Characterization of three different transgenic mouse lines that carry human poliovirus receptor gene--influence of the transgene expression on pathogenesis

Three transgenic mouse lines, ICR-PVRTg1, ICR-PVRTg5, and ICR-PVRTg21, which are susceptible to poliovirus, have been established by introducing the human gene for poliovirus receptor (PVR) into the genome of mouse strain ICR. Genetic characterizations of the PVR gene were carried out on these mouse lines to define the approximate copy number, insertion site, and expression of the transgene in the central nervous system (CNS). The transgene was integrated in the chromosome 4, 12, and 13 of ICR-PVRTg1, ICR-PVRTg5 and ICR-PVRTg21 mice, respectively, and was stably transmitted to progeny mice. ICR-PVRTg1 appeared to have the most abundant copy numbers of the transgene and showed the highest level of PVR mRNA and membrane associated PVR protein in the CNS among the three mouse lines. Those in ICR-PVRTg21 and ICR-PVRTg5 were at intermediate and lowest levels, respectively. In the CNS, PVR mRNA was detected at high levels only in neurons of the spinal cord and brain stem where poliovirus can replicate, suggesting that the PVR mRNA expression confers cell specificity to poliovirus in the CNS. ICR-PVRTg1 and ICR-PVRTg5 showed the highest and the lowest sensitivity to poliovirus, respectively, whereas ICR-PVRTg21 was in-between. These results may suggest that poliovirus sensitivity of the mice is attributed to relative levels of PVR expression.

Two human homologues of Saccharomyces cerevisiae SWI2/SNF2 and Drosophila brahma are transcriptional coactivators cooperating with the estrogen receptor and the retinoic acid receptor

A set of genes (SWI1, SWI2/SNF2, SWI3, SNF5 and SNF6) in Saccharomyces cerevisiae are required for transcription of a variety of yeast genes. It was recently reported that the mammalian glucocorticoid receptor failed to activate transcription when transiently expressed in swi1-, swi2- or swi3- yeast strains. We report here that two highly related human cDNAs, hSNF2 alpha and -beta, encode amino acid sequences homologous to both the yeast SWI2/SNF2 and the Drosophila brahma. Similar to their yeast and Drosophila counterparts, both human cDNAs contain helicase motifs, a bromodomain, a highly charged C-terminal sequence and an N-terminal sequence rich in proline, glutamine and glycine. Tissue distribution of the mRNAs varied slightly. Transcriptional activation by the estrogen receptor and the retinoic acid receptor was enhanced by co-expression of either hSNF2 cDNA. No enhancement was observed for promoters which do not respond to nuclear receptors. We suggest that global transcriptional coactivators equivalent to the yeast SWI/SNF complex exist in mammalian cells.

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 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 Gly73, Ser74, Gln82, Leu99-Glu102, or Gln130-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 Arg98 and Leu99) 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.

Tissue tropism and species specificity of poliovirus infection

Paralytic poliomyelitis is caused by the destruction of neurons in the central nervous system during the lytic replication of poliovirus. The human gene for the poliovirus receptor has been cloned recently, and transgenic mice generated that are susceptible to this virus. This has allowed the critical amino acids that determine the specificity of poliovirus to be identified.

Molecular biology and cell-free synthesis of poliovirus

Poliovirus is a small icosahedral particle consisting of only five species of macromolecules: 60 copies each of the capsid protein VP1-4; and one copy of single-stranded RNA, approximately 7500 nt long. The genome, linked at the 5' end to a small protein VPg and 3' polyadenylylated, is of plus strand polarity. After receptor-mediated uptake of the virus and release of the RNA into the cytoplasm, the genome serves as mRNA, encoding only a single polypeptide, the polyprotein. The polyprotein is cleaved co-translationally into numerous polypeptides by its own, internal proteinases 2Apro, 3Cpro and 3CDpro. Initiation of translation is mediated by a novel genetic element, called internal ribosomal entry site (IRES). IRES elements, which are 400 nt long RNA segments located within the 5' non-translated region of the viral genome, are common to all picornaviruses. Their function renders translation of picornavirus mRNAs cap- and 5'-independent, an observation that has upset the dogma of cap-dependent translation in eukaryotic cells. IRES elements have also been used to genetically dissect the viral genome and to construct novel expression vectors. Genome replication is not fully understood, the major conundrum being the initiation of RNA synthesis by the primer-dependent viral RNA polymerase 3Dpol, a process leading to VPg-linked RNA products. Nearly all non-structural proteins appear to be involved in initiation, the proteinases 2Apro and 3CDpro included. A HeLa cell-free system has been developed that, on programming with plasmid-transcribed viral RNA, will perform viral translation, protein processing, RNA replication, and assembly of capsid protein and newly made genomic RNA. The final yield is infectious poliovirus.(ABSTRACT TRUNCATED AT 250 WORDS)

Temperature-sensitive mouse cell factors for strand-specific initiation of poliovirus RNA synthesis

Two cell lines, TgSVA and TgSVB, were established from the kidneys of transgenic mice carrying the human gene encoding poliovirus receptor. The cells were highly susceptible to poliovirus infection, and a large amount of infectious particles was produced in the infected cells at 37 degrees C. However, the virus yield was greatly reduced at 40 degrees C. This phenomenon was common to all mouse cells tested. To identify the temperature-sensitive step(s) of the virus infection cycle, different steps of the infection cycle were examined for temperature sensitivity. The results strongly suggested that the growth restriction observed at 40 degrees C was due to reduced efficiency of the initiation process of virus-specific RNA synthesis. Furthermore, this restriction appeared to occur only on the synthesis of positive-strand RNA. Virus-specific RNA synthesis in crude replication complexes was not affected by the nonpermissive temperature of 40 degrees C. In vitro uridylylation of VPg seemed to be temperature sensitive only after prolonged incubation at 40 degrees C. These results indicate that a specific host factor(s) is involved in the efficient initiation process of positive-strand RNA synthesis of poliovirus and that the host factor(s) is temperature sensitive in TgSVA and TgSVB cells.

Pharmacological profile of FR139317, a novel, potent endothelin ETA receptor antagonist

The effects of FR139317 on the cardiovascular system were investigated in cultured cells, isolated organs and whole animals. FR139317 inhibited the specific binding of [125]endothelin(ET)-1 to porcine aortic microsomes in a concentration-dependent, monophasic fashion with an IC50 of 0.53 nM. In contrast, FR139317 showed low affinity for [125I]ET-1 specific binding sites in porcine kidney (IC50, 4.7 microM). In isolated rabbit aorta, FR139317 shifted the ET-1-induced concentration-contractile response curve to the right with a pA2 value of 7.2 and lacked agonist activity. A single (i.v.) bolus dose of FR139317 completely inhibited the pressor response to ET-1 in vivo, but had no effect on the initial depressor response in conscious normotensive rats. These data indicate that FR139317 is a potent, highly specific ETA receptor antagonist. In addition, FR139317 also inhibited ET-1 induced [3H]thymidine incorporation in cultured vascular smooth muscle cells from rat aorta (IC50, 4.1 nM), suggesting that ET-1-induced mitogenesis is mediated only by the ETA receptor. FR139317 could become a useful tool for investigating the physiological and pharmacological actions of ET.

Antigenicities of Group I and II hepatitis C virus polypeptides--molecular basis of diagnosis

Comparative nucleotide sequence studies on the putative NS3 and NS4 regions of the genomes of hepatitis C viruses (HCV) have revealed that there are at least two groups of HCV, group I and group II. The cDNA clone E, corresponding to a boundary between the NS3 and NS4 (NS3-4) region of the group II HCV genome, encodes antigens that react to antibodies specific to group II HCV (Tsukiyama-Kohara et al. (1991) Virus Genes 5, 243-254). To understand the molecular basis of the group-specific antigenicity of HCV peptides, the predicted amino acid sequences around the NS3-4 region of our group II HCV cDNAs were compared with those of other HCV isolates. The analysis revealed the presence of group-specific amino acids in this peptide region. Evolutionary analysis of nucleotide sequences within this region of these HCV isolates also led to the same classification. A similar result was obtained by sequence analysis of cloned cDNAs corresponding to the core region. A cDNA of the group II HCV core region was prepared by polymerase chain reaction from the cDNA synthesized with group II-specific primer complementary to the NS3-4 region. The products directed by the cDNA of the core region did not have group-specific antigenicity. The NS3 peptide region also appeared not to carry group-specific antigens. Our results indicate that most HCV isolates can be classified into either group I or II, and that the existence of two groups of HCV does not disturb HCV diagnosis as long as core and/or NS3 peptides are used to detect HCV antibodies.

A second gene for the African green monkey poliovirus receptor that has no putative N-glycosylation site in the functional N-terminal immunoglobulin-like domain

Using cDNA of the human poliovirus receptor (PVR) as a probe, two types of cDNA clones of the monkey homologs were isolated from a cDNA library prepared from an African green monkey kidney cell line. Either type of cDNA clone rendered mouse L cells permissive for poliovirus infection. Homologies of the amino acid sequences deduced from these cDNA sequences with that of human PVR were 90.2 and 86.4%, respectively. These two monkey PVRs were found to be encoded in two different loci of the genome. Evolutionary analysis suggested that duplication of the PVR gene in the monkey genome had occurred after the species differentiation between humans and monkeys. The NH2-terminal immunoglobulin-like domain, domain 1, of the second monkey PVR, which lacks a putative N-glycosylation site, mediated poliovirus infection. In addition, a human PVR mutant without N-glycosylation sites in domain 1 also promoted viral infection. These results suggest that domain 1 of the monkey receptor also harbors the binding site for poliovirus and that sugar moieties possibly attached to this domain of human PVR are dispensable for the virus-receptor interaction.

Adenovirus E1A proteins stimulate inositol phospholipid metabolism in PC12 cells

To study the influence of nuclear oncogenes on inositol phospholipid metabolism, we examined the various parameters of inositol phospholipid metabolism in PC12 cells expressing adenovirus type 12 or adenovirus type 5 E1A. Although the inositol 1,4,5-trisphosphate content was increased only slightly, the diacylglycerol content was 2.4-fold higher in E1A-expressing PC12 cells. Furthermore, we found that the activity of phospholipase C, one of the key enzymes in inositol phospholipid metabolism, was increased at least five- to eightfold. Diacylglycerol kinase activity in the membrane fraction was 10 to 15% of that in parental PC12 cells. Overall protein kinase C activities in E1A-expressing PC12 cells were decreased, but the activity of membrane-bound protein kinase C was significantly increased. These observations clearly indicate that inositol phospholipid metabolism is stimulated in cells producing E1A and suggest that nuclear oncogene E1A has the ability to stimulate inositol phospholipid metabolism.

Internal ribosome entry site within hepatitis C virus RNA

The mechanism of initiation of translation on hepatitis C virus (HCV) RNA was investigated in vitro. HCV RNA was transcribed from the cDNA that corresponded to nucleotide positions 9 to 1772 of the genome by using phage T7 RNA polymerase. Both capped and uncapped RNAs thus transcribed were active as mRNAs in a cell-free protein synthesis system with lysates prepared from HeLa S3 cells or rabbit reticulocytes, and the translation products were detected by anti-gp35 antibodies. The data indicate that protein synthesis starts at the fourth AUG, which was the initiator AUG at position 333 of the HCV RNA used in this study. Efficiency of translation of the capped methylated RNA appeared to be similar to that of the capped unmethylated RNA. However, a capped methylated RNA showed a much higher activity as mRNA than did the capped unmethylated RNA in rabbit reticulocyte lysates when the RNA lacked a nucleotide sequence upstream of position 267. The results strongly suggest that HCV RNA carries an internal ribosome entry site (IRES). Artificial mono- and dicistronic mRNAs were prepared and used to identify the region that carried the IRES. The results indicate that the sequence between nucleotide positions 101 and 332 in the 5' untranslated region of HCV RNA plays an important role in efficient translation. Our data suggest that the IRES resides in this region of the RNA. Furthermore, an IRES in the group II HCV RNA was found to be more efficient than that in the group I HCV RNA.

Crystallization and preliminary X-ray diffraction studies of coxsackievirus B1

Preparations of coxsackievirus B1 (CVB1) derived from an infectious cDNA clone have been crystallized in multiple crystal forms. Using high intensity synchrotron radiation, an orthorhombic form of the crystals was shown to diffract X-rays to at least 2.9 A resolution. The unit cell has a primitive lattice with dimensions a = 323 A, b = 450 A, and c = 522 A. A crystallographic asymmetric unit of these CVB1 crystals probably contains an entire virus particle, implying the presence of 60-fold non-crystallographic redundancy. This CVB1 crystal form appears to be suitable for high-resolution structure determination by X-ray crystallography.

Role of inflammatory responses in initiation of atherosclerosis: effects of anti-inflammatory drugs on cuff-induced leukocyte accumulation and intimal thickening of rabbit carotid artery

Immediately after a cuff-sheathing of rabbit carotid artery, a large number of leukocytes adhered to injured endothelium then infiltrated into the media. These inflammatory responses were followed by an atherosclerotic change, intimal thickening, of the artery. A simultaneous injection of dexamethasone (10 mg/kg i.m.) inhibited the leukocyte accumulation by 74% when evaluated 18 h thereafter. Similarly, 39% inhibition was obtained with the same dose of FR110302, a potent 5-lipoxygenase inhibitor. On the other hand, the same dose of indomethacin, a cyclooxygenase inhibitor, had little effect on the leukocyte accumulation. The intimal thickening which was evaluated 3 weeks after the cuff-treatment was attenuated by a daily dose (10 mg/kg i.m.) of dexamethasone or FR110302 but not by one of indomethacin. The inhibition by the two former drugs were 91 and 58%, respectively. In vitro, the three drugs in concentrations up to 10 microM hardly affected endothelial adhesion of PMN which was induced by LPS or IL-1. Though 10 microM of FR110302 and indomethacin significantly decreased PMN chemotaxis induced by LTB4, the decreases were less than that at 10 microM dexamethasone. These results confirm a possible linkage between inflammation and atherosclerosis, and suggest that 5-lipoxygenase products contribute to the initiation and development of atherosclerosis.

Nucleotide sequences important for translation initiation of enterovirus RNA

An infectious cDNA clone was constructed from the genome of coxsackievirus B1 strain. A number of RNA transcripts that have mutations in the 5' noncoding region were synthesized in vitro from the modified cDNA clones and examined for their abilities to act as mRNAs in a cell-free translation system prepared from HeLa S3 cells. RNAs that lack nucleotide sequences at positions 568 to 726 and 565 to 726 were found to be less efficient and inactive mRNAs, respectively. To understand the biological significance of this region of RNA, small deletions and point mutations were introduced in the nucleotide sequence between positions 538 and 601. Except for a nucleotide substitution at 592 (U----C) within the 7-base conserved sequence, mutations introduced in the sequence downstream of position 568 did not affect much, if any, of the ability of RNA to act as mRNA. Except for a point mutation at 558 (C----U), mutations upstream of position 567 appeared to inactivate the mRNA. In the upstream region, a sequence consisting of 21 nucleotides at positions 546 to 566 is perfectly conserved in the 5' noncoding regions of enterovirus and rhinovirus genomes. These results suggest that the 7-base conserved sequence functions to maintain the efficiency of translation initiation and that the nucleotide sequence upstream of position 567, including the 21-base conserved sequence, plays essential roles in translation initiation. A deletion mutant whose genome lacks the nucleotide sequence at positions 568 to 726 showed a small-plaque phenotype and less virulence against suckling mice than the wild-type virus. Thus, reduction of the efficiency of translation initiation may result in the construction of enteroviruses with the lower-virulence phenotype.

A second group of hepatitis C viruses

cDNA clone 11-7 was isolated by immunoscreening a cDNA library that was prepared from a pooled plasma of non-A non-B hepatitis (NANBH) patients using expression vector lambda gt11. This cDNA corresponds to known nucleotide positions 3983-4745 of the genome of hepatitis C virus (HCV). This clone was used as a probe for screening the HCV-related cDNAs in a cDNA library similarly prepared by using lambda gt10. As a result, six more cDNA clones were isolated and analyzed for their nucleotide sequences. The results strongly suggested that there are at least two groups of HCV, group I and group II. According to our classification, the prototype HCV and clone 11-7 belong to group I HCV, and their nucleotide and deduced amino acid sequences were diverged from those of group II HCV. Genetic variation observed in the nucleotide and the amino acid sequences between the two groups resembles that in the NS3 region of the genome between Japanese encephalitis virus and West Nile fever virus. Polypeptides produced in Escherichia coli carrying a clone 11-7 or a group II cDNA clone E reacted with antibodies in the blood of 12 or 4 out of 14 individual chronic NANBH patients, respectively. Our data clearly indicate the existence of a second group of HCV.

Functional domains of the poliovirus receptor

A number of mutant cDNAs of the human poliovirus receptor were constructed to identify essential regions of the molecule as the receptor. All mutant cDNAs carrying the sequence coding for the entire N-terminal immunoglobulin-like domain (domain I) confer permissiveness for poliovirus to mouse L cells, but a mutant cDNA lacking the sequence for domain I does not. The transformants permissive for poliovirus were able to bind the virus and were also recognized by monoclonal antibody D171, which competes with poliovirus for the cellular receptor. These results strongly suggest that the poliovirus binding site resides in domain I of the receptor. Mutant cDNAs for the sequence encoding the intracellular peptide were also constructed and expressed in mouse L cells. Susceptibility of these cells to poliovirus revealed that the entire putative cytoplasmic domain is not essential for virus infection. Thus, the cytoplasmic domain of the molecule appears not to play a role in the penetration of poliovirus.

Identification of mutations that occurred on the genome of Japanese encephalitis virus during the attenuation process

The total nucleotide sequences of the genomes of two Japanese encephalitis virus (JEV) strains, the attenuated vaccine strain SA14-14-2 and its parental virulent strain SA14, were determined by using the molecular cloning technique. The sequence analysis revealed that both virion RNA molecules were 10,976 nucleotides long with 95 and 585 flanking bases at the 5' and 3' untranslated sequences, respectively. A single, long open reading frame spanning 10,296 nucleotides was observed to encode a polyprotein of 3432 amino acid residues. When these sequences were compared with each other, 57 nucleotide substitutions were found to be scattered all over the genome. Of these, 24 resulted in amino acid changes within viral proteins. Structural proteins C and E contain one and eight amino acid changes, respectively. Of the nonstructural proteins, NS1 contains three, NS2a two, NS2b two, NS3 four, NS4a one, NS4b one, and NS5 two amino acid substitutions. The 5'- and 3'-terminal untranslated regions contain one- and two- point mutations, respectively. These data and comparative studies with other JEV strain genomes provide a molecular basis for investigating attenuation mechanisms of JEV.

Preparation of Japanese encephalitis virus nonstructural protein NS1 obtained from culture fluid of JEV-infected Vero cells

The Japanese encephalitis virus (JEV) nonstructural protein NS1 was released efficiently into culture fluid of JEV-infected Vero cells. The JEV NS1 protein in the infected culture fluid was found almost as a high-molecular-weight form, probably a dimer form of NS1, and was converted to a monomer by boiling. Large amounts of NS1 protein were accumulated in the infected culture fluid. The NS1 protein, separated from JE virions by centrifugation through sucrose layer, could be obtained in large quantities.

Transgenic mice susceptible to poliovirus

Poliovirus-sensitive transgenic mice were produced by introducing the human gene encoding cellular receptors for poliovirus into the mouse genome. Expression of the receptor mRNAs in tissues of the transgenic mice was analyzed by using RNA blot hybridization and the polymerase chain reaction. The human gene is expressed in many tissues of the transgenic mice just as in tissues of humans. The transgenic mice are susceptible to all three poliovirus serotypes, and the mice inoculated with poliovirus show clinical symptoms similar to those observed in humans and monkeys. Rabbit antipoliovirus serum detects the antigens mainly in motor neurons in the anterior horn of the spinal cord and in nerve cells in the medulla oblongata and pons of the paralyzed transgenic mice. Therefore, cell types sensitive to poliovirus in the central nervous system of the transgenic mice appear to be identical to those of humans and monkeys. Furthermore, many more doses of oral poliovirus vaccine strains than of the virulent strains are required to cause paralysis in the transgenic mice. This may reflect the observation that the virulent strain multiplies more efficiently in the central nervous system than the attenuated strain. Thus, the transgenic mice may become an excellent new animal model to study molecular mechanisms of pathogenesis of poliovirus and to assess oral poliovirus vaccines.

The poliovirus receptor protein is produced both as membrane-bound and secreted forms

Both genomic and complementary DNA clones encoding poliovirus receptors were isolated from genomic and complementary DNA libraries prepared from HeLa S3 cells, respectively. Nucleotide sequence analysis of these cloned DNAs revealed that the poliovirus receptor gene is approximately 20 kb long and contains seven introns in the coding region, and that at least four mRNA isoforms referring to the coding sequence are generated by alternative splicing and appear to encode four different molecules, that is, PVR alpha, PVR beta, PVR gamma and PVR delta. The predicted amino acid sequences indicate that PVR alpha and PVR delta, corresponding to the previously described cDNA clones H20A and H20B, respectively, are integral membrane proteins while the other two molecules described here for the first time lack a putative transmembrane domain. Mouse cell transformants carrying PVR alpha were permissive for poliovirus infection, but those carrying PVR beta were hardly permissive. In contrast to PVR alpha, PVR beta was not detected on the surface of the mouse cell transformants but was detected in the culture fluid by an immunological method using a monoclonal antibody against poliovirus receptor. Three types of splicing products for PVR alpha, PVR beta and PVR gamma were detected by polymerase chain reactions using appropriate primers in poly(A)+ RNAs of the brain, leukocyte, liver, lung and placenta of humans; the choice of primers used did not permit detection of PVR delta. In situ hybridization using a cDNA fragment as a probe demonstrated that the PVR gene is located at the band q13.1----13.2 of human chromosome 19.

Construction of less neurovirulent polioviruses by introducing deletions into the 5' noncoding sequence of the genome

Viral attenuation may be due to lowered efficiency of certain steps essential for viral multiplication. For the construction of less neurovirulent strains of poliovirus in vitro, we introduced deletions into the 5' noncoding sequence (742 nucleotides long) of the genomes of the Mahoney and Sabin 1 strains of poliovirus type 1 by using infectious cDNA clones of the virus strains. Plaque sizes shown by deletion mutants were used as a marker for rate of viral proliferation. Deletion mutants of both the strains thus constructed lacked a genome region of nucleotide positions 564 to 726. The sizes of plaques displayed by these deletion mutants were smaller than those by the respective parental viruses, although a phenotype referring to reproductive capacity at different temperatures (rct) of viruses was not affected by introduction of the deletion. Monkey neurovirulence tests were performed on the deletion mutants. The results clearly indicated that the deletion mutants had much less neurovirulence than with the corresponding parent viruses. Production of infectious particles and virus-specific protein synthesis in cells infected with the deletion mutants started later than in those infected with the parental viruses. The rate at which cytopathic effect progressed was also slower in cells infected with the mutants. Phenotypic stability of the deletion mutant for small-plaque phenotype and temperature sensitivity was investigated after passaging the mutant at an elevated temperature of 37.5 degrees C. Our data strongly suggested that the less neurovirulent phenotype introduced by the deletion is very stable during passaging of the virus.

In vitro construction of poliovirus defective interfering particles

To construct poliovirus defective interfering (DI) particles in vitro, we synthesized an RNA from a cloned poliovirus cDNA, pSM1(T7)1, which carried a deletion in the genome region corresponding to nucleotide positions 1663 to 2478 encoding viral capsid proteins, by using bacteriophage T7 RNA polymerase. The RNA was designed to retain the correct reading frame in nucleotide sequence downstream of the deletion. HeLa S3 monolayer cells were transfected with the deletion RNA and then superinfected with standard virus as a helper. The DI RNA was observed in the infected cells after three passages at high multiplicity of infection. The sequence analysis of RNA extracted from the purified DI particle clearly showed that this DI RNA had the same deletion in size and location as that in the RNA used for the transfection. Thus, we succeeded in construction of a poliovirus DI particle in vitro. To gain insight into the mechanism for DI generation, we constructed poliovirus cDNAs pSM1(T7)1a and pSM1(T7)1b that, in addition to the same deletion as that in pSM1(T7)1, had insertion sequences of 4 bases and 12 bases, respectively, at the corresponding nucleotide position, 2978. The RNA transcribed from pSM1(T7)1a was not a template for synthesis of poliovirus nonstructural proteins and therefore was inactive as an RNA replicon. On the other hand, the RNA from pSM1(T7)1b replicated properly in the transfected cells. Superinfection of the transfected cells with standard virus resulted in production of DI particles derived from pSM1(T7)1b and not from pSM1(T7)1a. These observations indicate that deletion RNAs that are inactive replicons have little or no possibility of being genomes of DI particles suggesting the existence of a nonstructural protein(s) that has an inclination to function as a cis-acting protein(s). The method described here will provide a useful technique to investigate genetic information essential for poliovirus replication.