Synthetic peptides derived from the deduced amino acid sequence of the E-glycoprotein of Murray Valley encephalitis virus elicit antiviral antibody

VIROLOGY

171,49-60

Synthetic

(1989)

Peptides Derived from the Deduced Amino Acid Sequence of the E-Glycoprotein of Murray Valley Encephalitis Virus Elicit Antiviral Antibody J. T. ROEHRIG,*,’ A. R. HUNT,* A. J. JOHNSON,* AND R. A. HAWKESt

*immunochemistry Branch, Division of Vector-Borne Viral Diseases, Center for Infectious Diseases, Centers for Disease Control, Public Health Service, U.S. Department of Health and Human Services, P. 0. Box 2087, Ft. Collins, Colorado 80522, and tSchool of Microbiology, University of New South Wales, P.O. Box 1, Kensington, New South Wales, Australia, 2033 Received January 11) 1989; accepted March 9, 1989 We used computer analysis to study hydrophilicity, homology, surface accessibility, molecular flexibility, and secondary structure of the deduced amino acid sequence of the flavivirus envelope (E)-glycoprotein. Using the results, we modified the E-glycoprotein antigenic structure proposed by Nowak and Wengler (1987, Virology, 156, 127-l 37). Our model predicts considerable overlaps in the previously defined domains. We have prepared 11 synthetic peptides from the deduced amino acid sequence of the E-glycoprotein of Murray Valley encephalitis (MVE) virus and analyzed their immunogenicity. Peptides derived from the redefined Rl and R2 domains elicit antiviral antibody. Nine of these peptides are recognized by polyclonal antiviral antibodies; however, none are consistently recognized by monoclonal antibodies. Peptides derived from the Rl domain demonstrate MVE virus specificity, and 1 peptide elicited low-level virus neutralizing antibody. Spatial overlap of the domains was defined by competitive binding assays between antipeptide antisera and radioactive monoclonal antibodies. These results indicate that synthetic peptides aid in defining flavivirus antigenic structure, and may serve as possible type-specific diagnostic reagents. o iS8SAcademic PWSS. IK.

INTRODUCTION

antibody. Domain B contained four epitopes (Bl , B2, B3, and B4) which appeared to be resistant to reduction-denaturation. Later studies determined that these epitopes were stabilized by an autoforming disulfide bond, because epitope reactivity was destroyed if the reduced protein was alkylated (Heinz et a/., 1983b; Winkler et al., 1987). Some domain B epitopes elicited neutralizing antibody; however, these antibodies were not as effective as domain A antibodies at protecting animals from TBE virus challenge (Heinz et a/., 1983a). Domain C contained one epitope which was resistant to reduction denaturation. Nowak and Wengler (1987) recently analyzed disulfide bridging in the E-glycoprotein of West Nile encephalitis (WNE) virus and proposed three antigenic domains (Rl, R2, and R3). The Rl domain (amino acids 1 to 12 1) was stabilized by four disulfide bridges and contained the largest number of hydrophilic regions. The R2 domain (amino acids 186 to 333) was stabilized by two disulfide bridges which formed two loops of 99 and 32 amino acids and had no predicted hydrophilic sequences. The R3 domain (amino acids 367 to 486) had no stabilizing disulfide bridges but had one predicted hydrophilic sequence. The significance of these disulfide bridges was revealed when sequence information indicated that all E-glycoprotein cysteine residues were completely conserved in all flaviviruses (Rice et a/., 1985; Dalgarno et a/., 1986; Deubel et al.,

The surface envelope (E)-glycoprotein (50-60 kDa) is the most immunologically important flavivirus structural antigen and has been extensively studied by a number of laboratories (for reviews see Heinz, 1986; Roehrig, 1986). Monoclonal antibody (MAb) analysis of the E-glycoprotein of several flaviviruses has identified two to five antigenic domains (Heinz eta/., 1983a; Roehrig et a/., 1983; Henchal et al., 1985). Anti-E-l c MAbs elicited by both Murray Valley encephalitis (MVE) and St. Louis encephalitis (SLE) viruses are the most efficient virus neutralizing and protective antibodies (Mathews and Roehrig, 1984; Hawkes et a/., 1988). Because of these results, we hypothesized that this type-specific epitope was critical to virus infectivity. Epitopes shared by other flaviviruses elicit neutralizing and protective antibodies; however, they are not as potent as type-specific anti-E-l antibodies. Similar results have been determined for other flaviviruses (Henchal et al,, 1985; Cammack and Gould, 1986; Kimura-Kuroda and Yasui, 1986). Studies with tick-borne encephalitis (TBE) virus identified three unlinked antigenic domains (Heinz et a/., 1983a). Domain A contained three epitopes (Al, A2, and A3) sensitive to reduction-denaturation. Epitope A3 elicited the most potent neutralizing and protective ’ To whom reprint requests should be addressed.

49

0042.6822/89

$3.00

CopyrIght 0 1989 by Academc Press. Inc. All rights of reproducmn in any form reserved.

50

ROEHRIG ET AL.

1986; Trent et al., 1986; Zhao et al., 1986; McAda et a/., 1987; Chu et al., 1988; Coia et al., 1988; Mandl et al., 1988). These results indicate that although the Eglycoprotein demonstrated significant sequence microheterogeneity, this microheterogeneity probably occurs on a background of conserved conformational structure. Recent attempts to correlate epitope reactivity with amino acid sequence or chemical characteristics have centered on (1) identifying MAb reactive peptide fragments from enzyme digested E-glycoprotein, (2) cloning and expressing E-glycoprotein deletion mutants, or (3) isolating and sequencing MAb neutralization (N) escape variants. Heinz et al. (1983b) identified a 9-kDa fragment from chymotrypsin- or thermolysin-digested E-glycoprotein that was stabilized by an autoforming disulfide bridge. All TBE epitopes defined in the B domain appeared to reside in this fragment. Unfortunately, the many chymotrypsin and thermolysin cleavage sites in the TBE virus amino acid sequence makes placing this peptide difficult until its amino terminus is sequenced. Deletion mutants of Japanese encephalitis (JE) virus identified a region of the R2-R3 junction (Mason et a/., 1987) in which antigenicity appears to depend on retaining the disulfide bridge between amino acids 304 and 335. A similar analysis of dengue (DEN) virus identified an antigenic domain identical to that of JE virus and one additional domain from amino acids 76 to 93 in the Rl (P. Mason, personal communication). Nescape variants of yellow fever (YF) virus prepared with YF virus-specific MAbs varied in sequence at one of three residues: 71, 72, or 125 (Lobigs et a/., 1988). These residues flank a highly conserved, hydrophobic amino acid sequence. All of these results suggest that the type-specific neutralizing flavivirus epitopes are located in the disulfide-stabilized Rl domain, consistent with characteristics of the TBE virus A domain. We report here our synthesis of peptides derived by computer modeling of the amino acid sequence of the MVE virus E-glycoprotein. Our results are consistent with previous epitope definitions, and they are the first report of the successful use of synthetic peptides to elicit antiflavivirus immunity. MATERIALS Computer

AND METHODS

analysis

The E-glycoprotein sequences used in the computer analysis were for the following viruses: DEN-l (Chu et al., 1988) DEN-2 (Deubel et a/., 1986) DEN-4 (Zhao et a/., 1986) JE (McAda et a/., 1987) Kunjin (KUN) (Coia et al., 1988) MVE (Dalgarno et a/., 1986) SLE (Trent et al., 1986) TBE (Mandl et a/., 1988) WNE (Nowak and

Wengler, 1987) and YF (l7D) (Rice et al., 1985). The sequences were aligned for maximum homology, and a consensus sequence was determined using the most prevalent amino acid for each residue. A homology number was assigned to each residue to correspond with the number of viruses sharing the amino acid of the consensus sequence. The consensus sequence was analyzed for homology (using the assigned homology values), hydrophilicity, and accessibility using a moving average with a window of six amino acids. Hydrophilicityvalues were those of Parker eta/. (1986). Accessibility of sequence was determined using the values of Janin (1979) as modified by Parker et al. (1986). Molecular flexibility was determined with the algorithm of Karplus and Schultz (1985). Secondary structure was analyzed by the method of Chou and Fasman (1978a, b,c), as modified by Rose (1978) using the PROSIS computer program (LKB, Bromma, Sweden).’ Peptide synthesis Peptides from the MVE (original strain) sequence (Dalgarno et a/., 1986) were synthesized on an automated peptide synthesizer (Model 430A, Applied Biosystems, Foster City, CA), using conventional solidphase t-BOC chemistry. Peptides were cleaved from the resins using anhydrous hydrofluoric acid in the presence of suitable scavengers (Applied Biosystems’ custom peptide synthesis group). The average coupling efficiency was 99.0% monitored with the quantitative ninhydrin test (Sarin et a/., 1981). The purity of each peptide was analyzed by reverse-phase high-performance liquid chromatography. Whenever possible, peptide sequences were adjusted to include a cysteine residue at the amino or carboxy terminus to facilitate later coupling to a carrier molecule. When no terminal cysteine was located in the sequence, one was added at the carboxy terminus. A negative control peptide was derived from the amino acid sequence of the E2 of Venezuelan equine encephalomyelitis (VEE) virus (VE2 peptide 09). Antibodies The MAbs were from a variety of sources and have been previously described (Roehrig et al., 1983; Hawkes et al., 1988). A MAb prepared against VEE virus (lA3A-9) was used as a negative control and has been described previously (Roehrig et a/., 1982). ’ Use of trade names is for identification only and does not imply endorsement by the Public Health Service or by the U.S. Department of Health and Human Services.

FLAVIVIRUS PEPTIDES

Mouse hyperimmune ascitic fluids (HIAF) were produced as previously described (Chappell et a/., 1974). The HIAF used were prepared with the following strains: SLE (MSI-7) MVE (Original or 1 1A), KUN (MRM-16) WNE (Egyptian 101) Stratford (STR, Ausc338), Alfuy (ALF, MRM-3929) JE (Nakayama), Kokobera (KOK, AusMRM32) YF (17D), and DEN-l (Hawaii). A normal ascitic fluid was included as a negative control. Antibody

assays with synthetic

peptides

Reactivity of both polyclonal and monoclonal antibodies with free peptide was determined by enzymelinked immunosorbent assay (ELISA). Wells of ImmunIon 2 microtiter plates (Dynatech, Alexandria, VA) were coated with 1 pg of peptide overnight at 4”, and the peptide was rinsed from the plate with PBS-Tween 20 (0.05% v:v) and blocked by a I-hr incubation with 3Ob BSA in PBS. The rest of the assay followed procedures described earlier(Roehrig eta/., 1982). In an alternative assay, various concentrations of peptides (50 to 0.4pg) were used to block binding of 5 pg purified MAb to virus. Briefly, MAbs were incubated with peptides overnight at 4’. The reaction mixtures were added to 96well lmmunlon 2 plates previously coated with MVE virus (1 pg/well) and blocked for 1 hr at 37“ with 3% BSA. Peptide-antibody mixtures were allowed to react with antigen for 15 min at 37”. The plates were rinsed and the amount of bound antibody was detected with goat anti-mouse alkaline phosphatase. Activity

of anti-peptide

antibody

BALB/c mice were immunized with 50 pg of free peptide on Days 0, 14, and 42. Animals were bled at 10, 21, and 56 days postimmunization and sera were tested for antiviral activity. Initially both RIBI adjuvant (Hamilton, MT, for peptides 1 through 7) and Freund’s complete adjuvant (FCA, for peptide 1, 6, and 7) were used; but because of the comparable response to both adjuvants, FCA was later used exclusively (for peptides 8, 14, 16, and 17). Freund’s incomplete adjuvant was used on Days 14 and 42 with animals that received FCA in the first inoculation. Reactivity of antipeptide antibody with both MVE virus and peptide was determined by ELISA. Assays for virus N and hemagglutination inhibition (HI) were performed as previously described (Roehrig et al., 1983). Fifty plaque-forming units were used in the N assay, and 80% endpoints are reported. HI tests were done with four to eight hemagglutination units at a pH of 6.6.

51

Antipeptide antibody assay (CBA)

competition

binding

Localization of epitopes to various peptides was analyzed by CBA. MAbs were purified from murine ascitic fluids using protein A-Sepharose column chromatography and were tritiated using 1 mCi /I-succinimidyl [2,3-3H]propionate(Kummereta/., 1981). Labeled antibody avidity (Roehrig et al., 1982) was determined by standard radioimmunoassay (RIA) using purified MVE virus coated onto lmmunlon 2 Remove-a-well strips (Dynatech Laboratories, Alexandria, VA). Virus-coated strips were blocked for 30 min with a 39/o solution of BSA in PBS prior to assay. After avidities were determined, a standard amount of MAb marker was competed for virus binding by pools of MVE virus-reactive antipeptide antibodies. Competition with antipeptides was compared to competition with positive and negative controls. Due to the low avidity of antipeptide antibodies as compared to MAbs, dilutable competition similar to that with HIAF (>20%) was considered positive. RESULTS Computer

analysis of the flavivirus

E-glycoprotein

The consensus sequence of 10 flaviviruses is shown in Fig. 1 and average homology values in Fig. 2A. Regions with less than 509/o homology (shaded areas in Fig. 2A) were candidates for type-specific domains, and regions of high homology were candidates for group-reactive domains. We found 12 homology minima (Fig. 2A), 9 of which were localized in the amino-terminal half of the glycoprotein. Minima 1 through 3 were clustered at amino acids 43 to 88. This cluster preceded a highly conserved sequence (amino acids 106 to 120). Minima 4 through 9 clustered at the carboxy-terminal side of the highly conserved region between amino acids 122 and 200. Minima 10 was located at amino acids 270 to 280,ll at 385 to 395, and 12 at 405 to 410. Homology minima 1, 2, 3, 4, 5, 8, and 12 contained regions of significant hydrophilicity (Fig. 2B); however, specific analysis indicated that homology minima 2 was hydrophobic and homology minima 11 was hydrophilic in MVE virus (data not shown). The highly conserved sequence (98 to 120) was still predicted to be hydrophobic. Predicted surface accessibility correlated with hydrophilicity (Fig. 2C). Although molecular flexibility also appeared to correlate with hydrophilicity and accessibility, some specific hydrophobic and inaccessible sequences demonstrated predicted molecular flexibility

52

ROEHRIG ET AL.

1 FNcLO”SNR0 1 .m..... 5.. 1 . . . . . T . ..-

“0L”LEOOSC . . . . . ..D.. I . . . . _. . . . . . . . . ..D..

"RIF"noGrr .G"....S.O . . . . . ..s.o . . . . . ..P.. . . . . . ..P.. .t.....r.. IRROL."CRK .I"T..TOOP I".TP.SOEE .""T..NOO. .K"EP.TODY

FzI;‘EEF ..HOSPD”“:

CONS. ““E SLE Y” KU”

JE

YF MN1 DE”2 DE”, WE

461 +63 433 459 463 462 4% 457 67 4% 458

tRL”r”HPF” ..“..R..Y. . . . . . . . ..I

STSTRN-K”L RS....l?K.. ..OO..“K.” .“R...SK.. . . . . . . . . . . .“....RK.. . . . . . . . . . . R..S..SK.. .I . . . . . . IR .-NODE--.. rclKEKP--.” . ..I.“..*. . ..I _... 1. TEKDSP--.” . .IISST.LA EWTYSV--TN R”.I.P..TX LNNSOO--4

GLLGRLLLY” GINRRORS~S . . . . . . . . . . .“....K..A . . . ..-.... -LO....... . . . . . . . . . . . . . . . . . ..R . . .._..... . . . . . . . ..R ..“....... .“.......R urn..u.x.v . ..r.“nm. IGX.I..T.L .L.S.ST.L. I.I.VIIT.1 .“.S.ST.L. I.I.F.“..I .T.S.“T.nn L...“R.R.L .L.“.wT”.

~NLREVRKYC . . ..L..“.. .E..T..... R...D..S.. R......S.. “0 . . . . . 5.. ORP.....“. ..P.“L..L. KOP..L.... KE”.LL.T.. ENP.W.E..

H SESHD”YsTP .T........ .T......E. "....----K "....".F.. ..N.....R. O.NY"rDIKHP"-..ET.E "NV-..OTW )I""-..DTSN "RRNEi-H-O-

SOEb(lL”kL RrKPS”“RL0 . . . . . . . . . . . . . . . ..a.. . . ..RPR... . . ..r..... . . . . . . . . . . . . . . . ..o.. . . . . . ..o.. . . . .._ I... . . . . . . . . . . . . ..o..a.. .RTII.L... “...S.K... .K..E..“.. . . . ..n.... .K..O..“.. . . . ..n.... .KR.O.W.E . . . ..n.s.. O.T.".LK.. .“.“O.Y”..

..T---.... .““---.... ..R---.... GG”---..E” . ..oEr.wo O.OOS”.IOK O..E”“.“YK No”---.“.”

“NSLNDLTW . . . . ..n... T.“L”..... . . . . . . . ... . . . . . . ... . . . ..n... WD.TRRI”K W-EKO”.O”

Y .: LDVELnKiER ..IR”.NI.. ..FK”..“.. I..K”.““.. I..K”.““.. . ..Rnx”I.. ..ISLET”RI .-I.-L...” . ..=..I.... ..F..T..T. “..Y.ORIYP

“TI”RKoKP, I....“.... ..“..PE... . . ..s..... . . ..s..... L....“.... ..“..P...S ..T....... ..T...N... ..T..OG... . ..T.EG..S

RORIP---“E . . . ..---.. . . . ..---NT . . . ..---..

IELEPPFOOS “.I . . . . . . . ..“.......

FIG. 1. Flavivirus E-glycoprotein

KLOO---SE .n..---... ““.L---..T . ..E---... . ..E---... . . ..---... EFIO---..K O-T.---..” E.TG---..r ..P.---... r* LT”OE . . 0

LDrERYY”nr VrLDCEPRSo ..“E...... .H........ ..I...“... I”..D...F. ..v....... I..S.....S ..“....... I..5 . . . . . . . . . . . . . . . . -H-.-F.... R..E.P”PTR “.FGNS.IRE ..F”R”“LL. L....S..T. ..“E.S..T. ..FNE”“LLO L......OLE .YFNE”IL.K V.LRPT”1l.E .S.L.R”R..

.. .. L .. .. ..

.KLS.F.... .KLN.F.... .n...FR... .S.P.FO...

.......... . ..r...... NT........ or........ .......... . ..T......

. . ..L....S .I”.L....S . . ..L..... . . ..L..... ..I.LS.S.S ..“O.K”SKG

.E”.E.(I... .EI.E.P... .E”.E.O... RR.T.S..O.

.L”O.K.E.. I.IR...E.O T.“K.K.E.R ..“..TFS..

LF06”SYITO . . . . . . ..s. .-..-..... ._........

. . . .._._..

I......“.. .TYP...E..

“.“....... ..“W...... ..R.....E. ..R....... . . ..R.-L.. ..“(lL.P..Y

:v n “TFLRVGGVL LFLITNVLIR I.“...,-.... . ..R...“. L.L.....I. I..R.S... . . . . . . . . . . . ..S”..H. L......... . ..S”..“. LR...T.... V.-R...“. .snIL..“I” “..SLO.O. ..rx...n.r .Y..““... “sL”L..“.T .Y..R”... ..CI....IT . . ..FT... .S..LR..L” .R”TLG.G.

RT....Ll..P R.....“.P.

0”LKCR”K”E KLKLKOTTYO FSSWLKLTS .......... . . . . . . . . . . .......... .... ..R.L D .“.I ...... “.....i-.P. .......... ..o ....... m..“.“.... . . . . . . . . . . ..o....... . . . . ..L..O ..R....... ..“S....LS R.T....S.K . . . . ..L..O ..r...ns.u . . . . ..LR.D ..o...ns.s . . . ..K.P.. ..ar..ns.r ..“T.E.BL. . ..“..L..T

I......E 1-Y.. I......S I......5

. . . . . . . .._

I...I.NIT. LH.K....K”

. . . .

. ..“..R.R ..on _... . ..“..S.Y ..R”.OK.K

n

........ I. .... ..S.A. ..... L.-I. ...... ..L. E ......... EN......“.

. ..“...... .F...“..EI . . . ..“..LI . . . . . . . ..L .L...L.... .FLS......

.. .. .. .. ..

. . ..LN...K ..S.“..T”K R.S.“..T”K “...“..“I” I...“OFLPY

*99 SO1 SO1 l 97 SO1 SO0 493 495 495 m 4%

consensus

(cons.) sequence and flavivirus sequence comparisons

(Fig. 20). For example, the amino acid sequence 20 to 50 appears to be hydrophobic and inaccessible; however, two flexible regions could be identified. The most conserved sequence in the Rl region (amino acids 98 to 120) appears to have one flexible region despite its predicted hydrophobicity and inaccessibility. Analysis of the predicted ,&turns revealed regions of homologous secondary structure (Table 1 and Fig. 2E). Three regions of high predicted secondary structure homology were identified among amino acids 80 to 120, 200 to 275, and 377 to 425. The first region corresponded to the area of highest sequence homology and the third occurred on the carboxy-terminal flanking region of homology minima 11. Data on MVE virus E-glycoprotein hydrophilicity and secondary structure were used to develop a model of the MVE virus E-glycoprotein (Fig. 3A), similar to the WNE virus model proposed by Nowak and Wengler (1987) but with some significant differences. The Rl domain proposed in our model contains sequences

from the amino-terminal one-third of the glycoprotein, and sequences from the R3 domain brought into proximity with Rl by predicted protein folding. Unlike the WNE virus analysis, the MVE virus analysis predicts that the Ll domain (interjacent between Rl and R2) may be antigenically important due to its high sequence heterogeneity and hydrophilicity. The Ll domain is included in the Rl/R3 domain in the MVE virus model. Similar to the WNE virus model, the MVE virus model predicts that the R2 domain is spatially distal from the Rl/R3 domain; however, with MVE virus, this R2 domain is considerably more hydrophilic than that reported for WNE virus. Derivation Using prepared previous epitopes

of synthetic

peptides

these computer analyses we designed and 1 1 synthetic peptides (Table 2). Because our investigations determined that type-specific are most efficient at neutralization and protec-

FLAVIVIRUS

53

PEPTIDES

100 80 60 40 20 0 2

6 i

I

-I

I

1

0

-1

-2

AMINO

ACID

NUMBER

FIG. 2. Computer-generated analyses of the flavivirus E-glycoprotein consensus sequence. The data from Fig. 1 were analyzed by five computerized algorithms as described under Materials and Methods. (A) Homology; darkened areas indicate regions less than 50% homologous, Location of synthetic peptides within the proposed Rl , R2. and R3 regions are underlined (Nowak and Wengler, 1987). (B) Parker hydrophilicity analysis; hydrophilic areas are darkened. (C) Accessibility analysis with surface accessible regions darkened. (D) Molecular flexibility analysis with flexible regions darkened. (E) Conserved predicted p-turns.

tion, and because MVE virus-specific were localized over tides 1 and 8 were ous success with

we were interested in developing antigens, most of these peptides hydrophilic homology minima. Pepsynthesized because of our previN-terminal peptides and the pre-

dieted molecular flexibility of this region (Fig. 2D). Peptides 2, 3,4, 5, and 17 were synthesized to correspond to the homology minimas 1, 3, 4 and 5, 9, and 11, respectively. Peptide 6 was synthesized because it was a major hydrophilic domain in R2, highly conserved, ac-

ROEHRIG ET AL.

54

TABLE1 PREDICTEDLOCATIONSOF ~-TURNS IN FLAVIVIRUSE-GLYCOPROTEINS Virus(amino Turnno.

MVE

SLE

JE

9

9

27

27

acid residue)

WN

KUN

YF

DEN1

DEN2

9

9 18

9

9 18

9 17

2 52

48

23

z;

DEN4

TBE

39 E

63

;;

::

i;:

:;f

99 103 110 123 127 133

El 103 110 123 127

::

8 72

10 79

;: 88

ii 103 110 123 20

133 157 171 192 194

it 99 103 110 123 128 133

E 99 103 110

ii 103 110

iz 103 110

zi 103 110 127

128 133

128

157 161 171 194 199

157 174 192 194

209 215

209 215

188 190

190 194

205 211 215 223

227

227 231 235 244 250 257

257

276 283 299 307 313

276 283 299 307 313

235 244

40

227 235 244 250 257 260 276 282 298 306 313

231 240 246 253 272 279 295 303

227 230 235 244 250 257

154 166 168 177 192 201 207 211 219 222 236 249 252

147 153 166

153 160 168

148

169

169

187

187

187

190

204 210

203 210

203 209

204 211 217

222 229

229

233 248 255

234 242 248 255

234

232 237

247 254

259

271

271 276 283 299 307

331 334

330 333

275 281 302 309

330 334

50

60

423 440 448 462

338 346 361

359

377 38i 388 391 399 402 406

125

146 153

304 310

304

328 339 343 349 361 363 371

331

375 384

375 385

394 396

394 396

285 304 309

326 331 334

E 99 103 110

141

30 215

ii 99 103 110

369

368

377 381 388

376 380 387 390

399 402 406 414 423 440 448 462

401 405

365 373 377 384 388 393 398 402

377 381 391 391 399 402 406

369 373 380 383 394

439 447 461

419 436 444 458

334 344

371

432 440

70 479

479

478

475

479

41

41

38

35

38

370 382 392 394 402

369 378 380 390 395 397 405 408

417

423 440 448 462

350 361

363

402 417

422

316 321

434

470

470

39

39

418 432

435 443 473 484

Total turns

37

30

38

Sharedturns

FLAVIVIRUS

PEPTIDES

55

A

FIG. 3. Molecular model of the MVE virus E-glycoprotein. (A) Modified Rl , R2, and R3 domains. (B) MVE synthetic peptides. Hydrophilic regions are underlined. Darkened amino acid circles are only for sequence continuity clarity. Arrows in Rl region indicate sequence reading direction, The figure is drawn from a ball and wire model derived using data from Table 1, MVE virus hydrophilicity analysis, and known disulfide linkages (Nowak and Wengler, 1987).

cessible, and flexible. We synthesized peptides 7, 14, and 16 because previous observations with JE virus indicated that this region may include relevant antigenic determinants (Mason et a/., 1987) and because our MVE model predicted significant overlap with the Rl region. As previously mentioned, homology minima 2, 6,7, 10, and 12 were not synthesized dueto their overall hydrophobic nature within the MVE virus sequence (data not shown). A graphic representation of these peptides with respect to the computer analyses (Fig.

2A) and placement on the MVE model (Fig. 3B) is shown. For comparison, the modified regions of antigenie importance (Rl, R2, and R3) are included (Fig. 3A). The region of highest homology (amino acids 98120) was not synthesized due to its highly hydrophobic nature. Reactivity

of peptides with polyclonal

antisera

A variety of virus HIAF were used at a 1 :lOO dilution to test the cross-reactivity of peptide expressed epi-

TABLE 2 SYNTHETICPEPTIDESFROMTHEE-GLYCOPROTEINOFMVEVIRUS

Peptide

Length

AA No.

01 02 03 04 05 06 07 08 14 16 17

16 17 22 21 20 23 16 21 17 19 22

1-15 35-50 77-97 122-141 179-197 230-251 305-319 13-33 289-305 336-354 356-376

a Cys indicates addition of nonencoded

terminal cysteine.

Sequence N-FNCLGMSSRDFIEGA-Cys-C” N-AADKPTLDIRMMNIEA-Cys-C N-TGESHNTKRADHNYLCKRGVT-Cys-C N-SNSAAGRLILPEDIKYEVGV-Cys-C N-KMGDYGEVTVECEPRSGLN-Cys-C N-STEWRNREILVEFEEPHATKQS-Cys-C N-CTEKFTFSKNPADTG-Cys-C N-EGASGATWVDLVLEGDSCITI-C N-RVKMEKLKLKGTTYGMC-C N-CKIPISSVASLNDMTPVGR-C N-VTANPYVASSTANAKVLVEIE-Cys-C

Average coupling efficiency 98.82 99.00 99.01 98.70 99.02 99.01 99.01 99.53 98.43 99.08 98.97

ROEHRIG ET AL.

56

TABLE 3 ELISA REACTIVITY OF MVE VIRUS PEPTIDES WITH VARIOUS ANTI-VIRAL POLYCLONAL ANTIBODY lmmunrzing virusa Peptide

MVE

1 2 3 4 5 6 7 8 14 16 17 Control

+b + ++ + +++ +++ ++ +++

Alf

JE

SLE

Kun

Kok

Str

WN

Dl

YF ++

+

+++

++ +++

++ ++

++ ++

++ ++

++ ++

+ ++

+ ++

++

++

+

++

++

++

+

+

+

+

++

+

a Virus abbreviations are as follows: Murray Valley encephalitis (MVE; either strain 11A or original), Alfuy (Alf), Japanese encephalitis Louis encephalitis (SLE), Kunjin (Kun), Kokobera (Kok), Stratford (Str), West Nile (WN), dengue 1 (Dl), and yellow fever (YF). b Optical densities of 1: 100 ascites dilution were converted as follows: ++$ = > 1.5, ++ = 0.5-l .5, + = 0.25-0.50, blank = <0.25.

topes (Table 3). Nine peptides were recognized by one or more of the HIAF after background reactivity on a negative control peptide was subtracted. Peptides 14 and 16 failed to react with any of the HIAF. Six peptides (1, 2, 3, 4, 7, and 17) appeared to be MVE virus-specific, except for cross-reactivity of peptide 1 with YF virus. Peptide 3 may be strain specific because it failed to react with the MVE 11A elicited HIAF (data not shown). Peptides 5, 6, and 8 were cross-reactive. These results are consistent with the homology analysis in Fig. 2A. An interesting observation was the high reactivity of anti-Alfuy virus HIAF with peptide 6. Antipeptide

antibody

response

Ten peptides elicited antipeptide antibody with one or both adjuvants (Table 4). Antiviral antibody was detected in 2 peptides (8 and 14) that elicited a poor antipeptide response. Highest antiviral antibody titers were elicited by peptides 2, 6, and 17. Peptide 4 was not immunogenic in BALB/c mice with this immunization protocol. Peptide 2 elicited highest titered virus neutralizing antibody. Epitope definition

Norm

with synthetic

peptides

We attempted to localize previously defined epitopes in three ways. First we probed peptides for MAb reactivity in ELISA; however, consistent and convincing binding to any peptide could not be identified (data not shown). Since MAbs might not bind to peptides attached to polystyrene plates, we attempted to com-

(JE), St,

pete MAb virus interaction by preincubating MAb with peptide. No peptide could compete binding of any MAb, corroborating the direct binding results (data not shown). We then used an alternate approach whereby we competed MAb binding to virus with pools of virus reactive antipeptide antibody (Table 5). Because of the high avidity of the MAbs, only pools of the higher titered antipeptide antisera (antipeptides 2, 3, 6, 7, 17) were used in the CBA. Dilutable competition greater than 20% with a 1: 10 or 1:20 dilution of competitor was considered positive. Binding of E-lc MAb was competed by antipeptides 2 and 7. E-ld, and E-6 MAbs were competed by antipeptides 3 and 7. Binding of E-4a MAb was partially inhibited by antipeptide 7. E-3 MAb binding was competed by antipeptides 3, 7, and 17. E5 MAb was competed by all tested antipeptides. No antipeptide competed E-4b or E-la MAbs binding. These results appeared to link the amino acid sequences of peptides 2 and 3 with the spatially discontinuous sequence of peptide 7 and 17. DISCUSSION Computer modeling of the E-glycoprotein structure and analysis of published amino acid sequences allows speculation about the relationship between structure and function of protein antigenic domains. The complete conservation of cysteine residues throughout all sequenced flavivirus E-glycoproteins indicates that the overall E-glycoprotein structure is conserved. Serologic and pathologic variability probably resides in

FLAWVIRUS PEPTIDES

57

TABLE4 SEROLOGIC ACTIVITIES OF ANTI-PEPTIDE ANTIBODIES~

Days postinoculation Anti-peptide Peptide/adjuvant

Mouse

Ol/RIBI 2 3 4 OWFCA

02/RIBI 2 3 4 03/RIBI 2 3 4 04/RIBI 2 3 4 05/RIBI 2 3 4 06/RIBI 2 3 4 OG/FCA 2 3 4 14/FCA

07/RIBI 2 3 4 lG/FCA

17lFCA 2 3 4 5 PEWRIBI PWFCA VEEOl/FCA 2 "Seeunder “Notdone.

Materialsand

0

titer

Anti-MVE titer

21

0

21

56

HI

N

40 40

160 80 160 80

40 40

160 80 80 80

40 40 40 40

20 40

<40 40

40 40 40 40 40

40 80

80 160 40 40 80

40 160 40 80 40



40 20

1280 2560 2560 320

40 40

320 160 80 80

>1280 40 320 160

110 110 110 110

80 40 160 160

20 40

320 320 80 160

40 40

320 320 80 80

40 40 40 40

10 20 20

20 20

20 40 20 40

40 20

40 40 40 40

ND ND ND ND



20 <20

640 40 320 20

40 40

160 80 80 80

40 40 40 40

110

40 40

1280 1280 640 640

<40 40

320 >640 80 160

320 160 80 80



80 40

80 320 320 80


10 20 10
20 20

>2560 >2560 22560 >2560

>1280 >1280 >1280 21280

40 40

40 40 40 40 40

<40 <40

80 80 320 80 80

40 40 160 40 160


10 10 160 80
<20 <20

160 40 80 160

40 40

80 160 80 160

40 40 160 160


20 80 20 20

40 <40

160 160 80 80 80

80 40

40 40 80 40 160

160 40 80 40 80

110
10 80 20 40 40

40 40

40 160 320 >1280 80 ND ND ND ND

>1280 80 640 >1280 640


20 20 20 40 40

40 40 40 40



<40 40

NDb ND ND ND

Methodsforexperimentaldetails.

21280 >1280 >1280 >1280 >1280 120 20 t20 20

40 40 40 40

ROEHRIG ET AL.

58

TABLE 5 COMPETITIONOF ANTI-PEPTIDEAND MABs FORVIRUS BINDING Antipeptide Epitope E-ld

MAb 4B5A-2

Specificity Type

CBA group

02

03

06

07

17

NMAF

HIAF

1

-B

+ -

-

+ -

-

++

-

-

++ ++

++

+

++

+

-

++

++

+

+

2 2

4B3B-6

Type Type Subcomplex

-

3

4B6B-10

Subgroup

4

+ -

+ -

-

-

-

-

E-la E-lc

4B6A-2 4B6C-2

E-5 E-6 E-3 E-4b

2B5B-3 6B6C-1

Subgroup Group

5 5

E-4a

2868-Z

Group

6

a Competitive

competitors

+

-

+

+ -

+

-

+

-

++ ++

-

++

results are as follows: ++ = >40%, + = ZO-40%, - = 920%.

microvariation within well-defined protein secondary structure. MAb analysis of SLE, TBE, and MVE viruses indicated that type-specific epitopes were located in a disulfide stabilized domain, and should demonstrate considerable sequence heterogeneity. Our modified Rl domain contains seven major areas of sequence variability (minima 1 through 5, 16, and 17) and is stabilized by four intrachain disulfide bonds (Fig. 2A). It also displays the antigenic characteristics of hydrophilicity, flexibility, and surface accessibility (Figs. 2B, 2C, and 20). Variants of yellow fever virus which escape neutralization by protective, neutralizing, and type-specific MAbs change in amino acid sequence within the Rl domain at amino acids 7 1, 72, and 125 (Lobigs, ef al., 1988). This is consistent with the hypothesis that the Rl domain contains epitopes which elicit type-specific, protective, and neutralizing antibody. Our observations that synthetic peptides 1, 2, 3, 4, and 17, derived from Rl domain sequences, demonstrate MVE virus type specificity when screened with a panel of polyclonal antisera (Table 3), and that only peptide 2 consistently elicited virus neutralizing antibody (Table 4), are also consistent with this hypothesis. The low reactivity of antiviral antibody with these peptides (except 17) may be due to intrachain disulfide bond requirement for true epitope expression. The low neutralizing response may also occur because of disulfide bond requirement for true epitope expression, or because peptide 2 appears to be distal to the binding sites defined by N escape. Embedded within the highly variable Rl domain is an amino acid sequence (98 to 120) that appears to be extremely important to aH flaviviruses. It is the most conserved region of the E-glycoprotein in both amino acid sequence and presence of predicted B-turns. We hypothesize that it is the fusion sequence for the flavi-

virus E-glycoprotein. It displays sequence homologies (high glycine content) and chemical characteristics (hydrophobicity and surface inaccessibility) consistent with fusion sequences of other viral proteins (Skehel and Water-field, 1975; Gething et al., 1978). It is also molecularly flexible, a necessary characteristic for the flavivirus fusion region which undergoes a conformational change when the pH shifts within endocytic vesicles (Gollins and Porterfield, 1986). The E-glycoprotein may be functionally and structurally similar to other characterized animal virus surface proteins. The highly conserved fusion sequence may be buried in a canyon of the glycoprotein three-dimensional structure (Wiley and Skehel, 1987). Epitopes that interact with neutralizing MAbs (amino acids 71,72, and 125) probably occur on the glycoprotein surface (Fig. 2) proximal to and surrounding the hydrophobic fusion domain. Antibody binding to this region might interfere with conformational changes required for intracellular fusion. This explanation of the E-glycoprotein interaction with neutralizing antibody is consistent with the recently defined mechanism of flavivirus neutralization, which appears to be a postattachment phenomenon (Gollins and Porter-field, 1986). Neutralizing antibodies appear to inhibit intracellular membrane fusion by blocking conformational shifts. These results are consistent with observations in our laboratory that show that unlike alphaviruses, flavivirus neutralizing MAbs are not efficient at blocking viral attachment (unpublished observation). Clear definition if this model requires synthesis of the highly conserved sequence. Antibodies to this peptide should be useful in further defining its biological function and morphological transformations. These experiments are currently in progress in our laboratory. Peptides derived from the R2 domain of the glycoprotein are also immunogenic. Peptides 6 and 7 elicit

FLAVIVIRUS

MVE virus antibody. The immunogenicity of peptides 7 and 17 is consistent with the immunogenic domain defined on JE virus by protein expression in Escherichia co/i (Mason et al., 1987). Actual virus biological function of the R2 domain remains to be defined. The type-specificity of peptides 1, 2, 3, 4, 7, and 17, and the spatial linking of peptides 3 and 7 seems to indicate that while these sequences are quite distal from each other in the protein backbone, they appear to be proximal to each other in the glycoprotein threedimensional structure. These results require cautious interpretation, but clearly an overlap of the Rl, R2, and R3 domains is implied. We are currently producing monoclonal antipeptide antibodies to further define these spatial arrangements. We are also investigating the protective capacity and serologic cross-reactivity of these synthetic peptides. We hope that these results will be the basis of refined, rapid diagnostic tests and highly defined target vaccines.

REFERENCES CAMMACK, N., and GOULD, E. A. (1986). Topographical analysis of epitope relationships on the envelope glycoprotein of yellow fever 17D vaccine and the wild type Asibi parent virus. virology 150, 333-341. CHAPPELL, W., BUCCA, M., WHITE, L., and GAMBLE, W., Eds. (1974). Arbovirus reagents. ln “Production Manual: Viral, Rickettsial, Chlamydial, and Mycoplasmal Reagents,” 5th ed., pp. 4-6. U.S. Department of Health, Education, and Welfare, Centers for Disease Control, Atlanta. CHOU, P. Y., and FASMAN, G. D. (1978a). Conformational parameters for amino acids in helical, B-sheet, and random coil regions calculated from proteins. Biochemistry 13, 21 l-222. CHOU, P. Y., and FASMAN, G. D. (197813). Prediction of protein conformation. Biochemistry 13, 222-245. CHOU, P. Y., and FASMAN, G. D. (1978c). Empirical predictions of protein conformation. Annu. Rev. Biochem. 47, 251-276. CHU, M., O’ROURKE, E. J., and TRENT, D. W. (1988). Genetic relatedness among structural genes of dengue-1 virus strains. 1. Gen. Viral., in press. COIA, G., PARKER,M. D., SPEIGHT, G., BYRNE, M. E., and WESTAWAY, E. G. (1988). Nucleotide and complete amino acid sequences of Kunjin virus: Definitive gene order and characteristics of the virusspecified proteins. 1. Gen. Viral. 96, l-2 1. DALGARNO, L., TRENT, D. W., STRAUSS,J. H., and RICE, C. M. (1986). Partial nucleotide sequence of the Murray Valley encephalitis virus genome. /. Mol. Biol. 187, 309-323. DEUBEL, V.. KINNEY, R. M., and TRENT, D. W. (1986). Nucleotide sequence and deduced amino acid sequence of the structural proteins of dengue 2 virus, Jamaica genotype. Virology 155,365-377. GETHING, M. J.. WHITE, J. M., and WATERFIELD.M. D. (1978). Purification of the fusion protein of Sendai virus: Analysis of the NH2-terminal sequence generated during precursor activation. Proc. Sot. Nat/. Acad. Sci. USA 75,2737-2740. GOLLINS. S. W., and PORTERFIELD,J. S. (1986). A new mechanism for the neutralization of enveloped viruses by antiviral antibody. Nature (London) 321,244-246.

PEPTIDES

59

HAWKES,R. A., ROEHRIG,J. T., HUNT, A. R., and MOORE, G. A. (1988). Antigenic structure of the Murray Valley encephalitis virus E-glycoprotein. J. Gen. W-o/. 69, 1 105-l 109. HEINZ, F. X. (1986). Epitope mapping of flavivirus glycoproteins. Adv. Virus. Res. 31, 103-l 86. HEINZ, F. X., BERGER,R., TUMA, W., and KUNZ, C. (1983a). Location of immunodominant antigenic determinants on fragments of the tickborne encephalitis virus glycoprotein: Evidence for two different mechanisms by which antibodies mediate neutralization and hemagglutination inhibition. Virology 130, 485-501. HEINZ, F. X., BERGER,R., TUMA, W., and KUNZ, C. (198313). A topological and functional model of epitopes on the structural glycoprotein of tick-borne encephalitis virus defined by monoclonal antibodies. Virology 126, 525-537. HEINZ, F. X., TUMA, W., GUIRAKHOO, F., BERGER, R., and KUNZ, C. (1984). lmmunogenicity of tick-borne encephalitis virus glycoprotein fragments: Epitope-specific analysis of the antibody response. virology65, 1921-1929. HENCHAL, E. A., MCCOWN. J. M., BURKE, D. S., SEGUIN, M. C., and BRANDT, W. E. (1985). Epitopic analysis of antigenic determinants on the surface of Dengue-2 virions using monoclonal antibodies. Amer. J. Trop. Med. Hyg. 341, 162-169. HOPP, T. P., and WOODS, K. R. (1981). Prediction of protein antigenic determinants from amino acid sequences. froc. Nat/. Acad. Sci. USA 78,3824-3828. JANIN, J. (1979). Surface and inside volumes in globular proteins. Nature (London) 277, 491-492. KARPLUS,P. A., and SCHULZ, G. E. (1985). Prediction of chain flexibility in proteins. Naturwissenschaften 72,212-213. KIMURA-KURODA,J.. and YASUI, K. (1986). Antigenic comparison of envelope protein E between Japanese encephalitis virus and some other flaviviruses using monoclonal antibodies. J. Gen. Viral. 67, 2663-2672. KUMMER, U.. THIEL, E., DOXIADIS, I., EULITZ, M., SLADOLJEV,S., and THIERFELDER,S. (1981). Tritium radiolabeling of antibodies to high specific activity with N-succinimidyl (2,3-3H) propionate: Use in detecting monoclonal antibodies. /. lmmunol. Methods 42, 367-374. LOBIGS, M., DALGARNO, L., SCHLESINGER,J. J., and WEIR, R. C. (1988). Location of a neutralization determinant in the E protein of yellow fever virus (17D vaccine strain). Hrology 161, 474-478. MANDL, C. W., HEINZ, F. X., and KUNZ, C. (1988). Sequence of the structural proteins of tick-borne encephalitis virus (Western Subtype) and comparative analysis with other flaviviruses. Virology 166,197-205. MASON, P. W., MCADA, P. C., DALRYMPLE,J. M., FOURNIER,M. J., and MASON, T. L. (1987). Expression of Japanese encephalitis virus antigens in Escherichia coli. wfology 158, 36 l-372. MATHEWS,1. H.. and ROEHRIG,J. T. (1984). Elucidation of the topography and determination of the protective epitopes on the E-glycoprotein of Saint Louis encephalitis virus by passive transfer with monoclonal antibodies. /. Immunol. 132, 1533-l 537. MCADA, P. C., MASON, P. W., SCHMALJOHN. C. S., DALRYMPLE.J. M., MASON, T. L., and FOURNIER. M. J. (1987). Partial nucleotide sequence of the Japanese encephalitis virus genome. Virology 158, 348-360. NOWAK, T., and WENGLER,G. (1987). Analysis of disulfides present in the membrane proteins of the West Nile flavivirus. Virology 156, 127-137. PARKER,J. M. R., Guo. D., and HODGES, R. S. (1986). New hydrophilicity scale derived from high-performance liquid chromatography peptide retention data: Correlation of predicted surface residues with antigenicity and x-ray-derived accessible sites. Biochemisrry 25,5425-5432.

60

ROEHRIG ET AL.

RICE, C. M., LENCHES. E. M., EDDY, S. R., SHIN, S. J., SHEETS, R. L., and STRAUSS,J. H. (1985). Nucleotide sequence of yellow fever virus: Implications for flavivirus gene expression and evolution. Science 229,726-733. ROEHRIG,J. T. (1986). The use of monoclonal antibodies in studies of the structural proteins of togaviruses and flaviviruses. In “The Togaviridae and Flaviviridae” (S. Schlesinger and M. 1. Schlesinger, Eds.), pp. 251-278. Plenum, New York/London. ROEHRIG,J. T., DAY, J. W., and KINNEY,R. M. (1982). Antigenic analysis of the surface glycoproteins of a Venezuelan equine encephalomyelitis virus (TC-83) using monoclonal antibodies. virology 118, 269-278. ROEHRIG,J. T., MATHEWS,J. H., and TRENT, D. W. (1983). Identification of epitopes on the E-glycoprotein of Saint Louis encephalitis virus using monoclonal antibodies. Virology 128, 1 18-l 26. ROSE, G. D. (1978). Prediction of chain turns in globular proteins on a hydrophobic basis. Nature (London) 272, 586-590. SARIN, V. K., KENT, S. B. H., TAM, J. P., and MERRIFIELD,R. B. (1981).

Quantitative monitoring of solid-phase peptide synthesis by the ninhydrin reaction. Anal. Biochem. 117, 147-l 57. SKEHEL,1. J., and WATERFIELD,M. D. (1975). Studies on the primary structure of the influenza virus hemagglutinin. Proc. Alar/. Acad. Sci. USA 72, 93-97. TRENT, D. W., KINNEY,R. M., JOHNSON,B. J. B., VORNDAM, A. V.. GRANT, 1. A., DEUBEL, V.. RICE, C. M., and HAHN, C. (1986). Partial nucleotide sequence of St. Louis encephalitis virus RNA: Structural proteins, NSl, ns2a, and ns2b. Virology 156, 293-304. WILEY, D. C., and SKEHEL,J. 1. (1987). The structure and function of hemagglutlnin membrane glycoprotein of influenza virus. Annu. Rev. Biochem. 56,365-394. WINKLER, G., HEINZ, F. X., and KUNZ, C. (1987). Characterization of a disulfide bridge-stabilized antigenic domain of tick-borne encephalitis virus structural glycoprotein. J. Gen. Viral. 68, 2239-2244. ZHAO, B., MACKOW, E., BUCKLER-WHITE,A., MARKOFF, L., CHANOCK, R. M., LAI, C.-J., and MAKINO. Y. (1986). Cloning full-length dengue type 4 DNA sequences: Analysis of genes coding for structural proteins. \/irology 155,77-88.