VIROLOGY
188,388-390
(1992)
Comparative Analysis of Sequence Variation in gpl16 and gp55 Components of Glycoprotein B of Human Cytomegalovirus SUNWEN CHOU Medical and Research Services, VA Medical Center; and Division of Infectious Oregon Health Sciences University, Port/and, Oregon 9720 1 Received November
Diseases,
15, 199 1; accepted January 22, 1992
Sequence variation in the gpl16 component of cytomegalovirus envelope glycoprotein B was examined in 11 clinical strains and compared with variation in gp55. The peptide variation in gpl16 was found to be strongly clustered at codons 27-67,440-460, and to a lesser extent at codons 181-l 94, 311-317, and 387-397. Strains adopted one of three to four peptide configurations at these loci, usually consistent with their gp55 sequence configuration. Two instances were observed of a sequence variation arising from recombination within gB. The limited, largely group-specific nature of variation in both gpl16 and gp55 facilitates functional and immunologic analyses. o 1992 Academic Press.
Inc.
The major envelope glycoprotein (gB) of human cytomegalovirus (CMV) is expressed as a precursor molecule of about 906 amino acids which is glycosylated and cleaved at a specific site (codon 461) to form a disulfide-linked complex of gp55 and gpl 16 (I, 13, 14). Glycoprotein gp55, encoded by the 3’ portion of the gB sequence, has a hydrophobic transmembrane segment towards its carboxyl terminus. Human and murine neutralizing antibodies recognize epitopes in both gp55 and gpl16 (6, 10, 13-15). Partial codihg sequence data (codons 384-7 17) revealed that clinical strains of CMV adopted one of four group-specific peptide sequence configurations in the vicinity of the gB cleavage site at codons 448-480 (4). Because of the biologic and immunologic importance of gB, additional sequences from clinical CMV strains are reported here in order to compare variation in gpl 16 with that observed in gp55. Efforts were aimed at providing a complete gB coding sequence for at least one clinical strain from each of the four gp55 groups and a gpl16 coding sequence (codons l-460) for at least two strains from each group. Eleven strains were selected based on diversity of restriction enzyme digest profiles of amplified gB sequences. Partial gB sequences (codons 384 to 7 17) of nine of these strains have been reported (4). The two new ones (CO82 and C076) were grouped by restriction analysis (4) and subsequently confirmed to have peptide sequences at codons 384-717 identical to strains C327 (group 1) and C359 (group 3) respectively. Therefore, four of the sequenced strains were in group 1, two in group 2, three in group 3, and two in group 4. They were all genetically distinct (3) and were 0042.6822/92 $3.00 CopyrIght 0 1992 by Academic Press, Inc. All rights of reproduction I” any form resewed.
obtained from epidemiologically unrelated organ trans plant recipients. DNA was extracted for analysis within 12 culture passages. Sequencing was performed using PCR-generated templates attached to streptavidin beads as previously described (4). Overlapping coverage of sequences was achieved by using primers of conserved sequence spaced approximately 300 bases apart, representing both DNA strands of the region sequenced. New sequences were aligned with gB sequences of laboratory strains Towne and AD1 69 (13). Observed gB coding sequences varied in length from 905 to 907 codons, because of codon deletions at one or more of seven sites. The codon numbering used here is from the Towne strain (907 codons); this leaves two codons (inserted after codons 28 and 36) which are present in some other strains but not in Towne. Figure 1a shows the nonuniform distribution of variable codons which affect peptide encoding. Two major loci of variation were observed at codons 27-67 and at codons 448-480. Lesser variant loci in gpl16 were noted at codons 181-l 95, 312-316, and 387408. The remaining variable codons usually main tained amino acid identity within groups of strains. Isolated mutations which were specific to a single strain were few and conservative; the number of such mutations which affected peptide encoding ranged from 0 to 6, with most strains being at the low end of this range. In gp55, a high degree of peptide conservation was observed beyond codon 500, along with extensive nucleotide conservation beyond codon 700. There were 1 1 cysteine residues in gpl 16 (four in the signal sequence, codons l-23) and five cysteine residues in gp55. All these residues were conserved, although an 388
SHORT COMMUNICATIONS
389
(4
Group 1 Towne c327
SSSSTXRGTSATHSXHHSSHTTSAAHSRSGSVSQRVTSSQTVSHGVNETIYNTTSHSQCYSSYSRVIAG ________________________________________--------------------_____-__
CO82 Cl78 C325
---X-SHA--S--NGS-T-R----QTX--V-X--H----EA---RA--------
---X----------------------------------------------------X--------------------------------------------------
___-___________ -_---_-________ ----------___--
RPNSALE _______ _ _ _ __ _ _ _ _ _ __ _ _ _ _ _ __ - _
Group 2 -----SHA--S--NGS-T-R----QTX--VyX--H----EA---RA------------SHA--S--NGS-T-R----QTX--V-X--H----EA---RA------------SHA--S--NGS-T-R----QTX--V-X--H----EA---RA--------
,(FA----------GKFA----------GKFA----------G-
---A-P---A-P---A-P-
-----SHA--SA-NGS-T-R----QTX--V-X--H----EA---RA------------SHA--S--NGS-T-R----QTX--V-X--H----EA---RA------------SHA--S--NGS-T-R----QTX--V.X--H----EA---RA--------
R-------------R-------------R--------------
-A---P-A---P-A---P-
Cl94 Cl28
---X-S-A--X--NGN-T------QTX-------EA---RA----------X-S-A--X--NGS-I------QTX-------EA---RA--------
_______________ R- _ _ _ _ _ _ _‘__ _ _ _ _
__- _- - _ _ _ _ __ _
Codon
24
AD169 C336 C338
Group 3 CO76 c359 c354
Group 4
FIG. 1. (a) Map of gB codons showing variation at peptide level. sometimes in one strain only (shorter lines). Closed circles indicate cleavage site (codon 461). (b) Amino acid sequence alignments of sites of variation are shown. Strains are segregated according to
75
181
195
311
317
Variation usually occurred in more than one strain (longer vertical lines), but potential glycosylation sites (Asn-X-Ser or Asn-X-Thr). Arrow indicates gp55 clinical strains and strain AD1 69 with Towne strain (top line). Three separate gp55 group (see text). X = Deletion.
alternative codon was used by different strains in several locations. There were 21 potential glycosylation sites (Asn-X-Ser or Asn-X-Thr, Fig. 1a). Each strain had 20 sites as a result of deletion of a site at either codon 37 or 456. Figure 1b shows the amino acid alignments of three variant loci in gpl16. In the major variant locus (codons 27-67) only three peptide configurations were discernable; strains of gp55 groups 2 and 3 were indistinguishable at this locus. All except one of the gp55 group 1 strains (including Towne) adopted a configuration quite different from group 2/3 strains, with 19 of 41 amino acid residues affected, as well as one potential glycosylation site. The configuration of group 4 strains was quite similar to that of group 213 strains. In the variant loci beyond codon 67, assortment into groups was also apparent, although groups 1 and 4 were very similar at codons 181-l 95 and 311-317. However, groups 2 and 3 were first differentiated at these loci. Definitive features identifying the four groups first appeared beyond codon 448. This major variable locus (4) spans the gpll6/gp55 cleavage site and has maintained its group character across the site in all 16 strains for which we have examined the sequence to date.
The only gp55 group 1 strain not observed to have a typical group 1 sequence at codons 27-67 was found to be a recombinant. Strain C325 adopted a nucleotide sequence close to AD1 69 (and other group 2/3 strains) from codons 1 through 69 and thereafter crossed over to a group 1 configuration. Strain C325 was plaque purified directly from a urine specimen and has remained stable through multiple passages (2). Strain C327 (group 1) was also plaque purified from the same specimen but is distinct at the nucleotide level and is unrelated to C325. New sequence data from beyond codon 717 revealed that laboratory strain Towne, classified as gp55 group 1, was in fact a recombinant which crossed over to group 4 at a site between codons 501 and 520 and thereafter maintained a group 4 sequence to the end of gB. Although the gp55 groups differ little in peptide encoding beyond codon 500, the amino acid residues at codons 768,780,802, and 853 of Towne strain are characteristic of group 4. Recombinant strains which cross over from one variant group to another indicate that it is not essential for strains to adopt a specific group character through all of gB. However, we have not encountered crossover within any variant locus of gB, particularly the two major loci 27-67 and448-480.
390
SHORT COMMUNICATIONS
The relatively limited gB variation encountered in a herpesvirus as widespread and long established in human populations as CMV prompts the question of whether similar variants are seen worldwide, or whether our strains reflect local epidemiology. Further study of isolates from diverse sources will be required, but evidence is accumulating that the gB groups we have seen are widespread, if not necessarily in the same relative distribution. Recently, Lehner eT a/. (7) reported partial gB coding sequences from five European clinical strains, covering the two major variant loci. Although the data provided did not fully establish group relationships, the five strains had peptide configurations at those loci which placed them in groups 1 (two strains), 3 (two strains), and 4, similar to what could be expected from five random isolates from our area. Maintenance of gB variant groups is probably the result of functional requirements and host factors, such as receptor structure and host immune responses. Epitopes of gB relevant to neutralizing antibody include one at codons 68-76 recognized by a human neutralizing monoclonal antibody (IO), one at codons 609-625 recognized by a mouse monoclonal antibody (15), and ones at codons 589-645 and 703906 recognized by human immune sera (6). All these epitopes are relatively well conserved among CMV strains. A T-cell clone, which apparently recognized the variant locus at codon 181, showed reduced reactivity to heterologous strains (8). The degree of crossimmunity to CMV strains of different gB groups is currently unknown, but likely to be high. Functional consequences of gB variation will not be clear until aspects of pathogenesis are further explored. This major envelope glycoprotein appears to bind a 31 -kDa receptor on the cell surface and is probably important in viral entry into host cells, although other viral proteins such as gH may also be involved (12). Suggestive evidence of the importance of gB in such matters as tissue tropism and virulence are the
differential neurovirulence of strains of herpes simplex virus with variant gB sequences (5, 76) and the recent reports of differential permissiveness of brain cell cultures to different strains of CMV (9, 11). Further understanding of the role of CMV gB in these phenomena can now be obtained by using strains of known gB configuration in experimental systems. ACKNOWLEDGMENTS This work was supported by Department of Veterans Affairs research funds. Karen Dennison, Taylor Dunn, Marcia Kennedy, and Gail Marousek provided technical assistance. Sequence data from this article have been deposited with the EMBVGenbank Data Libraries under Accession Nos. M60924, M60925, M60926, M60927, M60929, M60931, M60932, M60933, M60934, M85228 and M85229.
REFERENCES 1. 2. 3. 4. 5. 6. 7. 8.
9. 10. 11. 12. 13. 14. 15. 16.
B~n-r, W. J., and VULGER. L. G.. /. tirol. 63, 403-10 (1989). CHOU, S., J. Infect. Dis. 160, 11-15 (1989). CHOU, S., /. Infect. Dis. 162, 738-742 (1990). CHOU, S., and DENNISON, K. M., /. Infect. Dis. 163, 1229-1234 (1991). GOODMAN, J. L., and ENGEL,J. P., J. Viral. 65, 1770-l 778 (1991). KNIESS, N., MACH, M., FAY, J., and BRIIT, W. J., J. Viral. 65, 138146 (1991). LEHNER, R., STAMMINGER, T., and MACH, M., J. Clin. Microbial. 29,2494-2502 (1991). LIU, Y. N., OPITZ, G., KLAUS, A., COOPER, J., and GEHRZ, R. C., In “Progress in Cytomegalovirus Research” (M. P. Landini, Ed.), pp. 187-90. Elsevier, Amsterdam, 199 1. MCCARTHY, M., RESNICK, L., TAUB, F., STRNART, R. V., and DIX, R. D., J. Neuropathol. fxp. Neural. 50, 441-450 (1991). MEYER, H.. MASUHO, Y., and MACH, M., /. Gen. Viral. 71,24432450 (1990). PULLIAM, L., /. Neuropathol. fxp. Neural. 50, 432-440 (1991). RASMUSSEN, L., RESTA, S., and MERIGAN, T., Transplant. froc. 23(Suppl 3), 60-63 (1991). SPAETE, R. R.. THAYER, R. M., PROBERT,W. S., eta/., Virology 167, 207-215 (1988). SPAETE, R. R., SAXENA, A., SCOTT, P. I., et a/., J. Viral. 64, 29222931 (1990). UTZ, U., BRITT, W., VUGLER, L., and MACH, M., J. Viral. 63, 19952001 (1989). WEISE, K., KAERNER,H. C., GLORIOSO,J., and SCHRODER,C. H., /. Gen. Viral. 68, 1909-l 919 (1987).
188,388-390
(1992)
Comparative Analysis of Sequence Variation in gpl16 and gp55 Components of Glycoprotein B of Human Cytomegalovirus SUNWEN CHOU Medical and Research Services, VA Medical Center; and Division of Infectious Oregon Health Sciences University, Port/and, Oregon 9720 1 Received November
Diseases,
15, 199 1; accepted January 22, 1992
Sequence variation in the gpl16 component of cytomegalovirus envelope glycoprotein B was examined in 11 clinical strains and compared with variation in gp55. The peptide variation in gpl16 was found to be strongly clustered at codons 27-67,440-460, and to a lesser extent at codons 181-l 94, 311-317, and 387-397. Strains adopted one of three to four peptide configurations at these loci, usually consistent with their gp55 sequence configuration. Two instances were observed of a sequence variation arising from recombination within gB. The limited, largely group-specific nature of variation in both gpl16 and gp55 facilitates functional and immunologic analyses. o 1992 Academic Press.
Inc.
The major envelope glycoprotein (gB) of human cytomegalovirus (CMV) is expressed as a precursor molecule of about 906 amino acids which is glycosylated and cleaved at a specific site (codon 461) to form a disulfide-linked complex of gp55 and gpl 16 (I, 13, 14). Glycoprotein gp55, encoded by the 3’ portion of the gB sequence, has a hydrophobic transmembrane segment towards its carboxyl terminus. Human and murine neutralizing antibodies recognize epitopes in both gp55 and gpl16 (6, 10, 13-15). Partial codihg sequence data (codons 384-7 17) revealed that clinical strains of CMV adopted one of four group-specific peptide sequence configurations in the vicinity of the gB cleavage site at codons 448-480 (4). Because of the biologic and immunologic importance of gB, additional sequences from clinical CMV strains are reported here in order to compare variation in gpl 16 with that observed in gp55. Efforts were aimed at providing a complete gB coding sequence for at least one clinical strain from each of the four gp55 groups and a gpl16 coding sequence (codons l-460) for at least two strains from each group. Eleven strains were selected based on diversity of restriction enzyme digest profiles of amplified gB sequences. Partial gB sequences (codons 384 to 7 17) of nine of these strains have been reported (4). The two new ones (CO82 and C076) were grouped by restriction analysis (4) and subsequently confirmed to have peptide sequences at codons 384-717 identical to strains C327 (group 1) and C359 (group 3) respectively. Therefore, four of the sequenced strains were in group 1, two in group 2, three in group 3, and two in group 4. They were all genetically distinct (3) and were 0042.6822/92 $3.00 CopyrIght 0 1992 by Academic Press, Inc. All rights of reproduction I” any form resewed.
obtained from epidemiologically unrelated organ trans plant recipients. DNA was extracted for analysis within 12 culture passages. Sequencing was performed using PCR-generated templates attached to streptavidin beads as previously described (4). Overlapping coverage of sequences was achieved by using primers of conserved sequence spaced approximately 300 bases apart, representing both DNA strands of the region sequenced. New sequences were aligned with gB sequences of laboratory strains Towne and AD1 69 (13). Observed gB coding sequences varied in length from 905 to 907 codons, because of codon deletions at one or more of seven sites. The codon numbering used here is from the Towne strain (907 codons); this leaves two codons (inserted after codons 28 and 36) which are present in some other strains but not in Towne. Figure 1a shows the nonuniform distribution of variable codons which affect peptide encoding. Two major loci of variation were observed at codons 27-67 and at codons 448-480. Lesser variant loci in gpl16 were noted at codons 181-l 95, 312-316, and 387408. The remaining variable codons usually main tained amino acid identity within groups of strains. Isolated mutations which were specific to a single strain were few and conservative; the number of such mutations which affected peptide encoding ranged from 0 to 6, with most strains being at the low end of this range. In gp55, a high degree of peptide conservation was observed beyond codon 500, along with extensive nucleotide conservation beyond codon 700. There were 1 1 cysteine residues in gpl 16 (four in the signal sequence, codons l-23) and five cysteine residues in gp55. All these residues were conserved, although an 388
SHORT COMMUNICATIONS
389
(4
Group 1 Towne c327
SSSSTXRGTSATHSXHHSSHTTSAAHSRSGSVSQRVTSSQTVSHGVNETIYNTTSHSQCYSSYSRVIAG ________________________________________--------------------_____-__
CO82 Cl78 C325
---X-SHA--S--NGS-T-R----QTX--V-X--H----EA---RA--------
---X----------------------------------------------------X--------------------------------------------------
___-___________ -_---_-________ ----------___--
RPNSALE _______ _ _ _ __ _ _ _ _ _ __ _ _ _ _ _ __ - _
Group 2 -----SHA--S--NGS-T-R----QTX--VyX--H----EA---RA------------SHA--S--NGS-T-R----QTX--V-X--H----EA---RA------------SHA--S--NGS-T-R----QTX--V-X--H----EA---RA--------
,(FA----------GKFA----------GKFA----------G-
---A-P---A-P---A-P-
-----SHA--SA-NGS-T-R----QTX--V-X--H----EA---RA------------SHA--S--NGS-T-R----QTX--V-X--H----EA---RA------------SHA--S--NGS-T-R----QTX--V.X--H----EA---RA--------
R-------------R-------------R--------------
-A---P-A---P-A---P-
Cl94 Cl28
---X-S-A--X--NGN-T------QTX-------EA---RA----------X-S-A--X--NGS-I------QTX-------EA---RA--------
_______________ R- _ _ _ _ _ _ _‘__ _ _ _ _
__- _- - _ _ _ _ __ _
Codon
24
AD169 C336 C338
Group 3 CO76 c359 c354
Group 4
FIG. 1. (a) Map of gB codons showing variation at peptide level. sometimes in one strain only (shorter lines). Closed circles indicate cleavage site (codon 461). (b) Amino acid sequence alignments of sites of variation are shown. Strains are segregated according to
75
181
195
311
317
Variation usually occurred in more than one strain (longer vertical lines), but potential glycosylation sites (Asn-X-Ser or Asn-X-Thr). Arrow indicates gp55 clinical strains and strain AD1 69 with Towne strain (top line). Three separate gp55 group (see text). X = Deletion.
alternative codon was used by different strains in several locations. There were 21 potential glycosylation sites (Asn-X-Ser or Asn-X-Thr, Fig. 1a). Each strain had 20 sites as a result of deletion of a site at either codon 37 or 456. Figure 1b shows the amino acid alignments of three variant loci in gpl16. In the major variant locus (codons 27-67) only three peptide configurations were discernable; strains of gp55 groups 2 and 3 were indistinguishable at this locus. All except one of the gp55 group 1 strains (including Towne) adopted a configuration quite different from group 2/3 strains, with 19 of 41 amino acid residues affected, as well as one potential glycosylation site. The configuration of group 4 strains was quite similar to that of group 213 strains. In the variant loci beyond codon 67, assortment into groups was also apparent, although groups 1 and 4 were very similar at codons 181-l 95 and 311-317. However, groups 2 and 3 were first differentiated at these loci. Definitive features identifying the four groups first appeared beyond codon 448. This major variable locus (4) spans the gpll6/gp55 cleavage site and has maintained its group character across the site in all 16 strains for which we have examined the sequence to date.
The only gp55 group 1 strain not observed to have a typical group 1 sequence at codons 27-67 was found to be a recombinant. Strain C325 adopted a nucleotide sequence close to AD1 69 (and other group 2/3 strains) from codons 1 through 69 and thereafter crossed over to a group 1 configuration. Strain C325 was plaque purified directly from a urine specimen and has remained stable through multiple passages (2). Strain C327 (group 1) was also plaque purified from the same specimen but is distinct at the nucleotide level and is unrelated to C325. New sequence data from beyond codon 717 revealed that laboratory strain Towne, classified as gp55 group 1, was in fact a recombinant which crossed over to group 4 at a site between codons 501 and 520 and thereafter maintained a group 4 sequence to the end of gB. Although the gp55 groups differ little in peptide encoding beyond codon 500, the amino acid residues at codons 768,780,802, and 853 of Towne strain are characteristic of group 4. Recombinant strains which cross over from one variant group to another indicate that it is not essential for strains to adopt a specific group character through all of gB. However, we have not encountered crossover within any variant locus of gB, particularly the two major loci 27-67 and448-480.
390
SHORT COMMUNICATIONS
The relatively limited gB variation encountered in a herpesvirus as widespread and long established in human populations as CMV prompts the question of whether similar variants are seen worldwide, or whether our strains reflect local epidemiology. Further study of isolates from diverse sources will be required, but evidence is accumulating that the gB groups we have seen are widespread, if not necessarily in the same relative distribution. Recently, Lehner eT a/. (7) reported partial gB coding sequences from five European clinical strains, covering the two major variant loci. Although the data provided did not fully establish group relationships, the five strains had peptide configurations at those loci which placed them in groups 1 (two strains), 3 (two strains), and 4, similar to what could be expected from five random isolates from our area. Maintenance of gB variant groups is probably the result of functional requirements and host factors, such as receptor structure and host immune responses. Epitopes of gB relevant to neutralizing antibody include one at codons 68-76 recognized by a human neutralizing monoclonal antibody (IO), one at codons 609-625 recognized by a mouse monoclonal antibody (15), and ones at codons 589-645 and 703906 recognized by human immune sera (6). All these epitopes are relatively well conserved among CMV strains. A T-cell clone, which apparently recognized the variant locus at codon 181, showed reduced reactivity to heterologous strains (8). The degree of crossimmunity to CMV strains of different gB groups is currently unknown, but likely to be high. Functional consequences of gB variation will not be clear until aspects of pathogenesis are further explored. This major envelope glycoprotein appears to bind a 31 -kDa receptor on the cell surface and is probably important in viral entry into host cells, although other viral proteins such as gH may also be involved (12). Suggestive evidence of the importance of gB in such matters as tissue tropism and virulence are the
differential neurovirulence of strains of herpes simplex virus with variant gB sequences (5, 76) and the recent reports of differential permissiveness of brain cell cultures to different strains of CMV (9, 11). Further understanding of the role of CMV gB in these phenomena can now be obtained by using strains of known gB configuration in experimental systems. ACKNOWLEDGMENTS This work was supported by Department of Veterans Affairs research funds. Karen Dennison, Taylor Dunn, Marcia Kennedy, and Gail Marousek provided technical assistance. Sequence data from this article have been deposited with the EMBVGenbank Data Libraries under Accession Nos. M60924, M60925, M60926, M60927, M60929, M60931, M60932, M60933, M60934, M85228 and M85229.
REFERENCES 1. 2. 3. 4. 5. 6. 7. 8.
9. 10. 11. 12. 13. 14. 15. 16.
B~n-r, W. J., and VULGER. L. G.. /. tirol. 63, 403-10 (1989). CHOU, S., J. Infect. Dis. 160, 11-15 (1989). CHOU, S., /. Infect. Dis. 162, 738-742 (1990). CHOU, S., and DENNISON, K. M., /. Infect. Dis. 163, 1229-1234 (1991). GOODMAN, J. L., and ENGEL,J. P., J. Viral. 65, 1770-l 778 (1991). KNIESS, N., MACH, M., FAY, J., and BRIIT, W. J., J. Viral. 65, 138146 (1991). LEHNER, R., STAMMINGER, T., and MACH, M., J. Clin. Microbial. 29,2494-2502 (1991). LIU, Y. N., OPITZ, G., KLAUS, A., COOPER, J., and GEHRZ, R. C., In “Progress in Cytomegalovirus Research” (M. P. Landini, Ed.), pp. 187-90. Elsevier, Amsterdam, 199 1. MCCARTHY, M., RESNICK, L., TAUB, F., STRNART, R. V., and DIX, R. D., J. Neuropathol. fxp. Neural. 50, 441-450 (1991). MEYER, H.. MASUHO, Y., and MACH, M., /. Gen. Viral. 71,24432450 (1990). PULLIAM, L., /. Neuropathol. fxp. Neural. 50, 432-440 (1991). RASMUSSEN, L., RESTA, S., and MERIGAN, T., Transplant. froc. 23(Suppl 3), 60-63 (1991). SPAETE, R. R.. THAYER, R. M., PROBERT,W. S., eta/., Virology 167, 207-215 (1988). SPAETE, R. R., SAXENA, A., SCOTT, P. I., et a/., J. Viral. 64, 29222931 (1990). UTZ, U., BRITT, W., VUGLER, L., and MACH, M., J. Viral. 63, 19952001 (1989). WEISE, K., KAERNER,H. C., GLORIOSO,J., and SCHRODER,C. H., /. Gen. Viral. 68, 1909-l 919 (1987).
