VIROLOGY
191, 523-528
Homology
(1992)
of the Envelope
Glycoprotein
B of Human Herpesvirus-
and Cytomegalovirus
SUNWEN Q-IOU’ AND GAIL I. MAROUSEK Medical and Research Services, VA Medical Center, and Division of Infectious Oregon Health Sciences University, Portland, Oregon 9720 1 Received July 7. 1992; accepted August
Diseases,
IO, 7992
The envelope glycoprotein B (gB) coding sequences of two strains of human herpesvirus(HHVG GS and 229) were determined by sequencing a 2.5-kb open reading frame adjacent to the DNA polymerase sequence. The deduced primary translation product is 830 amino acids in length and is 964/o conserved between the two divergent strains with no localized hypervariability noted. It contains the expected signal and transmembrane sequence motifs as well as a putative site of protease cleavage. There was 39% amino acid identity with the gB of human cytomegalovirus (CMV, strain AD1 69). HHVG-CMV gB peptide homology was evident through the entire sequence, but was especially strong in the amino-terminal portion of CMV gp55, which contains linear and conformational epitopes recognized by CMV-neutralizing antibodies. All 10 cysteine residues of HHV6 gB match corresponding residues of CMV gB. Sequence data suggest strong structural similarity and possible immunologic cross-reactivity of gB from the two viruses.
Human herpesvirus(HHVG) was originally isolated from individuals with lymphoproliferative disease (ZU), but is now known to be an extremely prevalent virus ( 13, 17,2 1). Primary infection usually occurs early in life and may be accompanied by roseola infantum (2, 19, 28); viral reactivation later in life appears to be common (13) but the disease syndromes attributable to reactivation remain to be established. HHV6 has a 170-kb dsDNA genome consisting of a long unique region bounded by terminal repeats (15). Restriction mapping and sequencing studies to date have shown that its genetic organization is similar to that of human cytomegalovirus (CMV) (12, 15). There is a colinear arrangement of genes in much of the long unique region of the two viruses (16), and there is peptide homology of the deduced translation products of corresponding genes. This has been demonstrated for the major capsid protein (12), DNA polymerase (26), and envelope glycoprotein H (9), among others. Glycoprotein B (gB) is a major envelope component with homologs in all known human herpesviruses. CMV-neutralizing antibodies recognize a number of linear and conformational epitopes on gB (3,4, 10,27). In immune human sera the bulk of neutralizing antibody activity appears to be directed at gB (10). The CMV gB is translated as a 130- to 160-kDa precursor which is cleaved by a cellular enzyme to form a disulfide-linked complex of gpl 16 and gp55 (24). There is significant interstrain variation in two localized regions of CMV gB,
with clinical strains adopting one of 2 to 4 alternative configurations at these loci (6, 7). Preliminary data indicate that two types of HHV6 strains (A and B) exist, as exemplified by strains GS and 229, respectively (1, 22). The strains differ in tropism for certain T cell lines, in reactivity with some monoclonal antibodies, and probably in epidemiology as well. The aim of this study was to determine the extent of relatedness of the gB’s from these HHV6 strains to each other and to human cytomegalovirus. HHV6 strain GS was obtained from D. V. Ablashi (1) (National Cancer Institute, Bethesda, MD), and strain 229 was obtained from J. A. Stewart (14) (Centers for Disease Control, Atlanta GA). Strain GS was propagated in the T cell line HSB-2, and 229 was propagated in cord ,blood mononuclear cells. After 7 to 10 days of culture, extracellularvirus was recovered by ultracentrifugation, and DNA was extracted using SDS and phenol. Cosmid CB4 was cloned from HHV6 strain U1102 by M. E. D. Martin et al. (15) and was obtained from the Public Health Laboratory Service, Porton Down, U.K. The sequence of the terminal 0.7-kb part of gB and the DNA polymerase sequence of strain Ul 102 has been published (26). This provided data for synthesis of downstream amplification primers for polymerase chain reaction (PCR) amplification (Fig. 1). In order to provide upstream primers for PCR, the BamHIIPstl proximal terminal fragment of cosmid CB4 was cloned into the Bluescript vector pBSIISK- (Stratagene, La Jolla, CA) and partly sequenced. From subsequent PCR product sizes it was estimated that this region is
’ To whom reprint requests should be addressed. 523
0042.6822/92
$5.00
SHORT COMMUNICATIONS
524
1
0
1
I
1
I
1
4
3
2
I
/
I
1
I
7
6
5
1
I
1
I
1
I
a /
kb v34).
BamHl
__.. ._
EC&l PStll Xhol
gB 4
Pol FIG. 1. Map of HHV6 gB region. Feature locations were calculated from new and existing (15, 26) data. Xhol site in gB is absent in strain 229. Part of cosmid CB4 was subcloned (hatched box) and partly sequenced to provide for upstream primers. gB sequencing templates were prepared by PCR using one of the biotinylated primers CB4AxB or HP28xB.
about 0.9 kb upstream of the start of the gB coding sequence. Oligonucleotide primers of approximately 20 bases were synthesized, including biotinylated primers HP28xB and CB4AxB (Fig. 1). DNA templates for sequencing were produced by PCR directly from viral DNA extracts, using one biotinylated primer and a primer from the opposite strand outside the limit of the region to be sequenced. The resulting unpurified product was agitated with streptavidincoated magnetic beads (Dynabeads, Dynal, Oslo) to attach the biotinylated strand to a solid phase. The nonbiotinylated strand was denatured off, and sequencing was subsequently accomplished by annealing a selected sequencing primer, followed by labeling and chain termination reactions as previously described (7). The sequence of gB of strain GS was filled in a stepwise manner from its 3’ end using the noncoding strand as template, by synthesizing a new sequencing primer approximately every 250 bases using information acquired from the previous primer. When the upstream end of the open reading frame was reached, the sequence information was verified by synthesis of additional primers and sequencing of the opposite (coding) strand. All overlapping sequence data were checked for consistency, and no discrepancy was noted from information generated from templates produced from separate PCRs. The gB sequence of strain 229, likewise derived from both DNA strands, was generated in a similar manner. The primers used for PCR and sequencing are shown in Table 1. The gB coding sequences of strains GS and 229 were aligned and loci of nucleotide and peptide variation were tallied. The translated peptide sequences were checked for homology with the gB coding sequence of CMV strain AD169, using a FASTP algo-
rithm (18) as implemented in the software package PROSE (Hitachi). Because of the importance of glycosylation and disulfide linkages in CMV gB, locations of cysteine residues and potential N-linked glycosylation sites were noted. Source sequences for CMV gB were as published (24). The complete gB coding sequence of strains HHV6 GS and 229 determined here are avail-
TABLE 1 AMPLIFICATION
Name/locus”
AND SEQUENCING PRIMERS
Sequence (Y-3’)
Direction
1330 1624 1875 2188
TATAATAGGGTGATGGCATG AGTTACTGTG-IXGAACGCC T-ITCCAGTGAGAACGTATAA TTCTACATCAACGTCTCTCA AGAGAATGAATCGGTGTGTA TAAGACTACGGGAGATTTGA ACGATAACGATGT-TGCACGA GATCATAGGACAGAGGAATG CTCATATAAGAGCGCACTAT CCGGTGACTACTGTGTCCAG
F F F F F F F F F F
72 314 538 778 1015 1233 1544 1696 1842 1994 2321 HP28xB
CGGATCACAATATATCAT-TA ACCACACCCACATCACGGTA AGGGTTCTTTCGTAGTGATA CATTCATACCAT-TCATTAGATC TAATACACGTI-TGTITCGGA GAGAATGTCATGTCTAGACTTA AAGGGCCGAT-TATAACACATCG TATACACGTGCGCATGATT-TCC AACGTTTGCTCTACTCAATTCGT GAGCCCAATCCAGTGCCTATAG CTTAC-I-TGCCCACTGACGTC ATTATACCGCGAGGAAGTATAC
R R R R R R R R R R R R
CB4AxB -50 250 567 852 1101
Note. CB4AxB and HP28xB shown on Fig. 1. F = forward (coding); R = reverse (noncoding) strand. ’ Location of 5’ end from start of gB coding sequence.
525
SHORT COMMUNICATIONS
PDHYIRAGYNHKYPFRICSIAKGTDLMRFDRDISCSPYKSNAKMSEGFFIIYKTNIETYTFPVRTYKKELTFQSSYR S-D--------------------+-+--------+-+---------------------------+--N-+---T---
GS 7.29
D"GVVYFLDRTVMGLAHPWEANLVNSHAQCYSA"~~~S~S~D-~TLN~FPLN~SITN~ITTKEPYFARGPLWLYSTSTSLNCI"T R----+---I----------y-----E--E------+-V-------------------+---~ _______------___-_-_-------
--------
100 200
EATAKAKYPFSYFALTTGEIVEGSPFFNISNGKHFAEFLEKLTILEN~THIEDLMN-GATTLVRKIAFLEKADTLFSWEIKEEN~SSVCMLK-TVTH -------+---------+---------D--------------+-----------------+---+-+---+--G---------------------+----
300
Gs 229
wDvKSRaDILWQLQnYDTLKDYINDALGNLAESWCLDQKRTITMLHELSKISPSSI"SEVYGRPISAQLBGDVLAISKCIEVN~SS"QLHKSMRVVD+-----------+-+------------------+--------+--+-----y-----+L--+----+-+-----++--------+-+--------++--
500
Gs z29
AKGVRSETMCYNRPLVTFSFTPE-GQLGLDNEILLGDHRTEECEIPSTKIFLSGNHAHVYTDYTHTNlSTPIEDIE~D~I~~IDPL~N~F~ +---------------------+-----L++-----+-----+-----------+-----+-+---+--------------------------+-+-+--+---L
600
GS 229
LDLYSPDELSRANVFDLENILREYNSYKSALYTIEAKIATNTPSYVNGIN +---+-------+---+----------------------+-++--+----
GS z29
!iiEi$*--+-------
GS z29
PSSSESRASKPSLIDRIRYRGYKSVNVEEA 830 ---+--------------------------
GS 229 GS 229
QGLGAIGTGLGSVISVTAGALGDIVGGWSF ----------------+---+-------------
NPFGGGLMLILAIV --------------
QRAVLSKPID~PYATNPVTTVSSVTGTTWKTPSVRDVDGGTSVAVSEKEEGHADVSGQVSDDEYSQ~~~IKSLDESYRRK --------------------+ ----- f -------+-- ------+-I-G------D---------+--------
100 800
FIG. 2. gB amino acid sequence alignment of HHV6 strains GS and 229. Variant amino acids of strain 229 are shown. -, identical codon; +, alternate encoding of same amino acid. Signal sequence, transmembrane region, and putative cleavage site are boxed. Potential glycosylation sites are marked 0.
able from GenBank (accession numbers M97928 and M97927). An open reading frame (ORF) of 2505 bp was identified by sequencing. The nearest ATG codon was 15 bases from the upstream end of the ORF. It meets consensus criteria for a translational start site (1 I), as does another ATG codon 9 bases downstream. Based on the first ATG codon, the predicted gB coding sequence is 2493 bp (counting the stop codon) and the translation product is 830 amino acids in length, with a calculated molecular weight (before glycosylation or signal peptide cleavage) of 93 kDa. These findings were the same with both strains GS and 229. As previously demonstrated (26), the HHV6 DNA polymerase coding sequence is immediately downstream of gB (Fig. 1). The ORF upstream of gB extends for 13 codons into the gB coding sequence. This is similar to CMV, where the ORF UL56 extends for several codons into UL55 (gB). Peptide sequence fragments deduced from the vicinity of upstream primer CB4AxB as well as the start of HHV6 gB show clear homologywith CMV UL56, which encodes an 850-aa product of unknown function (possibly a transport protein). As expected for a herpesvirus envelope glycoprotein, the deduced gB translation product contains a signal sequence of hydrophobic residues at the beginning, and another hydrophobic region around residue 700 which is in the putatative transmembrane region (Fig. 2). There are 10 (strain GS) or 8 (strain 229) Nlinked glycosylation sites (Asn-X-Ser or Asn-X-Thr), and 10 cysteine residues. In the vicinity of residue 400,
there is a cluster of basic amino acids that could represent a protease cleavage site. The gB sequences of strains GS and 229 showed 95.4% identity at the nucleotide level and 96.3% identity at the peptide level. Variation was sporadic throughout the coding sequence (Fig. 2); there was no strong clustering of gB variant codons as previously observed in CMV strains of different groups (6). Of the 31 codons showing peptide variation, about half of the substitutions involved some change in polarity. As with CMV, there was more peptide variation in the first half of the HHV6 gB; this region contains 23 of the 31 codons showing variation between strains GS and 229. HHV6 strains GS (A type) and 229 (B type) are known to differ in a number of respects and are epidemiologically divergent. Restriction analyses of their genomes readily reveal distinctive differences (I), including those in the gB/Pol region. Most HHV6 isolates from patients with roseola (22) and transplant recipients (our unpublished data) have been type B, with restriction patterns that resemble those of 229 rather than GS. The similarity in the gBs of strains GS and 229 suggests that the extent of interstrain variation in HHV6 gB may not be as great as in CMV, and that gB differences may not account for the different properties of type A and B HHV6 strains. Alignment of the gB peptide sequences of HHV6 (either strain) and CMV (strain AD1 69) revealed an overall 39% amino acid identity. Strong homology with CMV was noted throughout HHV6 gB except for the signal sequence (Fig. 3). Overall, the optimized FASTP/
SHORT COMMUNICATIONS
526
HHV6 CMV
(GS) (AD169)
HIiVC (GS) CMV
HHV6 CNV
(AD169)
(GS) (AD169)
IiHV6 (GS) CMV
HHV6
(AD169)
(GS)
CMV
(AD169)
HHV6
(GS)
CMV
HHV6 CMV
HHV6
(AD169)
(GS) (AD169)
(GS)
CNV
(AD169)
HHV6
(GS)
CNV
HHV6 CMV
HHVC CMV
HHV6 CMV
HHV6 CMV
HHVC CNV
HHVC CMV
(AD169)
(GS) (AD169)
(GS)
NSKMVVLF&FLt4NSV&yCDpDHyIiiGyNHKypF&SIAKGTDtiRpD~& I:I.:.tl.(.lt:I.*:.*.~ :.. SQHVT~~EAVS~aIYT~Y~%W~~~~VCS~ O$DLIRFERN#TS PG YKS--NAKMS~~FFIIYKTN~~TYTFPVRT&ELTFQS~i!tDVGWYF&Vt4GLAMpV :. :....:I..*.:: :: ..:: ::.:.: t:z..:: . ..:.:.... NKPItii;LDEGI IVMT:~~~QKVLT~~YAYIY~~~~SNTE'fPM NV?z?
.: t.
120 130 140 150 160 170 YEANLVNSHAQCYSAVAMKRPDGTVFSAFHEDNNKNNTLNLFPLNFKSITNKRFITTKEP .:::: :.: :. .:.:..:.: . .. . . ,.:..:.:.. ~I~~~~~~~:~-~VI~~~~YN~S~~~~NQLIPDD~~~THSTR wQ 180 190 200 210 220 230 YFARGPLWLYSTSTSLNCIVTEATRKAKYPFSYFALTTGEIVEGSPFFNGSN--GKHFAE * ':' . :::.,. .:::..: .::..:*s. .:: .::..: :::.::.I . . .. . . NNSRG;~yRETC~XJSLTITTAR~~PyHFFA~~~GDWYIS~~~NGTNRNAsY&GE 240 250 260 270 280 290 PLEKLTILENYTNIEDLNNGMNGATTLVRKIAFLEKADTLFSNEIKEENESVCNLKHNTT .:. :. :::...:. . :.:.. : .:::t.:z...::.:..:,. .: :. :.. NADKF~~~PNYTIVS~#iP-NA&PE~~UJAFLERAD~VISWDIQ~~~NVTCQLTFW&A 300 310 320 330 340 350 VTHGLRAETDETYHFISKELTAAFVAPKESLNLTDPKQTCIKDEFEKIINEWMSDYNDT . ...I.......:: : ..t:.:.. :...:..:. .*..tz . . . . . . ..::.: SERTIRS~~DSYHFS~A&MTATFLS~~ 8EVNNsDSAL~VRDEAIl??;QQIFNTSY$ 8T 360 370 380 YSNNGSYQIFKTTGDLILIWQPLVQKSLNFLE-I.. :. .z.:.:.:...:x . ::::. :: YEKYGNV~~ETSGGLVVE&NQGIKQK~~~LERS.J'.N 410 420 430 440 450 460 -----DILWQLQYLYDTLKDYINDAM;NLAESWCLDQKRTSKISPSSIVSEV . .. . . ..t. :::I..*:: t:...:t.::.z :.::.....:*:::.::.:.:.. MKSVIiN~~~AQLQFTY~;~RGYINRF+~ IAKAWCVD RRTLEVFKE&KINPSAU&I 8 5Po 470 480 490 500 510 520 YGRPISAQLHGDVLAISKCIEVNQSSVQLHKSMRWDAKGVRSETMCYNRPLVTFSFVNS :..::.:.. ::::....:...::.:z : .: . .. . . : ..::.::.:.:.:.:: YNKPIAAR$MGDVLGLAi$VTINQTSV --,~RDMNVKE-,s~RCYS~~~FNF~S 530 540 550 560 570 580 TPEVVPGQLGLDNEILLGDHRTEECEIPSTKIFLSGNHAWD .: :::: :::::t*.:t::*:..:: :::..:: :. :.:: . .. .. . . .. -SY$ YGQLGED~~;LLGNHRT~~~QLPSLKI~~~GNSAYEyV~~FS&.lID;~~ISTVD 8 590 600 610 620 630 640 AFIRLKIDPLENADFKVLDLYSPDELSRANVFDLENILREXNSYKSALYTIEAKIATNTP t :.::::::.z:.::.*::..zs..,s ::*::.:.::.:::: .:.a...
:
(AD169)
(GS) (AD169)
(GS) (AD169)
(GS) (AD169)
(GS) (AD169)
650 660 670 680 690 700 SYVNGINSFLQGLGAIGTGLGSVISVTAGALGDIVGGWSFLKNPFGGGLMLILAIVWV .:..:..... :::: :...: .:....::....:.::..:::::::. . . . .. . .. . . P"4~~LDDLNSG$GjAGW~~%AVGGAVASSGVATFL~~~FGAFTII~~~IAWI 710 720 730 740 750 760 IIIWFVRQRHVLSKPIDNMFPYATNPVTTVSSVTGTTWKTPSVKDVDGGTSVAVSEKE :. . . .. . . . . . ..:.. .::: . . . .t. . ..:. . . .. . . . . . . . :. "YL;YTRQRPLC; 8PLQNLFPYL;SADGTTVTS&sTKDTSL 0EPSYEESV-YNZGRKG 770 780 790 800 810 EGMADVSGQVSDDEYSQEAALKMLKAIKSLDESyRRKpSSSES---------HAsKPSLI : . . . . .. . :..:.: .:: :, .:: : . .. .. . . . . ::.:. PGP;$~DASTAAPZ";NEQAYQbU~ DAE RAQQNGTDS~~%QTGTQDKG KPNLL 850 0 8%0 820 830 DRIRYR--GYKSVNVKEA ::.:.: :t. . . . . DRLFt'd~NGYRHL;%~DE
FIG. 3. Alignment of gB peptide sequences of HHV6 and CMV. :, identical amino acids; boxed and region of highest homology is underlined.
FASTA homology score was 1760, higher than that usually noted between herpesvirus species and comparable to the homology score between herpes simplex (HSV) and varicella-zoster (VZV) which is 2200. By comparison, FASTP/FASTA scores of CMV gB align-
. conserved amino acids. Putative cleavage site is
ments with HSV, VZV, and Epstein-Barr viruses are in the range of 700 to 1100 (5). Scores for HHV6 gB alignments with HSV and VZV were in the range of 500 to 900. The CMV gB is about 76 a.a. longer than HHV6 gB; there is no homology with HHV6 in the region be-
SHORT COMMUNICATIONS
low CMV residue 88. This part of CMV gB shows considerable interstrain variation and is the site of strainspecific epitope(s) (3, 6). The other focus of interstrain variation in CMV, at codons 440-480 surrounding the protease cleavage site (7) is also a region of relatively low homology with HHV6 because of the deletion of a number of amino acids from HHV6 gB (Fig. 3). Thus, homology data appear to support the impression from GS-Z29 comparative data that HHV6 gB differs less among strains than CMV gB. Most of the key features of gB are in homologous locations in both HHV6 and CMV (Fig. 3). There is a run of arginine residues in HHV6 gB corresponding to the protease cleavage site of CMV gB. All 10 cysteine residues of HHV6 gB outside the signal sequence are matched to corresponding cysteine residues of CMV. CMV gB has an additional six cysteine residues, of which four are in the signal sequence. The N-linked glycosylation sites of HHV6 and CMV gB are likewise similar though not identically matched. This correspondence of features suggests a close structural similarity of gB’s from HHV6 and CMV. Peptide sequence homology is especially strong in the region following the putative cleavage site. HHV6 gB residues 412-630 are almost perfectly conserved (99%) between the two strains examined (Fig. 2) and have a 549/o amino acid identity with CMV gB residues 480-690 (Fig. 3). This region of CMV gB (gp55) is known to contain linear and conformational epitopes recognized by several different monoclonal antibodies with neutralizing activity (4, 10, 24, 27). Human immune sera recognize linear epitope(s) located between residues 589 and 645 (70). The corresponding HHV6 gB sequence contains a tract of 3 1 residues (530-560) where 24 are identical to those of CMV (Fig. 3). The close peptide sequence similarity of parts of gB between HHV6 and CMV may be relevant to the yet unresolved issue of cross-reactivity between the viruses (8, 2 1, 25). Strong rises in antibody to HHV6 are frequently observed in association with primary CMV infection (8) yet antigen absorption experiments have not suppressed the overall serologic reactivity to the other virus (21). Cross-reactivity of these two viruses may be limited to epitopes of certain proteins, similar to that observed between HSV and VZV (23). The main conclusions of this sequence analysis are as follows. HHV6 gB is closely related to CMV gB, much more so than usually observed between human herpesvirus species. Structural similarity is especially striking in a region beyond the putative cleavage site, corresponding to CMV gp55, where amino acid identity exceeds 50%. This region in CMV contains major
527
neutralization-related epitopes. With the observed degree of gB homology between HHV6 and CMV, partial immunologic cross-reactivity is plausible. Two divergent HHV6 strains (types A and B) had gB peptide sequences that differed little and were especially conserved in regions corresponding to known CMV gB neutralizing epitopes. The known biological differences of these HHV6 subtypes are therefore not likely to be due to gB variation.
ACKNOWLEDGMENT This work was supported search funds.
by Department
of Veterans Affairs re-
REFERENCES 1. ABLASHI, D. V., BALACHANDRAN,N., JOSEPHS,S. F., HUNG, C. L., KRUEGER,G. R., KRAMARSKY.B., SALAHUDDIN, S. Z., and GALLO, R. C., Vifology 184, 545-552 (1991). 2. ASANO, Y., YOSHIKAWA,T., SUGA, S., YAZAKI, T., HATA, T., NAGAI, T., KAJITA,Y., OZAKI, T., and YOSHIDA, S., J. Pediatr. 114, 535539 (1989). 3. BASGOZ, N., QADRI, I,, NAVARRO, D., SEARS, A., LENNETTE, E., YOUNGBLOM, J.. and PEREIRA, L., /. Gen. viral. 73, 983-988 (1992). 4. BANKS, T., Huo, B., KOUSOULAS,K., SPAETE, R., PACHL, C., and PEREIRA,L., J. Gen. Viral. 70, 979-985 (1989). 5 CHEE, M. S.. BANKIER,A. T., BECK, S., BOHNI, R., BROWN, C. M., CERNEY, R., HORSNELL. T., HUTCHISON, C. A., KOUZARIDES,T.. MARTIGNETTY,J. A., and BARRELL, B. G., Curr. Top. Microbial. Immunol. 154, 127-l 69 (1990). 6. CHOU, S., virology 188, 388-390 (1992). 7. CHOU, S., and DENNISON, K. M., J. lnfecf. Dis. 163, 1229-1234 (1992). 8. CHOU, S., and Scorr, K. M., J. C/in. Microbial. 28, 851-854 (1990). 9. JOSEPHS,S. F., ABLASHI, D. V., SALAHUDDIN, S. Z., JAGODZINSKI, L. L., WONG-ST&IL. F., and GALLO, R. C., J. Virol. 65, 55975604 (1991). 10. KNIESS,N., MACH, M., FAY, J., and BRIIT, W. J., /. Viral. 65, 138146(1991). 11. KOZAK, M., Nucleic Acids Res. 9, 5233-5252 (1981). 12. LAWRENCE, G. L., CHEE, M., CRAXTON, M. A., GOMPELS, U. A., HONES% R. W., and BARRELL, B. G., J. Viral. 64, 287-299 (1990). 13. LEW. J. A., FERRO, F., GREENSPAN,D., and LENNE‘TTE,E. T., Lancet335,1047-1050(1990). 14. LINDQUESTER,G. J., and PELLETT, P. E., virology 182, 102-l 10 (1991). 15. MARTIN, M. E. D., THOMSON, B. J., HONESS, R. W., CRAXTON, M. A., GOMPELS, U. A., LIU. M. Y., LITTLER, E., ARRAND, J. R., TEO, I., and JONES, M. D., J. Gen. Viral. 72, 157-168 (1991). 16. NEIPEL, F., ELLINGER,K., and FLECKENSTEIN,B., J. Gen. !/ire/. 72, 2293-2297(1991). 17. OKUNO. T., TAKAHASHI, K., BALACHANDFUI,K., SHIRAKI,K., YAMANI-
528
78. 79.
20.
21.
22.
SHORT COMMUNICATIONS SHI, K., TAKAHASHI, M., and BABA, K., J. C/in. Microbial. 27, 651-653 (1989). PEARSON,W. R., Methods Enzymol. 183, 63-98 (1990). PRUKSANANONDA,P., HALL, C. B., INSEL, R. A., MCINTYRE, K., PELLETT, P. E., LONG, C. E., SCHNAEIEL,K. C., PINCUS, P. H., STAMEY, F. R.. DAMEAUGH, T. R., and STEWART,J. A., N. Engl. /. Med. 326, 1445-l 450 (1992). SALAHUDDIN, S. Z., ABLASHI, 0. V., MARKHAM, P. D., JOSEPH& S. F.. STURZENEGGER,S., KAPLAN, M., HALLIGAN, G., BIBERFELD, P., WONG-STAAL. F., KRAMARSKY.B., and GALLO, R. C.. Science 234, 596-601 (1986). SAXINGER,C., POLESKY,H., EBY, N., GRUFFERMAN,S., MURPHY, R., TEGTMEIR,PAREKH,V.. MEMON. S., and HUNG, C.,/. Viral. Methods 21, 199-208 (1988). SCHIRMER, E. C., WYAIT, L. S., YAMANISHI, K., RODRIGUEZ,W. J.,
23. 24.
25. 26. 27.
and FRENKEL,N., Proc. Nat/. Acad. Sci. USA 88, 5922-5926 (1991). SCHMIDT, N. J., f. Med. Viral. 9, 27-36 (1982). SPAETE, R. R., THAYER, R. M., PROBERT,W. S., MASIARZ, F. R., CHAMBERLAIN, S. H., RASMUSSEN, L., and MERIGAN, T. C., PACHL, C., Virology 167, 207-215 (1988). SUTHERLAND,S., CHRISTOFINIS,G., O’GRADY, J., and WILLIAMS, R., J. Med. Viral. 33, 172-176 (1991). TEO, I. A., GRIFFIN, B. E., and JONES, M. D., J. Viral. 65, 46704680 (1991). UR. U., BRITT, W., VUGLER, L., and MACH, M., J. Viral. 63, 19952001 (1989).
28. YAMANISHI, K., OKUNO, T.. SHIRAKI, K., TAKAHASHI, M.. KONDO, T., ASANO. Y., and KUW\TA, T., Lancer 1,1065-l 067 (1988).
191, 523-528
Homology
(1992)
of the Envelope
Glycoprotein
B of Human Herpesvirus-
and Cytomegalovirus
SUNWEN Q-IOU’ AND GAIL I. MAROUSEK Medical and Research Services, VA Medical Center, and Division of Infectious Oregon Health Sciences University, Portland, Oregon 9720 1 Received July 7. 1992; accepted August
Diseases,
IO, 7992
The envelope glycoprotein B (gB) coding sequences of two strains of human herpesvirus(HHVG GS and 229) were determined by sequencing a 2.5-kb open reading frame adjacent to the DNA polymerase sequence. The deduced primary translation product is 830 amino acids in length and is 964/o conserved between the two divergent strains with no localized hypervariability noted. It contains the expected signal and transmembrane sequence motifs as well as a putative site of protease cleavage. There was 39% amino acid identity with the gB of human cytomegalovirus (CMV, strain AD1 69). HHVG-CMV gB peptide homology was evident through the entire sequence, but was especially strong in the amino-terminal portion of CMV gp55, which contains linear and conformational epitopes recognized by CMV-neutralizing antibodies. All 10 cysteine residues of HHV6 gB match corresponding residues of CMV gB. Sequence data suggest strong structural similarity and possible immunologic cross-reactivity of gB from the two viruses.
Human herpesvirus(HHVG) was originally isolated from individuals with lymphoproliferative disease (ZU), but is now known to be an extremely prevalent virus ( 13, 17,2 1). Primary infection usually occurs early in life and may be accompanied by roseola infantum (2, 19, 28); viral reactivation later in life appears to be common (13) but the disease syndromes attributable to reactivation remain to be established. HHV6 has a 170-kb dsDNA genome consisting of a long unique region bounded by terminal repeats (15). Restriction mapping and sequencing studies to date have shown that its genetic organization is similar to that of human cytomegalovirus (CMV) (12, 15). There is a colinear arrangement of genes in much of the long unique region of the two viruses (16), and there is peptide homology of the deduced translation products of corresponding genes. This has been demonstrated for the major capsid protein (12), DNA polymerase (26), and envelope glycoprotein H (9), among others. Glycoprotein B (gB) is a major envelope component with homologs in all known human herpesviruses. CMV-neutralizing antibodies recognize a number of linear and conformational epitopes on gB (3,4, 10,27). In immune human sera the bulk of neutralizing antibody activity appears to be directed at gB (10). The CMV gB is translated as a 130- to 160-kDa precursor which is cleaved by a cellular enzyme to form a disulfide-linked complex of gpl 16 and gp55 (24). There is significant interstrain variation in two localized regions of CMV gB,
with clinical strains adopting one of 2 to 4 alternative configurations at these loci (6, 7). Preliminary data indicate that two types of HHV6 strains (A and B) exist, as exemplified by strains GS and 229, respectively (1, 22). The strains differ in tropism for certain T cell lines, in reactivity with some monoclonal antibodies, and probably in epidemiology as well. The aim of this study was to determine the extent of relatedness of the gB’s from these HHV6 strains to each other and to human cytomegalovirus. HHV6 strain GS was obtained from D. V. Ablashi (1) (National Cancer Institute, Bethesda, MD), and strain 229 was obtained from J. A. Stewart (14) (Centers for Disease Control, Atlanta GA). Strain GS was propagated in the T cell line HSB-2, and 229 was propagated in cord ,blood mononuclear cells. After 7 to 10 days of culture, extracellularvirus was recovered by ultracentrifugation, and DNA was extracted using SDS and phenol. Cosmid CB4 was cloned from HHV6 strain U1102 by M. E. D. Martin et al. (15) and was obtained from the Public Health Laboratory Service, Porton Down, U.K. The sequence of the terminal 0.7-kb part of gB and the DNA polymerase sequence of strain Ul 102 has been published (26). This provided data for synthesis of downstream amplification primers for polymerase chain reaction (PCR) amplification (Fig. 1). In order to provide upstream primers for PCR, the BamHIIPstl proximal terminal fragment of cosmid CB4 was cloned into the Bluescript vector pBSIISK- (Stratagene, La Jolla, CA) and partly sequenced. From subsequent PCR product sizes it was estimated that this region is
’ To whom reprint requests should be addressed. 523
0042.6822/92
$5.00
SHORT COMMUNICATIONS
524
1
0
1
I
1
I
1
4
3
2
I
/
I
1
I
7
6
5
1
I
1
I
1
I
a /
kb v34).
BamHl
__.. ._
EC&l PStll Xhol
gB 4
Pol FIG. 1. Map of HHV6 gB region. Feature locations were calculated from new and existing (15, 26) data. Xhol site in gB is absent in strain 229. Part of cosmid CB4 was subcloned (hatched box) and partly sequenced to provide for upstream primers. gB sequencing templates were prepared by PCR using one of the biotinylated primers CB4AxB or HP28xB.
about 0.9 kb upstream of the start of the gB coding sequence. Oligonucleotide primers of approximately 20 bases were synthesized, including biotinylated primers HP28xB and CB4AxB (Fig. 1). DNA templates for sequencing were produced by PCR directly from viral DNA extracts, using one biotinylated primer and a primer from the opposite strand outside the limit of the region to be sequenced. The resulting unpurified product was agitated with streptavidincoated magnetic beads (Dynabeads, Dynal, Oslo) to attach the biotinylated strand to a solid phase. The nonbiotinylated strand was denatured off, and sequencing was subsequently accomplished by annealing a selected sequencing primer, followed by labeling and chain termination reactions as previously described (7). The sequence of gB of strain GS was filled in a stepwise manner from its 3’ end using the noncoding strand as template, by synthesizing a new sequencing primer approximately every 250 bases using information acquired from the previous primer. When the upstream end of the open reading frame was reached, the sequence information was verified by synthesis of additional primers and sequencing of the opposite (coding) strand. All overlapping sequence data were checked for consistency, and no discrepancy was noted from information generated from templates produced from separate PCRs. The gB sequence of strain 229, likewise derived from both DNA strands, was generated in a similar manner. The primers used for PCR and sequencing are shown in Table 1. The gB coding sequences of strains GS and 229 were aligned and loci of nucleotide and peptide variation were tallied. The translated peptide sequences were checked for homology with the gB coding sequence of CMV strain AD169, using a FASTP algo-
rithm (18) as implemented in the software package PROSE (Hitachi). Because of the importance of glycosylation and disulfide linkages in CMV gB, locations of cysteine residues and potential N-linked glycosylation sites were noted. Source sequences for CMV gB were as published (24). The complete gB coding sequence of strains HHV6 GS and 229 determined here are avail-
TABLE 1 AMPLIFICATION
Name/locus”
AND SEQUENCING PRIMERS
Sequence (Y-3’)
Direction
1330 1624 1875 2188
TATAATAGGGTGATGGCATG AGTTACTGTG-IXGAACGCC T-ITCCAGTGAGAACGTATAA TTCTACATCAACGTCTCTCA AGAGAATGAATCGGTGTGTA TAAGACTACGGGAGATTTGA ACGATAACGATGT-TGCACGA GATCATAGGACAGAGGAATG CTCATATAAGAGCGCACTAT CCGGTGACTACTGTGTCCAG
F F F F F F F F F F
72 314 538 778 1015 1233 1544 1696 1842 1994 2321 HP28xB
CGGATCACAATATATCAT-TA ACCACACCCACATCACGGTA AGGGTTCTTTCGTAGTGATA CATTCATACCAT-TCATTAGATC TAATACACGTI-TGTITCGGA GAGAATGTCATGTCTAGACTTA AAGGGCCGAT-TATAACACATCG TATACACGTGCGCATGATT-TCC AACGTTTGCTCTACTCAATTCGT GAGCCCAATCCAGTGCCTATAG CTTAC-I-TGCCCACTGACGTC ATTATACCGCGAGGAAGTATAC
R R R R R R R R R R R R
CB4AxB -50 250 567 852 1101
Note. CB4AxB and HP28xB shown on Fig. 1. F = forward (coding); R = reverse (noncoding) strand. ’ Location of 5’ end from start of gB coding sequence.
525
SHORT COMMUNICATIONS
PDHYIRAGYNHKYPFRICSIAKGTDLMRFDRDISCSPYKSNAKMSEGFFIIYKTNIETYTFPVRTYKKELTFQSSYR S-D--------------------+-+--------+-+---------------------------+--N-+---T---
GS 7.29
D"GVVYFLDRTVMGLAHPWEANLVNSHAQCYSA"~~~S~S~D-~TLN~FPLN~SITN~ITTKEPYFARGPLWLYSTSTSLNCI"T R----+---I----------y-----E--E------+-V-------------------+---~ _______------___-_-_-------
--------
100 200
EATAKAKYPFSYFALTTGEIVEGSPFFNISNGKHFAEFLEKLTILEN~THIEDLMN-GATTLVRKIAFLEKADTLFSWEIKEEN~SSVCMLK-TVTH -------+---------+---------D--------------+-----------------+---+-+---+--G---------------------+----
300
Gs 229
wDvKSRaDILWQLQnYDTLKDYINDALGNLAESWCLDQKRTITMLHELSKISPSSI"SEVYGRPISAQLBGDVLAISKCIEVN~SS"QLHKSMRVVD+-----------+-+------------------+--------+--+-----y-----+L--+----+-+-----++--------+-+--------++--
500
Gs z29
AKGVRSETMCYNRPLVTFSFTPE-GQLGLDNEILLGDHRTEECEIPSTKIFLSGNHAHVYTDYTHTNlSTPIEDIE~D~I~~IDPL~N~F~ +---------------------+-----L++-----+-----+-----------+-----+-+---+--------------------------+-+-+--+---L
600
GS 229
LDLYSPDELSRANVFDLENILREYNSYKSALYTIEAKIATNTPSYVNGIN +---+-------+---+----------------------+-++--+----
GS z29
!iiEi$*--+-------
GS z29
PSSSESRASKPSLIDRIRYRGYKSVNVEEA 830 ---+--------------------------
GS 229 GS 229
QGLGAIGTGLGSVISVTAGALGDIVGGWSF ----------------+---+-------------
NPFGGGLMLILAIV --------------
QRAVLSKPID~PYATNPVTTVSSVTGTTWKTPSVRDVDGGTSVAVSEKEEGHADVSGQVSDDEYSQ~~~IKSLDESYRRK --------------------+ ----- f -------+-- ------+-I-G------D---------+--------
100 800
FIG. 2. gB amino acid sequence alignment of HHV6 strains GS and 229. Variant amino acids of strain 229 are shown. -, identical codon; +, alternate encoding of same amino acid. Signal sequence, transmembrane region, and putative cleavage site are boxed. Potential glycosylation sites are marked 0.
able from GenBank (accession numbers M97928 and M97927). An open reading frame (ORF) of 2505 bp was identified by sequencing. The nearest ATG codon was 15 bases from the upstream end of the ORF. It meets consensus criteria for a translational start site (1 I), as does another ATG codon 9 bases downstream. Based on the first ATG codon, the predicted gB coding sequence is 2493 bp (counting the stop codon) and the translation product is 830 amino acids in length, with a calculated molecular weight (before glycosylation or signal peptide cleavage) of 93 kDa. These findings were the same with both strains GS and 229. As previously demonstrated (26), the HHV6 DNA polymerase coding sequence is immediately downstream of gB (Fig. 1). The ORF upstream of gB extends for 13 codons into the gB coding sequence. This is similar to CMV, where the ORF UL56 extends for several codons into UL55 (gB). Peptide sequence fragments deduced from the vicinity of upstream primer CB4AxB as well as the start of HHV6 gB show clear homologywith CMV UL56, which encodes an 850-aa product of unknown function (possibly a transport protein). As expected for a herpesvirus envelope glycoprotein, the deduced gB translation product contains a signal sequence of hydrophobic residues at the beginning, and another hydrophobic region around residue 700 which is in the putatative transmembrane region (Fig. 2). There are 10 (strain GS) or 8 (strain 229) Nlinked glycosylation sites (Asn-X-Ser or Asn-X-Thr), and 10 cysteine residues. In the vicinity of residue 400,
there is a cluster of basic amino acids that could represent a protease cleavage site. The gB sequences of strains GS and 229 showed 95.4% identity at the nucleotide level and 96.3% identity at the peptide level. Variation was sporadic throughout the coding sequence (Fig. 2); there was no strong clustering of gB variant codons as previously observed in CMV strains of different groups (6). Of the 31 codons showing peptide variation, about half of the substitutions involved some change in polarity. As with CMV, there was more peptide variation in the first half of the HHV6 gB; this region contains 23 of the 31 codons showing variation between strains GS and 229. HHV6 strains GS (A type) and 229 (B type) are known to differ in a number of respects and are epidemiologically divergent. Restriction analyses of their genomes readily reveal distinctive differences (I), including those in the gB/Pol region. Most HHV6 isolates from patients with roseola (22) and transplant recipients (our unpublished data) have been type B, with restriction patterns that resemble those of 229 rather than GS. The similarity in the gBs of strains GS and 229 suggests that the extent of interstrain variation in HHV6 gB may not be as great as in CMV, and that gB differences may not account for the different properties of type A and B HHV6 strains. Alignment of the gB peptide sequences of HHV6 (either strain) and CMV (strain AD1 69) revealed an overall 39% amino acid identity. Strong homology with CMV was noted throughout HHV6 gB except for the signal sequence (Fig. 3). Overall, the optimized FASTP/
SHORT COMMUNICATIONS
526
HHV6 CMV
(GS) (AD169)
HIiVC (GS) CMV
HHV6 CNV
(AD169)
(GS) (AD169)
IiHV6 (GS) CMV
HHV6
(AD169)
(GS)
CMV
(AD169)
HHV6
(GS)
CMV
HHV6 CMV
HHV6
(AD169)
(GS) (AD169)
(GS)
CNV
(AD169)
HHV6
(GS)
CNV
HHV6 CMV
HHVC CMV
HHV6 CMV
HHV6 CMV
HHVC CNV
HHVC CMV
(AD169)
(GS) (AD169)
(GS)
NSKMVVLF&FLt4NSV&yCDpDHyIiiGyNHKypF&SIAKGTDtiRpD~& I:I.:.tl.(.lt:I.*:.*.~ :.. SQHVT~~EAVS~aIYT~Y~%W~~~~VCS~ O$DLIRFERN#TS PG YKS--NAKMS~~FFIIYKTN~~TYTFPVRT&ELTFQS~i!tDVGWYF&Vt4GLAMpV :. :....:I..*.:: :: ..:: ::.:.: t:z..:: . ..:.:.... NKPItii;LDEGI IVMT:~~~QKVLT~~YAYIY~~~~SNTE'fPM NV?z?
.: t.
120 130 140 150 160 170 YEANLVNSHAQCYSAVAMKRPDGTVFSAFHEDNNKNNTLNLFPLNFKSITNKRFITTKEP .:::: :.: :. .:.:..:.: . .. . . ,.:..:.:.. ~I~~~~~~~:~-~VI~~~~YN~S~~~~NQLIPDD~~~THSTR wQ 180 190 200 210 220 230 YFARGPLWLYSTSTSLNCIVTEATRKAKYPFSYFALTTGEIVEGSPFFNGSN--GKHFAE * ':' . :::.,. .:::..: .::..:*s. .:: .::..: :::.::.I . . .. . . NNSRG;~yRETC~XJSLTITTAR~~PyHFFA~~~GDWYIS~~~NGTNRNAsY&GE 240 250 260 270 280 290 PLEKLTILENYTNIEDLNNGMNGATTLVRKIAFLEKADTLFSNEIKEENESVCNLKHNTT .:. :. :::...:. . :.:.. : .:::t.:z...::.:..:,. .: :. :.. NADKF~~~PNYTIVS~#iP-NA&PE~~UJAFLERAD~VISWDIQ~~~NVTCQLTFW&A 300 310 320 330 340 350 VTHGLRAETDETYHFISKELTAAFVAPKESLNLTDPKQTCIKDEFEKIINEWMSDYNDT . ...I.......:: : ..t:.:.. :...:..:. .*..tz . . . . . . ..::.: SERTIRS~~DSYHFS~A&MTATFLS~~ 8EVNNsDSAL~VRDEAIl??;QQIFNTSY$ 8T 360 370 380 YSNNGSYQIFKTTGDLILIWQPLVQKSLNFLE-I.. :. .z.:.:.:...:x . ::::. :: YEKYGNV~~ETSGGLVVE&NQGIKQK~~~LERS.J'.N 410 420 430 440 450 460 -----DILWQLQYLYDTLKDYINDAM;NLAESWCLDQKRTSKISPSSIVSEV . .. . . ..t. :::I..*:: t:...:t.::.z :.::.....:*:::.::.:.:.. MKSVIiN~~~AQLQFTY~;~RGYINRF+~ IAKAWCVD RRTLEVFKE&KINPSAU&I 8 5Po 470 480 490 500 510 520 YGRPISAQLHGDVLAISKCIEVNQSSVQLHKSMRWDAKGVRSETMCYNRPLVTFSFVNS :..::.:.. ::::....:...::.:z : .: . .. . . : ..::.::.:.:.:.:: YNKPIAAR$MGDVLGLAi$VTINQTSV --,~RDMNVKE-,s~RCYS~~~FNF~S 530 540 550 560 570 580 TPEVVPGQLGLDNEILLGDHRTEECEIPSTKIFLSGNHAWD .: :::: :::::t*.:t::*:..:: :::..:: :. :.:: . .. .. . . .. -SY$ YGQLGED~~;LLGNHRT~~~QLPSLKI~~~GNSAYEyV~~FS&.lID;~~ISTVD 8 590 600 610 620 630 640 AFIRLKIDPLENADFKVLDLYSPDELSRANVFDLENILREXNSYKSALYTIEAKIATNTP t :.::::::.z:.::.*::..zs..,s ::*::.:.::.:::: .:.a...
:
(AD169)
(GS) (AD169)
(GS) (AD169)
(GS) (AD169)
(GS) (AD169)
650 660 670 680 690 700 SYVNGINSFLQGLGAIGTGLGSVISVTAGALGDIVGGWSFLKNPFGGGLMLILAIVWV .:..:..... :::: :...: .:....::....:.::..:::::::. . . . .. . .. . . P"4~~LDDLNSG$GjAGW~~%AVGGAVASSGVATFL~~~FGAFTII~~~IAWI 710 720 730 740 750 760 IIIWFVRQRHVLSKPIDNMFPYATNPVTTVSSVTGTTWKTPSVKDVDGGTSVAVSEKE :. . . .. . . . . . ..:.. .::: . . . .t. . ..:. . . .. . . . . . . . :. "YL;YTRQRPLC; 8PLQNLFPYL;SADGTTVTS&sTKDTSL 0EPSYEESV-YNZGRKG 770 780 790 800 810 EGMADVSGQVSDDEYSQEAALKMLKAIKSLDESyRRKpSSSES---------HAsKPSLI : . . . . .. . :..:.: .:: :, .:: : . .. .. . . . . ::.:. PGP;$~DASTAAPZ";NEQAYQbU~ DAE RAQQNGTDS~~%QTGTQDKG KPNLL 850 0 8%0 820 830 DRIRYR--GYKSVNVKEA ::.:.: :t. . . . . DRLFt'd~NGYRHL;%~DE
FIG. 3. Alignment of gB peptide sequences of HHV6 and CMV. :, identical amino acids; boxed and region of highest homology is underlined.
FASTA homology score was 1760, higher than that usually noted between herpesvirus species and comparable to the homology score between herpes simplex (HSV) and varicella-zoster (VZV) which is 2200. By comparison, FASTP/FASTA scores of CMV gB align-
. conserved amino acids. Putative cleavage site is
ments with HSV, VZV, and Epstein-Barr viruses are in the range of 700 to 1100 (5). Scores for HHV6 gB alignments with HSV and VZV were in the range of 500 to 900. The CMV gB is about 76 a.a. longer than HHV6 gB; there is no homology with HHV6 in the region be-
SHORT COMMUNICATIONS
low CMV residue 88. This part of CMV gB shows considerable interstrain variation and is the site of strainspecific epitope(s) (3, 6). The other focus of interstrain variation in CMV, at codons 440-480 surrounding the protease cleavage site (7) is also a region of relatively low homology with HHV6 because of the deletion of a number of amino acids from HHV6 gB (Fig. 3). Thus, homology data appear to support the impression from GS-Z29 comparative data that HHV6 gB differs less among strains than CMV gB. Most of the key features of gB are in homologous locations in both HHV6 and CMV (Fig. 3). There is a run of arginine residues in HHV6 gB corresponding to the protease cleavage site of CMV gB. All 10 cysteine residues of HHV6 gB outside the signal sequence are matched to corresponding cysteine residues of CMV. CMV gB has an additional six cysteine residues, of which four are in the signal sequence. The N-linked glycosylation sites of HHV6 and CMV gB are likewise similar though not identically matched. This correspondence of features suggests a close structural similarity of gB’s from HHV6 and CMV. Peptide sequence homology is especially strong in the region following the putative cleavage site. HHV6 gB residues 412-630 are almost perfectly conserved (99%) between the two strains examined (Fig. 2) and have a 549/o amino acid identity with CMV gB residues 480-690 (Fig. 3). This region of CMV gB (gp55) is known to contain linear and conformational epitopes recognized by several different monoclonal antibodies with neutralizing activity (4, 10, 24, 27). Human immune sera recognize linear epitope(s) located between residues 589 and 645 (70). The corresponding HHV6 gB sequence contains a tract of 3 1 residues (530-560) where 24 are identical to those of CMV (Fig. 3). The close peptide sequence similarity of parts of gB between HHV6 and CMV may be relevant to the yet unresolved issue of cross-reactivity between the viruses (8, 2 1, 25). Strong rises in antibody to HHV6 are frequently observed in association with primary CMV infection (8) yet antigen absorption experiments have not suppressed the overall serologic reactivity to the other virus (21). Cross-reactivity of these two viruses may be limited to epitopes of certain proteins, similar to that observed between HSV and VZV (23). The main conclusions of this sequence analysis are as follows. HHV6 gB is closely related to CMV gB, much more so than usually observed between human herpesvirus species. Structural similarity is especially striking in a region beyond the putative cleavage site, corresponding to CMV gp55, where amino acid identity exceeds 50%. This region in CMV contains major
527
neutralization-related epitopes. With the observed degree of gB homology between HHV6 and CMV, partial immunologic cross-reactivity is plausible. Two divergent HHV6 strains (types A and B) had gB peptide sequences that differed little and were especially conserved in regions corresponding to known CMV gB neutralizing epitopes. The known biological differences of these HHV6 subtypes are therefore not likely to be due to gB variation.
ACKNOWLEDGMENT This work was supported search funds.
by Department
of Veterans Affairs re-
REFERENCES 1. ABLASHI, D. V., BALACHANDRAN,N., JOSEPHS,S. F., HUNG, C. L., KRUEGER,G. R., KRAMARSKY.B., SALAHUDDIN, S. Z., and GALLO, R. C., Vifology 184, 545-552 (1991). 2. ASANO, Y., YOSHIKAWA,T., SUGA, S., YAZAKI, T., HATA, T., NAGAI, T., KAJITA,Y., OZAKI, T., and YOSHIDA, S., J. Pediatr. 114, 535539 (1989). 3. BASGOZ, N., QADRI, I,, NAVARRO, D., SEARS, A., LENNETTE, E., YOUNGBLOM, J.. and PEREIRA, L., /. Gen. viral. 73, 983-988 (1992). 4. BANKS, T., Huo, B., KOUSOULAS,K., SPAETE, R., PACHL, C., and PEREIRA,L., J. Gen. Viral. 70, 979-985 (1989). 5 CHEE, M. S.. BANKIER,A. T., BECK, S., BOHNI, R., BROWN, C. M., CERNEY, R., HORSNELL. T., HUTCHISON, C. A., KOUZARIDES,T.. MARTIGNETTY,J. A., and BARRELL, B. G., Curr. Top. Microbial. Immunol. 154, 127-l 69 (1990). 6. CHOU, S., virology 188, 388-390 (1992). 7. CHOU, S., and DENNISON, K. M., J. lnfecf. Dis. 163, 1229-1234 (1992). 8. CHOU, S., and Scorr, K. M., J. C/in. Microbial. 28, 851-854 (1990). 9. JOSEPHS,S. F., ABLASHI, D. V., SALAHUDDIN, S. Z., JAGODZINSKI, L. L., WONG-ST&IL. F., and GALLO, R. C., J. Virol. 65, 55975604 (1991). 10. KNIESS,N., MACH, M., FAY, J., and BRIIT, W. J., /. Viral. 65, 138146(1991). 11. KOZAK, M., Nucleic Acids Res. 9, 5233-5252 (1981). 12. LAWRENCE, G. L., CHEE, M., CRAXTON, M. A., GOMPELS, U. A., HONES% R. W., and BARRELL, B. G., J. Viral. 64, 287-299 (1990). 13. LEW. J. A., FERRO, F., GREENSPAN,D., and LENNE‘TTE,E. T., Lancet335,1047-1050(1990). 14. LINDQUESTER,G. J., and PELLETT, P. E., virology 182, 102-l 10 (1991). 15. MARTIN, M. E. D., THOMSON, B. J., HONESS, R. W., CRAXTON, M. A., GOMPELS, U. A., LIU. M. Y., LITTLER, E., ARRAND, J. R., TEO, I., and JONES, M. D., J. Gen. Viral. 72, 157-168 (1991). 16. NEIPEL, F., ELLINGER,K., and FLECKENSTEIN,B., J. Gen. !/ire/. 72, 2293-2297(1991). 17. OKUNO. T., TAKAHASHI, K., BALACHANDFUI,K., SHIRAKI,K., YAMANI-
528
78. 79.
20.
21.
22.
SHORT COMMUNICATIONS SHI, K., TAKAHASHI, M., and BABA, K., J. C/in. Microbial. 27, 651-653 (1989). PEARSON,W. R., Methods Enzymol. 183, 63-98 (1990). PRUKSANANONDA,P., HALL, C. B., INSEL, R. A., MCINTYRE, K., PELLETT, P. E., LONG, C. E., SCHNAEIEL,K. C., PINCUS, P. H., STAMEY, F. R.. DAMEAUGH, T. R., and STEWART,J. A., N. Engl. /. Med. 326, 1445-l 450 (1992). SALAHUDDIN, S. Z., ABLASHI, 0. V., MARKHAM, P. D., JOSEPH& S. F.. STURZENEGGER,S., KAPLAN, M., HALLIGAN, G., BIBERFELD, P., WONG-STAAL. F., KRAMARSKY.B., and GALLO, R. C.. Science 234, 596-601 (1986). SAXINGER,C., POLESKY,H., EBY, N., GRUFFERMAN,S., MURPHY, R., TEGTMEIR,PAREKH,V.. MEMON. S., and HUNG, C.,/. Viral. Methods 21, 199-208 (1988). SCHIRMER, E. C., WYAIT, L. S., YAMANISHI, K., RODRIGUEZ,W. J.,
23. 24.
25. 26. 27.
and FRENKEL,N., Proc. Nat/. Acad. Sci. USA 88, 5922-5926 (1991). SCHMIDT, N. J., f. Med. Viral. 9, 27-36 (1982). SPAETE, R. R., THAYER, R. M., PROBERT,W. S., MASIARZ, F. R., CHAMBERLAIN, S. H., RASMUSSEN, L., and MERIGAN, T. C., PACHL, C., Virology 167, 207-215 (1988). SUTHERLAND,S., CHRISTOFINIS,G., O’GRADY, J., and WILLIAMS, R., J. Med. Viral. 33, 172-176 (1991). TEO, I. A., GRIFFIN, B. E., and JONES, M. D., J. Viral. 65, 46704680 (1991). UR. U., BRITT, W., VUGLER, L., and MACH, M., J. Viral. 63, 19952001 (1989).
28. YAMANISHI, K., OKUNO, T.. SHIRAKI, K., TAKAHASHI, M.. KONDO, T., ASANO. Y., and KUW\TA, T., Lancer 1,1065-l 067 (1988).
