Proc. Nati. Acad. Sci. USA Vol. 89, pp. 559-563, January 1992 Medical Sciences Characterization of a prototype strain of hepatitis E virus (anti-hepatitis E virus antibodies/PCR/genome sequence comparison) SERGEI A. TSAREV*tt, SUZANNE U. EMERSON*, GREGORY R. REYES§, TATIANA S. TSAREVA*, LLEWELLYN J. LEGTERSI, IFTIKHAR A. MALIK II, MUHAMMAD IQBAL II, AND ROBERT H. PURCELL* *Hepatitis Viruses Section, Laboratory of Infectious Disease, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD 20892; tShemyakin Institute of Bioorganic Chemistry, Moscow 117871, Russia, U.S.S.R.; §Molecular Virology Department, Genelabs Inc., Redwood City, CA 94063; 1Department of Preventive Medicine and Biometrics, Uniformed Services University of the Health Sciences, Bethesda, MD 20814; and 'lPakistan-U.S. Laboratory for Sero-Epidemiology, Army Medical College, Ralwalpind, Pakistan Contributed by Robert H. Purcell, October 10, 1991

ABSTRACT A strain of hepatitis E virus (SAR-55) impli- because the level of HEV in feces and bile is very low (5, cated in an epidemic of enterically transmitted non-A, non-B 11-19), the sensitivity of IEM is inadequate for complete hepatitis, now called hepatitis E, was characterized extensively. characterization of HEV infection. A more sensitive tech- Six cynomolgus monkeys (Macaca fascicularis) were infected nique, detection of the virus genome by reverse transcrip- with a strain of hepatitis E virus from Pakistan. Reverse tion-polymerase chain reaction (RT-PCR), was used in this transcription-polymerase chain reaction was used to determine study to correlate the presence of HEV in serum, bile, and the pattern of virus shedding in feces, bile, and serum relative feces ofan experimentally infected cynomolgus monkey with to hepatitis and induction of specific antibodies. Virtually the biochemical evidence of hepatitis and development of anti- entire genome of SAR-55 (7195 nucleotides) was sequenced. bodies to HEV (anti-HEV). In addition, we obtained struc- Comparison of the sequence of SAR-55 with that of a Burmese tural information about the virus genome through sequencing strain revealed a high level of homology except for one region of the PCR products themselves or their cloned deriva- encoding 100 amino acids of a putative nonstructural polypro- tives.** tein. Identification ofthis region as hypervariable was obtained by partial sequencing of a third isolate ofhepatitis E virus from MATERIALS AND METHODS Kirgizia. Virus Samples and Inoculation of Primates. Feces contain- Epidemics of enterically transmitted non-A, non-B hepatitis ing HEV SAR-55 were collected from a patient during a (hepatitis E) have been reported in Asia, Africa, and North hepatitis E outbreak in Sargodha, Pakistan (20, 21). Approx- America (for reviews, see refs. 1-4). Similar cases of spo- imately 0.5 ml per monkey ofa 10% (wt/vol) stool suspension of feces in fetal calf serum was used for intravenous inocu- radic hepatitis, presumed to be hepatitis E, account for up to lation of six cynomolgus monkeys (Macaca fascicularis). 90o of reported hepatitis in countries where hepatitis E is Blood samples from these monkeys were taken approxi- endemic. Furthermore, hepatitis E has been implicated in mately twice weekly before and after inoculation and tested fulminant hepatitis of pregnancy: up to 20o of pregnant by Metpath (Rockville, MD) for biochemical evidence of women with hepatitis E die oftheir disease. That a viral agent hepatitis by measuring levels of serum alanine aminotrans- was responsible for hepatitis E epidemics was first shown by ferase, isocitric dehydrogenase, and y-glutamyltransferase. Balayan et al. (5). It was provisionally called hepatitis E virus Fecal and bile samples were also collected from one monkey. (HEV) (2). On the basis of electron microscopy and other For daily collection of bile, implantation surgery was per- characterization it was proposed that HEV belongs to the formed on the 7th day after inoculation to establish an calicivirus family (6, 7). A partial sequence (8) and later the indwelling T-tube into the bile duct. entire sequence (9) of an HEV genome have been reported. Bile from a cynomolgus monkey infected with another Comparisons of HEV sequences with sequences available in strain of HEV (OSH-1852) was kindly provided by Michael computerized data bases, including those of caliciviruses Balayan (Institute of Poliomyelitis and Viral Encephalitis, (10), showed little similarity between HEV and any other Moscow). This cynomolgus monkey had been inoculated known viruses. with feces collected during a hepatitis E outbreak in 1988 in Several species of primates have been experimentally Osh, Kirgizia, U.S.S.R. Bile was collected on the 14th day infected with HEV (5, 11-19). Although HEV infection is less after infection. severe in monkeys than in humans, it is sufficiently similar Detection of Anti-HEV Antibodies. A modified ELISA that monkeys can serve as a reasonable and reproducible protocol (22) was used in this study. Recombinant HEV animal model to provide a more complete characterization of antigens for use in the ELISA were derived from Mexican this infection. and Burmese strains (23, 24) and were produced in the Determination of the pattern of HEV shedding in feces, pGEX1 vector system (25). Cynomolgus sera were diluted bile, and sera during infection is important for characteriza- 1:100 in 1% gelatin/phosphate-buffered saline (PBS). Alka- tion of the disease and elucidation of the epidemiology of line phosphatase-conjugated goat anti-human IgG was used HEV. In some cases, the presence of HEV in a sample has as a second antibody. been demonstrated by transmission to animals (5, 11-19) but such transmission studies cannot be used for routine inves- Abbreviations: HEV, hepatitis E virus; IEM, immune electron tigation. Until recently, the primary method available for microscopy; RT, reverse transcription; nt, nucleotide(s); ORF, open characterization of virus shedding was immune electron reading frame. microscopy (IEM) (for reviews, see refs. 1-4). However, tTo whom reprint requests should be addressed at: Laboratory of Infectious Disease, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Building 7, Room 200, 9000 The publication costs of this article were defrayed in part by page charge Rockville Pike, Bethesda, MD 20892. payment. This article must therefore be hereby marked "advertisement" **The sequences reported in this paper have been deposited in the in accordance with 18 U.S.C. §1734 solely to indicate this fact. GenBank data base (accession nos. M80581 and M81415). 559 Downloaded by guest on September 29, 2021 560 Medical Sciences: Tsarev et al. Proc. Natl. Acad. Sci. USA 89 (1992) Primers. Ninety-two primers, 21-40 nucleotides (nt) long, MgCl2, all four dNTPs (each at 0.2 mM), 50 pmol of direct and complementary to plus or minus strands of the genome primer, 50 pmol of reverse primer, 40 units of RNasin of a strain of HEV from Burma (BUR-121) (9) or the SAR-55 (Promega), 16 units of avian myeloblastosis virus reverse genome were synthesized using an Applied Biosystems transcriptase (Promega), 4 units of AmpliTaq (Cetus), under model 391 DNA synthesizer. For cloning of PCR fragments, 100 ,ul of light mineral oil. The mixture was incubated 1 h at EcoRI, BamHI, or Bgl II restriction sites preceded by 3-7 nt 42°C and then amplified by 35 PCR cycles: 1 min at 94°C, 1 were added to the 5' end of primers. min at 45°C, and 1 min at 72°C. The PCR products were For detection of the HEV genome in samples of bile, sera, analyzed on 1% agarose gels. and feces, two sets of "nested" primers were used that Cloning of PCR Fragments. PCR fragments containing represented sequences from the 3' region (ORF-2) of the restriction sites at the ends were digested with EcoRI and SAR-55 genome. Primers for RT and the first PCR were BamHI or EcoRI and Bgl II restriction enzymes and cloned 5'-GTATAACGGATCCACATCTCCCCTTACCTC-3' and in EcoRI/BamHI-digested pBR322 or pGEM-3Z (Promega). 5'-TACAGATCTATACAACTTAACAGTCGG-3' and for Alternatively, PCR fragments were cloned into pCR1000 the second PCR were 5'-GCGGCAGATCTCACCGACAC- using the TA cloning kit (Invitrogen, San Diego). CATTAGTAC-3' and 5'-TAACCTGGATCCTTATGCCGC- Sequencing of PCR Fragments and Plasmids. PCR frag- CCCTCTTAG-3'. ments were excised from 1% agarose gels and purified by Preparation of Virus RNA Template for PCR. Bile (10 sul), Geneclean (Bio 101, La Jolla, CA). Double-stranded PCR 20%6 (wt/vol) SDS (to a final concentration of 1%), proteinase fragments were sequenced by using Sequenase (United K (10 mg/ml; to a final concentration of 1 mg/ml), 1 A.l of States Biochemical) as described (26). Double-stranded plas- tRNA (10 mg/ml), and 3 A.l of 0.5 M EDTA were mixed in a mids purified through CsCl gradients were sequenced with a final volume of250 Ala and incubated for 30 min at 550C. Total Sequenase kit (United States Biochemical). nucleic acids were extracted from bile twice with phenol/ of chloroform, 1:1 (vol/vol), at 65TC and once with chloroform, Computer Analysis of Sequences. Nucleotide sequences then precipitated by ethanol, washed with 95% ethanol, and HEV strains were compared using the Genetics Computer used for RT-PCR. RT-PCR amplification ofHEV RNA from Group (Madison, WI) software package (27), version 7.5, on feces and especially from sera was more efficient when RNA a VAX 8650 computer (at the National Cancer Institute, was more extensively purified. Serum (100 ,l) or a 10%6 fecal Frederick, MD). This software package was also used to suspension (200 ,ul) was treated as above with proteinase K. generate the hydropathy plots. After a 30-min incubation, 300 ,ul of CHAOS buffer (4.2 M guanidine thiocyanate/0.5 N-lauroylsarcosine/0.025 M RESULTS Tris HCI, pH 8.0) was added. Nucleic acids were extracted twice with phenol/chloroform at 65°C followed by chloro- Biological, Serological, and Virological Characterization of form extraction at room temperature. Then 7.5 M ammonium Infection. Six cynomolgus monkeys were intravenously in- acetate (225 ,ul) was added to the upper phase and nucleic oculated with a 10% suspension of human feces containing acids were precipitated with 0.68 ml of2-propanol. The pellet the SAR-55 strain of HEV. Biochemical and serological was dissolved in 300 ,lA of CHAOS buffer and 100 ,ld ofwater assays confirmed that all six inoculated animals were in- was added. Chloroform extraction and 2-propanol precipita- fected. Although data are shown for only one monkey (Fig. tion were repeated. Nucleic acids were dissolved in water, 1), all six monkeys displayed an increase in alanine ami- precipitated with ethanol, washed with 95% ethanol, and notransferase, isocitric dehydrogenase, and 'y-glutamyltrans- used for RT-PCR. ferase activity, indicative of acute hepatitis. All six animals RT-PCR. The usual 100-Al RT-PCR mixture contained also developed antibody to recombinant-derived HEV anti- template, 10 mM Tris HCl (pH 8.4), 50 mM KCl, 2.5 mM gens. These data demonstrated that the SAR-55 strain of HEV genome - - + + + + + ------serum - -- -++ +---++ +++ + ++++++ ++ - - - - - feces Anti-HEV _ _ _ + + + +

125

100 _

FIG. 1. Evidence for SAR-55 -j 75 infection of Cyno-376. Serum ala- nine aminotransferase (ALT) lev- els are plotted as units/liter and the presence (+) or absence (-) of viral genomes or antibody to HEV 50 is indicated. Viral genomes in the feces and serum were detected by RT-PCR. The presence of total serum antibodies to HEV was monitored by ELISA using re- 25 - combinant antigens from both the -6 0 10 20 30 40 50 60 Mexican and Burmese strains of Day post-inoculation HEV (23, 24). Downloaded by guest on September 29, 2021 Medical Sciences: Tsarev et al. Proc. Natl. Acad. Sci. USA 89 (1992) 561

4 ., r..V

...... FIG. 2. IEM detection of SAR-55 virions in the bile of Cyno-376. IEM was performed as described (13). Viral aggregates were identified in the bile by reac- tion of diluted bile with anti-HEV from a chimpanzee that had been A. ,,e infected with a strain of HEV from India (13). (x 150,000.)

HEV was able to consistently infect and induce hepatitis E in sequenced to provide the remainder of the sequence and to cynomolgus monkeys. confirm the consensus sequence. The sequence of the entire To correlate the distribution of virus with an increase in 7.195-kilobase genome, with the exception of 30 and 27 nt of virus-specific antibody and alanine aminotransferase levels, the 5' and 3' termini, respectively, was obtained. a sensitive nested RT-PCR protocol was performed on fecal, Since one other isolate of HEV, BUR-121 (9), had been serum, and bile samples from cynomolgus monkey 376 totally sequenced, computer analyses were performed to (Cyno-376). Control experiments in which we assayed dilu- determine the relatedness of SAR-55 to BUR-121. Unique tions of feces of known infectivity demonstrated that on nucleotide insertions or deletions were not detected and the average the HEV genome could be detected by RT-PCR in same three ORFs identified in BUR-121 were found in feces, with a sensitivity approximately equal to that of an SAR-55. Overall, the genomes were quite similar and differed assay based on transmission ofHEV to cynomolgus monkeys by only 6.7% in nucleotide sequence and by 1% in deduced (unpublished results). All bile, serum, and fecal samples were amino acid sequence (Table 1). As might be expected, the tested by RT-PCR multiple times to ensure that positive most conserved region was located in the overlap of ORF-2 samples were identified. and ORF-3 (bases 5147-5477). Against this overall pattern of Although Cyno-376 was inoculated with at least 1 x 103 sequence relatedness, the region between bases 2011 and infectious doses intravenously, we were not able to detect 2325 in ORF-1 appeared unique (Fig. 3). Although the sum of virus in serum on the day of infection or for several days the nucleotide differences in this region was only 2% higher thereafter (Fig. 1). However, we were able to detect viral than the average for the entire genome, the sum ofamino acid genomes in feces and bile as early as day 6 and day 7, differences was 14% or 14 times higher than the average. respectively, confirming that excretion of virus was an early Because the sequence of this region in the SAR-55 genome indicator of infection. All samples of bile collected (days differed so greatly from that of BUR-121, it was also ampli- 7-41) were positive for HEV RNA by RT-PCR. Unfortu- fied by RT-PCR directly from the original human feces. The nately, due to the indwelling catheter method of bile collec- sequence obtained from virus in the human fecal sample was tion, residual contamination from earlier samples could not identical to that obtained from virus in the cynomolgus bile. be eliminated so the end point for virus presence in the bile Identification of a Hypervariable Region. To determine if could not be determined by PCR. However, when the less- this extreme divergence of sequence between nt 2011 and sensitive IEM technique was used in the analysis of the bile, 2325 was unique to these two isolates or had a more universal virus was detected sporadically from day 26 to day 37 with a significance, sequences from two regions of a third strain of peak on day 32 (data not shown). The HEV particles resem- HEV were derived. Fragments of cDNA from the RT-PCR bled those described previously (Fig. 2). amplification of the OSH-1852 strain of HEV corresponding Viral excretion in feces could be documented more pre- to 2002-2424 and 4424-4800 the genome were cisely by PCR, and excretion began on day 6 and ended by positions of day 35 (Fig. 1). Three samples, taken between days 10 and 15, sequenced. Both the nucleotide and amino acid identities in were repeatedly negative for viral RNA. The inability to the region (positions 4424-4800) encoding the putative RNA- detect viral RNA in these three samples suggested that the directed RNA polymerase were high and did not differ level of excreted particles was low. Virus was first detected substantially among the three strains (Table 2 and Fig. 3). In in the serum on day 9, appeared to be present at higher titer contrast, between nt 2002 and 2424 there was significant (based on the quantity of PCR product) from days 14 to 23, divergence in nucleotide sequences and even greater diver- then disappeared by day 28 (Fig. 1). A similar pattern of gence in amino acid sequences among the three strains. viremia in a cynomolgus monkey infected with HEV was Table 1. Comparison of BUR-121 and SAR-55 sequences observed by Uchida et al. (28). Molecular Characterization of the SAR-55 Genome. RT- Nucleotides Amino acids PCR amplification of bile from Cyno-376 was employed to Length, Identity, Length, Identity, generate six cDNA fragments encompassing the entire ge- Region no. % no. % nome of SAR-55. The PCR-generated fragments were either sequenced directly or were cloned into individual plasmids ORF-1 5079 93.1 1693 98.8 and sequenced subsequent to amplification in Escherichia ORF-2 1980 93.8 660 99.4 coli. Direct sequencing of the PCR product provided 45% of ORF-3 369 98.9 123 100 the genome sequence as a consensus sequence. Both strands Total 7138* 93.3 2476 99.0 of cloned cDNAs representing 87% of the genome were *Length of the primers is not included. Downloaded by guest on September 29, 2021 562 Medical Sciences: Tsarev et al. Proc. Nati. Acad. Sci. USA 89 (1992) Non structural region (nucleotide binding protein?)

H T F 1 7Y 7 A O R L S L T G H F W F H 2002 GCCCAGCGCCTTTCGCTGACCGGTAATTTTTGGTTCCAT 2040 A T C A A L L 7 I P EG L L G P F A P F S P G H V W E S A N P F C G E S T L Y T R T W S E V D A V 2160 A A T C a T C C A T G G T

S R A P P L L L P P

P S P A O P D L G F T SE P SI P S R AA T P T P A A P L P P P A P D P S P T L 2280 T G T Tr. G T C cc C T C G T c T C T TT T C C

LA S - G S AP A R G E P A P G A T A R A P A I T HO0 T AR H R R L L F T Y P D G S KV F 2400 C T C T G a T C C A C GT Tr. C C T T A T

A G S L F E S T GCCGGCTCGCTGTTTGAGTCGACA 2424 C

FMA-dependlent EWA polymerase region

S T O NHF 4423 TCCACCCAGAATAATTTT 444 C

S L G L E C A I M E EC G M PO0W L I R L Y H L I R S AW I L 0 A P K E S L R G 4560 TC T C T CT T A C

L 7 F W K K H S G E P G T L L W H T V W H M A V I T H C Y 0 F R D L O VA A F KIGD TTTTGAAGAACCTCCGTGACCCGCACCTTCGTGAATATGTTGGACATGCCGTATCCCCATGTATGATTCGCGACTGCGGTGCTGCTTTAAGTG 46808 T A T T C T T T A T T C

D S I V L C S E Y ROQ S P G A AV L I A G C G L K L K V D F R P IG L Y A G V V 4800 T T A CC-- T C A C C T T T C T

V A P G L G AL P D V V R F AG R L T E K H W G P G GTGGCCCCCGGCCTTGGCGCGCTTCCTGATGTCGTGCGCTTCGCCGGTCGGCTTACTGAGAAGAATTGGGGCCCTGGC 4878 C T C C T C C

FIG. 3. Partial cDNA nucleotide sequence and the deduced amino acid sequence of SAR-55. Nucleotides and amino acids that differ in the BUR-121 (boldface type) (9) or OSH-1852 (boldface underlined type) strains are indicated.

Therefore, the region between nt 2011 and 2325 probably agreement with previous studies in which virus was detected represents a relatively hypervariable region of the HEV by IEM in feces and bile as early as 9 days after infection (5, genome. 11, 19). Although the monkey was inoculated intravenously with the virus, we were not able to detect virus in serum on

DISCUSSION Table 2. Comparison of SAR-55, BUR-121, and OSH-1852 HEY Infection of Cynomolgus Monkeys. Previous studies sequences in two regions of ORF-1 have shown that cynomolgus monkeys are useful for exper- imental HEV infection (5, 11-13, 16, 18, 19). Extension of our Identity, % original study demonstrated that all six cynomolgus monkeys Nucleotides Amino acids inoculated with human feces containing HEV SAR-55 devel- Region BUR-121 OSH-1852 BUR-121 OSH-1852 oped hepatitis E, as defined by liver enzyme elevation, and developed anti-HEV antibodies. Since this virus was able to nt 2002-2424 consistently infect cynomolgus monkeys and cause hepatitis, SAR-55 92.0 92.9 89.4 91.5 it appeared to be suitable for selection as a prototype virus for BUR-121 100 91.0 100 92.2 biological and molecular characterization. Therefore, infec- nt 4423-4878 tion of a cynomolgus monkey with the SAR-55 strain of HEV SAR-55 93.2 96.3 99.3 100 was intensively monitored. Detection of the viral genome in BUR-121 100 92.6 100 99.3 bile and feces suggested that the virus replicated in the liver Regions compared are nt 2002-2424 (predicted nonstructural pro- during the first week after infection. This finding is in good teins) and nt 4423-4878 (predicted RNA-directed RNA polymerase). Downloaded by guest on September 29, 2021 Medical Sciences: Tsarev et al. Proc. Natl. Acad. Sci. USA 89 (1992) 563

-3 RNA polymerase region could be used for designing PCR primers to detect a spectrum of HEV strains with different origins. 11I We are grateful to Mrs. Marianne Lewis, Mr. Terry Popkin, and Ms. Katherine Gabor for technical assistance and to Drs. Michael Balayan, Joe Bryan, and Stephen Denny for providing material containing HEV. Excellent animal care was provided by the staff of Bioqual, Gaithersburg, MD.' We acknowledge the National Cancer Institute for allocation of computing time and staff support at the Advanced Scientific Computing Laboratory ofthe Frederick Cancer Research and Development Center. This study was supported in part by a grant from the World Health Organization Programme for Vaccine Development and Contrict N01-AI-05069. 1. Balayan, M. S. (1987) Virol. Rev. 2, Sect. E, 235-261. -3 2. Purcell, R. H. & Ticehurst, J. R. (1988) in Viral Hepatitis and Liver Disease, ed. Zuckerman, A. J. (Liss, New York), pp. 131-137. FIG. 4. Hydropathy plot of the hypervariable region shown in 3. Bradley, D. W. (1990) Br. Med. Bull. 46, 442-461. Fig. 3 was plotted for each strain using the method of Kyte and 4. Ticehurst, J. R. (1991) in Viral Hepatitis and Liver Disease, eds. Doolittle (30) and profiles were superimposed for Hollinger, F. B., Lemon, S. M. & Margolis, H. S. (Williams & Wilkins, comparison. SAR- Baltimore), pp. 501-513. 55, bold line; BUR-121, dotted line; OSH-1852, dashed line. 5. Balayan, M. S., Andjaparidze, A. G., Savinskaya, S. S., Ketiladze, A. S., Bragipsky, D. M., Savinov, A. P. & Poleschuk, V. F. (1983) the day ofinfection or thereafter until 2-3 days after virus was Intervirology 211, 23-31. first detected in bile and feces. Virus excretion started by 1 6. Bradley, D. W. & Baiayan, M. S. (1988) Lancet i, 819. 7. Bradley, D. W., Andjaparidze, A. G., Cook, P. H., McCaustland, K., week after infection and 3 weeks before evidence of disease Balayan, M., Steeler, H., Velazques, O., Robertson, B., Humphrey, C., was observed. The amount of virus excreted at this time was Kane, M. & Weisfuse, I. (1988) J. Gen. Virol. 69, 731-738. not great, which probably explains why we were not able to 8. Reyes, G. R., Purdy, M. A., Kim, J., Luk, K.-C., Young, L. M., Fry, detect HEV genome K. E. & Bradley, D. W. (1990) Science 247, 1335-1339. the in fecal samples taken 12-14 days 9. Tam, A. W., Smith, M. M., Guerra, M. E., Huang, C., Bradley, D. W., after infection even though samples taken prior to and after Fry, K. E. & Reyes, V4. R. (1991) Virology 185, 120-131. this time were positive. Viremia was first observed on the 9th 10. Neil, J. D. (1990) Virus Res. 17, 145-160. day of the infection and appeared to intensify between 2 and 11. Andjaparidze, A. G., Balayan, M. P., Savinov, A. P., Braginsky, D. M., Poleschuk, V. F. & Zamyatina, N. A. (1986) Vopr. Virusol. 31, 73-81. 3 weeks after infection. Elevation of liver enzymes was first 12. Bradley, D. W., Krawczynski, Cook, E. H., Jr., McCaustland, K. A., noted as viremia ceased. It is probable that viremia is the Humpherey, C. D., Spelbring, J. E., Myint, H. & Maynard, J. E. (1987) result of viral replication in the liver although nonhepatic Proc. Nati. Acad. Sci. USA 84, 6277-6281. sites of replication have not been ruled out. 13. Arankalle, V. A., Ticehursf, J. R., Sreenivasan, M. A., Kapikian, A. Z., Popper, H., Payri, K. M. & Purcell, R. H. (1988) Lancet i, 550-554. Comparison of SAR-55 Genome with That of Other Strains. 14. Gupta, H., Tandon, B. N., Sriramachpri, S., Joshi, Y. K. & Iyenger, B. We found that the sequence of SAR-55 was quite similar to (1990) Indian J. Med. Rex. Sect. A 91, 87-90. that of BUR-121 over the entire length of the genome (Table 15. Krawczynski, K. & Bradley, D. W. (1989) J. Infect. Dis. 159,1042-1049. 16. Soe, S., Uchida, T., Suzuki, K., Komatsu, K., Azumi, J., Okuda, Y., 1). This result is not unexpected since the two viruses were fida, F., Shikata, T., Rikihisa, T., Mizuno, K., Win, K. M. & Tin, K. M. isolated from geographically contiguous regions at about the (1989) Liver 9, 135-145. same time. Differences between the two sequences were 17. Panda, S. K., Datta, R., Kaur, J., Zuckerman, A. J. & Nayak, N. C. distributed approximately uniformly throughout the genome (1989) Hepatology 10, 466-472. 18. Uchida, T., Win, K. M., Suzuki, K., Komatsu, K., fida, F., Shikata, T., except for one hypervariable domain located in the putative Rikihisa, T., Mizuno, K., Soc, S., Myint, H., Tin, K. M. & Nakane, K. nonstructural ORF (9) (Fig. 3 and Table 2). This region of (1990) Jpn. J. Fxp. Med. 60, 13-21. hypervariability preceded the nucleotide triphosphate bind- 19. Uchida, T., Suzuki, K., Komatsu, K., fida, F., Shikata, T., Rikihisa, T., Mizuno, K., Soe, S., Win, K. M. & Nakane, K. (1990) Jpn. J. Exp. Med. ing domain and also the RNA-directed RNA polymerase 60, 23-29. motifs recognized in ORF-1 (8, 9, 29). That this region was 20. Iqbal, M., Ahmed, A., Qamar, A., Dixon, K., Duncan, J. F., Islam, hypervariable was confirmed when another strain from Asia N. U., Rauf, A., Bryan, J. P., Malik, I. A. & Legters, L. J. (1989) Am. was partly sequenced. The genome of the OSH-1852 strain J. Trop. Med. Hyg. 40, 438-443. from approx- 21. Ticehurst, J., Popkin, T. J., Bryan, J. P., Innis, B. L., Duncan, J. F., Kirgizia differed in the hypervariable region to Ahmed, A., Iqbal, M., Malik, I., Kapikian, A. Z., Legters, L. J. & imately the same extent from the Pakistani and Burmese Purcell, R. H., J. Med. Virol., in press. strains as those strains differed from each other (Table 2). 22. Skidmore, S. J., Yarbough, P. O., Gabor, K. A., Tam, A. W., Reyes, However, all three strains did not differ significantly from G. R. & Flower, A. J. E. (1991) Lancet 337, 1541. each in RNA 23. Reyes, G. R., Yarbough, P. b., Tam, A. W., Purdy, M. A., Huang, C., other the region coding for RNA-directed Kim, J., Bradley, D. W. & Fry, K. E. (1991) Gastroenterol. Jpn. 26, polymerase. The finding of a hypervariable region in a 142-147. presumed nonstructural protein is surprising, since such 24. Yarbough, P. O., Tam, A. W., Fry, K. E., Krawczynski, K., McCaust- regions are usually found in structural proteins of viruses, land, K. A., Bradley, D. W. & Reyes, G. R. (1991) J. Virol. 65, 5790- 5797. where they are exposed to selection by the immune response 25. Smith, D. B. & Johnson, K. S. (1988) Gene 67, 31-40. ofthe host. A hydropathy analysis of the region revealed that 26. Winship, P. R. (1989) Nucleic Acids Res. 17, 1260. all three strains displayed a similar profile in spite of their 27. Devereux, J., Haeberly, P. & Smithies, 0. (1984) Nucleic Acids Res. 12, extensive sequence differences (Fig. 4). However, the sig- 387-395. pattern to 28. Uchida, T., Suzuki, K., Iida, F., Shikata, T., Araki, M., Ichikawa, M., nificance of this remains be determined. We Rikihisa, T., Mizuno, K., Soe, S. & Win, K. M. (1991) in Viral Hepatitis predict that the existence of this hypervariable region will and Liver Disease, eds. Hollinger, F. B., Lemon, S. M. & Margolis, allow even closely related strains to be distinguished by the H. S. (Williams & Wilkins, Baltimore), pp. 526-527. application of molecular epidemiological approaches. Pre- 29. Fry, K., Tam, A. W., Smith, N.; M., Kim, J. P., Luk, K.-E., Young, L. M., Piatak, M., Feldman, R. A., Yun, K. Y., Purdy, M. A., Mc- sumably, the sequence of the hypervariable region could be Caustland, K., Bradley, D. W. & Reyes, G. R. (1991) Virus Genes, in used for more detailed molecular classification of HEV press. strains, and the sequence from the much more conserved 30. Kyte, J. & Doolittle, R. F. (1982) J. Mol. Biol. 157, 105-132. Downloaded by guest on September 29, 2021