VIROLOGY

176,604-6

19 (1990)

Discrimination

of Hepatitis B Virus (HBV) Subtypes Using Monoclonal PreSl and PreS2 Domains of the Viral Envelope

LARRY T. MIMMS,’

Antibodies

to the

MARCO FLOREANI, JOAN TYNER, ERIC WHIl-TERS, ROBERT ROSENLOF, LARRY WRAY, ANDREW GOETZE, VIREN SARIN, AND KIM EBLE Hepatitis/AIDS

R&D, Abbott Laboratories,

Received December

Abbott Park, Illinois 60064

18, 1989; accepted February

1, 1990

We report the production and characterization of murine anti-PreS2 and anti-PreSl monoclonal antibodies (mAb) and demonstrate their utility in discriminating hepatitis B virus (HBV) subtypes. On the basis of Western blotting and reciprocal competition binding to HBV virions, at least five distinct epitopes have been identified in the PreS domain: two within the PreSl region and three within the PreS2 region. All PreSP mAb bind M protein (gp33 and gp36) but only one group binds strongly to M and L proteins (p39 and gp42). This group determinant was mapped to peptide residues 120-145. The second group bound to an endoglycosidase F-sensitive epitope which is defined by a mannose-rich glycan at ASN 123 in the PreSP region. The third group was mapped to peptide residues 150-l 74 and was reactive with the M envelope proteins but not L or S proteins on Western blots. All PreSl mAb bind L protein but not M protein on Western blots. Using these mAb, HBV subtype assays were developed allowing evaluation of the Paris (1975) HBsAg subtype panel members along with other HBsAg-positive specimens. All Paris subtype members (except ayw, and ayw,) could be easily distinguished by differential PreS2 mAb reactivity. The Paris subtypes, adw,, adw,, and adr, could be classified as distinct groups by PreS2 and PreSl mAb binding. Specimens from Hong Kong and the United States classified as adw, in the S region fell into two groups based on PreS2 mAb binding: one having reactivity similar to Paris adw, subtype and the other having identical reactivity to Paris ayw, subtype. Furthermore, some specimens classified as adr in the S region gave similar reactivity to the Paris ayr subtype in the PreSP and PreSl regions. One complicating factor in this approach toward subtyping was the discovery that some HBsAg positive sera may contain factors which block PreS epitopes. Grouping of HBV subtypes by PreSl, PreS2, and S mAb reactivity may allow better o 1990Academic PWSS, IN. correlation with groupings based on HBV DNA sequence homology.

morphological forms, 22-nm spherical particles and filaments of 22 nm diameter and variable length, lack capsid or DNA and are produced in huge excess over HBV virions. The protein compositions of these morphological forms differ considerably. Up to 20% of total envelope protein in HBV virions is L protein compared to only about 1% of total protein in 22-nm spherical particles (Heerman et a/., 1984; Stibbe and Gerlich, 1983). Some investigators have hypothesized that the L protein may play an essential role in the secretion and morphogenesis of complete virions (Ou and Rutter, 1987) and in viral pathogenesis (Hess eta/., 1987). Furthermore, the PreSl region of the L protein contains a specific receptor for the human hepatocyte (Neurath et al., 1986; Pontisso et a/., 1989). Specific binding of the PreS2 region to polymerized human serum albumin has also been proposed as a possible mechanism for association with the hepatocyte plasma membrane (lmai et al., 1979; Pontisso et a/., 1989). Regardless of the role of PreS-containing proteins in viral structure or function, PreSl and PreS2 Ag have been shown to be highly immunogenic in mice and their presence together with S may potentiate the immune response to S (Milich et a/., 1985, 1986).

INTRODUCTION The hepatitis B viral (HBV) envelope is composed of three proteins and their glycosylated forms which are derived from one continuous open reading frame (ORF), divided into the S, PreS2, and PreSl regions (Heermann et a/., 1984). The S gene starts at the third or fourth start codon of this ORF depending on the HBV subtype and codes for the major HBV envelope protein (S protein), ~24, and its glycosylated form gp27 (Fig. 1). The PreS region contains two additional initiation sites in frame with the S gene allowing the expression of two larger hepatitis B surface antigen molecules: M protein (55 amino acids of PreS2 + S) and L protein (108 or 1 19 amino acids of PreSl + PreS2 + S). M protein consists of singly and doubly glycosylated forms, gp33 and gp36, respectively. L protein has an unglycosylated and a singly glycosylated form, p39 and gp42, respectively. In the blood of HBV-infected persons only a very small portion of the total hepatitis B surface antigen exists as complete virions or Dane particles. Two other ’ To whom reprint requests should be addressed. 0042.6822/90

$3.00

Copyright 0 1990 by Academic Press, Inc. All rights of reproduction I” any form reserved

604

mAb TO HEPATITIS

HBV subtypes have been distinguished previously by differential reactivity to monoclonal and monospecific polyclonal antibodies against the S gene product (BenPorath el al., 1985; Courouce el a/., 1976, 1982; Usuda et a/., 1986). The S gene product of HBV envelope carries the common group-specific determinant a which is conformation dependent. In addition, the S gene contains one member of each of the two pairs of mutually exclusive subtype determinants d/y and w/r. These four major subtypes of HBAg, adw, adr, ayw, and ayr, were proposed to represent the phenotypic expression of four major HBV genotypes (Courouce et al., 1982). These subtypes differed in their epidemiological distribution and were used in tracking the route of HBV infection among populations and among individuals in cases of accidental and perinatal transmission (Courouce Pauty et a/., 1983). Okamoto et al. (1988) have concluded that an HBV subtype classification scheme based on anti-S reactivity does not reflect true genotypic variation and have suggested instead that HBV strains fall into four groups based on DNA sequence homology: group A (adw), group B (adw), group C (adw, adr, ayr), and group D (ayw). They had shown previously that d toy and w to rsubtypic changes could be induced by an A to G point mutation at nucleotides 365 and 479 in the S gene, respectively (Okamoto et a/., 1987). Furthermore most or all of the S antigenic determinants probably lie within a stretch of only 50 amino acids (Howard et a/., 1988). Anti-PreSl and PreS2 mAb produced previously have not allowed discrimination of HBV subtypes (Budkowska et al., 1986; Gerken et al., 1987; Neurath et a/., 1987; Okamoto et al., 1985) and it is unknown which changes in DNA sequence among HBV strains results in phenotypic (antigenic) variation. We have developed a large collection of mAb against the HBV envelope proteins and have identified at least five distinct epitopes within the PreS domain; three within the PreS2 and two within the PreSl region. Many of these mAb show unique HBV subtype reactivity and have been used to evaluate the Paris HBsAg subtype panel and HBV specimens from the United States and Hong Kong. Our results demonstrate that PreS mAb can differentiate all Paris subtype members from each other except ayw, and ayw, and can detect phenotypic variability arnong HBsAg subtypes not previously recognized by S mAb alone. MATERIALS

AND METHODS

Materials CNBr-activated Sepharose CL-4B, Sepharose S300, and all other chromatographic gels were from Pharmacia (Piscataway, NJ). Electrophoresis and

B PreSl

AND PreS2

605

Western blot reagents were obtained from Bio-Rad, (Richmond, CA). Horseradish peroxidase (HRP) was from Toyoba (NY). Na1251was obtained from Amersham (Arlington Heights, IL). The nine-member HBsAg subtype panel was from Courouce (from the International Workshop of HBsAg Subtypes, Paris, April 1975). All glycosidases were obtained from Sigma: neuraminidase (Type X, Clostridium perfringens), P-galactosidase (Jack bean), /3-N-acetylglucosaminidase (Jack bean), endoglycosidase H (Strepfomyces griseus), and endoglycosidase F (Flavobacterium meningosepticum). Hepatitis B markers were determined using commercially available kits from Abbott Laboratories (Abbott Park, IL): AUSRIA or Monoclonal Auszyme for HBsAg and AUSAB for anti-HBs detection. Human plasma (positive for HBsAg) was obtained from Pyramid Biologicals (Van Nuys, CA). Dane particle purification Human plasma with a high HBsAg titer was clarified by centrifugation at 650 g in a GSA rotor for 20 min. The supernatant was spun for 8 hr at 132,000 g in a SW28 rotor. The resulting pellet was resuspended in 0.01 M Tris, 1 mn/r EDTA, pH 7.6 (Buffer A), and this resuspension mixture was layered on 10% sucrose in buffer A which had been overlayered on a 2-ml 62% sucrose/Buffer A pad. After centrifugation for 8 hr at 132,000 g in a SW28 rotor, fractions were collected and assayed for HBcAg. Peak fractions of HBcAg found at the 620/,/l Oslosucrose interface were pooled, diluted 1:4 in Buffer A, overlayed on a continuous 1562% sucrose gradient, and spun for 24 hr at 132,000 g. Fractions were collected and assayed for HBsAg and evaluated for the presence of S, M, and L proteins by Western blot. mAb production Female 8-week-old BALB/c mice (Dominion Labs) were immunized with 10 pg of purified Dane particles [ad or ay subtypes emulsified in adjuvant (MPl and TDM R-700, RIBI lmmonochem Research, Inc.)] three to fourtimes at 3-week intervals. The RIBI adjuvant was selected because of ease of use and because it gave lower animal toxicity. Serum samples were drawn 2 weeks after the last boost and reactivity with Dane particles was evaluated by microtiter EIA and by Western blotting. Responding mice were rested 2-4 months then given a prefusion i.v. boost of 10 rg Dane particles in phosphate-buffered saline (PBS). Three days later the splenocytes were fused 1: 1 with the SP2/0 myeloma line using standard protocols with slight modifications (Koehler and Milstein, 1975). The fusion pellet was dispersed with 1 ml 50% polyethylene glycol

606

MIMMS

ET AL.

Schematic Representation of HBV Envelope Protein

t

I

108-119AA

9C

1 I

1

1 ’ 55AA

I ’

P39

Gm

I

GP36

Gm

II

GP33

GP42

L-protein 226AA Gc

I M-protein I

$2

vGP27

S-protein -

P24

FIG. 1. Schematic representation of HBV envelope proteins. The three envelope glycoproteins are the product of a single open reading frame and arise from different translation initiations at each of three in frame AUG codons. An additional upstream initiation codon exists in some HBV subtypes (adw,, adr, and ayr) producing a PreSl gene product 11 amino acids longer than other subtypes. Glycosylation sites are indicated by G, (high mannose glycan) in the PreS2 region and G, (complex glycan) in the S region.

(American Type Culture Collection, MW 1450) and centrifuged in 20 ml media, and cells were resuspended in HAT (hypoxanthine-aminopterin-thymidine)-selective IMDM (Iscove’s modification of Dulbecco’s MEM) to be plated at 1.5 X 1OE5 lymphocytes per well in 96well tissue culture plates. To promote hybrid survival, STM mitogen (RIBI) was added into the initial plating media at a 1 to 200 dilution. The mitogen was not used in subsequent media changes and feedings. Hybridoma supernatants which were found positive on the anti-Dane immunoassay were evaluated by Western blotting to identify specific banding patterns. EIA and Western blot positive hybrids were cloned by limiting dilution. Established clones were grown up in T flasks and injected into pristane-primed BALB/c mice (1 x lOE7 cells/mouse) to generate mAb-containing ascites. Antibodies from IgG-cloned lines were purified from mouse ascites fluid using an Affi-Gel Protein A MAPS II kit (Bio-Rad). Radioiodination

and conjugation

MAb protein concentration in milligrams per milliliter was determined from the A280 divided by the extinction coefficient, 1.38. Purified mAb were radioiodinated using a chloramine T method (Greenwood and Hunter, 1963) to a specific activity of 20-30 $CiIpg. Free lz51 was separated from iodinated protein by passage of the reaction mixture over a Sephadex G-50 column. MAb were conjugated to horseradish peroxidase (HRP) using the method of Nakane and Kawaoi (1974).

Immunoassay

to detect HBcAg

Neat specimens or those diluted into human plasma nonreactive for HBsAg, anti-HBs, and anti-HBc were incubated with Auszyme II beads for 2 hr at 40”. Beads were washed and then incubated with 200 ~1 of PBS containing 5% Tween 20 for 30 min at 40”. Beads were washed and further incubated with 200 ~1 of HRP-conjugated monoclonal anti-HBc (Abbott Laboratories) for 2 hr at 40”. Beads were washed a third time then incubated with OPD substrate solution [0.3 ml of 0.39/o ophenylene diamine-2-HCI in 0.1 lvI citrate-phosphate buffer (pH 5.5) containing 0.029/o H202]. The enzymatic reaction was allowed to proceed for 30 min at room temperature then stopped with the addition of 1 ml of 1 N H,SO, and absorbance at 492 nm was measured. Immunoassay

for the detection

of anti-Dane Ab

To coat beads (i inch diameter) or microtiter wells, HBV virions were first denatured in 2% 2-mercaptoethanol and 1% SDS, and boiled for 3 min. Denatured virions were diluted to 2 pg/ml in PBS and then coated onto beads or wells by an overnight incubation at room temperature. Beads or wells were washed with distilled water and blocked for 30 min with 200 ~1of 3% bovine serum albumin (BSA) in PBS. Beads or wells were washed again and allowed to air-dry. In a typical assay, the specimen was incubated 1 hr at room temperature in the wells, the incubation wells were washed, and HRP-conjugated goat anti-mouse lg (Kirkegaard and

mAb TO HEPATITIS

607

AND PreS2

to appropriate wells for a 1 hr incubation at room temperature. After a second washing, 50 PI/well of a 1: 1000 dilution of the kit isotype-specific conjugate were added to each sample for a 30-min incubation. After the final wash, chromagen was added as described above.

Perry Laboratories) was added and incubated for 30 min. After the final wash, 100 ~1 of an OPD substrate solution (Abbott Laboratories) was added to each well. The color reaction was stopped by addition of 1 N sulfuric acid. Competitive

El PreSl

binding to Dane particles HBV subtyping

One hundred microliters of [“51]labeled anti-PreS mAb was incubated with 100 ~1 of negative human plasma (NHP) containing varying concentrations of purified unlabeled mAb for 30 min at 20”. A polystyrene bead coated with native HBV virions was added and incubated with this mixture for 2 hr at 40” or 18-22 hr at 20”. The bead was washed with distilled water and counted for radioactivity.

Specimens containing HBsAg were diluted into recalcified human plasma nonreactive for HBsAg and anti-HBs, as measured by AUSRIA and AUSAB, respectively. Diluted specimen (0.2 ml) was then incubated for 18 hr at room temperature or 2 hr at 40” with Auszyme II beads (Abbott Laboratories). After washing with water, the beads were incubated with 0.2 ml of ‘251-labeled or HRP-conjugated mAb for 2 hr at 40”. The beads were washed with water and counted for radioactivity in a gamma counter for the RIA or incubated with OPD substrate solution for the EIA.

lsotype determination MAb isotype was determined with the SBA Clonotyping System III kit (Southern Biotechnology Associates, Inc.) with slight modifications. EIA 96-well microplates were coated overnight at room temperature with 100 PI/well of a 1 :lOOO dilution of goat anti-mouse IgG + M (H + L) (Kirkegaard and Perry Laboratories) as described for Dane particles above. Plates were washed and 50 ~1 of clone supernatants were added PreS2:

PreSl:

peptides

Synthetic peptides were produced using an ABI solid phase peptide synthesizer (Applied Biosystems). The following peptide sequences were made and then purified by high pressure liquid chromatography: (120-145)

adw,subtype

(150-l 74) ayw subtype.

GTNLSVPNPLGFFPDHQLDPAFGANSNNPDWDFNPVKDD

Peptide inhibition Synthetic peptides coding for Pre-Sl residues 1351 and Pre-S2 residues 120-l 45 and 150-l 74 were used to capture PreS antibody in inhibition assays. Microtiter EIA plates were coated with 2 pglml denatured HBV virions in PBS overnight at room temperature. Prior to the addition of the sample to the EIA plates, antibodies at 10 pg/ml were preincubated with 10 pg of either the 13-51, 120-l 45, or 150-l 74 peptide for 20 min. Fifty microliters of sample was added per well, incubated for 20 min, and washed. A 1: 1000 dilution of HRP-conjugated goat anti-mouse lg (Kirkegaard and Perry Laboratories) was added at 100 PI/well, incubated 30 min and washed, and color was developed with the OPD substrate. Polyacrylamide

Synthetic

MQWNSTAFHQTLQDPRVRGLYLPAGG TVNPVLTTASPLSSIFSRIGDPALN

gel electrophoresis/Western

blots

Analytical polyacrylamide gel electrophoresis was performed with a Bio-Rad slab gel apparatus with 12Ob running gel and 4% stacking gel. The gel and buffer

assays

(13-51) adw,subtype.

formulations were those of Laemlli (1970). A specimen was boiled for 5 min in 1O/O SDS, 2.5% 2-mercaptoethanol in Tris buffer prior to addition to the gel. Western blotting was conducted essentially as described by Towbin and Gordon (1984). After transfer of protein to nitrocellulose, the nitrocellulose strips or sheets were soaked in blocking buffer (1% bovine hemoglobin, 0.10/o, v/v, Tween 20 in PBS). Nitrocellulose strips were incubated with mAb to be tested at 0.5 to 1 pglml in blocking buffer for l-2 hr, then washed with PBS, and incubated with HRP-conjugated goat anti-mouse IgG or IgM (0.5-l .O pg/ml). Strips were washed then developed with 4-chloro-1 -naphthol/hydrogen peroxide substrate. Other reagents Recombinant PreSl (rPreSl)and recombinant PreS2 (rPreS2) were gifts from A. Sarthy (Abbott Laboratories). Recombinant PreSl is a fusion protein whose amino terminal portion is CMP-KDO synthetase (CKS)

608

MIMMS

ET AL

HBsAg -

PreS2Ag

-40

Fraction

i I

number

FIG. 2. Purification of Dane particles by sucrose gradient centrifugation. The final step of Dane purification was centrifugation gradient (1562%). Centrifugation conditions and assay procedures are described under Materials and Methods.

and carboxy terminus contains PreSl residues 12120. The rPreS2 is a fusion protein whose amino terminal portion is CKS and the carboxy terminus contains PreS2 residues 123-l 74. rPreS1 and rPreS2 were subtype ayw and adw, respectively. Two mAb, 18/7, reactive against PreSl region (Heermann eta/., 1984), and Q19/10, reactive against PreS2 region of M protein (Gerken eta/., 1987), were kind gifts from W. Gerlich. RESULTS mAb production To develop mAb against multiple HBV envelope determinants, purified Dane particles (subtype ad and ay) were chosen as mouse immunogens. These Dane particles were enriched in L and M envelope protein (1 O30% total protein) compared to HBsAg preparations which generally contained less than 10% M protein and 1o/oL protein. Dane particles were separated from 22-nm spherical particles by centrifugation on a 1562% continuous sucrose gradient. As shown in Fig. 2, Dane-containing fractions gave high HBcAg and PreSl Ag reactivity compared to the bulk of HBsAg activity found in lower density fractions which contained 22nm HBsAg particles. Whereas previous immunization with purified HBsAg particles (largely 22 nm diameter) resulted primarily in

on a sucrose

production of mAb against S but not against PreS2 and PreSl (Peterson et a/., 1984), all 45 Dane immunized mice developed detectable immune response against L and M proteins as measured by Western blot and reactivity with SDS-denatured Dane particles coated onto a solid phase. Some mice were given a final boost with synthetic PreS2 peptide (120-l 45) prior to splenectomy. Four of five mice immunized exclusively with synthetic peptides administered either as free peptide in adjuvant or coupled to carrier proteins (key hole limpet hemocyanin (KLH) or BSA) produced low anti-Dane titers and were not used for hybridoma production. Characterization

of anti-PreSl

mAb

Purified mAb were characterized by reactivity with purified Dane particles ad and ay subtype on Western blots, binding to synthetic peptides and recombinant PreSl proteins, and reciprocal competition studies. A summary of these results obtained with PreSl mAb is shown in Table 1. As observed by Western blotting (Fig. 3A), all of the epitopes defined by these PreSl mAb were retained after treatment of Dane particles with heat, SDS, and reductant. Therefore, their determinants may be linear epitopes or alternatively are conformational ones which can reform after transfer of protein to nitrocellulose. No cysteines are present in either PreSl or

mAb TO HEPATITIS

B PreSl

609

AND PreS2

TABLE 1 CHARACTERIZATIONOF Pre-Sl mAb Western blot reactivity Group

lsotype

13-5 1 Peptide inhibition

Reactivity with rPre-Sl

ad

ay

116-9 116-11 116-29 116-80 113-39

IgGl IgG2a IgGl IgGl IgGl

+ + + + +

+ + + + +

+ + + + +

+ + + + +

115-16 116-72 116-109 116-192 116-86

IgGl IgGl IgGl IgGl IgGl

-

+ + + + +

+ + + + +

+ + + + +

Clone no.:’

a lmmunogen for 1 13. 1 15, and 1 16 series of mAb was purified Dane particles (subtype ad).

PreS2 regions so that proper reforming of disulfide bonds within these regions is not necessary to retain native conformation. This is in distinct contrast to most antigenic epitopes of S (~24 and gp27) which are highly dependent on disulfide bond formation and proper assembly into particles. As shown in Fig. 3A, all PreSl mAb reacted exclusively with L protein (p39, gp42) and not with M (gp33, gp36) or S (~24, gp27) protein. These mAb also reacted with rPreS1 expressed in Escherichia co/i (data not shown). PreSl mAb defined at least two distinct epitopes based on reciprocal competition assays in which mAb competed for bihding to Dane particle coated beads (Figs. 4A and 48). Group 1 mAb, but not group 2, specifically bound synthetic peptide 13-51. The group 1 mAb also competed with mAb 18/7 which has been shown to react exclusively with L protein in Western blots (Heermann et al., 1984). Further characterization

of the epitopes recognized by these mAb is presented in another paper in this issue (Kuroki et a/., 1990).

Characterization of PreS2 mAb The PreS2 mAb could be divided into at least three groups based on reciprocal competition studies (Tables 2 and 3) and Western blot patterns (Fig. 3B). The PreS2 group 1 mAb showed reactivity with a linear epitope on both M and L proteins by Western blotting (Fig. 3B), a synthetic peptide (120-l 45), and a rPreS2 fusion protein containing residues 123-l 74. Group 2 and 3 PreS2 mAb strongly bound only the M protein by Western blotting, but showed no reciprocal competition upon binding to immobilized Dane particles (Table 3). The PreS2 group 2 mAb were subdivided into groups 2a and 2b based on reciprocal competition studies. The group 2a mAb (116-34) conipletely blocked the

GP42,

GP42 GP39 GP33lGP36

GP42 GP39 GP36 GP33

GP27 P24

GP27 P24 ABCDEFG

H

I

J

GP39c GP99GP33c GP27P24-

ABCDEFGH

I

J

ABCDEF

ABCDEF

FIG. 3. Western blot reactivity of S, PreS2, and PreSl mAb to purified HBV virions. (A) Western blotting was performed as described under Materials and Methods using purified HBV virions of subtype ay (left) and ad (right). Different mAb were used for detection in each lane as follows: Anti-S H 166 (lane A), anti-PreS2 50-80 (lane B), lanes C-J were anti-PreSl-1 16-9 (lane C), 1 16-l 1 (lane D), 116-29 (lane E), 116-80 (lane F), 1 15-l 6 (lane G), 1 16-72 (lane H), 1 16-l 92 (lane I), and 1 16-86 (lane J). (B) The following mAb were used for detection: Anti-S H 166 (lane A), lanes B-F were anti-PreS2-50-80 (lane B), 25-l 9 (lane C), 1 16-34 (lane D), 1 15-32 (lane E), and 128-603 (lane F).

610

16000 -

A

16000 14000 z

12000 -

2 Q I

-

116-29

+

11516

10000 8000 6000 4000 2000 -

1

[Competing

10

100

antibody]

nglml

-

[Competing

antibody]

'Ly 10000

1000

116-29

rig/ml

FIG. 4. Reciprocal competition of anti-PreSl mAb for binding to HBV virion coated beads was performed as described under Materials Methods. Competition of specified unlabeled mAb with ‘251-labeled 1 16-29 (group 1) (A) and ‘251-labeled 115-l 6 (group 2) (B) are shown.

binding of group 2b mAb (115-32) but group 2b mAb showed no ability to block group 2a mAb binding to Dane particle coated beads even at 120 pg/ml which represented a 2000-fold excess of competing mAb. This differential inhibition did not result from differences in mAb affinity for Dane particles because some group 2a mAb (e.g., 116-94) had equivalent or lower affinities than 1 15-32 as estimated using the method of Van Heyningen (1986) (data not shown). Group 2a or 2b mAb also differed in their reactivity to a synthetic

and

peptide (120-l 45) and in their reactivity to L protein on Western blots (Fig. 3B). Exclusive binding of group 2a mAb to M protein may be explained either by a dependence upon a free PreS2 amino terminus or upon the carbohydrate moiety at asparagine 123. To determine the role of this glycan moiety in mAb binding, Dane particles were treated with a variety of glycosidases. Treatment with neuraminidase alone, neuraminidase + ,&galactosidase, or neuraminidase + P-galac-

mAb TO HEPATITIS

B PreSl

611

AND PreS2

TABLE 2 CHARACTERIZATIONOF

Clone no.a

lsotype

120-145 Peptide inhibition

1

50-80 116-183 55-392 25-19 128-410

IgG2a IgGl IgGl IgGl lgG2a

+ + + + +

-

2(a)

116-34 116-94

IgGl IgGl

-

nd

IgGl

+I-

Group

(b)

150-174 Peptide inhibition

-

128-603

IgG2b

Western blot reactivity Reactivity with rPre-S2

ad gp33/36/39/42 gp33/36/39/42 gp33/36/39/42 gp33/36/39/42 gp33/36/39/42

w gp33/36/39/42 gp33/36/39/42 gp33/36/39/42 gp33/36/39/42 gp33/36/39/42

gp33l36 gp33/36

gp33l36 gp33l36

-

gp33l36 Weak gp39l42

gp33/36/39/42

+

gp33l36

gp33l36

115-32 3

Pre-S2 mAb

Note. nd, not done. B Primary immunogen was SDS-denatured Dane particles (subtype ad + ay) for the 50 and 55 series, native Dane particles (subtype ad) for the 1 16 and 1 15 series, and native Dane particles (subtype ay) for the 128 series.

tosidase + /3-/I/-acetylglucosaminidase had little or no significant effect on group 2a reactivity (Fig. 5A) or any other PreS2 mAb reactivity. Endoglycosidase F treatment completely abolished group 2a mAb (1 16-34) reactivity, but had no effect on group 1 mAb binding (Fig. 56). The commercial endoglycosidase F contains two glycosidase activities: endo F and peptide-l\l-glycosidase F (PNGase F: E.C. 3.5.1.52). Endo F, like endoglycosidase H (E.C. 3.2.1.96), cleaves the oligosaccharide chain between the di-N-acetylchitobiose moiety of some asparagine-linked glycans, primarily those of the high mannose type. PNGase F hydrolyzes at the glyco-

l-ABLE 3 RECIPROCALCOMPETITIONOF PreS2 mAb FOR BINDINGTO DANE PARTICLES Radioiodlnated

PreS2 mAb

50-80

25519

128-410

116-34

115-32

128-603

++ ++ ++ -

++ ++ ++ -

++ ++ ++ -

+ -

+ -

-

-

++ -

++ ++ -

-

Competing clonea 50-80 (57) 25-19 (113) 128-410(40) 116-34 (45) 115-32 (30) 128-603(25)

++

a Final competing mAb concentration in pglml is given in parentheses. >95% competition is indicated by ++. 1O-40% inhibition is indicated by + and ~10% inhibition is indicated as -. Anti-PreS2 competition studies were conducted as described under Materials and Methods.

sylamine linkage and generates a carbohydrate-free peptide chain (Maley et al., 1989). Since the endoglycosidase H had little or no effect on 116-34 binding to Danes, it is likely that the PNGase F activity in the commercial endoglycosidase F produced the cleavage resulting in loss of mAb binding. No attempt was made to distinguish which activity or whether the combination of enzyme activities was necessary to remove carbohydrate. The carbohydrate sequence of this PreS2 bound glycan has not been determined. To evaluate the effect of endoglycosidase F treatment on the apparent molecular weight of M proteins after SDS-PAGE, treated and untreated Dane particles were analyzed by Western blot (Fig. 6). The M protein bands of untreated Dane particles were strongly reactive for all PreS2 mAb. After treatment, no reactivity to Dane proteins was observed with group 2a mAb; however, bands at 28,30,35, and 40 kDa were reactive for group 1 mAb. The 28- and 30-kDa bands likely correspond to deglycosylated M protein. The appearance of higher molecular weight bands was unexpected, but may represent anomalous mobility of deglycosylated or aggregated M protein. These data indicated that the presence of bound glycan was essential for Group 2a mAb reactivity with PreS2, but not for group 1. Up to 90% of group 2b PreS2 mAb reactivity could be destroyed by endoglycosidase F treatment, but even after extended incubation of Dane particles some residual binding to deglycosylated M protein remained (Fig. 5B). The reactivity of 1 15-32 with deglycosylated M protein was confirmed by Western blot (Fig. 6). Q19/10, an anti-PreS2

612

MIMMS

ET AL.

180000 160000

-E-

Endcglycosidase

F

b

Endcglycosidase

ii

b

Neuramlnidase(NA)

-t

B-Galactosidase(Gal)+NA

-

N-AceFllglucosamlntdase+Gal+NA

4

2

8

6

Time

(hours)

0.8

-c-

50-80

--A-

116-183

-

116-34

+

115-32

0 r:.g.,.,.,.:., 4

8

12

Incubation

16

time

20

24

28

(hours)

FIG. 5. Effect of glycosidase treatment of HBV virions on their reactivity with anti-PreS2 mAb. (A) Dane particles at 2.2 mg/ml in 500 pl PBS were incubated with 10 ~1 of each glycosidase individually or together. According to the manufacturer (Sigma Chemical Co.), the enzymatic activity in each of the glycosidases was as follows: neuraminidase-50 U/ml. endoglycosidase H-O.5 U/ml, fi-N-acetylglucosaminidase-50 U/ml, @galactosidase-6.8 U/ml, and endoglycosidase F-5 U/ml. Reaction mixtures were incubated at 37” and aliquots were removed at indicated times, diluted 1OOO-fold into negative human plasma, and placed at 4’ to stop the reaction. These specimens were assayed for antiPreS2 mAb binding by first incubating with Auszyme II beads for 18 hr at 20”. Beads were washed, then incubated with ‘*51-labeled 1 16-34

mAb TO HEPATITIS

tp36bP33*

-ga -P z8 q19/10

116-34

115-32

50-80

FIG. 6. Western blot of endoglycosidase F-treated Dane particles. Dane particles purified from human plasma were incubated at 37” in the presence (+) or absence (-) of endoglycosidase F for 8 hr as described in the legend to Fig. 5. Western blotting of treated and untreated Danes were developed with four different anti-PreS2 mAb as indicated and as described under Materials and Methods. Apparent molecular weights were estimated from mobility of prestained markers.

mAb which reacts strongly with M protein (Gerken et a/., 1987) also bound weakly to the deglycosylated M protein (Fig. 6). The group 3 PreS2 mAb (128-603) did not reciprocally compete with any other PreS2 mAb and was strongly reactive with M protein but unreactive with L protein of Dane particles (ad and ay) on Western blots (Fig. 3B). This mAb bound to a synthetic peptide (150174) subtype ayw (Table 2), but showed no reactivity with the 120-l 45 (adwz sequence) peptide, with rHBsAg containing only the S gene produced in mouse L cells (Mimms et al., 1989) or with rPreS2 produced in E. co/i which contained residues 123-l 70 (sequence adw). This lack of 128-603 binding to the rPreS2 protein may be explained by its unique HBV subtype specificity, as described below. mAb competition

binding

PreSl mAb did not significantly inhibit any PreS2 mAb binding to Dane particles; however, PreS2 mAb, 50-80, at greater than 20 pg/ml showed partial inhibition of group 1 and 2 PreSl mAb binding to Danes (2134%). All other PreS2 mAb showed negligible (~10%) inhibition of PreSl binding. Interestingly 50-80 at very high concentration (>50 pglml) was able to compete 30-389/o of group 2 PreS2 mAb binding. H 166, an antiS mAb, gave 259/o inhibition of group 1 PreSl binding but virtually no inhibition for group 2 PreSl mAb. These data suggest a close proximity of the S, PreS2, and PreSl epitopes on the surface of Dane particles. mAb reactivity (Paris, 1975)

with HBsAg subtype panel

Significant differences in the HBsAg subtype panel (Paris) reactivity were observed among mAb even

B PreSl

AND PreS2

613

within the same competition group. Binding of S, PreS2, and PreSl mAb to the Paris panel members is shown in Table 4. Pairwise comparisons of S, PreS2, or PreSl mAb were performed to try to normalize binding within each region (Table 5). We chose not to normalize binding to HBsAg concentration determined by commercial immunoassay, because these assays measure S antigen concentration and the relative amounts of M and L compared to S may vary significantly among individuals. Using binding ratios of anti-S mAb, ad subtypes could be easily distinguished from ay subtypes, adw, distinguished from adr and adw,, and ayw, differentiated from other aywsubtypes. All Paris subtype members except ayw, and ayw, could be differentiated from one another on the basis of PreS2 mAb reactivity (Table 6). In group 1 PreS2 mAb, 50-80 and 116-l 83 gave little or no reactivity with subtypes ayr, adw,, or adr; however, their binding to ayw, subtype differed significantly. 50-80 binds strongly to ayw, but 116-l 83 shows virtually no binding to this subtype (Tables 4-6). Another group 1 mAb, 128-410, is similar to 116183 in reactivity for all subtypes except adw,. In fact 128-410 binds with high affinity only to ayw,_, subtypes. A fourth group 1 mAb, 25-l 9, binds to all Paris subtype members, but showed a significantly lower affinity for subtypes ayw, and adw,. In contrast to the group 1 mAb binding, group 2a mAb (1 16-34) shows high affinity binding to all Paris subtypes. Comparison of group 2a mAb (116-34) and group 2b mAb (115-32) showed that the ayr subtype was significantly more reactive to group 2a, although both 115-32 and 116-34 binding is dependent upon glycan at asparagine 123. The group 3 PreS2 mAb 128-603, gave subtype reactivity unlike any other PreS2 mAb. This mAb does not show significant binding to subtypes adw,, adw,, or ayw, and has low affinity for ayw, compared to ayw,, ayw,, ayr, and adr (Tables 4 and 6). Since the PreSl region contains no bound carbohydrate, all mapped epitopes are defined by amino acid sequence alone. The most striking subtype difference is observed between two group 2 PreSl mAb, 116-86 and 115-16. 115-16 binds weakly to ayw,, ayr, and adw, compared to 116-86 (Tables 4 and 5). The simplest explanation for this reactivity difference would be one amino acid change in this epitope common to all three subtypes. Another interesting comparison can be made between the reactivity of PreSl group 2 (116-

mAb for 2 hr, and washed a second time and counted for radioactivity. (B) Dane particles were incubated at 37” with 10 ~1of endoglycosidase F (50 U/ml) and treated Dane particles were assayed for reactivity with various ‘251-labeled PreS2 mAb as described above. Binding ratios were calculated as cpm of endoglylcosidase F-treated Dane particles/cpm of untreated Dane particles at each time point for each PreS2 mAb.

MIMMS ET AL

86) and group 1 (116-80) mAb. The ratio of 116-86 to 116-80 binding is greater for the a&v, subtype than the ratio for all other subtypes (Table 5). Two explanations exist: the group 1 epitope which is nearer to the amino terminus of PreSl region than the group 2 epitope has been preferentially lost or destroyed (e.g., by proteolysis) or a&v, has a unique sequence in the group 1 PreSl region. Several PreSl mAb from different hybridomas show little or only subtle changes in reactivity with Paris HBsAg subtypes, suggesting that the PreSl group 1 and group 2 epitopes are relatively conserved. In addition to the eight major subtypes in Table 6, three other subtypes were identified in Paris (1975). One designated ayw; was indistinguishable from ayw, and ayw, on the basis of S and PreS mAb binding (Table 5). Another subtype, designated a&-, showed almost identical reactivity to adr [alternatively designated as adrq+ by Courouce-Pauty et al. (1983)] (Table 5). Subtle differences between these subtypes were previously based on differential reactivity with polyclonal or affinity adsorbed polyclonal antisera. The Paris subtype designated adyw was from a patient in which the ad and ay subtypes occur on separate particles (Paul et al., 1986). S and PreS mAb reactivity was consistent with a mixture of adwand ayw particles, but no attempt was made to separate ad from ay particles which would have allowed detailed characterization of the individual subtypes. mAb reactivity

to HBsAg-positive

specimens

To determine the utility of these mAb for subtyping, we tested 50 HBsAg-positive specimens from Hong Kong and Los Angeles along with the Paris panel. Most specimens showed reactivity indistinguishable either from adw, or aywzm3 subtypes; however, two specimens from Hong Kong, HK7 and HKl 1, classified as adw, on the basis of S mAb binding were identical in reactivity to the Paris ayw, subtype in the PreS2 region (Table 7). Surprisingly, HK7 and HKl 1 also reacted strongly with PreSl mAb 1 15-l 6 unlike either adw, or ayw, Paris subtype members, but similar to aywzm4, adw,, and adr. Two other interesting specimens, HK3 and HK19, classified as adrladw, based on S mAb binding, gave reactivity similar to Paris ayr subtype with PreS 1 and PreS2 mAb (Table 7). These data suggest that ayr and adr subtypes may be closely related and that the rare ayr subtype could have resulted from one amino acid change in the S region of the more prevalent adr subtype. None of the 50 specimens tested gave reactivity similar to Paris ayw, or ayr with all of S, PreS2, and PreS 1 mAb. Two specimens with high concentrations of S subtype adw, (LA3 1 and LA34) showed virtually no reactiv-

mAb TO HEPATITIS

B PreSl

615

AND PreS2

TABLE 5 DIFFERENTIALmAb REACTIVITYWITH PARISSUBTYPEPANEL EXPRESSEDAS SELECTEDmAb BINDING RATIOS Parissubtypes Ratiosof mAb binding SmAb PreS2 mAb

PreSl mAb

H95/HlO 50-80/116-183 25-19/50-880 50-80/116-34 116-34/25-19 116-34/128-410 116-343/115--32 116-34/128-603 116-80/116--86 116-86/115-16

ayw

ww3

avwi

0.0 2.1 0.8 1.4 0.9 1.9 0.9 0.5 0.2 0.9

0.0 2.2 0.9 1.4 0.8 1.8 1.0 0.6 0.4 0.9

0.0 1.8 1.1 1.5 0.6 3.4 0.7 0.8 0.5 0.9

am

ww

adw,

adr

0.0 2.8 0.8 1.4 0.9 2.1 1.1 42.5 0.6 0.8

0.2 748.3 0.1 1.5 11.6 781.1 1.1 788.7 0.8 50.0

564.4 2.6 0.6 1.9 0.9 2.5 1.3 101.9 0.7 48.1

19.0 1.6 393.7 0.0 1.4 1949.5 1.6 0.8 0.4 0.9

adrq-

adw,

adyw

25.5 1.4 9.6 0.0 20.7 328.6 0.7 238.1 0.1 1.2

0.4 2.2 0.8 1.5 0.8

w

25.2 1.6 381.9 0.0 1.1 1159.0 1.2 1.4 0.4 0.7

0.0 0.8 284.4 0.0 1.0 398.4 23.4 0.8 0.5 64.1

0.8 0.4 1.4

Note. To normalize binding, S/N value of S or PreS mAb binding to Paris subtype specimen was divided by the S/N value of another mAb binding within the same region. Values in boldface are significantly different from those of ayw, subtype. S/N values were determined as described under Materials and Methods.

itywith any PreSl or PreS2 mAb (Table 8). These specimens contained a factor which could block the reactivity of exogenously added Dane particles. Furthermore, when these specimens were subjected to SDS-PAGE and Western blotted, M and L proteins could be detected. This blocking activity does not bind to a Protein A affinity column and its nature is unknown.

have been precisely mapped to residues 27-35 and 72-78 by deletion analysis as described in another paper in this issue by Kuroki et a/. (1990). The PreS2 epitopes have not been precisely mapped; however, one group of PreS2 mAb (group 1) binds to a peptide spanning residues 120-l 45. Although these group 1 PreS2 mAb gave reciprocal competition, they may bind overlapping but not necessarily identical epitopes, since they react differently with various subtypes. Detailed mapping of the epitopes represented by the group 1 PreS2 mAb will be determined using shorter synthetic peptides. Group 1 PreS2 mAb vary considerably in their ability to bind different HBsAg subtypes (Table 4) indicating significant amino acid heterogeneity in 120-l 45 region. Pairwise comparison of mAb binding in this re-

DISCUSSION mAb were generated which defined at least two distinct epitopes in the PreSl region and three distinct epitopes within the PreS2 region of the HBV envelope as demonstrated by reciprocal competition studies and Western blotting of HBV virions. The PreSl epitopes

TABLE 6 SUMMARY OF RELATIVE REACTIVW

OF PreS2 AND PreSl mAb WITH THE PARIS SUBTYPEPANEL (1975)

Group

Cloneno.

ayw2

ww3

ayw

awl

adw,

adr

PreS2mAba la lb

50-80 116-183 25-19 128-410 116-34 115-32 128-603

++ ++ ++ ++ ++ ++ ++

++ ++ ++ ++ ++ ++ ++

++ ++ ++ ++ ++ ++ +

++ -

-

-

-

++ -

++ -

+ -

++ ++ -

++ ++ ++ + ++ ++ -

++ ++ ++

++ + ++

++ ++ -

116-80 116-86 115-16

+ ++ ++

++ ++ ++

++ ++ ++

++ ++ -

++ ++ -

++ ++ ++

++ ++ -

++ ++

2a 2b 3 PreS1mAbb 1 2

'++.S/N b++.SIN

> 100.+,5

Discrimination of hepatitis B virus (HBV) subtypes using monoclonal antibodies to the PreS1 and PreS2 domains of the viral envelope.

We report the production and characterization of murine anti-PreS2 and anti-PreS1 monoclonal antibodies (mAb) and demonstrate their utility in discrim...
2MB Sizes 0 Downloads 0 Views