key: cord-301755-fxfsr9bj
authors: Wang, F.-I.; Fleming, John O.; Lai, Michael M.C.
title: Sequence analysis of the spike protein gene of murine coronavirus variants: Study of genetic sites affecting neuropathogenicity
date: 1992-02-29
journal: Virology
DOI: 10.1016/0042-6822(92)90041-m
sha: 
doc_id: 301755
cord_uid: fxfsr9bj

Abstract Mouse hepatitis virus (MHV), a coronavirus, causes encephalitis and demyelination in susceptible rodents. Previous investigations have shown that the MHV spike (S) protein is a critical determinant of viral tropism and pathogenicity in mice and rats. To understand the molecular basis of MHV neuropathogenesis, we studied the spike protein gene sequences of several neutralization-resistant variants of the JHM strain of MHV, which were selected with monoclonal antibodies (MAbs) specific for the S protein. We found that variant 2.2-V-1, which was selected with MAb J.2.2 and primarily caused demyelination, had a single point mutation at nucleotide (NT) 3340, as compared to the parental JHM virus, which predominantly caused encephalitis. This site was in the S2 subunit of the S protein. In contrast, variant 7.2-V-1, which was selected with MAb J.7.2 and primarily caused encephalitis, had two point mutations at NT 1766 and 1950, which were in the S1 subunit. Finally, the double mutant 2.2/7.2-V-2, which was selected with both MAbs J.2.2 and J.7.2, and was attenuated with respect to both virulence and the ability to cause demyelination, had a deletion spanning from NT 1523 to 1624 in the S1 and a point mutation at NT 3340 in the S2. We conclude that at least two regions of the S protein contribute to neuropathogenicity of MHV. We have also isolated a partial revertant of 2.2-V-1, which was partially resistant to MAb 1.2.2 but retained the same neuropathogenicity as the variant 2.2-V-1. This revenant retained the mutation at NT 3340, but had a second-site mutation at NT 1994, further confirming that NT 3340 contributed to the pathogenic phenotype of MHV. By comparing these results with MHV variants isolated in other laboratories, which had mutations in other sites on the S gene and yet retained the demyelinating ability, we suggest that the ability of JHM viruses to induce demyelination is determined by the interaction of multiple sites on the S gene, rather than the characteristics of a single, unique site. Our study also revealed the possible presence of microheterogeneity of S gene sequence, particularly in the S1 region, in these viruses. The sequence microheterogeneity may also contribute to the differences in their biological properties.

Mouse hepatitis virus (MHV) is a prototypic member of the Coronaviridae. The virus is enveloped and contains a single-stranded, positive-sensed, and nonsegmented RNA genome of 31 kb in length (Lai and Stohlman, 1978; Pachuk et al., 1989; Lee er al., 1991) . The virus particle contains three envelope glycoproteins (S, M, and HE) and an internal nucleocapsid protein (N). The S (spike) protein of 180 kDa forms peplomers on the virion surface and is responsible for the viral attachment to cellular receptors, the elicitation of neutralizing antibodies, and the induction of cell-to-cell fusion (Collins eta/., 1982; Sturman eta/., 1985) . Alterations in the S protein resulted in changes in viral pathogenicity, as demonstrated by neutralization-escape mutants se-' Current address: Departments of Neurology and Medical Microbiology and Immunology, School of Medicine, University of Wisconsin at Madison, Madison, WI 53792.

' To whom correspondence and reprint requests should be addressed. Q 1992 Academic Press, Inc. lected with monoclonal antibodies (MAbs) specific for the S protein (Fleming et a/., 1986 (Fleming et a/., , 1987 Dalziel et al., 1986; Wege et a/., 1988) . Furthermore, passive immunization with S-specific monoclonal antibodies also altered viral neuropathogenicity (Buchmeier et a/., 1984) .

The S protein is cleaved into two 90-kDa glycoprotein subunits (Sl and S2). Its cleavage is carried out by a cellular protease and is probably required for the virusinduced cell fusion (Storz et a/., 1981; Yoshikura and Tajima, 1981; Sturman et al., 1985) . The S protein is translated from mRNA 3, which represents the transcript of the 4.2-kb gene (S gene), located at roughly 6-10 kb from the 3'-end of the RNA genome (Lai, 1990) . The complete sequence of the S gene has been reported for three strains of MHV, namely, A59 (Luytjes et a/., 1987) JHM (Schmidt et a/., 1987) , and MHV-4 (equivalent to JHM used in our laboratory) (Parker et a/., 1989 ). It appears that the S gene sequences of different MHV strains differ significantly, with various degrees of deletion and substitutions. Infection of the central nervous system (CNS) of rodents by MHV, particularly the JHM strain, has been used as a model system for studying human demyelinating diseases, such as multiple sclerosis. Parental JHM strain shows tropism for both neurons and glial cells, causing both encephalitis and chronic demyelination (Weiner, 1973) . The tropism of MHV is mainly determined by the properties of the S protein, since it is the outermost component of the viral envelope, which interacts with the cellular receptors, and it is also the immunodominant component which interacts with the host immune system (Sturman and Holmes, 1983) . The role of the other structural proteins in viral pathogenesis is less well understood; passive immunization with MAbs specific for N and M (membrane) proteins also changed viral pathogenicity (Fleming et a/., 1989; Nakanaga et a/., 1986) .

Previously, we have isolated several neutralizationresistant variants of JHM using MAbs specific for the S protein (Fleming et a/., 1983) . These variants showed strikingly different capabilities from that of the parental JHM virus in causing demyelination or encephalitis in rodents (Fleming eta/., 1986 (Fleming eta/., , 1987 . While the parental virus caused encephalitis, the variants selected with one neutralizing MAb (J.2.2) had lower virulence and caused demyelination, and the variants selected with two MAbs (J.2.2 and 1.7.2) caused neither encephalitis nor demyelination (Fleming et al., 1986 (Fleming et al., , 1987 . Based on these studies, we hypothesized that MAb J.2.2 identified a specific encephalitis-determining site and MAb J.7.2 identified a specificdemyelination-determining site on the S protein. Alternatively, we considered the possibility that a single linear determinant might not be the necessary and sufficient cause of JHM-induced demyelination, and that demyelination would depend on several regions of S or even on other viral genes (Fleming et a/., 1987) . Our present sequence analysis indicates that the latter hypothesis is correct; that is, the neuropathogenicity of JHM depends on the characteristics of at least two sites on the S gene. Complex interaction of different regions of S protein likely influences viral pathogenesis by changing the degree of viral attenuation in viva.

The viruses used in this study included several JHM variants which were selected for their resistance to neutralizing MAbs specific for the S protein (Fleming et a/., 1983) . Their virulence and neuropathogenicity in mice are listed in Table 1 . The parental virus JHM-DL was a large-plaque isolate of wild-type JHM strain and caused marked encephalitis with chronic demyelination in survivors (Stohlman et al., 1982) . The 2.2-V-l Note. MAb, Monoclonal antibody; LD,,, 50% lethal dose in C57/ B6 mice on day 7 after intracerebral inoculation; PFU, plaque forming units; E, predominantly encephalitis; D. predominantly demyelination. (Fleming et al., 1986 (Fleming et al., , 1987 . variant was selected from JHM-DLwith MAb J.2.2, and caused marked demyelination with little encephalitis (Fleming et al., 1986) . The 2.2/7.2-V-2 variant was selected from 2.2-V-l with another MAb J.7.2 and was nonpathogenic, causing only minimal neuropathology (Fleming eta/., 1987) . The 7.2-V-l variant was selected directly from JHM-DL with J.7.2 and remained encephalitic (Fleming et a/., 1986 ). An astrocytoma cell line (DBT cell) derived from mice (Hirano et a/., 1974) was used throughout the experiment. All of these viruses were initially plaque-purified three times before propagation as virus stock. Early passage viruses (before passage 3) were used for preparing large amounts of virus for biochemical studies. The viruses of the same passage level were previously used for pathogenicity studies (Fleming et a/., 1986 (Fleming et a/., , 1987 .

Variant viruses were grown in the presence of the respective MAb used for their selection. Viruses were mixed with an equal amount of MAb tissue culture supernatants and incubated for 45 to 60 min at 37" before inoculating onto confluent DBT cell monolayer. After virus adsorption for 1 hr at 37", Dulbecco's minimal essential medium containing 1% fetal calf serum was added. The supernatants from infected cell cultures were harvested when the cytopathic effects reached 90-l 00%. The harvested viruses were examined for their resistance to the selecting MAbs to ensure that the viruses used for biochemical studies maintained the original phenotype. The JHM-DL virus was grown in the same way except that no antibodies were used.

Viruses were purified according to a procedure modified from Makino et a/. (1983) . Briefly, the supernatants were centrifuged at 8000 rpm for 30 min to clear cell debris and were precipitated with 50% saturated am-monium sulfate for at least 1.5 hr at 4", followed by centrifugation at 8000 rpm for 20 min. The pellets were suspended in NTE (0.1 M NaCI, 0.01 MTris-HCI, pH 7.4, and 1 mM EDTA), and the suspensions were centrifuged at 2000 rpm for 10 min to remove undissolved debris. The supernatants were then subjected to two rounds of centrifugation in discontinuous sucrose gradient consisting of 20, 30, 50, and 60% sucrose in NTE at 26,000 rpm for 3.5 hr at 4" in a Beckman SW28 rotor. Viruses at the 30-500/o interface were harvested and diluted with NTE and pelleted by centrifugation at 40,000 rpm for 90 min in a Beckman SW41 rotor. The virus pellets were resuspended in NTE.

Isolation of partial revertants of neutralizationescape variant 2.2-V-l Variant 2.2-V-l was passaged in DBT cells for 10 times in the absence of any antibodies. These viruses were plaque-assayed, and plaques were randomly picked and further purified by three rounds of plaque purification. Several plaques were examined for their sensitivity to MAb J.2.2. Large pools of revertants were prepared and purified following the same procedure as described above. These plaque-purified revertants were designated as Vl 0. cDNA cloning For viral RNA extraction, the virus suspension in NTE was mixed with an equal volume of proteinase K buffer (2% SDS, 0.025 M EDTA, 0.2 M Tris-HCI, pH 7.5, and 0.3 M NaCI) and proteinase K (200 pg/mI), and incubated for 45 min at 37". RNA was extracted twice with phenol/chloroform (1 :l) and precipitated with ethanol. The pelleted RNA was resuspended in distilled H,O, and stored at -70".

For cDNA synthesis, viral RNA was first denatured with 10 mM methylmercury at room temperature for 10 min and then incubated at 42" for 45 to 60 min in 50 ~1 pM of dNTPs, l-3 units of Taq polymerase (Cetus), and 0.1-l pM of synthetic oligonucleotide primers (first and second primers). The reaction mixture was overlaid with 60 ~1 of mineral oil and subjected to 20-40 cycles of amplification, each cycle consisting of 94" for 1 min, 37-42" for l-2 min, and 72" for 2-3 min. These PCR conditions were adjusted slightly for different reactions, depending on the size of target sequence to be amplified and the pair of primers used.

The PCR product was resolved on a 1 Oh agarose gel in 1X TBE buffer (0.089 MlTris-HCI, pH 8.3, 0.089 M boric acid, and 2 mM EDTA) and stained with ethidium bromide. The specific product band was cut out, and the DNA was extracted from the gel with phenol and chloroform, precipitated with ethanol, and the pellet suspended in distilled H,O. The PCR product was treated with 10 units of polynucleotide kinase (Boehringer-Manheim) in buffer containing 0.07 M Tris-HCI, pH 8.0, 0.01 M MgCI,, 5 mM DlT, 0.1 mM spermidine, and 1 mM ATP and incubated at 42" for 45 min. After phenol/chloroform extraction and ethanol precipitation, the PCR product was blunt-ended with 1 unit of T4 DNA polymerase (Boehringer-Manheim) at 37" for 30 min in buffer containing 0.05 MTris-HCI, pH 7.5, 7 mM MgCI,, 1 mM DlT, 0.25 mM of dNTPs, followed by a booster reaction with an additional 0.5 unit of T4 DNA polymerase and 0.05 mM of dNTPs at 37" for another 30 min. After phenol/chloroform extraction, the PCR product was ligated using 1 unit of T4 DNA ligase (Boehringer-Manheim) into the Smal site of plasmid pTZ18U (U.S. Biochemicals).

The ligation buffer contained 0.05 M Tris-HCI, pH 7.5, 0.01 M MgCI,, 2 mM DTT, and 1 mM ATP. The ligation mixture was used to transform competent cell DH5a (BRL) following the manufacturer's direction.

DNA sequencing DNA sequencing was performed using doublestranded DNAof plasmid pTZ18U. Dideoxychain termination sequencing was performed following the procedure outlined in the Sequenase version 2.0 DNA sequencing kit from U.S. Biochemical. The reaction product was analyzed in 69/o polyacrylamide gel containing 7 M urea.

The whole S gene, approximately 4.2 kb in length, was cloned into 2 to 3 overlapping fragments after amplification by PCR using appropriate synthetic primers. Since we were dealing with several viral isolates with presumably only minor differences in S gene, crosscontamination was a major concern. To avoid this problem, whenever possible, different pairs of primers were used in the cloning of different viral RNAs so that distinguishable gene fragments were obtained for each viral isolate. In cases where alternative primers were not available, new aliquots of reagents were used for each isolate. Another concern was that the enzymes used, namely, reverse transcriptase, Taq polymerase, and Sequenase may generate errors. To avoid this problem, clones from different cDNA synthesis events, different PCR events, and multiple clones were sequenced. For each isolate, the full-length sequence was initially obtained from two different DNA clones. Differences in sequence identified by this initial sequencing were further studied by sequencing more clones. Only consistent changes were considered as bona fide mutations. Some of the sequence variations were confirmed by direct RNA sequencing using purified virion RNA (Banner et a/., 1990) .

The complete sequences of S gene of the parental JHM-DL and their neutralization-escape variants are summarized schematically in Fig. 1 . The S gene of JHM-DL is nearly identical to that of MHV-4 (Parker et a/., 1989) with only 6 nucleotide differences (3 amino acid differences) ( Table 2 ). It has a coding capacity of 1376 amino acids. Identical to those in other MHVs (Schmidt et al., 1987; Luytjes et al., 1987) , the predicted proteolytic cleavage site is between amino acid 769 (Arg) and 770 (Ser). The S gene of 2.2-V-1, which was derived from JHM-DL by selection with MAb J.2.2, appeared to consist of two populations. All of the clones contained a nucleotide difference between JHM-DL and 2.2-V-1, which was located at nucleotide (NT) 3340, changing the codon from CTC to TX (Leu to Phe). Besides this mutation, a minority of clones contained four additional mutations at the N-terminal 

Note. MHV-4 sequence was from Parker et al. (1989) . Note. Sequence variations in 2.2-V-l were found consistently in two cDNA clones, which were derived from different reverse transcription events and polymerase chain reactions. In contrast, sequence variations in JHM-DL were found randomly in different clones. Frequency indicates the number of clones showing the alternative nucleotide out of the total number of clones that were sequenced.

end of the S gene (Table 3 ). These were located at NT 731, 901, 950, and 1439, changing the codon from lTT to TCT (Phe to Ser), from CCG to TCG (Pro to Ser), from GGA to GAA (Gly to Glu), and from CCG to CTG (Pro to Leu), respectively. These mutations were not likely to be due to enzyme errors, because these mutations were observed on two clones derived from different reverse transcription and PCR events. Sequencing of multiple JHM-DL-derived clones also showed heterogeneity at several different nucleotide positions (Table  3) , which were different from the sequence variations observed in 2.2-V-l. Since the only consistent difference between JHM-DL and 2.2-V-l was NT 3340, this mutation was most likely responsible for the resistance of 2.2-V-l to MAb J.2.2. However, the contribution of sequence microheterogeneity (Table 3) to the phenotype of viruses could not be ruled out. It should be noted that most of the minor sequence variations were in the N-terminal half, which was the more divergent part of the S gene (de Groot et a/., 1987).

The variant 2.2/7.2-V-2 was derived from 2.2-V-l by selection with MAb J.7.2 (Fleming et a/., 1987) . Sequence analysis showed that 2.2/7.2-V-2 had a 102nucleotide deletion at the N-terminal half between NT 1523 and 1624, corresponding to a 34-amino acid deletion. Accompanying the deletion, the two amino acids arginine and tyrosine at the junction of deletion were converted into histidine. As expected, NT 3340 in this variant remained as TTC, similar to that in 2.2-V-l. Also, similar to 2.2-V-1, a small number of clones contained several additional point mutations at the up-stream and downstream regions of the deletion (data not shown). These sequence variations were different from those observed in the variant 2.2-V-l. Since the 102-nucleotide deletion was the only consistent change observed in all clones of 2.2/7.2-V-2, it was most likely responsible for the phenotype differences.

To confirm that the deletion in the N-terminal half of S was responsible for the viral resistance to MAb J.7.2, we sequenced the S gene of 7.2-V-1, which was derived directly from JHM-DL by selection with MAb J.7.2 (Fleming et a/., 1986) . To our surprise, 7.2-V-l did not have a deletion. Instead, there were two point mutations consistently detected downstream of the 2.2/ 7.2-V-2 deletion site, one at NT 1766, changing the codon from TCA to TTA (Ser to Leu) and the other one at NT 1950, changing the codon from AAG to AAT (Lys to Asn). All of these changes were localized around the hypetvariable region (de Groot et al., 1987; Parker et a/., 1989; Banner et a/., 1990) of the S protein. These results suggested that the mutations, including deletions and point mutations, in this region were most likely responsible for the viral resistance to MAb J.7.2. NT 3340 of 7.2-V-l remained as CTC, similar to that in JHM-DL, consistent with the fact that 7.2-V-l was directly derived from JHM-DL.

To determine whether the mutation at NT 3340 or the microheterogeneity in Sl was responsible for the viral resistance to MAb J.2.2 and for the pathogenicity of the virus, we attempted to isolate revertants of 2.2-V-l. Variant 2.2-V-l was passaged in DBT cells in the absence of any MAb, and viruses were randomly selected from plaques. Since the parental JHM-DL grew better than the neutralization-resistant variants (unpublished observations), we reasoned that the revertants would become a predominant population during serial viral passages in the absence of selection pressure. Indeed, most of the plaque-purified viruses isolated this way were partially sensitive to MAb J.2.2 (data not shown). Their titers were reduced by MAb J.2.2 by approximately 2.7 log,,,. However, none of the viruses examined (N = 7) completely reverted to the level seen in parental JHM-DL, whose titer was reduced by MAb J.2.2 by approximately 4 loglo (Fleming et al., 1986 (Fleming et al., , 1987 . Sequence analysis of multiple clones from all seven partial revertants showed that all of them have retained the sequence of 2.2-V-l at NT 3340, which was the only consistent difference observed between JHM-DL and 2.2-V-l. One of these revertants was chosen for complete sequencing; this virus had a consistent change at a second site at NT 1994, changing the codon from AAT (Asn) to AGT (Ser). Since none of the revertants completely restored the phenotype to that of the parental JHM-DL, the most logical explanation was that different regions of the S protein may cooperate to contribute to the resistance of the virus to neutralization by MAb J.2.2. The finding that NT 3340 mutation was retained in all of the partial revet-tants suggested that this site had a major contribution to the resistance to MAb J.2.2. Preliminary viral pathogenicity study by intracerebral inoculation in mice revealed that VlO had an LD,, of approximately 1 O3 PFU, similar to that of 2.2-V-1, but in contrast to JHM-DL, which had an LD,, of 4 PFU (Table 1) (Fleming et a/., 1986) . Similar to 2.2-V-1, the partially revertant virus primarily caused demyelination with little encephalitis. These results suggested that the second site mutation in the Sl region did not alter the pathogenicity of the virus and that the mutation at NT 3340 was most responsible for the neuropathogenicity of the JHM viruses.

Our sequence analysis of MAb-escape mutants of JHM indicated that a mutant selected with MAb J.2.2 had a point mutation in the S2 region of the spike protein. By contrast, when MAb J.7.2 was used to select mutants, point mutations or a deletion in Sl within NT 1500-l 950 were observed. Previous attempts at mapping the binding sites of these MAbs have failed (Weismiller eta/., 1990). Our study suggested that MAb J.7.2 would bind to Sl while MAb J.2.2 would bind to S2. These findings are consistent with a recent study measuring the direct binding of the MHV-specific MAb, which also indicated that J.2.2 bound to S2 (P. Talbot, personal communication). The binding site of J.2.2 is close to other 52 neutralizing epitopes identified in other laboratories (Luytjes et al., 1989; Routledge et al., 1991) . Interestingly, the mutation at NT 3340, which changed the coded amino acid from leu to phe affected one of the three amino acids which are critical for the fusion function of the S protein (Gallagher et al., 1991) . However, the variant 2.2-V-l induced fusion when grown on DBT cells (data not shown); this was expected since the loss of the fusion activity required the simultaneous mutations of all the three amino acids (Gallagher er al., 199 1) . Nevertheless, these findings show that this amino acid plays an important role in both fusion activity and neuropathogenicity of MHV. The region of Sl with deletion or point mutations in 2.2/7.2-V-2 or 7.2-V-l mutants corresponds to a previously identified hypervariable site on Sl (Parker et a/., 1989; Banner et a/., 1990) . This site underwent frequent deletions or point mutations upon viral passages in tissue culture or animals (Dalziel eta/., 1986; Banner et a/., 1990; Gallagher et a/., 1990) . Although the precise sites of deletion or point mutations were different among these variants, they were in the same general area, i.e., near the C-terminus of Sl Previously, Buchmeier's laboratory also has reported different types of mutation among the neutralization-escape MHV-4 variants, which were derived by use of a monoclonal antibody different from the one used in our study (Parker er a/., 1989; Gallagher et al., 1990) . All of these variants have either deletion or point mutations in this area of Sl These results together suggested that this region represented an immunodominant and highly variable site, although the binding site of MAb J.7.2 has not been directly determined.

Taken together with previous studies, our analysis indicates that at least two sites on S play critical roles in MHV neuropathogenesis:

(1) the hypervariable Sl region within NT 1500-l 950 and (2) a second region on S2 around NT 3340. Preservation of the parental (wildtype) sequences at both of these regions appears to be required for the ability of virus to cause fatal encephalitis, since mutations at either site reduced neurovirulence.

The mutation in S2 (2.2-V-l) converted the parental JHM virus, which was highly encephalitic, into a highly demyelinating virus, and an additional mutation (deletion) in Sl (2.2/7.2-V-2) led to an essentially non-demyelinating virus (Fleming et al., 1987) (Table 1) . This result was consistent with the interpretation that the site on Sl was tightly associated with demyelination, while the site on S2 was responsible for encephalitis. However, this interpretation is unlikely, since other neutralization-resistant mutants with deletions in the same domain of Sl still caused demyelination (Dalziel et al., 1986; Gallagher et a/., 1990) . Thus, we hypothesize that the virtual inability of 2.217.2-V-2 to cause demyelination or encephalitis (Fleming et a/., 1987) was due to viral attenuation that occurred when both the Sl and S2 determinants were lost or mutated. Such an interpretation is consistent with the finding that the mutation in S2 of the variant 2.2-V-l and the mutation in Sl of other mutants (Dalziel et al., 1986) all led to a similar demyelinating phenotype. This is also consistent with our current understanding of the biological activities of the S protein.

The S protein forms peplomers which project from the surface of the MHV virion, and is thus in an ideal position to influence pathogenesis by altering viral attachment to cellular receptors, by interaction with the immune system, by fusion of cellular membranes (Collins et a/., 1982; Sturman et a/., 1985) and perhaps by other mechanisms, such as influencing viral stability, uncoating, or spread. After intracerebral inoculation, parental JHM-DL rapidly gains entry to glia and neurons throughout the CNS and thus is able to cause fatal panencephalitis (Weiner, 1973; Knobler et a/., 1981) . By contrast, MAb-selected JHM mutants (2.2-V-1 and 2.2/7.2-V-2) are less able to spread to neurons, and virulence is thus drastically reduced (Fleming et a/,, 1986) . However, these viruses are able to infect glial cells of the white matter of the brain and spinal cord. Presumably the requirements of viral entry or replication in glial cells are less stringent than those for neuronal infection. This differential entry may reflect the presence of different receptors utilized by wild-type virus versus selected variants. In addition, the susceptibility of different neural cells to JHM is known to be affected by the developmental stages of different cell species and host factors (Pasick and Dales, 1991) . In the case of infection by JHM mutants, the immune response clears virus from the white matter by 12 days postinoculation but has the deleterious consequence of inducing myelin loss (Wang eT al., 1990) . Thus, the apparent increase in ability to cause demyelination shown by these and similar variants (Dalziel et a/., 1986; Wege et a/., 1988) is probably due to the unmasking of potentialities of the parental virus which are usually hidden by its virulence.

The ability of mutations at different sites to affect the same biological properties of the S protein was further demonstrated by the sequence analysis of a partial revertant of 2.2-V-l. The second site mutation which partially restored the MAb sensitivity of this virus was localized on Sl, in contrast to the original mutation on S2. This study suggested that two sterically separated sites (one in the globular portion, and the other in the heptad repeat region of the stalk portion according to the computer modeling [de Groot et al., 19871) on the two different S subunits could interact with each other. This effect has also been illustrated by the antibody sensitivity of 2.2/7.2-V-2 and 2.2-V-l. The former was roughly one loglo more resistant than the latter to MAb J.2.2 (Fleming eta/., 1987) , suggesting that the second mutation further altered the conformation at the first site (in S2). Interestingly, the partial revertant of 2.2-V-l retained the same pathogenic properties of the parent variant 2.2-V-l. Furthermore, 7.2-V-l variant retained the ability to cause encephalitis (Table 1) . These results suggest that the antibody binding and pathogenic properties of the S protein can be dissociated.

Our study also suggested the possible presence of sequence microheterogeneity in the MHVs (Table 3 and data not shown). All of these viruses were initially plaque-purified but had been amplified by limited passages to obtain sufficient virus titer before molecular cloning and sequencing. Some of these sequence variations were likely the results of PCR and cloning errors. However, several nucleotide variations were consis-tently noted in repeated PCR products. Furthermore, the sequence microheterogeneity was located mostly in the Sl hypervariable region (Banner et a/., 1990; de Groot et a/., 1987) , suggesting that the observed sequence variations represented variations in the RNA. Since the plaque-purified viruses need to be propagated for either biochemical or pathogenicity studies, this sequence microheterogeneity is likely an inherent property of any virus population. The possible contribution of the sequence microheterogeneity to viral neuropathogenicity is not clear at the present time. In conclusion, our studies suggest that at least two regions of the S gene play critical roles in JHM neuropathogenesis. These regions are near NT 1500-l 950 (Sl segment) and NT 3340 (S2 segment). Antigenic studies indicate that these sites may be closely linked on the S protein and may influence each other by spatial proximity or steric effects (Fleming et al., 1986) . Viruses with the parental JHM sequences may infect all cell types and regions of the CNS, resulting in fatal panencephalitis. When either site is altered, viral infection is limited to glial cells of the white matter, and subsequent immune attack (Wang et al., 1990) or virus-induced cytolysis will lead to demyelination. When both sites are affected, the virus shows very limited distribution and presumably is contained by the immune system at an early stage of pathogenesis. It is probable that other S domains may also contribute to the viral pathogenicity. In addition, although the S gene is considered to be the most important in influencing pathogenicity, other viral genes such as hemagglutinin-esterase (HE) (La Monica et a/., 1991, Yokomori et al., 1991) and those localized to the 3' portion of the genome (Lavi et a/., 1990) may also, to some extent, contribute to viral pathogenicity.

A clustering of RNA recombination sites adjacent to a hypervariable region of the peplomer gene of murine coronavirus

Murine hepatitis virus-4 (strain JHM) induced neurologic disease is modulated in viva by monoclonal antibody. I/iro/ogy

Monoclonal antibodies to murine hepatitis virus-4 (strain JHM) define the viral glycoprotein responsible for attachment and cell-cell fusion

Site-specific alteration of murine hepatitis virus type 4 peplomer glycoprotein E2 results in reduced neurovirulence

Evidence for a coiled-coil structure in the spike proteins of coronaviruses. 1

Monoclonal antibodies to the matrix (El) glycoprotein of mouse hepatitis virus protect mice from encephalitis

Antigenic relationships of murine coronaviruses: Analysis using monoclonal antibodies to JHM (MHV-4) virus

Pathogenicity of antigenic variants of murine coronavirus JHM selected with monoclonal antibodies

Experimental demyelination induced by coronavirus JHM (MHV-4): Molecular identification of a viral determinant of paralytic disease

Alteration of the pH dependence of coronavirus-induced cell fusion: Effect of mutations in the spike glycoprotein

Neutralization-resistant variants of a neurotropic coronavirus are generated by deletions within the amino-terminal half of the spike glycoprotein. 1. viral

Replication and plaque formation of mouse hepatitis virus (MHV-2) in mouse cell line DBT culture

Mouse hepatitis virus type 4 (JHM strain)-induced fatal central nervous system disease. I. Genetic control and the murine neuron as the susceptible site of disease

Coronavirus: Organization, replication and expression of genome

RNA of mouse hepatitis virus

Localization of extensive deletions in the structural genes of two neurotropic variants of murine coronavirus JHM

Determinants of coronavirus MHV pathogenesis are localized to 3' portions of the genome as determined by ribonucleic acid-ribonucleic acid recombination

The complete sequence (22 kilobases) of murine coronavirus gene 1 encoding the putative proteases and RNA polymerase

Amino acid sequence of a conserved neutralizing epitope of murine coronavirus

Primary structure of the glycoprotein E2 of coronavirus MHV-A59 and identification of the trypsin cleavage site

Characterization of small plaque mutants of mouse hepatitis virus

Protective effect of monoclonal antibodies on lethal mouse hepatitis virus infection in mice

Molecular cloning of the gene encoding the putative polymerase of mouse hepatitis coronavirus strain A59

Sequence analysis reveals extensive polymorphism and evidence of deletions within the E2 glycoprotein gene of several strains of murine hepatitis virus

Infection by coronavirus JHM of rat neurons and oligodendrocyte-type-2-astrocyte lineage cells during distinct developmental stages. I. viral

Analysis of murine coronavirus surface glycoprotein functions by using monoclonal antibodies

Primer-directed enzymatic amplification of DNA with a thermostable DNA polymerase

Nucleotide sequence of the gene encoding the surface projection glycoprotein of coronavirus MHV-JHM

Murine coronaviruses: isolation and characterization of two plaque morphology variants of the JHM neurotropit strain

Enhancement of plaque formation and cell fusion of an enteropathogenic coronavirus by trypsin treatment

The molecular biology of coronaviruses

Proteolytic cleavage of the E2 glycoprotein of murine coronavirus: Activation of cell fusing activity of virions by trypsin and separation of two different 90K cleavage fragments

Demyelination induced by murine hepatitis virus JHM strain (MHV-4) is immunologically mediated

The peplomer protein E2 of coronavirus JHM as a determinant of neurovirulence: definition of critical epitopes by variant analysis

Pathogenesis of demyelination induced by a mouse hepatitis virus (JHM virus). Acfa Neural

Monoclonal antibodies to the peplomer glycoprotein of coronavlrus mouse hepatitis virus identify two subunits and detect a conformational change in the subunit released under mild alkaline conditions

Heterogeneity of gene expression of the hemagglutinin-esterase (HE) protein of murine coronaviruses

Role of protease in mouse hepatitis virus-induced cell fusion

We thank Drs. Y-C. Chao and C-Z. Lee for technical advice. We also thank Daphne Shimoda for the manuscript preparation. This work was supported by U.S. Public Health Services Grants NS1816 from the National Institutes of Health and RG 2283-A-2 of the National Multiple Sclerosis Society. F.I.W. is a recipient of postdoctoral training fellowship from NIH. M.M.C.L. is an Investigator of Howard Hughes Medical Institute.