Amino Acid Residues on Human Poliovirus Receptor Involved in Interaction with Poliovirus*

We have previously demonstrated that the N-terminal immunoglobulin-like domain (domain 1; 115 amino acids) of human poliovirus receptor (hPVR) is essential for poliovirus binding and infection to cells. To identify amino acids involved in the interaction with poliovirus, we constructed a number of cDNAs encoding mutant hPVRs whose domain 1 was partially derived from mouse PVR (mPVR) homolog, which does not serve as a binding site for poliovirus. Poliovirus binding and infec- tion assays were performed on mouse L cells that ex-press these chimera cDNAs. Anti-hPVR monoclonal an- tibodies were employed to confirm the presence of mutant PVRs on the surface of mouse cells and to know conformational alteration of these PVRs. A significant decrease in efficiency of both poliovirus binding and infection to the cells was observed when one or a few amino acids of hPVR at G~Y’~, Ser74, Glns2, L e ~ ~ ’ - G l u ~ ~ ~ , or Gln’30-Ser132 were substituted by the corresponding amino acids of mPVR. Similar results were obtained when a 2-amino acid insertion of mPVR, which was miss-ing in hPVR, was introduced at the corresponding site (between and Leugg) of


Amino Acid Residues on Human Poliovirus Receptor Involved in
Interaction with Poliovirus* (Received for publication, October 11, 1993, andin revised form, December 29, 1993) Junken AokiSg, Satoshi Koike8, Iku IseS, Yasuko Sato-Yoshida+fl, and Akio NomotoSn From the $Department of Microbiology, the Tokyo Metropolitan Znstitute of Medical Science, Honkomagome, Bunkyo-ku, Tokyo 113, Japan and the IDepartment of Microbiology, Institute of Medical Science, the University of Tokyo,Shirokanedai,Tokyo 108,Japan We have previously demonstrated that the N-terminal immunoglobulin-like domain (domain 1; 115 amino acids) of human poliovirus receptor (hPVR) is essential for poliovirus binding and infection to cells. To identify amino acids involved in the interaction with poliovirus, we constructed a number of cDNAs encoding mutant hPVRs whose domain 1 was partially derived from mouse PVR (mPVR) homolog, which does not serve as a binding site for poliovirus. Poliovirus binding and infection assays were performed on mouse L cells that express these chimera cDNAs. Anti-hPVR monoclonal antibodies were employed to confirm the presence of mutant PVRs on the surface of mouse cells and to know conformational alteration of these PVRs. A significant decrease in efficiency of both poliovirus binding and infection to the cells was observed when one or a few amino acids of hPVR at G~Y'~, Ser74, Glns2, L e~~' -G l u~~~, or Gln'30-Ser132 were substituted by the corresponding amino acids of mPVR. Similar results were obtained when a 2-amino acid insertion of mPVR, which was missing in hPVR, was introduced at the corresponding site (between and Leugg) of hPVR. These amino acids were highly conserved in functional PVRs of primates but not in unfunctional PVRs of rodents. These results indicate that the amino acids identified may have important roles in interaction of PVR with poliovirus that leads to the establishment of the virus infection. In the three-dimensional model of the domain 1 of hPVR, these amino acids are located on one side of the molecule. This suggests that the interaction with poliovirus occurs on this side of the domain 1.
Poliovirus, known to be the causative agent of poliomyelitis, is a human enterovirus that belongs to the

Picornaviridae.
Poliovirion, a n icosahedral nonenveloped particle, is composed of 60 copies each of four capsid proteins, VP1, -2, -3, and -4, and a single-stranded RNA genome of positive polarity (1). A precise three-dimensional structure of the poliovirion has been elucidated by an x-ray crystallographic study (2, 3). Depressions, called "canyon," were observed on the surface of the virion and were suggested to be attachment sites for specific cellular receptors (4).
:r This work was supported in part by special coordination funds from the Science and Technology Agency of the Japanese government, re-Japan, and from the Ministry of Health and Welfare of Japan. The costs search grants from the Ministry of Education, Science, and Culture of of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence($ reported in this paper has been submitted to the GenBankTM/EMBL Data Bank with accession numberfs) 026107.
Poliovirus infection is initiated by binding of the virus to a specific cell surface molecule that serves as a poliovirus receptor (PVR).I Interaction between poliovirus and PVR destabilizes the virion particle (5). Indeed, the binding leads to the formation of "A-particles" that do not contain the capsid protein VP4. These A-particles are considered to be intermediates during the virus-uncoating process. Thus, PVR appears to have an important role in the uncoating process of the virus as well as in recognition and binding of the virus. Elucidation of the interaction of poliovirus with PVR, therefore, must provide insights into mechanisms of early infection processes of poliovirus.
The genomic and complementary DNAs for human PVR (hPVR) have been isolated from HeLa cells (6,7). Human PVR, a member of the immunoglobulin (Ig) superfamily, has three extracellular Ig-like domains, followed by a membrane-spanning domain and a cytoplasmic domain. Some hPVR mRNA isoforms, generated by alternative splicing, lack a nucleotide sequence encoding a transmembrane-spanning domain, and therefore encode soluble types of hPVR. Only membrane-bound forms (hPVRa and hPVRG) are functional receptors for poliovirus (7). Antibodies against PVR detected them as membrane glycoproteins of approximately 75 kDa, although molecular masses of hPVRa and hPVRG calculated from the deduced amino acid sequences were approximately 45 and 43 kDa, respectively. There are eight putative N-linked glycosylation sites in the extracellular domain (6, 7). Recent molecular genetic analysis has revealed that the poliovirus binding site resides in the N-terminal Ig-like domain (domain 1) (8,9) and that sugar moieties possibly attached to this domain are dispensable for the virus-receptor interaction leading to the infection (10, 11). These results suggest that protein moiety of the domain 1 is important for the interaction between poliovirus and PVR.
The host range of most poliovirus strains is restricted to primates. This host range restriction appears to be determined by cellular receptors accessible to poliovirus. Indeed, the mouse PVR (mPVR) homolog was proved not to serve as a functional PVR (12). It is therefore possible that experiments involving "homolog scanning mutagenesis" identify the amino acids important for the interaction between poliovirus and its cellular receptor.
Here, we describe the identification of several amino acid residues on hPVR as important residues for the virus-receptor interaction. All of these amino acids are located on one side of the three-dimensional structure proposed for the domain 1 of hPVR. The key amino acids identified here are highly conserved in primates but not in rodents. This suggests that these The abbreviations used are: PVR, poliovirus receptor; mPVR, mouse poliovirus receptor homolog; mAb, monoclonal antibody; PBS, phosphate-buffered saline; hPVR, human poliovirus receptor; PCR, polymerase chain reaction.

8431
amino acids are involved in determination of species specificity of poliovirus.

EXPERIMENTAL PROCEDURES
Cloning of mPVR cDNA-AcDNAlibrary was prepared from poly(A)+ RNA of the brain of mouse strain C57/BL6, using vector A g t l O as described previously (7), except that the NotI-BamHI-EcoRI adapter (Takara Shuzo Co.) was employed. An AatII-EcoRV fragment of hPVR cDNA (nucleotide positions 8 2 4 0 6 ) labeled with 32P was used a s a probe. Plaque hybridization was carried out as described (71, except that the membranes were washed with 3 x SSC (0.45 M NaCl and 0.045 M sodium citrate) and 0.1% SDS. Insert DNA, excised by Not1 digestion, was subcloned into pBluescript KS plasmid vector. Nucleotide sequences were determined by the dideoxy method (13) using Sequenase Version 2.0 kit (U. S. Biochemical Corp.). In some cases, synthetic oligonucleotides were used as primers for the dideoxy method. The nucleotide sequence data reported in this paper will appear in the GSDB, DDBJ, EMBL, and NCBI nucleotide sequence data bases with the following accession number D26107.
Cloning of PVR Gene from Various Animal Species-The genomic DNA segment encoding a part of the domain 1 of the PVR or its homolog was amplified by PCR from DNAs isolated from cynomolgus monkey (Macaca fascicularis), white-lipped tamarin (Sagunis labiatus), and Wister rat. Cloning of the gene from cynomolgus monkey was performed as described previously (10). Polymerase chain reaction (PCR) primers used for the amplification of white-lipped tamarin or Wister rat DNAs Recombinant PVR cDNAs of hPVR a n d mPVR-To construct a series of cDNAs encoding recombinant molecules between hPVR and mPVR with regard to the domain 1, restriction enzyme sites were introduced in both hPVR and mPVR cDNAs. As shown in Fig. 2, a fragment of hPVR cDNA (nucleotide positions 82-411) was amplified by PCR. In this step, Sal1 and BamHI sites were generated at the ends for cloning. BstEII and/or BglI sites were introduced in this fragment by a method previously reported (14). They were subcloned into the SalI/BamHI site of a plasmid vector pUC119 and designated phPVR, phPVR-BstEII, phPVR-BglI, and phPVR-BstEII-BglI. As to the domain 1 of mPVR cDNA, PCR primers were designed to amplify nucleotide position 8 9 4 3 1 and to generate Sal1 and AatII sites and EcoRV and BamHI sites at 5' and 3' ends, respectively. Removal of the BglI site a t nucleotide position 212 without affecting amino acid sequence and introduction of ApaI and/or EcoRI sites were carried out as above. These DNA fragments were subcloned into pUC119 and designated pmPVR, pmPVR-ApaI, pmPVR-EcoRI, and pmPVR-Apal-EcoRI. Thus the restriction enzyme map is common to both of the cDNAs with regard to cleavage sites for AatII, BstEII, ApaI, EcoRI, BglI, and EcoRV. Allele replacement experiments were carried out on these plasmids with these restriction enzymes. Region numbers are indicated in Fig. 2. An AatII-EcoRV DNA fragment (domain 1 region) of pSV2PVRa (7) was replaced by the corresponding fragments of various recombinant cDNAs thus constructed. Nucleotide sequences of all recombinant PVR cDNAs were confirmed as described above.
Site-directed Mutagenesis of hPVR cDNA-A HindIII-KpnI DNA fragment of pSV2PVRa (nucleotide positions, 206-735) was subcloned into M13tv19. Site-directed mutagenesis (15) was performed with the single-strand DNA template using the Mutan G site-directed mutagenesis system (Takara Shuzo Co.). Twenty-nine oligonucleotides of 25-48 nucleotides long were synthesized and used as primers to generate the mutants listed in Table I. Mutated HindIII-EcoRV fragments were ligated to a larger HindIII-EcoRV fragment of pSV2PVRa. Nucleotide sequences of all mutant cDNAs were confirmed as described above.
Monoclonal Antibodies against hPVR-Monoclonal antibodies (mAbs) p403 (IgGl), p216 (IgGl), p235 (IgG2a), p275 (IgG2a), and p242 (IgGl), which recognized hPVR in a conformation-dependent manner were prepared as follows. BALB/c mice (8 weeks, P ) were immunized with hPVR-human Ig fusion protein (5), and the spleen cells were fused with mouse myeloma cells (PAI) as described elsewhere (16). MAbs that reacted with La cells (mouse L cells expressing hPVRa cDNA) (7) but not with mouse L cells were chosen a s positive clones. All of these mAbs inhibit poliovirus infection to HeLa cells and react to a glycoprotein of 75 kDa in HeLa cell lysates. However they failed to react to the antigen in Western blot analysis. Competition analysis revealed that these mAbs were classified into five groups that are different in epitope recognition (data not shown). MAb p403 binds to PVRd4 (81, and other mAbs bind to PVRd3 (81, indicating that the epitopes of p403 and other mAbs reside in domain 1 and domain 1 and/or 2, respectively. MAbs that recognized hPVR in a conformation-independent manner were developed a s follows. BALB/c mice (8 weeks, P ) were immunized with purified soluble hPVR (extracellular domain of hPVR expressed by baculovirus expression system) using a n intrasplenic injection method (171, and spleen cells were fused with mouse myeloma cells (PAI). Western blot analyses were employed to choose positive clones. These mAbs recognize a glycoprotein of about 75 kDa in HeLa cell membrane extracts in Western blot analysis. These mAbs bound to PVRdl and PVRd3 ( 8 ) but not to PVRd2 ( 8 ) (data not shown). The observation indicates that epitopes for mAbs 4C6 and 4E12 reside in domain 2. 7hnsfection-L cells a t about 60% confluency were harvested for transfection using trypsin. The cells (5 x 10" cells) were transfected with 0 (mock) or 50 pg of plasmid DNA by the electroporation method. The cells were reseeded at a density of lo5 cells/cm2 in 24-well plates for poliovirus binding and infection assay and 96-well microplates for mAb binding assay. Assays were performed 48 h after the transfection.
mAh Binding Assay-To examine the binding activities of mAbs 4C6 and 4E12, cells transfected with chimera PVR cDNAs were fixed with 10% paraformaldehyde in PBS at room temperature for 30 min. For other mAbs, cells were used without fixation. MAbs were added to the cells and incubated at 4 "C for 3 h. After washing with PBS three times, bound mAbs were detected by incubation with anti-mouse IgG conjugated with horseradish peroxidase followed by color development reaction using o-phenylenediamine as peroxidase substrate. Absorbance a t 490 nm was measured.
Poliovirus Binding Assay-Poliovirus (type 1 Mahoney strain) labeled with [""Slmethionine was added to the cells in serum-free Dulbecco's modified minimum essential medium (100,000 cpdwell) and incubated a t 4 "C for 150 min (8). After washing with PBS four times, the cells were solubilized with 0.5% SDS, and the radioactivity was measured in a scintillation counter. Mock-transfected cells were used to estimate the background level of the radioactivity. All assays were performed in triplicate at least three times. The relative amount of bound poliovirus was normalized by using relative amounts of bound mAb 4C6, because mAb 4C6 recognizes the domain 2 of denatured hPVR and therefore binds to all mutant PVRs at an equal level of efficiency.
Poliovirus Infection-Light-sensitive poliovirus (10" plaque-forming units) was added to lo5 cells transfected with mutant cDNAs (7). After adsorption of poliovirus a t room temperature for 10 min, the cells were incubated at 37 "C for 60 min in the dark. The cells were washed with serum-free Dulbecco's modified minimum essential medium twice under a fluorescent lamp and further incubated a t 37 "C for 24 h. Cells were then disrupted by three cycles of freezing and thawing, and virus titers were determined by plaque assay using African green monkey kidney cells (18). Poliovirus strains used were type 1 Mahoney and Sabin 1 strains, type 2 Lansing and Sabin 2 strains, type 3 Leon and Sabin 3 strains.

RESULTS
Cloning of mPVR cDNA-A cDNA clone that hybridized to the hPVR cDNA probe was obtained from 5 x lo5 plaques of a mouse cDNA library. A 2.4-kilobase cDNA insert of this phage clone was subcloned into a plasmid vector pBluescript KS. Nucleotide sequence analysis revealed that the cDNA had a nucleotide sequence encoding a peptide of 530 amino acids (Fig. L A ) followed by a part of the 3' noncoding region of 874 base pairs. The nucleotide sequence in the first 1015 nucleotides of this cDNA was identical to that previously reported (12). However, the nucleotide sequence downstream of position 1016 was unique ( Fig. L A ) , resulting in the loss of a region encoding the putative membrane-spanning domain (Fig. 1B ). The coding sequence of this cDNA carried 4 additional exons (exons 7-10) in the downstream of exon 6 of the mPVR gene (data not shown). The blanch point of the sequence divergence was one of the splicing sites (12). Northern blot analysis revealed that a probe    was amplified, and BstEII and/or BglI sites were introduced by PCR a s described under "Experimental Procedures." A corresponding region of mPVR (open box) was amplified by PCR to generate AatII and EcoRV sites at the 5' and 3' ends, respectively. The BglI site at position 212 was removed without changing the amino acid sequence, and ApaI and/or EcoRI sites were introduced by PCR as described under "Experimental Procedures." The AatII-EcoRV fragments of hPVR and mPVR thus obtained were subcloned into pUC 119. Humadmouse chimera PVR cDNAs were constructed by using the common restriction enzyme sites. Numbers in parentheses indicate nucleotide positions of restriction enzyme sites.

G~CCCTGAGTCTCGAAGATGAGGAGGAAGATGATGAGGAGGAAGACTTCCTGGAT~TCAACCCTAT~ATGATGCCCTGTCCTAC V S L S L E D E E E D D E E E D F L D K I N P I Y D
of the first 1015 nucleotide sequence hybridized both to 2-and Regions of hPVR Influencing Interaction with Poliovirus-3-kilobase bands (data not shown). A probe that had nucleotide Amino acid sequences in domain 1 of hPVR and mPVR are sequence 1016-2464 hybridized only to the 3-kilobase band. shown in Fig. 2 A . In this region of mPVR, 55 amino acid sub-These data strongly suggest that cDNA obtained in this study stitutions, two deletions, and two insertions are observed comcorresponds to mRNA for a soluble form of mPVR, generated by pared with that of hPVR, resulting in 52% sequence homology alternative splicing from the primary transcript.
in this domain between hPVR and mPVR. To identify regions  Fig. 1).
critical for the virus-receptor interaction, allele replacement experiments were carried out with regard to 5 regions in domain l (Fig. 2, A and B ) . This domain is classified into V-type (19) and is composed of nine antiparallel p-strands. According to the consensus sequence of the V-type domain, it is roughly estimated that region 1 corresponds to A and B strands, rebeon 2 to C and C', region 3 to C", region 4 to D and a part of E, and region 5 to the rest of E, F, and G (Fig. 2).
Mouse L cells transfected with chimera cDNAs were examined for their capacity to find poliovirus and their susceptibility to the virus. Various mAbs against hPVR were used to examine for the existence of mutant PVRs on the cell surface and their conformation. The results are summarized in Fig. 3. Binding of mAbs (4C6 and 4E12) was observed in all mouse cell cultures transfected with chimera cDNAs used, indicating that these chimeras were expressed on the surface of mouse L cells.
When all 5 regions of hPVR werc replaced with those of mPVR (M, in Fig. 31, neither poliovirus binding nor infection was detected as expected previously (12). Similar results were obtained for chimera PVRs in which two or more regions were replaced by those of mPVR except for R234 (Fig. 3). Mouse L cells carrying R234 showed both binding activity to poliovirus and susceptibility to the virus, albeit to a lesser extent as compared with those carrying wild type hPVR (Fig. 3). This suggests that regions 2, 3, and 4 of hPVR harbor important sites for the virus-hPVR interaction.
As for chimera PVRs in which only one region of hPVR was replaced by that of mPVR, only R1234 seems to have receptor function with regard to both binding and infection of poliovirus. A relatively low amount of progeny virus was recovered from cells carrying R1235, R1245, or R2345. However, a significant amount of bound poliovirus was not detected in these cells (Fig.  3). This indicates that these chimeras have low activities as a poliovirus receptor. Cells carrying R1345 showed neither bind-of Poliovirus Receptor a435 ing activity nor susceptibility to poliovirus. Thus region 2 of hPVR seemed to Ire most important for maintaining receptor function, although every other region influences the function. R234 retained the receptor function to some extent as mentioned above and also retained binding capacity for mAbs that recognized hPVR in a conformation-dependent manner (data not shown). R2345 showed receptor activity lower than R234 both in virus binding and infection. Furthermore none of conformation-d[:pendent mAbs used recognized R2345. I t is therefore possiblc that exchange of region 1 caused global alteration in the conformation of domain 1 and that an additional exchange of rc$on 5 provides the initial conformation to some extent to the domain. Thus regions 1 and 5 appear to interact with each other. Indeed the two regions closely exist in a threedimensional model of hPVR (Fig. 4A).
Similar results were obtained from experiments using Sabin 1, Lansing, Sabin 2, Leon, and Sabin 3 strains of poliovirus (data not stlowll). These results suggest that most poliovirus strains recognize similar parts of hPVH.
Amino Acids ( I r i t i c n l for Virus-Receptor Interaction-Single or several amino acid substitutions were introduced in domain 1 of hPVR to iderltify the amino acid residues involved in the virus-receptor intcraction. Because region 2 of hPVR seemed to be most important for both poliovirus binding and infection, a number of single amino acid substitutions were introduced in region 2. Poliovirus and mAb binding assays and poliovirus infection assay w c x performed, and the results are summarized in Tahlv I. All mutant I'Vlis were present on their cell surface as judged by binding of r11Abs 4C6 and 4E12 (Table I). Eight mutants, i.e. S1, S5, S11, S17, 821, S24 (S21 and S24, insertion mutants), S25, and S28, showed significantly reduced activities in poliovirus binding (Table I ) ( t o approximately 4 0 % of wild type hPVR). The Si2 mutant also showed a significant reduction of virus binding activity (63% of hPVR). The results suggest that the amino acids replaced in these mutants are important in virus-receptor int,ernction. However, considerable amounts of virus production were observed in cells transfected with every mutant cDNA usod. It was impossible to find mutant PVRs that did not mcdiatc the infection but bound to poliovirus. Thus poliovirus bindlng activity appears to be always associated with functional I'VR activity.
Amino acid substitutions in the nine mutants described above might he critical contact residues for poliovirus.
It is possible, howevc,r, that the substitutions significantly alter the conformation o f the binding site recognized by the virus. Five mAbs with dif'ferent epitopes that recognized hPVR in a conformation-dependent nlanner were employed to examine the possibility (Table I). Significant reduction of binding of all five groups of mAbs was observed in the case of S1, S5, and S21 (Table I). Thus thcsc mutations seemed to induce gross conformational changes in mutant PVRs. In addition, it is impossible to identify contact amino acid residues for poliovirus from the experiment involving mutant S24, because this mutant is an insertion mutant of hPVR. Consequently, mutated amino acids in five mutants S11, S12, S17, S25, and S28 are thought to be directly involved in the interaction with poliovirus. It should be noted that no individual mutations can completely abolish poliovirus binding. It is therefore possible that the sum of the plural mutations in mPVR may result in a loss of interaction with poliovirus. This possibility is supported by a n observation that the S30 mutant PVR shows no receptor function for poliovirus, although mutants S21 and S24 retain binding activity to some extent and confer cell-permissiveness for the poliovirus infection (Table I).
Key Arnirlo Acids on Three-dimensional Model of PVR-A three-dimensional structure of PVR domain 1 has been pre-   * Binding of conformation-dependent mAbs were shown as follows: +++, 75-100%; +, 1040%; -, 0 -% binding of wild type. dicted by computer modeling by using the coordinates for Ig tion mutant), 525, and 528 are shown in Fig. 4B. Interestingly, V-type domain as a reference protein (5). Nine antiparallel these amino acid residues were located on one side of this p-strands of Ig V-type and possible C-a: trace of PVR domain 1 three-dimensional model. The observation supports that PVR are shown in Fig. 4, A and B, respectively. The location of interacts with poliovirus on this side. This is consistent with mutated amino acids in mutants 511, 512, 517,524 (an inser-our computer-aided prediction of how the domain 1 of PVR binds to the canyon of poliovirus (5). Key Amino Acids during Evolution-To examine if species specificity of poliovirus infection can be explained by the key amino acids, PVR homologs from other mammals species were examined for conservation of the key amino acids. We have already reported two types of cDNAs that encode functional PVRs ofAfrican green monkey (Ceropithecus aethiops) (cDNAs for AGMal and AGMa2) (10). In addition, genomic DNA clones for PVR domain 1 of the cynomolgus monkey, white-lipped tamarin, and Wister rat were prepared as described under "Experimental Procedures." Amino acid sequences of hPVR and PVR homologs from various animal species are shown in Fig.  2 A . Two types of PVR genes were cloned from the cynomolgus monkey. This may be due to the gene duplication that occurred only in old-world monkeys as described (10). Of these species, human, African green monkey, and cynomolgus monkey are susceptible to poliovirus, while mice and rats are not. Although susceptibility of tamarin to poliovirus is not known, a chimera PVR cDNA in which nucleotide sequence of positions 108-251 and the remaining sequence are derived from tamarin PVR homolog and hPVR, respectively, confer mouse L cells permissiveness for poliovirus (data not shown). This strongly suggests that tamarin is also susceptible to poliovirus.
Amino acid sequence homology in domain 1 among PVRs of human, African green monkey (AGMal and AGMa2), cynomolgus monkey (CYN1 and CYNB), and tamarin were 92, 89, 88, 87, and 73570, respectively. All key amino acids identified on hPVR were conserved among PVRs of primates examined, except for amino acid position 73 in a tamarin PVR homolog where Gly is replaced by k g . A number of substitutions other than the key amino acids observed in tamarin PVR homolog did not seem to affect the PVR function as described above.
On the other hand, rat PVR homolog and hPVR showed a 47% identity in the amino acid sequence of domain 1. The rat sequence had two deletions a t positions 58-59 and 70, and two insertions at positions between 89 and 90, and 98 and 99 of hPVR just as mPVR. The amino acid sequences of rat PVR homolog and mPVR showed a 96% identity. Amino acid residues a t positions 73, 74, and 82, which corresponded to the key amino acid positions of hPVR, were identical to those of mPVR (Fig. 2). These results support the idea that the host range of poliovirus infection may be determined by the key amino acids identified in this study. Thus accessibility of PVRs to poliovirus may be influenced by multiple amino acid residues on domain 1 of PVRs including the key amino acids proposed here.

DISCUSSION
The mouse PVR homolog gene reported previously (12) has been considered to encode equivalent cellular molecules to hPVR, because 1) nucleotide sequence of the cDNA is most homologous to that of hPVR cDNA, 2) the gene organization was identical t o that of the hPVR gene, and 3) the mouse gene was mapped to the locus in the proximal region to centromere of the mouse chromosome 7,2 which includes the corresponding region of q13.1-13.2 of the human chromosome 19, the locus of the hPVR gene (7). We demonstrated here the presence of the soluble form of PVR homolog in mouse a s well as in humans. Tissue distribution of mPVR mRNA detected by Northern blot hybridization was similar to that of hPVR mRNA in human tissues (20). These two additional observations further support the assumption that this gene encodes a mouse homolog for PVR.
In experiments involving homolog scanning mutagenesis performed in this study, mAbs that recognized hPVR in a con-' M. Niwa, H. Yonekawa, J. Anki, and s. Koike, unpublished observation. formation-dependent manner were effectively used to detect conformational alterations of PVR induced by mutations. Thus mutations in S1, S5, and S21, which appeared to cause conformational change in PVR, were eliminated from the key amino acids, although we had no evidence to deny that amino acid residues mutated in these mutants were contact residues for poliovirus. Consequently, amino acids Gly73, Glnsz, Leu99-Glu102, and Gln130-Ser132 of hPVR were proposed a s possible contact residues for poliovirus. It should be noted that there may be some other amino acid residues directly involved in the virus-hPVR interaction in amino acid residues conserved between hPVR and mPVR.
PVR has a dual function, which is binding to poliovirus and destabilization of the virus (initiation of uncoating). Mason et al. (21) and Aarnes et al. (22) showed that infection does not occur when poliovirus bound to the cell surface without PVR. This supports the fact that two functions of PVR are necessary for poliovirus infection. It might be possible to separate the two functions if the PVR has two catalytic regions. However, no mutants were obtained that have poliovirus binding activity without conferring the permissiveness to the cells, suggesting that both binding and uncoating activities reside on the same site in domain 1. Thus binding and initiation of uncoating failed to be separated by this strategy.
As for human immunodeficiency virus and CD4 interaction, a specific amino acid sequence (amino acid positions 41-59) of CD4 receptor plays a major role in human immunodeficiency virus binding (23,24,25). On the contrary, several amino acid residues were identified in I C A " 1 a s contact residues for a major group of human rhinovirus (26,27). The interaction of poliovirus with hPVR resembles the latter case. This observation appeared to be consistent with the fact that a canyon structure is involved in the interaction of poliovirus and rhinovirus with their receptors, whereas gp120/160 is involved in the case of human immunodeficiency virus. The important amino acids for poliovirus-hPVR interaction are located in a wide area of one side of domain 1 on the three-dimensional model of the PVR. Thus the interaction seems to occur between wide areas of the canyon and hPVR. Specific conformation of hPVR, therefore, may be essential for recognition by the virus. Because of high species specificity of poliovirus, it appears that the virus has evolved together with primates.