a$$$183 02-04-97 03:43:30 vira AP: Virology REPLICATION PHENOTYPES AND IMMUNOGENICITY OF HIV-1 ENVS 275

gp140 form of HXB-2 raised lower titers of antibody than the gp120 form. However, these titers were ‘‘high’’ (1.6 and 3.3 mg/ml) when compared with those raised by patient gp140s (£0.6 mg/ml) (Table 2).

DISCUSSION Our data demonstrate that the immunogenicity of gp120 and gp140 forms of Env from primary isolates correlates with the patient from which the Env was cloned and not the cytotropism or syncytium-inducing potential of the Env (Tables 1 and 2). For all of the Envs, the gp120 form was more immunogenic than the corresponding gp140 form (Table 2). The gp120 forms of the patient 5 primary Envs were as immunogenic as the gp120 form of HXB-2, a T-cell line-adapted Env (Table 2). In contrast, the gp140 forms of the primary Envs were less immunogenic than the gp140 form of HXB-2 (Table 2).

Influence of source and form of Env on immunogenic potential Within the two series of primary Envs, the source (pa- tient 5 or patient 6) and the form (gp120 versus gp140) of the Env had the most effect on the ability to raise antibody (Table 2 and Figs. 5 and 6). The gp120 forms of patient 5, but not patient 6 Envs, had good immunoge- nicity. With the exception of one low responding rabbit FIG. 6. Analysis of the ability of the gp120 and gp140 forms of the immunized with the 5A gp120, antibody titers elicited by patient 6 Envs to raise antibody in NZW rabbits. Gene gun DNA immuni- the gp140 forms of both patient 5 and 6 Envs were lower zations are indicated as dotted lines. Symbols used for rabbits follow: 6A gp120, R46(᭺), R47(ᮀ), R48(᭝); 6A gp140, R31(᭺), R32(ᮀ), R33(᭝); than the corresponding gp120s. These results suggest 6C gp120, R49(᭺), R51(ᮀ); 6C gp140, R64(᭺), R65(ᮀ), R66(᭝); 6D that immunodominant epitopes on both patient 5 and 6 gp120, R52(᭺), R53(ᮀ), R75(᭝); 6D gp140, R34(᭺), R35(ᮀ), R36(᭝). Envs were exposed on the monomeric gp120 form, but sequestered on the gp140 form that is likely to be oligo- meric (see Moore and Sodroski, 1996; for reviews, see per milliliter. The low level of expression of the 5B gp120 Moore and Ho, 1995; Kent and Robinson, 1996). These in COS cells did not appear to limit its immunogenicity results also suggest that the immunogenicity of the (Tables 1 and 2). ‘‘oligomer-masked’’ epitopes on the patient 5 and patient For both patients, the gp140 forms of Env were less 6 gp120s were not related to the replication phenotypes immunogenic than the corresponding gp120 forms (Ta- of these Envs (Table 1). ble 2). In the patient 5 series, this was reflected in both The immunogenicities of the patient gp120s over- lower frequencies of seroconversion (four of six rabbits lapped that of the T-cell line-adapted HXB-2 gp120. In immunized with gp140-expressing plasmids as opposed contrast, the immunogenicities of the patient gp140s to five of five rabbits immunized with gp120-expressing were consistently lower than that of the HXB-2 gp140. plasmids), and lower titers of antibody in rabbits that did The HXB2 gp140 may be more immunogenic because of seroconvert (up to 0.6 mg/ml for gp140 versus up to 8.6 the inherent instability of Env spikes from T-cell line- mg/ml for gp120) (Table 2). For the less immunogenic adapted viruses. Indeed T-cell-line-adapted strains of patient 6 Envs, only one of nine rabbits immunized with HIV-1 are known to shed monomeric gp120 more easily plasmids expressing the gp140 form seroconverted. This than primary isolates (for review see Moore and Ho, rabbit developed a lower titer (0.3 mg/ml) than the one 1995). This instability would lead to exposure of the epi- rabbit seroconverting with the gp120 plasmid (1.0 mg/ml) topes on gp140 that are masked in the oligomer causing (Table 2). the antigenicity of HXB-2 gp140 to approach that of HXB- The most distinctive difference in the immunogenicity 2 gp120. of the primary Envs and the HXB-2-T-cell-line-adapted All of the ELISAs in this study used a IIIb-Env as the Env was the relatively high immunogenicity of the gp140 solid-phase antigen. This common Env could have bi- form of HXB-2 (Table 2). As with the patient Envs, the ased our results by failing to detect type-specific determi-

AID VY 8445 / 6a2a$$$183 02-04-97 03:43:30 vira AP: Virology 276 MUSTAFA ET AL. nants in the patients 5 and 6 series. However, we con- and 6D Envs (all ‘‘rapid/high’’), amply demonstrate that sider a strong bias unlikely as HIV-1 Envs (even those the ‘‘rapid/high’’ phenotype does not necessarily make from different clades) cross-react in binding assays an Env immunogenic. (Moore et al., 1996). Attempts to use FACS analyses of infected cells to titer rabbit sera revealed that the titers Low immunogenic potential of Env of anti-Env IgG were below the threshold of detection for The DNA-mediated immunizations with the primary this assay. Envs, as well as the T-cell line-adapted HXB-2 Env, elic- Lack of correlation between the growth potential and ited quite low titers of anti-Env IgG (Figs. 5 and 6, Table the immunogenic potential of patient Envs 2). This low immunogenicity is likely to be related to the highly glycosylated nature of the HIV-1 Env (Leonard et One of the hypotheses behind this study was that the al., 1990; Myers et al., 1994). HIV-1 Envs have as many replication phenotype of an Env would influence the im- as 25–30 N-linked glycosylation sites on each monomer munogenicity of the Env. ‘‘Slow/low’’ Envs may sequester (ú75 glycosylation sites per oligomer). This level of gly- epitopes needed for entry; a sequestration that could cosylation is 5 to 6 times higher than that for the affect in turn the ability of these Envs to raise neutralizing and glycoproteins. Single DNA antibody. immunizations with these can raise con- Interestingly, marked differences in replication poten- valescent titers of neutralizing antibody (Robinson et al., tial and syncytium-inducing potential did not influence 1995; Yang et al., in press). Thus the immunodeficiency the immunogenicity of the patient 5 and 6 Envs (Tables viruses, which maintain chronic infections by continuous 1 and 2). Within the patient 5 and 6 series, the Envs that replication in the face of strong immune responses (Ho supported rapid replication were no more immunogenic et al., 1995; Wei et al., 1995), may have been selected for than the Envs that supported low replication. This sug- highly glycosylated envelope glycoproteins that minimize gests that conformational changes which expose epi- the immunogenicity of viral particles. topes involved in CD4 binding and fusion, do not affect This study, as well as less extensive prior studies us- overall immunogenicity. Assays for neutralizing activity ing DNA to raise antibody to HIV-1 Envs (Wang et al., have also not revealed a correlation between replication 1993a; 1993b; Fuller and Haynes, 1994; Lu et al., 1995; potential and the ability to raise neutralizing antibody Okuda et al., 1995), amply demonstrate that DNA-based (manuscript in preparation). immunizations do not overcome the low immunogenicity of immunodeficiency envelope glycoproteins. The chal- Role of Env in replication phenotypes lenge for future studies will be to develop constructs and protocols that can overcome this low immunogenicity. The NL4-3/env recombinants exhibited a spectrum of replication potentials, cytotropisms and abilities to in- duce syncytia (Table 2). These data are in agreement ACKNOWLEDGMENTS with a large body of data correlating mutations in Env We are indebted to Drs. J. Arthos and J. I. Mullins for design and with effects on growth phenotypes. However, for the pa- provision of pJW4303, Dr. J. Sodroski for the provision of psvIIIenv; Drs. tient 5 series of Envs, the growth characteristics con- P. Earl and C. Broder for the provision of vCB-14; and the AIDS Reposi- ferred on NL4-3 were the opposite of the original isolates tory (Rockville, MD) for the provision of T-cell lines and the plasmids pNL4-3 and pHXB-2. We thank Drs. J. Sullivan and M. Somasundaran (Table 1, von Gegerfelt et al., 1991). The 5A Env from a for provision of PBMCs. This work was supported by Public Health ‘‘slow/low’’ isolate conferred ‘‘rapid/high’’ growth on the Service Grant RO1 A1 34241. NL4-3/env recombinant; whereas the 5B and 5C Envs from ‘‘rapid/high’’ isolates conferred poor growth. This REFERENCES was not due to cloning of nonrepresentative Envs from Adachi, A., Gendelman, H. E., Koenig, S., Folks, T., Willey, R., Rabson, A., the PBMC cultures used to propagate the isolates (Fig. and Martin, M. A. (1986). Production of acquired immunodeficiency 3). Factors that could have contributed to the unexpected syndrome-associated in human and non-human cells growth phenotypes for the molecularly cloned Envs may transfected with an infectious molecular clone. J. Virol. 59, 284–291. include: (i) selection in culture (prior to cloning) for more A˚ sjo¨, B., Albert, J., Karlsson, A., Morfeldt-Manson, L., Biberfeld, G., Lid- rapidly growing variants of the 5A ‘‘slow/low’’ isolate, (ii) man, K., and Fenyo¨, E. M. (1986). Replicative properties of human immunodeficiency virus from patients with varying severity of HIV the introduction of mutations during cloning that ad- infection. Lancet 2, 660–662. versely affected the function of the ‘‘rapid/high’’ 5B and Bou-Habib, D. C., Roderiquez, G., Oravecz, T., Berman, P. W., Lusso, P., 5C Envs, and (iii) poor function of chimeric env, , or and Norcross, M. A. (1994). Cryptic nature of envelope V3 region sequences in the NL4-3/5B and NL4-3/5C recombinants epitopes protects primary monocytotropic human immunodeficiency (see Fig. 1 for chimeric proteins). Given that poor function virus type 1 from antibody neutralization. J. Virol. 68, 6006–6013. Chapman, B. S., Thayer, R. M., Vincent, K. A., and Haigwood, N. L. of the NL4-3/5B recombinant could have obscured an (1991). Effect of intron A from human cytomegalovirus (Towne) imme- actual ‘‘rapid/high’’ phenotype for the 5B Env, the very diate-early gene on heterologous expression in mammalian cells. low (undetectable) immunogenicity of the patient 6B, 6C, Nucleic Acids Res. 19, 3979–3986.

AID VY 8445 / 6a2a$$$183 02-04-97 03:43:30 vira AP: Virology REPLICATION PHENOTYPES AND IMMUNOGENICITY OF HIV-1 ENVS 277

Cheng-Mayer, C., Seto, D., Tateno, M., and Levy, J. A. (1988). Biologic Lu, S., Arthos, J., Montefiori, D. C., Yasutomi, Y. Manson, K., Mustafa, features of HIV-1 that correlate with virulence in the host. Science F., Johnson, E., Santoro, J. C., Wissink, J., Mullins, J. I., Haynes, J. R., 240, 80–82. Letvin, N. L., Wyand, M., and Robinson, H. L. (1996). Simian immuno- Delwart, E. L., Sheppard, H. W., Walker, B. D., Goudsmit, J., and Mullins, deficiency virus DNA vaccine trial in macaques. J. Virol. 70, 3978– J. I. (1994a). Human immunodeficiency virus type 1 evolution in vivo 3991. tracked by DNA heteroduplex mobility assays. J. Virol. 68, 6672– Moore, J. P., and Ho, D. D. (1995). HIV-1 neutralization: The conse- 6683. quences of viral adaptation to growth on transformed t cells. AIDS Delwart, E. I., Shpaer, E. G., Louwagie, J., McCutchan, F. E., Grez, M., 9(suppl. A), S117–S136. Rubsamen-Waigmann, H., and Mullins, J. I. (1994b). Genetic relation- Moore, J. P., and Sodroski, J. (1996). Antibody cross-competition analy- ships determined by a DNA heteroduplex mobility assay: Analysis sis of the human immunodeficiency virus type 1 gp120 exterior enve- of HIV-1 env genes. Science 262, 1257–1261. lope glycoprotein. J. Virol. 70, 1863–1872. Earl, P. L., Broder, C. C., Long, D., Lee, S. A., Peterson, J., Chakrabarti, Moore, J. P., Cao, Y., Leu, J., Qin, L., Korber, B., and Ho, D. D. (1996). S., Doms, R. W., and Moss, B. (1994). Native oligomeric human immu- Inter- and intraclade neutralization of human immunodeficiency virus nodeficiency virus type 1 envelope glycoprotein elicits diverse mono- type 1: Genetic clades do not correspond to neutralization serotypes clonal antibody activities. J. Virol. 68, 3015–3026. but partially correspond to gp120 antigenic serotypes. J. Virol. 70, Eisenbraun, M. D., Fuller, D. H., and Haynes, J. R. (1993). Examination 427–444. of parameters affecting the elicitation of humoral immune responses Myers, G., Wain-Hobson, S., Henderson, L. E., Korber, B., Jeang, K.-T., by particle bombardment-mediated genetic immunization. DNA Cell and Pavlakis, G. N. (1994). ‘‘Human and AIDS 1994,’’ Biol. 12, 791–797. Theoretical Biology and Biophysics Group, Los Alamos National Lab- Fenyo¨, E. M., Albert, J., and A˚ sjo¨, B. (1989). Replicative capacity, cyto- oratory, Los Alamos, NM. pathic effect and cell tropism of HIV. AIDS 1989 3(suppl. 1), S5-S12. Okuda, K., Bukawa, H., Hamajima, K., Kawamoto, S., Sekigawa, K.-I., Fenyo¨, E. M., Albert, J., and McKeating, J. (1996). The role of the humoral Yamada, Y., Tanaka, S.-I., Ishii, N., Aoki, I., Nakamura, M., Yamamoto, immune response in HIV infection. AIDS 1996 10(suppl. A), S97– H., Cullen, B. R., and Fukushima, H. (1995). Induction of potent hu- S106. moral and cell-mediated immune responses following direct injection Fenyo¨, E. M., Morfeldt-Manson, L., Chiodi, F., Lind, B., von Gegerfelt, of DNA encoding the HIV type 1 env and rev gene products. Aids A., Albert, J., Olausson, E., and Asjo, B. J. (1988). Distinct replicative Res. Hum. Retroviruses 11, 933–943. and cytopathic characteristics of human immunodeficiency virus iso- Pertmer, T. M., Eisenbraun, M. D., McCabe, D., Prayaga, S. K., Fuller, lates. J. Virol. 62, 4414–4419. D. H., and Haynes, J. R. (1995). Gene gun-based nucleic acid immuni- Fernandez-Larsson, R., Srivastava, K. K., Lu, S., and Robinson, H. L. zation: Elicitation of humoral and cytotoxic T lymphocyte responses (1992). Replication of patient isolates of human immunodeficiency following epidermal delivery of nanogram quantities of DNA. Vaccine virus type 1 in T cells: A spectrum of rates and efficiencies of entry. 13, 1427–1430. Proc. Natl. Acad. Sci. USA 89, 2223–2226. Reiter, J. N., Sergel, T., and Morrison, T. G. (1995). Mutational analysis Fuller, D. H., and Haynes, J. R. (1994). A qualitative progression in HIV of the leucine zipper motif in the Newcastle disease virus fusion type 1 glycoprotein 120-specific cytotoxic cellular and humoral im- protein. J. Virol. 69, 5995–6004. mune responses in mice receiving a DNA-based glycoprotein 120 Richmond, J. F. L., Mustafa, F., Lu, S., Santoro, J. C., Weng, J., O’Connell, vaccine. AIDS Res. Hum. Retroviruses 10, 1433–1441. Gibbs, J. S., Regier, D. A., and Desrosiers, R. C. (1994). Construction and M., Fenyo, E. M., Hurwitz, J. L., Montefiori, D. C., and Robinson, H. L., in vitro properties of SIVmac mutants with deletions in ‘‘nonessential’’ submitted. genes. AIDS Res. Hum. Retroviruses 10, 333–342. Robinson, J. E., Holton, D., Lui, J., McMurdo, H., Murciano, A., and Helseth, E., Kowalski, M., Gabuzda, D., Olshevsky, U., Haseltine, W., Gohd, R. (1990). A novel enzyme-linked immunosorbent assay and Sodroski, J. (1990). Rapid complementation assays measuring (ELISA) for the detection of antibodies to HIV-1 envelope glyco- replicative potential of human immunodeficiency virus type 1 enve- proteins based on immobilization of viral glycoproteins in microti- lope glycoprotein mutants. J. Virol. 64, 2416–2420. ter wells coated with concanavalin A. J. Immun. Methods 132, Ho, D. D., Neumann, A. U., Perelson, A. S., Chen, W., Leonard, J. M., 63–71. and Markowitz, M. (1995). Rapid turnover of plasmid virions and CD4 Robinson, H. L., Feltquate, D. M., and Morin, M. L. (1995). DNA vac- lymphocytes in HIV-1 infection. Nature (London) 373, 123–126. cines: A new approach to immunization. In ‘‘Vaccines 95’’ (F. Brown, Kabat, D., Kozak, S. L., Wehrly, K., and Chesebro, B. (1994). Differences R. Chanock, and H. Ginsberg, Eds.), pp. 65–70. Cold Spring Harbor in CD4 dependence for infectivity of laboratory-adapted and primary Press, Cold Spring Harbor, NY. patient isolates of human immunodeficiency virus type 1. J. Virol. 68, Stamatos, L., and Cheng-Mayer, C. (1995). Structural modulations of 2570–2577. the envelope gp120 glycoprotein of human immunodeficiency virus Kent, K. A., and Robinson, J. (1996). Antigenic determinants on HIV-1 type 1 upon oligomerization and different V3 loop epitope exposure envelope glycoproteins: a dickens of a time with oligomer twist. AIDS of isolates displaying distinct tropism upon virion-soluble receptor 1996 10(suppl. A), S107–S114. binding. J. Virol. 69, 6191–6198. Koot, M., Vos, A. H. V., Keet, R. P. M., de Goede, R. E. Y., Dercksen, Sullivan, N., Sun. Y., Li, J., Hoffmann, W., and Sodroski, J. (1995). Replica- M. W., Terpstra, F. G., Coutinho, R. A., Miedema, F., and Tersmette, M. tive function and neutralization sensitivity of envelope glycoproteins (1992). HIV-1 biological phenotype in long-term infected individuals from primary and T-cell line-passaged human immunodeficiency vi- evaluated with an MT-2 cocultivation assay. AIDS 1992 6, 49–54. rus type 1 isolates. J. Virol. 69, 4413–4422. Leonard, C. K., Spellman, M. W., Riddle, L., Harris, R. J., Thomas, J. N., Tersmette, M., de Goede, R. E. Y., Al, G. J. M., Winkel, I. N., Gruters, R. A., and Gregory, T. J. (1990). Assignment of intrachain disulfide bonds Cuypers, H. T., Huisman, H. G., and Miedema, F. (1988). Differential and characterization of potential glycosylation sites of the type 1 syncytium-inducing capacity of human immunodeficiency virus iso- recombinant human immunodeficiency virus envelope glycoprotein lates: Frequent detection of syncytium-inducing isolates in patients (gp120) expressed in Chinese hamster ovary cells. J. Biol. Chem. with acquired immunodeficiency syndrom (AIDS) and AIDS-related 265, 10373–10382. complex. J. Virol. 62, 2026–2032. Lu, S., Santoro, J. C., Fuller, D. H., Haynes, J. R., and Robinson, H. L. Tersmette, M., Gruters, R. A., de Wolf, F., de Goede, R. E. Y., Lange, (1995). Use of DNAs expressing HIV-1 Env and noninfectious HIV-1 J. M. A., Schellenkens, P. T. A., Goudsmit, J., Huisman, H. G., and particles to raise antibody responses in mice. Virology 209, 147– Miedeman, F. (1989). Evidence for a role of virulent human immuno- 154. deficiency virus (HIV) variants in the pathogenesis of acquired immu-

AID VY 8445 / 6a2a$$$184 02-04-97 03:43:30 vira AP: Virology 278 MUSTAFA ET AL.

nodeficiency syndrome: Studies on sequential HIV isolates. J. Virol. M., Dang, K., Agandjanan, M., Gilbert, L., Ugen, K. E., Williams, W. V., 63, 2118–2125. and Weiner, D. B. (1993b). DNA inoculation induces neutralizing im- von Gegerfelt, A., Albert, J., Morfeldt-Manson, L., Broliden, K., and Fenyo¨, mune responses against human immunodeficiency virus type 1 in E. M. (1991). Isolate-specific neutralizing antibodies in patients with mice and nonhuman primates. DNA Cell Biol. 12, 799–805. progressive HIV-1 related disease. Virology 185, 162–168. Wei, X., Ghosh, S. K., Taylor, M. E., Johnson, V. A., Emini, E. A., Deutsch, Wang, B., Ugen, K. E., Srikantan, V., Agadjanyan, M. G., Dang, K., Refaeli, P., Lifson, J. D., Bonhoeffer, S., Nowak, M. A., Hahn, B. H., Saag, M. S., Y., Sata, A. I., Boyer, J. Williams, W. V., and Weiner, D. B. (1993a). and Shaw, G. M. (1995). Viral dynamics in human immunodeficiency Gene inoculation generates immune responses against human im- virus type 1 infection. Nature (London) 373, 117–122. munodeficiency virus type 1. Proc. Natl. Acad. Sci. USA 90, 4156– Yang, K., Mustafa, F., Valsamakis, A., Santoro, J. C., Griffin, D. E., and 4160. Robinson, H. L., (1996). Early studies on DNA-based immunizations Wang, B., Boyer, J., Srikantan, V., Coney, L., Carrano, R., Phan, C., Merva, for measles virus. Vaccine, in press.

AID VY 8445 / 6a2a$$$184 02-04-97 03:43:30 vira AP: Virology