Hepatitis E Virus: Identification of Type ... - Journal of Virology

JOURNAL OF VIROLOGY, Nov. 1991, p. 5790-5797

Vol. 65, No. 11

0022-538X/91/115790-08$02.00/0 Copyright © 1991, American Society for Microbiology

Hepatitis E Virus: Identification of Type-Common Epitopes PATRICE 0. YARBOUGH,' ALBERT W. TAM,' KIRK E. FRY,' KRZYSZTOF KRAWCZYNSKI,2 KAREN A. McCAUSTLAND,2 DANIEL W. BRADLEY,2 AND GREGORY R. REYES"* Molecular Virology Department, Genelabs Incorporated, Redwood City, California 94063,1 and Hepatitis Branch,

Division of Viral Diseases, Centers for Disease Control, Atlanta, Georgia 303332 Received 29 May 1991/Accepted 8 August 1991

Large epidemic outbreaks of enterically transmitted non-A, non-B viral hepatitis (ET-NANBH) have been documented in developing countries. A molecular clone derived from the causative agent, the hepatitis E virus (HEV), has recently been described (G. R. Reyes, M. A. Purdy, J. P. Kim, K.-C. Luk, L. M. Young, K. E. Fry, and D. Bradley, Science 247:1335-1339, 1990). We now report the isolation, by serologic screening, of two cDNA clones derived from a fecal sample collected during a 1986 outbreak of ET-NANBH in Telixtac, Mexico. The cDNA clones encode epitopes that specifically reacted with acute- and convalescent-phase sera collected during five different ET-NANBH epidemics and represent the initial cloning of the Mexico strain of HEV. Recombinant fusion proteins expressed from these clones were also recognized by antibodies from cynomolgus macaques experimentally infected with HEV. The cDNA clones were shown to be derived from HEV by their specific hybridization to the previously recognized full-length genomic RNA transcript of -7.5 kb. In addition, however, subgenomic polyadenylated transcripts of -2.0 and -3.7 kb were also identified in HEV-infected cynomolgus monkey liver. Sequences homologous to the epitope clones were isolated from the Burma strain of the virus, and these demonstrated reactivity comparable to that seen with the Mexico strain epitopes. When compared with the available full-length sequence of the Burma strain of HEV, it was discovered that the cDNA clones were encoded in different open reading frames (ORFs). The comparison between Mexico and Burma HEV strains indicated amino acid homologies of 90.5 and 73.5% for these epitope-encoding clones derived from ORF2 and ORF3, respectively. The identification of these clones not only has provided insight into the expression strategy of HEV but has also resulted in a source of recombinant protein useful in the diagnosis of HEV-induced hepatitis. Hepatitis resulting from infection with viruses other than hepatitis A virus (HAV) and hepatitis B virus was previously referred to as non-A, non-B hepatitis (NANBH) (13, 15, 23). The diagnosis of NANBH was one of exclusion: i.e., demonstrating the absence of serologic markers for acute infection with HAV or hepatitis B virus. At least two epidemiologically distinct forms of NANBH were recognized and further substantiated by transmission studies in primates (1, 5, 8, 24). These different viruses were transmitted by either the parenteral or fecal-oral routes. The designation of NANBH therefore formerly represented a diverse collection of uncharacterized but etiologically distinct agents. The first characterized NANBH agent was that responsible for parenterally transmitted NANBH, or what is now called hepatitis C virus (10). This was followed by the cloning of a portion of the fecal-orally transmitted agent, the hepatitis E virus (HEV) (19). The latter form of virus-induced hepatitis has been referred to as enterically transmitted NANBH (ETNANBH). Epidemics of ET-NANBH have been recognized worldwide but occur principally in developing countries. They have been reported in southeast Asia, central Asia, Africa, Mexico, and Central America (6, 7, 16). In these areas, contaminated water has been implicated as the principle vehicle of virus transmission. Although HEV and HAV are transmitted in a similar manner, there are major differences in the clinical, pathological, and epidemiological courses of these two viruses. In particular, the mortality rate for HEV infection is 1 to 2%, or approximately 10-fold greater than that seen for HAV. Infection with HEV is particularly fatal *

Corresponding author. 5790

for pregnant women, for whom the mortality rate can be as high as 10 to 20% (6, 16). Practical serologic assays for the determination of HEV are presently not available. Currently, the diagnosis is made by clinical presentation and the lack of a serologic response to HAV and hepatitis B virus or by other predisposing conditions, exposures, or infections that can cause hepatitis. An immune response to viruslike particles (VLPs) in acutephase stool samples has been observed by using immunoelectron microscopy (7, 8, 16). However, most clinical specimens do not have sufficient VLPs in order for immunoelectron microscopy to be a useful tool for clinical or epidemiologic surveys. The serial passage of HEV in cynomolgus macaques has been reported previously (2, 3, 5, 8). In this primate model, the disease typically results in elevated liver enzymes and the detection of VLPs in both stool and bile. The development of a competent animal model for studies of human HEV has greatly assisted in the serologic characterization of HEV. It has been found that sera from geographically distinct outbreaks and experimentally infected cynomolgus monkeys will aggregate VLPs in human case stool specimens, thereby suggesting that one major virus is responsible for the majority of ET-NANBH observed worldwide (8). We demonstrate that cDNA clones isolated from a lambda gtll expression library are derived from the HEV genome and are specifically reactive with acute- and convalescentphase sera collected from individuals infected in the course of ET-NANBH epidemics. The serologic data presented here provide further evidence for HEV as the predominant etiologic agent of ET-NANBH. These results indicate the existence of type-common viral epitopes which appear to be

VOL. 65, 1991

HEV: IDENTIFICATION OF TYPE-COMMON EPITOPES

shared by divergent geographic isolates from Asia and North America. MATERIALS AND METHODS Human sera. Human serum samples were collected during ET-NANBH outbreaks in Pakistan, Mexico, Burma, Borneo, the USSR, and Somalia (7). Control sera were from individuals in the United States with no known history of exposure to HEV or travel to areas in which ET-NANBH is endemic. Acute-phase sera were collected between 1 and 12 days after the onset of symptoms. Convalescent-phase sera were collected from asymptomatic individuals between 30 and 90 days after the onset of jaundice. Any nonspecific reactivity in human sera against bacterium-derived proteins was removed by incubating appropriate serum dilutions with nitrocellulose filters prepared by using nonrecombinant lambda phage plated as described in the screening protocols

below. Cynomolgus macaque sera. Wild-caught cynomolgus macaques were quarantined for 6 to 8 weeks after importation (from the Philippines) and bled twice weekly for approximately 2 months to establish baseline liver enzyme profiles prior to infection (8). In brief, cynomolgus monkeys were intravenously infected with a 10% filtered stool suspension prepared from human stool samples collected during HEV outbreaks in Burma, Pakistan, and Mexico (7, 8). Serial passage of HEV in animals was by intravenous inoculation with a filtered 10% suspension of early-phase or pre-acutephase cynomolgus macaque stool filtrates. Cynomolgus monkey stools were analyzed by immunoelectron microscopy for the presence of disease-associated 27- to 34-nm VLPs. Acute-phase sera were collected within a 2-week period during the peak of alanine aminotransferase activity. Convalescent-phase sera were collected between 30 and 60 days after the return of alanine aminotransferase activity to preinoculation levels. Any possible seroreactivity to Escherichia coli-associated antigens was removed as described above. Library construction. A human stool sample collected during an ET-NANBH outbreak in Mexico (Mex#14) (7) was used for cDNA library construction. A 10% fecal suspension was clarified by centrifugation and extracted for viral RNA by phenol-chloroform extraction. After ethanol precipitation of RNA, cDNA was synthesized by random priming of the first strand by using a cDNA synthesis kit (Boehringer Mannheim). AB linker-primer oligonucleotides were ligated to the blunt-ended cDNA. The design of the AB linker-primer permits sequence-independent single-primer amplification of the heterogeneous cDNA population by repetitive rounds of annealing, extension, and denaturing in the presence of the A strand of the AB double-stranded oligonucleotide (18). After cleavage at the AB linker-primer EcoRI restriction site and purification of the cDNA from oligonucleotide fragments, cDNA libraries were constructed in the phage expression vector lambda gtll (27). Screening for HEV recombinant proteins. Recombinant cDNA libraries were screened for HEV specific antigenproducing clones by plating on E. coli KM392 (14) at 10,000 phage per 150-mm-diameter plate (27). After incubation at 42°C for 5 h, a dry nitrocellulose filter was overlaid and the plates were further incubated overnight at 370C. The filters were removed, washed in TBST buffer (10 mM Tris, pH 8.0, 150 mM NaCl, 0.05% Tween 20) and blocked for 1 h with antibody incubation buffer (10 mM Tris [pH 8.0], 150 mM NaCl, 1% gelatin). The filters were washed again with TBST

5791

and incubated overnight with purified immunoglobulin G (0.33 mg/ml in antibody incubation buffer) from F387-C, a convalescent human serum sample collected during the 1986 Telixtac, Mexico, outbreak (7). The filters were washed twice to remove any unbound first antibody and incubated for 2 h with alkaline-phosphatase-conjugated anti-human immunoglobulin G (0.13 ,ug/ml in antibody incubation buffer). After a final washing, the filters were developed in an indicator-substrate medium containing 0.33 mg of Nitro Blue Tetrazolium per ml and 0.16 mg of 5-bromo-4-chloro-3indolyl phosphate per ml in 100 mM Tris (pH 9.5)-100 mM NaCl-5 mM MgCl2. Clones producing immunoreactive antigen were identified by the presence of reacted substrate that had precipitated over the plaque. Positive plaques were picked and subjected to a second round of screening. The plaque-purified clones were prepared as phage stocks. Recombinant protein purification. Lysogens of plaquepurified lambda gtll recombinant clones were made with E. coli BNN103 (27). The lysogenized host cells were grown in liquid culture at 30°C to an A6. of 0.5, whereupon phage production was induced by temperature shift to 44°C for 15 min. The temperature was then reduced to 37°C for 1 h, and the cells were harvested by centrifugation. Cell pellets were resuspended in 10 mM Tris (pH 7.4)-2% Triton X-100-50 pLg of phenylmethylsulfonyl fluoride per ml-1% aprotinin and then quick-frozen in liquid nitrogen. The crude extracts were thawed and treated with DNase I (10 ,ug/ml) for 1 h at 37°C. The P-galactosidase fusion proteins were identified by immunoscreening of electrophoretically transferred proteins

(26).

Northern (RNA) hybridization. Liver tissue from uninfected and infected cynomolgus macaques was disrupted with a tissue homogenizer. RNA was solubilized with 5 M guanidinium thiocyanate, precipitated with 4 M LiCl, extracted with phenol-chloroform, and collected by ethanol precipitation (9). Selection of polyadenylated species was by two rounds of oligo(dT) chromatography (4). The RNAs were fractionated by electrophoresis on a 1.5% agarose gel containing 2.2 M formaldehyde (14) and transferred to Zeta Probe membranes (Bio-Rad). Northern blots were hybridized according to conditions outlined by Church and Gilbert (11). Lambda gtll cDNA clone inserts were 32p labeled by the DNA polymerase random priming procedure according to instructions of the supplier (Boehringer Mannheim). Sequence analysis. Standard procedures were used for restriction digestion and plasmid construction. Restriction enzymes and T4 DNA ligase were used according to the instructions of the supplier (BRL). Mexico strain HEV [HEV(M)] cDNA fragments were purified from agarose gels and subcloned into pBluescript II KS+ phagemid vectors (Stratagene) and M13 mpl8 and mpl9 vectors (NEB). Double- or single-stranded templates were sequenced by the dideoxynucleotide method (21) using T7 sequencing primers or M13 universal sequencing primers (Pharmacia). PCR amplification of HEV(B). The nucleotide sequences of HEV(M) clones identified by immunoscreening were compared with the sequence derived from the HEV Burma strain [HEV(B)] isolate (25). Comparable epitopes from HEV(B) were isolated by polymerase chain reaction (PCR) using the following Burma strain-derived 5' and 3' primers: to isolate the HEV(B) clone corresponding to HEV(M) 406.4-

2(M), 5'-CCGGAATTCGCCAACCCGCCCGACCAC-3' (5' primer) and 5'-GGGGAATTCCCGCGGTTAGCGGCGCG G-3' (3' primer); and to isolate the HEV(B) clone corresponding to HEV(M) 406.3-2(M), 5'-CCGGAATTCACCT TGGACTACCCTGCCCGC-3' (5' primer) and 5'-GGGAAT

5792

YARBOUGH ET AL. C OOM A S S IE

1

2

... .

.-

3

5

4

.'

w.

'.'

w _

.. ...

*_ _

. .s

3

4

7

1

6 7

1

5 6

2

3

4

5 6 7

3

4

5

::

CYNO 1 44-conv

CYNO 144-acute

CYNO 1 44-pre 2

2

:.

..

1

*'~. ~ ~.

..

F 387-C

NORMAL

1

6 7

J. VIROL.

6

7

1

2

3 4 5

2

4 5 6 7

3

..

*

....... .....

".: ''.',,,....... ..

1f''G'$

*

'~~~~~~~~~c

?

.

:." ...: :.,.:..::.:.

FIG. 1. HEV recombinant protein characterization. HEV(M) and HEV(B) lambda gtll recombinants, 406.4-2(M), 406.3-2(M), 406.4-2(B), and 406.3-2(B), were lysogenized in E. coli BNN103 as described in Materials and Methods. Lysogenized cells were grown under conditions that favor high production of the P-galactosidase fusion proteins. Crude extracts containing the recombinant antigens were prepared as described in Materials and Methods. The fusion proteins were resolved on SDS-10% polyacrylamide gels, transferred to nitrocellulose, and immunoscreened with human and cynomolgus monkey sera. Lanes 1 to 4 contained crude extracts of E. coli BNN103 infected with 406.3-2(B) (lanes 1), 406.4-2(B) (lanes 2), 406.4-2(M) (lanes 3), and 406.3-2(M) (lanes 4). Lanes 5 contained crude extracts of nonlysogenized E. coli BNN103. Lanes 6 contained 1 ,ug of purified P-galactosidase; lanes 7 contained molecular weight markers (1 ,ug each). Indicated is a Coomassie blue stain of the protein gel; the other panels are Western blots of identically prepared gels that were tested with various human sera (normal and F387-C) and cynomolgus macaque sera (as noted above each panel).

TCTCGAGTTTTTTTTTTTTTTTTTTTT-3' (3' primer). HEV(B) cynomolgus monkey liver cDNA clones containing the regions homologous to the identified HEV(M) epitopes were mixed with the appropriate primers and then subjected to 30 cycles of PCR amplification using Taq polymerase (20). These primers were modified to contain EcoRI restriction sites to facilitate in-phase subcloning. The PCR-amplified material was digested with proteinase K and cleaved with EcoRI, and the resulting DNA fragments were gel purified before ligation into lambda gtll. Insert-containing phage were identified by hybridization, and the production of HEV(B)-specific antigen was verified by immunoscreening as

described above. RESULTS

Isolation and characterization of HEV type-common epitopes. A portion of the HEV genome was first identified by differential hybridization of a cDNA library made from the bile of a Burma strain-infected cynomolgus macaque (19). The recognition of consensus sequences for the RNAdirected RNA polymerase had established the clone as derived from the nonstructural portion of the viral genome. Immunoscreening is an alternative approach to identification of clones from regions of the genome likely to encode structural genes. A human stool-derived lambda gtll library was therefore constructed from a stool sample collected during the Telixtac, Mexico, ET-NANBH outbreak (7).

Previous attempts at cloning HEV from stool samples had met with little success because of the low numbers of viral particles excreted in feces. In order to ensure that the cDNA library was representative of all extracted nucleic acid species present in feces, it would be necessary to expand the starting cDNA to facilitate the recovery of an HEV-specific clone. This was accomplished by the application of sequence-independent single-primer amplification to permit the nonselective enzymatic amplification of all cDNAs present in the initial starting material (18, 19). The resulting high-titer library was screened for HEVspecific antigen-producing clones by using a different convalescent human serum sample (F387-C) collected during the same Telixtac, Mexico, outbreak. Two positive immunoreactive plaques were isolated out of a total of 70,000 phage screened. These lambda gtll clones, 406.4-2(M) and 406.32(M), were plaque purified and lysogenized in E. coli BNN103, and ,-galactosidase recombinant fusion proteins were induced (see Materials and Methods). The molecular weights of the expressed 1-galactosidase fusion proteins for 406.4-2(M) (-120,000) and 406.3-2(M) (-123,000) indicated that they both had relatively small inserts compared with the nonrecombinant fusion protein (Fig. 1; see below). Crude, unpurified lysates from these lysogens were resolved by polyacrylamide gel electrophoresis, blotted onto nitrocellulose, and subsequently tested with human and cynomolgus monkey sera to confirm that the observed plaque reactivity was specific for the induced fusion protein (Fig. 1). The

VOL. 65, 1991

HEV: IDENTIFICATION OF TYPE-COMMON EPITOPES

5793

TABLE 1. HEV recombinant protein immunoreactivity with human seraa Serum

Source

Reactionc with:

Stageb 406.4-2(M)

FVH-21 FVH-8 B-IgG SOM-19 SOM-20 IM-35 IM-36 PAK-1 FFI-4 FFI-125 F387-C Normal

Burma Burma Burma Somalia Somalia Borneo Borneo Pakistan Mexico Mexico Mexico United States

Acute Acute Convalescent Acute Acute Acute Acute Acute Acute Acute Convalescent

406.3-2(M)

406.4-2(B)

406.3-2(B)

+ + + + + + + +

NT NT + + +

NT NT + + + NT NT NT + + +

+ + + + + + + + +

NT NT NT + +

+

a Two HEV(M) clones, 406.4-2(M) and 406.3-2(M), were isolated from lambda gtll expression libraries by immunoscreening with HEV-positive human serum F387-C. Two HEV(B) clones, 406.3-2(B) and 406.4-2(B), were generated by PCR amplification of an HEV(B) cDNA and cloned into lambda gtll expression vectors. Recombinant phage were plaque purified and first antibody immunoscreened with human sera collected during different ET-NANBH outbreaks. The reactions of the various sera with the indicated recombinant antigens were detected by alkaline-phosphatase conjugated anti-human immunoglobulin G after reaction with Nitro Blue Tetrazolium and 5-bromo4-chloro-3-indolyl phosphate in alkaline phosphatase buffer (see Materials and Methods). b Acute-phase sera were collected between 1 and 12 days after the onset of HEV-related symptoms. Convalescent-phase sera were collected between 30 and 90 days after the onset of jaundice. c +, reaction; -, no reaction; NT, not tested. All sera had no reaction with lambda gtll.

specificity of the convalescent human serum (F387-C) for 406.4-2(M) was confirmed in the Western blot (immunoblot) format. The 406.3-2(M) epitope was much less reactive and barely visible in this experiment. The specificity of both proteins for HEV-infected individuals and animals was, however, confirmed by the acute seroconversion of a cynomolgus macaque (no. 144) infected with HEV(M) (Fig. 1). The weak reactivity of the 406.3-2(M) protein in the denatured Western blot format (Fig. 1) suggested that the native protein conformation associated with the lambda plaque screening assay might provide better sensitivity in detecting specific anti-HEV antibody. The reactivity of the 406.4-2(M) and 406.3-2(M) recombinant clones was therefore tested with 12 human and 12 cynomolgus monkey serum samples in a plaque assay format. As shown in Table 1, both recombinant clones immunoreacted specifically with acuteand convalescent-phase human sera collected during outbreaks in different geographic locations (7). The 406.4-2(M)

clone immunoreacted with 9 of 11 and the 406.3-2(M) clone immunoreacted with 8 of 11 serum samples derived from HEV-infected individuals. Although the absolute number of serum samples tested was small, it is important to note that both 406.3-2(M) and 406.4-2(M) reacted with sera from five of five different epidemics. This screening strongly suggested that the identified epitopes were broadly reactive and represented HEV type-common epitopes. Both Mexico strain clones were also recognized by antibodies from acute- and convalescent-phase sera obtained from experimentally infected cynomolgus macaques (Table 2). As noted above, the 406.4-2(M) clone was recognized by sera from HEV-infected individuals from all five geographic locations; however, only sera from the HEV(M)-infected cynomolgus monkeys were found to be reactive. This contrasted with the cross-strain reactivity of the 406.3-2(M) clone to acute- and convalescent-phase sera from both Burma and Mexico strain-infected cynomolgus macaques.

TABLE 2. HEV recombinant protein immunoreactivity with cynomolgus macaque sera' Serum

Source

Stageb

C-113

Mexico

Preimmune Acute Convalescent Preimmune Acute Convalescent Preimmune Acute Convalescent Preimmune Acute Convalescent

C-144 C-130 C-30

Mexico Burma

Pakistan

Reactionc with: 406.4-2(B)

406.3-2(B)

406.4-2(M)

406.3-2(M)

+ +

+ +

+ +

+ +

+ +

+ +

-

-

+ +

+ +

+ +

-

-

-

a The Mexico and Burma clones were immunoscreened as described in footnote a of Table 1. HEV-infected primate sera were prepared after serial passage of HEV to cynomolgus macaques by intravenous inoculation of a 10o stool suspension of human stool samples infected with HEV. b Acute-phase sera were collected during peak alanine aminotransferase activities. Convalescent-phase sera were collected after the return of alanine aminotransferase activity to preinoculation levels. Control sera were preimmune sera from the corresponding HEV-infected cynomolgus monkey. c +, reaction; -, no reaction. All sera showed no reaction with lambda gtll.

5794

J. VIROL.

YARBOUGH ET AL.

The reactivity of the latter clone was therefore similar to that seen in humans. Sera from the Pakistan strain-infected animal did not react with either epitope-encoding clone, suggesting either strain divergence or individual cynomolgus monkey variability in its immune response to the Pakistan strain-derived inoculum. These observations suggested that any cross-species (i.e., human versus cynomolgus macaque) immunogenic response to HEV may be sensitive to the homology of sequences within the protein-encoding region of the HEV variant (see below). Epitope analysis of HEV(B). The fusion proteins derived from the HEV(M) clones were shown to encode epitopes specifically recognized by sera from patients with HEV infections (Fig. 1). Furthermore, the fusion proteins were recognized by both acute- and convalescent-phase sera from cynomolgus monkeys experimentally infected with the HEV(M) agent. In order to firmly establish that these regions of the virus genome encoded type-common epitopes, we decided to investigate whether the equivalent regions of the HEV(B) genome were as reactive and immunogenic as their

HEV(M) counterparts. A comparison of the sequence of the HEV(M) clones (see below) with that of the HEV(B) genome identified regions that were homologous to 406.4-2(M) and 406.3-2(M). The HEV(B) epitope equivalents were isolated by PCR amplification of corresponding sequences from HEV(B) cDNA clones, and their respective origins are indicated by the letter B following the clone designation. After subcloning into lambda gtll, insert-containing phage were immunoscreened for HEV antigen-producing clones (see Materials and Methods). Recombinant protein from these two Burma clones was induced from E. coli lysogens, and their reactivity was assessed in crude lysates (Fig. 1). The pattern of reactivity by Western blot was somewhat different from that seen before. The human convalescent antiserum F387-C was reactive with both clones; however, only 406.3-2(B) reacted with the cynomolgus macaque no. 144 serum sample. This observed variation might again be attributable to denaturation of the epitope and some dependence on the conformational nature of the encoded epitope; however, substantial sequence variation was found between 406.4-2(M) and its Burma equivalent, 406.4-2(B) (see below). Immunoreactivity for these Burma strain-derived clones was found to be identical to that for the original Mexico strain clones when they were tested with selected samples from the panel of sera from HEV-infected individuals (Table 1). These Burma strain clones were also recognized by antisera from cynomolgus macaques infected with the

HEV(B) agent (Table 2). The observation that the Burma 406.4-2(B) was recognized only by HEV(B)-infected cynomolgus monkeys indicates, however, that the 406.4-2 region of HEV(M) may bear some species specificity in its immunogenicity and in cynomolgus monkeys might react as a type-specific epitope. The Mexico clone 406.3-2(M) and the corresponding Burma clone, 406.3-2(B), immunoreacted with sera from both HEV(B)- and HEV(M)-infected animals. These data suggested that although both regions encoded broadly reactive epitopes when analyzed with human sera, the degree of sequence conservation affected the extent to which they were recognized in animals that had been challenged with the heterologous strain (see below). HEV(M) clone 406.4-2(M) identifies a novel open reading frame (ORF). The nucleotide sequences of clones 406.4-2(M) and 406.3-2(M) are shown in Fig. 2A. Comparison of these sequences with those in the GenBank data base (version 64) clone

indicated that these viral sequences were novel. As noted

406. 4-2 (M) 1 GAA TTC GCG GCC GCT CGC GCC AAC CAG CCC GGC CAC TTG GCT Q f a a a r A e N G L A P N 43 CCA CTT GGC GAG ATC AGG CCC AGC GCC CCT CCG CTG CCT CCC B G I R L S A L P P P P P P 85 GTC GCC GAC CTG CCA CAG CCG GGG CTG CGG CGC TGA CGG CTG R D L A G L R R L V P Q P Z 127 TGG CGC CTG CCC ATG ACA CCT CAC CCG TCC CGG ACG TTG ATT R L T H R T L I W P M P P S 169 CTC GCG GTG CAA TTC TAC GCC GCC AGT ATA ATT TGT CTA CTT Q F L A L A A I I L V Y S C

211 CAC CCC TGA CAT CCT CTG TGG CCT CTG GCA CTA ATT TAG TCC H L H L A L I P Z P W P Z S 253 TGT ATG CAG CCC CCC TTA ATC CGC CTC TGC CGC TGC AGG ACG Q

L

I

R

M

P

P

84 126 168 210

252 294

L

R C R T C 295 GTA CTA ATA CTC ACA TTA TGG CCA CAG AGG CCT CCA ATT ATG L I L T L M Q R I V W P P P C

42

336

337 CAC AGT ACC GGG TTG CCC GCG CTA CTA TCC GTT ACC GGC CCC L T G L H A L T G S P S V P

378

379 TAG TGC CTA ATG CAG TTG GAG GCT ATG CTA TAT CCA TTT CTT

420

L Q L E L A L F C M M Y P 421 TCT GGC CTC AAA CAA CCA CAA CCC CTA CAT CTG TTG ACA TGA G L K L H L L T Q Q S P P Z 463 ATT CCA GCT GAG CGC CGG TGC T 484 A E R I R P C

Z

462

406.3-2(M) 1 CAG GAA TTC GCC GGG GCG CGG GAT ACT TTT GAT TAT CCG GGG e f a g a r d T F G D Y P q

42

43 CGG GCG CAC ACA TTT GAT GAC TTC TGC CCT GAA TGC CGC GCT A R H T F D D F R C A C P E

84

85 TTA GGC CTC CAG GGT TGT GCT TTC CAG TCA ACT GTC GCT GAG G L L Q G C A F Q T A S V E

126

127 CTC CAG CGC CTT AAA GTT AAG GTT GGG TCA CTG ACT TCG GGC L Q R L s t s K V K V g 1 g

168

A 169 GTT ACT GAC TCC CAT GGT GTG ACA AGG AAT TCG ATA TC T V D S H G V T R N i S

406.4-2(M) 406.4-2(B)

206

20 30 10 ANQPGHLAPLGEIRPSAPPLPPVADLPQPGLRRZ

ANPPDHSAPLGVTRPSAPPLPHVVDLPQLGPRRZ 30 20 10

10

20

30

40

406.3-2(M) TFDYPGRAHTFDDFCPECRALGLQGCAFQSTVAELQRLKVKV

B 406.3-2(B)

TLDYPARAHTFDDFCPECRPLGLQGCAFQSTVAELQRLKMKV 10

20

30

40

FIG. 2. Sequence analysis of HEV epitope-encoding clones. (A) Nucleotide and amino acid sequences of 406.4-2(M) and 406.3-2(M). Amino acids in lowercase letters are derived from linker, vector, or random sequences. The epitope in 406.4-2(M) is believed to terminate at the UGA codon (Z) that marks the end of ORF3 (25). (B) Amino acid sequence alignments for the two different epitopeencoding regions from the Burma and Mexico strains of HEV. The ORF for 406.3-2(M) is 42 amino acids and has 90.5% identity with a region at the extreme 3' end of HEV(B) ORF2. The ORF for clone 406.4-2(M) contains 33 amino acids that have 78% identity to a region of HEV(B) ORF3.

above, HEV(M) sequences were also compared with the full-length sequence of HEV(B) (25) to isolate the equivalent regions. The nucleotide sequence comparison indicated 78.5% similarity for the 406.4-2 region and 79.8% similarity for the 406.3-2 epitope region. The amino acid similarity for 406.4-2(M) was only 73.5% (over 34 amino acids). This contrasted with the 90.5% identity of clone 406.3-2(M) with the equivalent HEV(B) sequence in a 42-amino-acid overlap (Fig. 2B). This sequence divergence in the respective epitope regions could account for the observed variation noted in the antigen profiling with cynomolgus macaque sera (see above). An independent cDNA isolation experiment indicated that

VOL. 65,

1991

HEV: IDENTIFICATION OF TYPE-COMMON EPITOPES

45 bp at the 3' end of clone 406.3-2(M) were not authentic HEV sequence (13b). This was confirmed by the lack of any sequence homology between the 3' end of the 406.3-2(M) clone and the full-length HEV(B) cDNA sequence (25). Sequencing of an independent HEV(M) clone overlapping the 3' end of 406.3-2(M) demonstrated the extraneous nature of these sequences (data not shown). The non-HEY nucleotide sequences and their translation are presented in lowercase letters in Fig. 2A. One principal advantage of the blind or random screening approach to epitope isolation is the identification of immunodominant epitopes on the basis of their reactivity with sera obtained from naturally infected, immunocompetent individuals (or experimental animals). One other advantage is the direct experimental verification of the open reading frames that are utilized by the virus for gene expression as a result of their 5'-end in-frame fusion with the bacterial P-galactosidase. The two independent HEV(M) epitope clones were separated by approximately -1,500 bp of sequence at the 3' end of the virus on the basis of their sequence comparison with HEV(B). The 406.3-2 epitope region mapped to the extreme 3' end of the genome in the second major ORF (ORF2) (25). It was surprising, however, to discover that the 406.4-2 epitope region mapped to ORF3. ORF3 overlaps ORFi at its 5' end (by 1 bp) and ORF2 at its 3' end, encompassing only 123 amino acids from the first methionine residue to its termination codon. It is apparent that the full-length sequence of the 406.4-2(M) cDNA insert does not encode protein; the first termination codon for ORF3 can be seen at nucleotide residue 120, with other termination codons (Z) present throughout the sequence (Fig. 2A). The epitope rescue of a comparable Burma strain sequence was directed to the initial 33 amino acids found in frame with ,-galactosidase. The utilization of ORF3 would not have been predicted were it not for its random selection by antibody screening. HEV cDNA clones identify subgenomic RNAs. It was previously determined that the full-length genome of HEV was a polyadenylated, positive-sense RNA of -7.5 kb (17, 19). In order to definitively confirm the viral origin of the cDNA epitope clones, they were used as hybridization probes against Northern blots of total and polyadenylated fractions of RNAs extracted from HEV(M)- and HEV(B)-infected as well as uninfected cynomolgus macaque livers. These same infected specimens were previously shown to be positive for HEV-specific transcripts (19). As shown in Fig. 3 (lane M), polyadenylated transcripts of -7.5, -3.7, and -2.0 kb were detected only in liver RNAs from HEV(M)-infected cynomolgus monkeys by using clone 406.4-2(M) as a probe. Identical hybridization results were obtained by using 406.32(M) as a probe (data not shown). The detection of subgenomic transcripts of -3.7 and -2.0 kb was surprising in that only the full-length genomic transcript of 7.5 kb was previously detected in these same samples (19). However, the probe utilized was located in what is considered the nonstructural region of the genome and 5' to that encoding the epitope clones (19). Under the stringent conditions utilized in this hybridization analysis, the HEV(M)-derived probe [78.5% homologous to the HEV(B) epitope] did not detect comparable transcripts in the Burma strain-infected animal (Fig. 3, lane B). Subgenomic transcripts were confirmed in HEV(B)-infected liver by hybridization using a cDNA clone that overlapped with the sequence represented by the 406.42(M) clone (25). The extreme 3' genomic position of the 406.3-2 epitope clone strongly suggested that these polyade-

5795

N M B kb

9.57.44.42.4 -

4

-4

1.3-

FIG. 3. Hybridization of poly(A)+ mRNAs from uninfected and infected cynomolgus monkey livers with cDNA clone 406.4-2(M). Lane N contained approximately 2 ±g of uninfected cynomolgus macaque liver poly(A)+ mRNA, lane M contained approximately 2 ,g of Mexico strain-infected cynomolgus macaque liver poly(A)+ mRNA, and lane B contained approximately 2 ,ug of Burma straininfected cynomolgus macaque liver poly(A)+ mRNA. The arrows indicate the positions of HEV-specific RNA transcripts at -7.5, -3.7, and -2.0 kb. Total RNAs were isolated from uninfected and HEV(M)- and HEV(B)-infected cynomolgus monkey livers. Polyadenylated mRNAs were selected by oligo(dT) column chromatography and fractionated on 1.5% formaldehyde agarose gels. After transfer of RNA to Zeta Probe, the Northern blot was prehybridized at 42°C in 7% sodium dodecyl sulfate-1% bovine serum albumin-1 mM EDTA-50 mM sodium phosphate (pH 7.2). The filter was hybridized with 2 x 106 cpm of 32P-labelled EcoRI fragment from cDNA clone 406.4-2(M). Hybridization proceeded overnight, and the filter was washed with 2x SSC (lx SSC is 0.15 M NaCl plus 0.015 M sodium citrate) at 42°C for 30 min, lx SSC at 50°C for 30 min, and 1 x SSC at 60°C for 30 min. The filters were air dried and exposed to X-ray film for 8 days at -70°C.

nylated subgenomic messages were coterminal with the 3' end of the viral genome. DISCUSSION HEV was first cloned by differential hybridization of cDNA libraries constructed from the bile of a cynomolgus macaque infected with HEV(B) (19). Subsequent molecular characterization led to the recognition that the clone was derived from the major causative agent of ET-NANBH worldwide (7, 16). The molecular epidemiology was in agreement with the previous studies on virus particle aggregation by heterologous sera that suggested that a single predominant agent was responsible for ET-NANBH (7, 8). The isolation of apparent immunodominant epitopes from two divergent strains and their cross-reactivity with both human and cynomolgus monkey sera further confirms the etiologic association of the isolated sequences with a unique virus responsible for ET-NANBH. The immunoscreening approach to clone isolation may be considered more than complementary to the plus/minus hybridization screening methods reported in initial cloning

5796

YARBOUGH ET AL.

efforts (19). The isolated cDNA clones are certainly derived from immunogenic viral proteins and may represent epitopes from the structural region of the virus. The two cDNA clones, 406.4-2(M) and 406.3-2(M), were identified directly from human stool by using human sera and therefore represent not only the initial isolation of HEV(M) strain clones but also the expression of recombinant epitopes recognized in the naturally infected host. The epitopes specifically reacted with human acute- and convalescent-phase sera and were also recognized by sera from experimentally infected cynomolgus macaques. Assays for the determination of HEV infection are presently unavailable, but their feasibility has been demonstrated here. Our interest has been in the molecular characterization of HEV, recognizing that the characterization of virus-encoded protein products may be useful in the diagnosis and treatment of what is believed to be the cause of the majority of epidemic and sporadic viral hepatitis cases seen in developing countries. Molecular analysis of the recently cloned viral genome for HEV(B) indicates that HEV is different from other RNA viruses known to cause hepatitis. The ORF analysis of HEV(B) (25) suggests a genomic organization different from that of the picornavirus HAV (12) or the flavivirus- or pestiviruslike virus, hepatitis C virus (10). The. blind immunoscreening approach described in this paper succeeded in isolating immunodominant epitopes of HEV and has aided in the elucidation of the expression strategy employed by the virus. The utilization of ORF3 was not predicted, but its expression in the infected liver was confirmed by the numerous sera that reacted with its product (406.4-2). The exact means by which this small ORF3 reading frame is expressed is not known but presumably involves the subgenomic transcripts that are coterminal with the 3' end of the virus. The Northern analysis of RNA isolated from HEV(M) indicated that there were two subgenomic, polyadenylated viral mRNAs of 2.0 and 3.7 kb. It is possible that the structural gene(s) of the virus is expressed by these subgenomic transcripts, as seen with the Alphavirus genus (e.g., Sindbis virus [22]). Although the genomic organization of HEV shows some similarity to alphaviruses, sequences which have been found to be highly conserved between alphaviruses have not been found in HEV (13a). Final classification of HEV will depend on the biochemical characterization of the protein products encoded by this virus and the genomic organization and expression strategy of its structural and nonstructural proteins. The understanding of the complete immune response to HEV infection will require delineation of the humoral and cellular immune responses to this agent. We have preliminary indications that the antibody response to these particular epitopes may have utility in the diagnosis of ETNANBH by confirming past and present exposure to HEV. Further studies will be required in order to determine whether these same epitopes elicit protective immunity or are involved in the cellular immune response to virus infection. ACKNOWLEDGMENTS

We gratefully acknowledge Randolph Allen Moeckli for countless discussions and suggestions about the immunoscreening. We thank Matt Smith for the isolation of cynomolgus macaque liver RNAs and Andrew Miller for the sequencing of the HEV(M) clones. We appreciate the assistance of John Fernandez and the Genelabs Visual Arts Department.

J. VIROL.

REFERENCES 1. Alter, H. J., R. H. Purcell, P. V. Holland, and H. Popper. 1978. Transmissible agent in non-A, non-B hepatitis. Lancet i:459463. 2. Andjaparidze, A. G., M. S. Balayan, A. P. Savinov, D. M. Braginsky, V. F. Poleschuk, and N. A. Zamyatina. 1986. Fecalorally transmitted non-A, non-B hepatitis induced in monkeys. Vopr. Virusol. 1:73-80. 3. Arankalle, V. A., J. Ticehurst, M. A. Sreenivasan, A. Z. Kapikian, H. Popper, K. M. Pavri, and R. H. Purcell. 1988. Etiological association of a virus-like particle with entericallytransmitted non-A, non-B hepatitis. Lancet i:550-554. 4. Aviv, H., and P. Leder. 1972. Purification of biologically active globin messenger RNA by chromatography on oligo-thymidylic acid-cellulose. Proc. Natl. Acad. Sci. USA 69:1408. 5. Balayan, M. S., A. G. Andzhaparidze, S. S. Savinskaya, E. S. Ketiladze, D. M. Braginsky, A. P. Savinov, and V. F. Poleschuk. 1983. Evidence for a virus in non-A, non-B hepatitis transmitted via the fecal oral route. Intervirology 20:23-31. 6. Bradley, D. W. 1990. Enterically-transmitted non-A, non-B hepatitis. Br. Med. Bull. 46:442-461. 7. Bradley, D. W., A. Andjaparidze, E. H. Cook, Jr., K. McCaustland, M. Balayan, H. Stetler, 0. Velasquez, B. Robertson, C. Humphrey, M. Kane, and I. Weisfuse. 1988. Aetiological agent of enterically-transmitted non-A, non-B hepatitis. J. Gen. Virol. 69:731-738. 8. Bradley, D. W., K. Krawczynski, E. H. Cook, K. A. McCaustland, C. D. Humphrey, J. E. Spelbring, H. Myint, and J. E. Maynard. 1987. Enterically transmitted non-A, non-B hepatitis: serial passage of disease in cynomolgus macaques and tamarins and recovery of disease-associated 27-34 nm virus-like particles. Proc. Natl. Acad. Sci. USA 84:6277-6281. 9. Cathala, G., J. F. Savouret, B. Mendez, B. L. West, M. Karin, J. A. Martial, and J. D. Baxter. 1983. A method for isolation of intact translationally active ribonucleic acid. DNA 2(4):329-335. 10. Choo, Q.-L., G. Kuo, A. J. Weiner, L. R. Overby, D. W. Bradley, and M. Houghton. 1989. Isolation of a cDNA clone derived from a blood-borne non-A, non-B viral hepatitis genome. Science 244:359-362. 11. Church, G. M., and W. Gilbert. 1984. Genomic sequencing. Proc. Natl. Acad. Sci. USA 81:1991-1995. 12. Cohen, J. 1., J. R. Ticehurst, R. H. Purcell, A. Buckler-White, and B. Baroudy. 1987. Complete nucleotide sequence of wildtype hepatitis A virus: comparison with different strains of hepatitis A virus and other picornaviruses. J. Virol. 61:50-59. 13. Feinstone, S., A. Z. Kapikian, R. H. Purcell, H. J. Alter, and P. V. Holland. 1975. Transfusion-associated hepatitis not due to viral hepatitis type A or B. N. Engl. J. Med. 292:767-770. 13a.Fry, K. E. Unpublished observation. 13b.Huang, C.-C. Unpublished data. 14. Maniatis, T., E. F. Fritsch, and J. Sambrook. 1982. Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. 15. Prince, A. M., G. F. Grady, and C. Hazzi. 1974. Long-incubation posttransfusion hepatitis without serological evidence of exposure to hepatitis B virus. Lancet ii:241. 16. Purcell, R. H., and J. R. Ticehurst. 1988. Enterically transmitted non-A, non-B hepatitis: epidemiology and clinical characteristics, p. 131-137. In A. J. Zuckerman (ed.), Viral hepatitis and liver disease. Alan R. Liss, Inc., New York. 17. Reyes, G. R., C.-C. Huang, P. 0. Yarbough, L. M. Young, A. W. Tam, R. A. Moeckli, K. Y. Yun, D. Nguyen, M. M. Smith, J. Fernandez, M. E. Guerra, J. P. Kim, K.-C. Luk, K. Krawczynski, M. A. Purdy, K. E. Fry, and D. W. Bradley. 1991. Hepatitis E virus (HEV): epitope mapping and detection of strain variation, p. 237-245. In T. Shikata, R. H. Purcell, and T. Uchida (ed.), Viral hepatitis C, D, and E. Elsevier Publishers BV, The Hague, The Netherlands. 18. Reyes, G. R., and J. P. Kim. Sequence-independent, singleprimer amplification (SISPA) of complex DNA populations. Mol. Cell. Probes, in press. 19. Reyes, G. R., M. A. Purdy, J. P. Kim, K.-C. Luk, L. M. Young, K. E. Fry, and D. Bradley. 1990. Isolation of a cDNA from the

VOL. 65, 1991

20.

21. 22.

23.

HEV: IDENTIFICATION OF TYPE-COMMON EPITOPES

virus responsible for enterically-transmitted non-A, non-B hepatitis. Science 247:1335-1339. Saiki, R. K., D. H. Gelfand, S. Stoffel, S. J. Scharf, R. Higuchi, G. T. Horn, K. B. Mullis, and H. A. Erlich. 1988. Primerdirected enzymatic amplification of DNA with a thermostable DNA polymerase. Science 239:487-491. Sanger, F., S. Nicklen, and A. R. Coulson. 1977. DNA sequencing with chain-terminating inhibitors. Proc. Natl. Acad. Sci. USA 74:5463-5467. Strauss, J. H., E. G. Strauss, C. S. Hahn, and C. M. Rice. 1987. The genomes of alphaviruses and flaviruses: organization and translation, p. 75-105. In D. J. Rowlands, M. A. Mayo, and B. W. J. Mahy (ed.), The molecular biology of the positive strand RNA viruses. Academic Press, Inc., New York. Tabor, E. 1985. The three viruses of non-A, non-B hepatitis. Lancet i:743-745.

5797

24. Tabor, E., R. J. Gerety, J. A. Drucker, L. B. Seeff, J. H. Hoofnagle, D. R. Jackson, M. April, L. F. Barker, and G. Pineda-Tamondong. 1978. Transmission of non-A, non-B hepatitis from man to chimpanzee. Lancet i:463-466. 25. Tam, A. W., M. M. Smith, M. E. Guerra, C.-C. Huang, D. W. Bradley, K. E. Fry, and G. R. Reyes. Hepatitis E virus (E): molecular cloning and sequencing of the full-length viral genome. Virology, in press. 26. Towbin, H., T. Staehelin, and J. Gordon. 1979. Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc. Natl. Acad. Sci. USA 76:4350-4354. 27. Young, R. A., and R. W. Davis. 1983. Efficient isolation of genes by using antibody probes. Proc. Natl. Acad. Sci. USA 80:11941196.