Construction and characterization of a poliovirus/rhinovirus antigenic hybrid.

In order to study the properties of foreign antigenic sites expressed on poliovirus a hybrid was constructed in which neutralization antigenic site IA of poliovirus type 1 strain Mahoney [PV1(M)] was replaced by neutralization immunogenic site IA (NImIA) of human rhinovirus 14 (HRV14). The resulting hybrid was viable, but growth was impaired in comparison to PV1(M). The hybrid expressed both PV1(M) and HRV14 antigenic determinants. When inoculated into rabbits it elicited neutralizing antibodies against both PV1(M) and HRV14. Furthermore, the hybrid was efficiently neutralized by polyclonal antisera specific for either PV1(M) or HRV14 and by three out of five monoclonal antibodies directed to NImIA. The monoclonal antibodies also blocked binding of the hybrid to the cellular receptor for poliovirus. One of them is thought to neutralize rhinovirus in this manner, and it appears that NImIA is expressed in a sufficiently favorable context on the hybrid for the same mechanism to be effective. This can be interpreted to mean that the interactions between the parental viruses and their respective cellular receptors are very similar.     

Host range of poliovirus is restricted to simians because of a rapid sequence change of the poliovirus receptor gene during evolution

Summary.  The host range of most poliovirus (PV) strains is restricted to simians. This host range specificity is believed to be determined by the interaction between PV and its receptor molecule. To elucidate the molecular basis of this species-specific infection of PV, we cloned orthologs of the PV receptor (PVR) gene (pvr) as well as those of PV receptor-related genes 1 and 2 (prr1 and prr2) from various mammalian species. These three genes are widely present in mammalian genomes including those of non-susceptible species. Comparison of the deduced amino acid sequences of PVR orthologs revealed that the NH₂-terminal immunoglobulin-like domain (domain 1), which is the virus binding site in the human PVR, is highly variable among species, whereas that of PRR1 is highly conserved. Domain 1 of the PVR orthologs for the ring-tailed lemur and rabbit, which are not susceptible to PV, show only 51 and 61% amino acid sequence identity to that of human PVR, respectively. Chimeric PVR proteins that have the domain 1 of the ring-tailed lemur and rabbit PVRs failed to serve as receptors for PV. These results suggest that rapid changes in the domain 1 sequence during mammalian evolution determined the host range restriction of PV. Note: Nucleotide sequence data reported are available in the DDBJ/EMBL/ GenBank databases under the following accession numbers; brown capuchin pvr: AB086124 to AB086131, ring-tailed lemur pvr: AB086132 to AB086137, rabbit pvr: AB086138 to AB086144, pvr gene fragment of squirrel monkey: AB086252, gorilla: AB086253, common marmoset: AB086254, chimpanzee: AB086255, prr1 gene fragment of African green monkey: AB086001, bovine: AB086002, brown capuchin: AB086003, squirrel monkey: AB086004, ring-tailed lemur: AB086005, rabbit: AB086006, gorilla: AB086007, dog: AB086008, common marmoset: AB086009, dolphin: AB086010, chimpanzee: AB086145, and mole: AB086146. prr2 gene fragment of African green monkey: AB086147, bovine: AB086148, brown capuchin: AB086149, squirrel monkey: AB086150, ring-tailed lemur: AB086151, rabbit: AB086152, gorilla: AB086153, dog: AB086154, common marmoset: AB086155, dolphin: AB086156, chimpanzee: AB086157, mole: AB086158. Received June 14, 2002; accepted August 22, 2002     

Mapping of protein domains of hepatitis A virus 3AB essential for interaction with 3CD and viral RNA

The small hydrophobic protein 3AB of the picornaviruses, encompassing the replication primer 3B, has been suggested to anchor the viral replication complex to membranes. For hepatitis A virus (HAV) 3AB, we have previously demonstrated its ability to form stable homodimers, to bind to membranes, and to interact specifically with RNA, implicating its multiple involvement in viral replication. In the present report, we show that HAV 3AB additionally interacts with HAV protein 3CD, a feature also described for the corresponding polypeptide of poliovirus. By assessing the interactions of three deletion mutants, distinct domains of HAV 3AB were mapped. The hydrophobic domain and the 3B moiety were found to be essential for the 3AB interaction with 3CD. Both electrostatic and hydrophobic forces are involved in this interaction. The cluster of charged amino acid residues at the C terminus of 3A seems to determine the specificity of 3AB interaction with RNA structures formed at either terminus of the HAV genome. Furthermore, our data implicate that 3A can interact with HAV RNA. Compared with poliovirus 3AB, which by itself is a nonspecific RNA-binding protein, HAV 3AB specifically recognizes HAV RNA structures that might be of relevance for initiation of viral RNA replication.     

Genetic analysis of an NTP-binding motif in poliovirus polypeptide 2C.

Poliovirus polypeptide 2C is a nonstructural protein involved in replication of the viral genome. Analysis of the primary amino acid sequence of 2C shows homology to a family of proteins which contain a nucleoside-triphosphate (NTP)-binding motif. This motif consists of elements "A" (2/5 hydrophobic stretch) G/AXXGXGKS/T, where X stands for any amino acid, and "B" (3/5 hydrophobic stretch) D or DD/E. To assess the significance of the consensus sequence in 2C, we have engineered point mutations into the most conserved residues in the A and B sites and tested their effect on viral RNA replication in vivo and translation in vitro. Whereas in vitro translation of synthetic RNAs carrying mutations in the NTP-binding motif showed efficient processing of all viral proteins, indistinguishable from that of the parental strain, transfection of the RNAs into HeLa cells did not give rise to infectious virus. No viral RNA replication could be detected in cells transfected with mutant RNAs. However, revertants to the wild-type genotype in the A and B sites were obtained which gave rise to wild-type RNA synthesis, but pseudorevertants or second-site suppressors were not observed. Thus, viral RNA synthesis is greatly reduced but not entirely abolished in cells transfected with mutant RNAs. These results strongly suggest a functional role for the proposed NTP-binding motif of 2C in RNA replication and proliferation of poliovirus.     

Pyrimidine-rich region mutations compensate for a stem-loop V lesion in the 5' noncoding region of poliovirus genomic RNA

Five revertants of a linker-scanning mutation adjacent to the stem-loop V attenuation determinant (X472) in the 5' noncoding region of poliovirus RNA were independently isolated from neuroblastoma cells and contained RNAs with seven nucleotide changes in the pyrimidine-rich region. Generation of the identical rare second-site mutations suggests the existence of a replicase-dependent mutagenesis mechanism during poliovirus replication. Enzymatic structure probing of the mutated pyrimidine-rich domain identified secondary structure changes between stem-loops V and VI. A consensus secondary structure model is presented for wild-type stem-loops V and VI and the pyrimidine-rich region located in the 5' noncoding region of poliovirus RNA. A pyrimidine-rich region mutant (X472-R4N) produced large plaques in neuroblastoma cells and small plaques in HeLa cells, but the plaque size differences were not due to cell-type differences in viral translation or RNA replication. Release of X472-R4N from HeLa cells was 10-fold lower than release from neuroblastoma cells, which may explain the small plaque phenotype of X472-R4N in HeLa cells. Wild-type poliovirus was also released more efficiently from neuroblastoma cells (approximately 4-fold increase compared with release from HeLa cells), indicating that poliovirus neurotropism may be influenced by the cell-type efficiency of virus release. Thermal treatment increased the levels of infectious X472-R4N virions but not wild-type virus particles; thus RNA sequence and structural changes in the mutated 5' noncoding region of X472-R4N may have altered RNA-protein interactions necessary for virus infectivity.     

Membrane-associated respiratory syncytial virus F protein expressed from a human rhinovirus type 14 vector is immunogenic

Human rhinovirus (HRV) replicons have the potential to serve as respiratory vaccine vectors for mucosal immunization in humans. However, since many vaccine immunogens of interest are glycosylated, an important concern is whether HRV replicons are capable of expressing glycosylated proteins. The human respiratory syncytial virus (RSV) fusion (F) protein was chosen as a model glycoprotein and the HRV replicon DeltaP1FVP3 was generated by inserting the F protein-coding sequence in frame and in lieu of the 5' proximal 1489 nucleotides of the capsid-coding segment in the HRV-14 genome. When transfected into H1-HeLa cells, DeltaP1FVP3 replicated and led to the expression of the F protein. Inhibition with guanidine demonstrated that F-protein expression was dependent on DeltaP1FVP3 replication and did not result from translation of input RNA. Although most of the F protein remained as an immature, glycosylated precursor (F0), a readily detectable fraction of the protein was processed into the mature glycosylated subunit F1, an event known to occur within the Golgi apparatus. Packaged DeltaP1FVP3 replicons were generated in transfected HeLa cells by coexpression of homologous HRV capsid proteins using the vaccinia virus/T7 RNA polymerase hybrid system. Packaged replicon RNAs were capable of infecting fresh cells, leading to accumulation of the F protein as in RNA-transfected cells. Mice immunized with HeLa cell lysates containing F protein expressed from DeltaP1FVP3 produced neutralizing antibodies against RSV. These results indicate that an HRV-14 replicon can express a foreign glycosylated protein, providing further support for the potential of HRV replicons as a vaccine delivery system.     

Growth phenotypes and biosafety profiles in poliovirus-receptor transgenic mice of recombinant oncolytic polio/human rhinoviruses

The use of oncolytic recombinant polioviruses has an important therapeutic potential in the treatment of human gliomas. This study was carried out to assess parameters of the utility of the oncolytic poliovirus/human rhinovirus type 2 chimeras (PV/HRV2). The prototype PV/HRV2 chimera was constructed containing the complete genome of wild-type PV type 1 (Mahoney) [PV1(M)] in which the cognate IRES was replaced with that of HRV2 [called PV1(RIPO)]. A derivative of PV1(RIPO) is PV1(RIPOS) in which the capsid coding region (P1) was replaced with the capsid-coding region of the PV type 1 (Sabin) [PV1(S)] vaccine strain. In addition, a third PV/HRV2 chimera was constructed containing the complete genome of PV1(S) in which the cognate IRES was replaced with that of HRV2 [termed PVS(RIPO)]. To analyze the growth phenotypes of PV/HRV2 recombinants [PV1(RIPO), PV1(RIPOS), PVS(RIPO)], one-step growth experiments were performed in four human cell lines at three different temperatures. To address the safety profile, PVS(RIPO) was injected into the brain of CD155 tg mice at the dose 10(7) PFU. Then, clinical signs, persistence of the virus in the CNS and genetic stability of PVS(RIPO) replicating in the CNS were evaluated. The data obtained in the present study suggest (i) a correlation between temperature-sensitive (ts) phenotype in both neuronal and non-neuronal cell lines and neuroattenuation in experimental animals, (ii) that PVS (RIPO) is genetically stable on replication in the CNS of poliovirus-susceptible mice. These findings highlight the safety of intracerebral inoculation of PVS(RIPO) for the treatment of human glioma.     

Testing the modularity of the N-terminal amphipathic helix conserved in picornavirus 2C proteins and hepatitis C NS5A protein

The N-terminal region of the picornaviral 2C protein is predicted to fold into an amphipathic alpha-helix that is responsible for the protein's association with membranes in the viral RNA replication complex. We have identified a similar sequence in the N-terminal region of NS5A of hepaciviruses that was recently shown to form an amphipathic alpha-helix. The conservation of the N-terminal region in two apparently unrelated proteins of two different RNA virus families suggested that this helix might represent an independent module. To test this hypothesis, we constructed chimeric poliovirus (PV) genomes in which the sequence encoding the N-terminal 2C amphipathic helix was replaced by orthologous sequences from other picornaviral genomes or a similar sequence from NS5A of HCV. Effects of the mutations were assessed by measuring the accumulation of viable virus and viral RNA in HeLa cells after transfection, examining membrane morphology in cells expressing chimeric proteins and by in vitro analysis of RNA translation, protein processing and negative strand RNA synthesis in HeLa cell extracts. The chimeras manifested a wide range of growth and RNA synthesis phenotypes. The results are compatible with our hypothesis, although they demonstrate that helix exchangeability may be restricted due to requirements for interactions with other viral components involved in virus replication.     

Complete sequence of the RNA genome of human rhinovirus 16, a clinically useful common cold virus belonging to the ICAM-1 receptor group

We report here the complete nucleotide sequence and predicted polyprotein sequence of HeLa cell-adapted human rhinovirus 16 (HRV16). This virus is more suitable than human rhinovirus 14 (HRV14) for clinical studies, and its growth and physical properties are favorable for biochemical and crystallographic analysis. The complete message-sense RNA genome of HRV16 is composed of 7124 bases, not including the poly(A) tail. An open reading frame, extending from base 626 to 7084 predicts a polyprotein containing 2152 amino acid residues. Comparison with other rhinovirus sequences shows HRV16 is much more representative of human rhinoviruses than HRV14. No apparent relationship was found between receptor group and amino acid sequence in VP1, the capsid protein bearing the binding site for the intercullular adhesion molecule-1 (ICAM-1) in both HRV14 and HRV16.Genbank accession number: L24917.     

Poliovirus translation initiation: differential effects of directed and selected mutations in the 5' noncoding region of viral RNAs.

We have analyzed the translational defects of a number of mutations in the 5' noncoding region of poliovirus type 1 RNA. These mutations fall into three categories: (1) two mutations which resulted in temperature sensitive (ts) viruses, (2) the second-site mutations responsible for the reversion of the two ts viruses, and (3) mutations which were lethal to virus production. RNAs containing either of the ts mutations translated in vitro at levels significantly lower than wild-type levels. RNAs containing the respective second-site reversions had corrected these translational defects to levels corresponding to their viral growth potentials. Unlike in vitro translation of wild-type poliovirus RNA, translation of the RNAs which gave rise to ts mutant viruses was not stimulated by the addition of an S10 fraction from an uninfected HeLa cell extract to a rabbit reticulocyte lysate (RRL). In vitro translation of the mutant RNAs (corresponding to the ts viruses) in a RRL was stimulated by factors present in a ribosomal salt wash (RSW) from a HeLa extract, although the levels of stimulation were only half those seen for wild-type. These results suggest that the stimulatory factors present in the RSW have a decreased affinity for the mutant RNA templates but can, to some extent interact, with such RNAs if provided in high enough concentration. The in vitro translation of RNAs containing either of the lethal mutations was not stimulated by factors present in the S10 or the RSW. Taken together, our data suggest a correlation between the ability of a genetically altered RNA to respond to translation stimulatory factors in vitro and the ability of that mutation to be recovered in infectious virus. In addition, we have identified the in vivo-selected reversion of translational defects for two different ts viruses.     

Subdomain specific functions of the RNA polymerase region of poliovirus 3CD polypeptide

The 3D polymerase domain of the poliovirus 3CD polypeptide plays a role in modulating its RNA binding and protein processing activities, even though the proteinase catalytic site and RNA binding determinants appear to reside within the 3C(pro) portion of the molecule. In this study, we have generated recombinant 3CD polypeptides that contain chimeric 3D polymerase domains representing suballelic sequence exchanges between poliovirus type 1 (PV1) and coxsackievirus B3 (CVB3) to determine which portions of the 3D domain are responsible for influencing these activities. By utilizing these recombinant protein chimeras in protein processing and RNA binding studies in vitro, we have generated data suggesting the presence of separate subdomains within the polymerase domain of 3CD that may independently modulate its RNA binding and protein processing activities. In predicting where our sequence exchanges map by utilizing the previously published three-dimensional structure of the PV1 3D polymerase, we present evidence that sequences contained within the RNA recognition motif of the polymerase are critical for 3CD function in recognizing the 5' RNA cloverleaf. Furthermore, our protein processing data indicate that at least some of the substrate recognition and processing determinants within the 3D domain of 3CD are separate and distinct from the RNA binding determinants in this domain.     

Identification of 50- and 23-/25-kDa HeLa cell membrane glycoproteins involved in poliovirus infection: occurrence of poliovirus specific binding sites on susceptible and nonsusceptible cells.

Glycoproteins in the range 50 and 23/25 kDa were identified as poliovirus specific binding sites on HeLa cells with the monoclonal antibody mAb 122. mAb 122 is characterized by its partial inhibiting effect on poliovirus reproduction and adsorption when prebound to HeLa cells. The binding sites are endocytosed in native cells and specific for poliovirus as mAb 122 did not interfere with the adsorption of human rhinovirus type 14 (HRV 14). The poliovirus binding sites are present also on nonprimate so called nonsusceptible cells, e.g., mouse L-cells, as could be shown with sensitive ELISA based binding assays and performance of binding studies with fixed cells at 37 degrees.     

Membrane fractions active in poliovirus RNA replication contain VPg precursor polypeptides

The poliovirus specific polypeptide P3-9 is of special interest for studies of viral RNA replication because it contains a hydrophobic region and, separated by only seven amino acids from that region, the amino acid sequence of the genome-linked protein VPg. Membraneous complexes of poliovirus-infected HeLa cells that contain poliovirus RNA replicating proteins have been analyzed for the presence of P3-9 by immunoprecipitation. Incubation of a membrane fraction rich in P3-9 with proteinase leaves the C-terminal 69 amino acids of P3-9 intact, an observation suggesting that this portion is protected by its association with the cellular membrane. These studies have also revealed two hitherto undescribed viral polypeptides consisting of amino acid sequences of the P2 and P3 regions of the polyprotein. Sequence analysis of stepwise Edman degradation show that these proteins are 3b/9 (Mr 77,000) and X/9 (Mr 50,000). 3b/9 and X/9 are membrane bound and are turned over rapidly and may be direct precursors to proteins P2-X and P3-9 of the RNA replication complex. P2-X, a polypeptide void of hydrophobic amino acid sequences but also found associated with membranes, is rapidly degraded when the membraneous complex is treated with trypsin. It is speculated that P2-X is associated with membranes by its affinity to the N-terminus of P3-9.     

Comparative sequence analysis of the 5'terminal noncoding regions of poliovirus vaccine strain Sabin 1, Sabin 2, and Sabin 3 genomes

Complementary DNAs (cDNAs) were synthesized with virion RNAS from immunologically distinct vaccine strains of poliovirus, that is, Sabin 1 (LSc, 2ab), Sabin 2 (p712, Ch, 2ab), and Sabin 3 (Leon 12a₁b). Restriction endonuclease HaeIII cleavage patterns of these cDNAs suggested striking difference in RNA sequences among these poliovirus serotypes. The virion RNA of Sabin 3 was shown to have a genome-linked protein (VPg) and physical properties of the VPg were identical to those of VPgs of the other poliovirus strains. Virion RNA from each serotype was labeled with ¹²⁵I-Bolton-Hunter reagent, after proteinase K treatment to remove all but a small peptide of the VPg which covalently linked to the 5′ terminus. The specifically 5′-end-labeled RNAs were analyzed by rapid RNA sequencing techniques. Sequences of the first 20 nucleotides were identical in Sabin 1, Sabin 2, and Sabin 3 strains, that is, VPg-pU-U-A-A-A-A-C-A-G-C-U-C-U-G-GG-G-U-U-G. Complete sequence homology was observed between virion RNAs of Sabin 1 and Sabin 2 strains up to the 43rd nucleotide from the 5′ end except for the 25th nucleotide by a “wandering spot” analysis of ¹²⁵I-labeled RNA mixture of these two strains. These results suggested that serotypes of poliovirus were derived by numerous mutations from a putative prototype poliovirus, and that the homologous sequence at 5′ termini, conserved through a long evolutional process, may have essential roles for replication and viral function of poliovirus.     

Substrate requirements of a human rhinoviral 2A proteinase.

The genetic information contained within the RNA genome of picornaviruses is expressed as a single large open reading frame; processing of the primary translation product begins while translation is still in progress. In rhinoviruses and enteroviruses, two picornavirus genera, the virally encoded proteinase 2A begins the processing cascade, cleaving between the C-terminus of VP1 and its own N-terminus. The natural variation in the amino acid sequences amongst rhinoviruses and enteroviruses at the cleavage site of the viral proteinase 2A served as the basis for a mutational analysis of the substrate specificity of the 2A proteinase of human rhinovirus 2. This enzyme was shown to have an unusual preference at the P1 site; out of eight amino acid substitutions made, only the branched amino acids Val and Ile were not readily accepted. The HRV2 2A was shown to process poorly the HRV89 2A cleavage site and to be unable to cleave at sites which included the P' region of poliovirus or HRV14. Furthermore, the 2A of HRV89 preferred the cleavage site of HRV2 to its own.     

A new antigenic site of poliovirus recognized by an intertypic cross-neutralizing monoclonal antibody

A monoclonal antibody (mAb 7J6) neutralizing poliovirus type 2 (PV2) and poliovirus type 1 (PV1) was obtained after immunization of BALB/c mice with infectious PV2, strain MEF-1. Preincubation of mAb 7J6 with PV1 inhibited its binding to PV2 and vice versa. Neutralization-resistant variants of PV2 and PV1 were selected. Nucleotide sequencing of the RNAs of some variants revealed mutations in the loop of amino acid residues 239 to 245 in VP2 and in the loop of amino acid residues 195 to 207 in VP3. This is the first evidence that these two loops contribute to a neutralization antigenic site (N-Ag) for poliovirus. Moreover, this new site on PV2 induced intertypic cross-neutralizing antibodies.     

Genotype-specific in vitro amplification of sequences of the wild type 3 polioviruses from Mexico and Guatemala.

The extensive nucleotide sequence heterogeneity among independent genotypes of wild polioviruses permits the systematic design of genotype-specific molecular reagents. We have prepared two sets of polymerase chain reaction (PCR) primer pairs specific for the genotype of wild poliovirus type 3 recently endemic to Mexico and Guatemala. Nucleotide sequences of a representative wild type 3 virus isolated in Mexico in 1989 differed from the corresponding Sabin 3 (Leon 12 a1b) sequences at 167 of 900 positions within the VP1 region. From the sequence data, wild virus-specific primer pairs were designed to complement regions of high mismatch (greater than 33%) with Sabin 3 templates. Primer binding sites were spaced along the genome so that the predicted amplification products (142 bp and 163 bp) could be easily resolved electrophoretically from the products generated with our Sabin strain-specific primers (Sabin 1: 97 bp; Sabin 2: 71 bp; Sabin 3: 53 bp). RNAs of all wild type 3 poliovirus isolates from Mexico and Guatemala obtained over a 13-year period (1977-1990) served as efficient templates for amplification of the 142-bp and 163-bp products. Genomic templates derived from vaccine-related polioviruses and most heterologous wild polioviruses were inactive under equivalent reaction conditions. Amplifications generating a 114-bp product with a broadly reacting primer pair, matching highly conserved sequences in the 5'-noncoding region, provided a positive control for the presence in samples of poliovirus (or enterovirus) RNAs. Selective amplification of wild Mexico-Guatemala type 3 poliovirus sequences was obtained with either primer set in reactions containing large stoichiometric excesses (up to 10(6)-fold) of vaccine-related RNAs. We have used wild genotype-specific PCR primer sets to facilitate identification of wild polioviruses present in both clinical and environmental samples.     

Comparative studies on the genomes of some picornaviruses: denaturation mapping of replicative form RNA and electron microscopy of heteroduplex RNA.

Methods for denaturation mapping of double-stranded RNA and for heteroduplex RNA analysis are described. Denaturation maps of encephalomyocarditis (EMC) virus, poliovirus type 1, and poliovirus type 3 replicative form RNAs are presented. The three maps possess distinct features. Electron microscopy of heteroduplex RNA molecules constructed from poliovirus type 1 and poliovirus type 3 complementary RNA chains reveals two regions of considerable nonhomology as well as relatively conserved regions. The position of loops in this heteroduplex roughly corresponds to the location of regions where the denaturation maps of two polioviruses differ from each other. There is no evidence indicating that differences in the RNA sequences of poliovirus type 1 and poliovirus type 3 are due to extensive deletions (insertions) of genetic material; rather, the evolutionary divergence of these viruses seems to result, at least largely, from point mutations or equal substitutions.     

A poliovirus hybrid expressing a neutralization epitope from the major outer membrane protein of Chlamydia trachomatis is highly immunogenic.

Trachoma and sexually transmitted diseases caused by Chlamydia trachomatis are major health problems worldwide. Epitopes on the major outer membrane protein (MOMP) of C. trachomatis have been identified as important targets for the development of vaccines. In order to examine the immunogenicity of a recombinant vector expressing a chlamydial epitope, a poliovirus hybrid was constructed in which part of neutralization antigenic site I of poliovirus type 1 Mahoney (PV1-M) was replaced by a sequence from variable domain I of the MOMP of C. trachomatis serovar A. The chlamydial sequence included the neutralization epitope VAGLEK. This hybrid was viable, grew very well compared with PV1-M, and expressed both poliovirus and chlamydial antigenic determinants. When inoculated into rabbits, this hybrid was highly immunogenic, inducing a strong response against both PV1-M and C. trachomatis serovar A. Antichlamydia titers were 10- to 100-fold higher than the titers induced by equimolar amounts of either purified MOMP or a synthetic peptide expressing the VAGLEK epitope. Furthermore, rabbit antisera raised against this hybrid neutralized chlamydial infectivity both in vitro, for hamster kidney cells, and passively in vivo, for conjunctival epithelia of cynomolgus monkeys. Because poliovirus infection induces a strong mucosal immune response in primates and humans, these results indicate that poliovirus-chlamydia hybrids could become powerful tools for the study of mucosal immunity to chlamydial infection and for the development of recombinant chlamydial vaccines.