Journal of General Virology (1990), 71, 2609-2614. Printed in Great Britain

2609

Location of epitopes on the major core protein p24 of human immunodeficiency virus J. P. M . Langedijk, 1 J. J. Schalken, 2 M . Tersmette, 3 J. G. H u i s m a n 3 and R. H. Meloen 1. 1Central Veterinary Institute, P.O. Box 65, 8200 A B Lelystad, 20rganon Teknica B.V., Boxtel and 3Central Laboratory o f the Netherlands R e d Cross Blood Transfusion Service, Amsterdam, The Netherlands

Antibody-binding sites were m a p p e d on all overlapping nonapeptides of the m a j o r core protein p24 of h u m a n immunodeficiency virus type 1 (HIV-1) using murine monoclonal antibodies (MAbs) and sheep and rabbit polyclonal antibodies raised against H I V - 1 / H 9 (strain IIIB) viral lysate and antibodies obtained f r o m h u m a n s infected with H I V - 1 . The binding sites were m a p p e d to various distinct regions o f this protein. After superim-

position of the antibody-binding sites on a p r o p o s e d model o f p24 o f H I V - 1 , these sites appeared to be located on the surface o f the protein on loops, turns and coils of p24 but, unexpectedly, not on the m a j o r part o f the predicted 'puff'. Little reaction was found with the inaccessible anti-parallel fl-barrel. These results are the first e x p e r i m e n t a l evidence for the validity o f the structure p r o p o s e d for p24 o f H I V - 1 .

Introduction

recently proposed by Argos (1989). This model is based upon the homology between p24 of H I V and the coat protein VP2 of the picornaviruses (Argos, 1989). The results show that the antibody-binding sites are located largely on loops, turns and coils of p24. Only peptide sequences that correspond with the N- and C-terminal part of the ' p u f f ' region bind antibodies.

A I D S is caused by the human immunodeficiency virus (H/V) (Barr6-Sinoussi et al., 1983; Gallo et al., 1984). Serum antibodies against the major core protein, p24, of HIV-1 appear early in infection together with antibodies against envelope proteins as serological markers (JollerJemelka et al., 1987). Disease progression is to some extent associated with higher extracellular core antigen levels and a decline in the level of antibodies against p24 in serum (Goudsmit et al., 1986; Lange et al., 1986). As the amino acid sequence of p24 is highly conserved, this protein has been used to develop screening assays to measure the serum antibodies to p24 for diagnostic and prognostic reasons (Broliden et al., 1989; Mathiesen et al., 1989). More and more synthetic peptides are being used to m a p antibody-binding sites of proteins used in diagnostic assays. Immunoassays which are based on synthetic peptides are in principle superior to assays based on whole protein, because such assays are less costly and, in theory, more specific than assays using whole protein. In this report immunoreactive peptides of p24 were selected with the aid of the Pepscan method (Geysen et al., 1984). With this method overlapping nonapeptides covering the complete amino acid sequence of p24 of HIV-1 were tested for their reactivity with polyclonal antibodies and monoclonal antibodies raised against viral lysate, as well as with patients' sera. The detected antibody-binding sites were superimposed on a structural model of p24,

0000-9629 © 1990 SGM

Methods Pepscan. Overlapping peptides of p24 of HIV-1 (strain HTLV-IIIB) were synthesized and tested for reactivity patterns of antisera and MAbs as described before (Geysen et al., 1984, 1985). In short, scanning for antibody-reactivepeptides required the synthesisof every overlapping peptide of the relevant protein sequence. For example, in the case of a nonapeptide scan, a protein of n residues can be read as ( n - 8) overlapping nonapeptides, in which peptide 1 consists of residues 1 to 9, peptide 2 of residues 2 to 10 and so on. The amino acid sequence was derived from the nucleotide sequence of HTLV-IIIB (Ratner et al., 1985). The peptides, still coupled to the solid supports, were then tested against the appropriate antibodies in a normal ELISA. Absorbance values were plotted against the position of the N-terminal amino acid of the peptide in the total sequence. If more than one peptide of a set of overlapping peptides reacted, the peptide that reacted best was considered to be the core of the antibody-binding site. When a set of overlapping peptides reacted equally, the shared amino acids in the reactive peptides were considered to constitute the antibody-binding site. Preparation of MAbs. Anti p-24 MAbs were obtained by immunization with purified HIV-1/H9 (strain IIIB) viral lysate as described previously (Tersmette et al., 1989). Primary screening was performed using viral lysate-coatedwells. The specific binding of MAbs OT37A, OT38B and OT39A to p24 was confirmedby immunoblotting(data not

2610

Epitopesof the p24 protein of HIV

shown) and the specificity of MAbs CLB-21, CLB-59 and CLB-19.7 was confirmed by immunoblotting and a radioimmunoprecipitation assay (Tersmette et al., 1989).

(a) CLB-21

P r e p a r a t i o n o f antisera. Rabbit anti-HIV sera were obtained by repeated immunization with 200 ~tg Triton X-100-disrupted purified HIV proteins (Tersmette et al., 1989). Sheep anti-HIV sera were obtained by the same procedure. Human

20

40

60

80

100

(b) CLB-19.7

(numbers 1 to 6) and six asymptomatic, HIV-l-infected individuals (numbers 7 to 12). All sera were positive for HIV-I. ....................... i.........................................................

20

140

160

180

200

g

sera. Sera were obtained from six patients with AIDS

0

120

40

60

.I

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

80

100

. . , . . . . . . . . . . . . . . . . L,..,.L . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

120

140

160

180

200

220

Results MAbs and polyclonal antibodies were applied in Pepscan analyses to determine the antigenic sites on the major core protein p24 of HIV-1. Representative results for all overlapping nonapeptides are shown in Fig. 1. The Pepscan data obtained with anti-p24 MAbs, rabbit antisera raised with viral lysate, sheep antisera raised with viral lysate and sera of humans infected with HIV-1 are summarized in Fig. 2. The antibody-binding sites of the MAbs and the antiserum of rabbit 2 indicated in Fig. 2 are superimposed on the proposed structure of p24 of HIV-1 (Argos, 1989) as shown in Fig. 3. Six different antibody-binding sites are detected using MAbs (Fig.2a, Table 1). Those sites which are part of the model are located on turns between the a-helix and strand D, strands D and E, and strands H and I. A majority of the MAbs reacted with amino acid sequences between amino acids 83 and 99 between the a-helix and strand D; this antibody-binding site appears to be immunodominant in mice. MAb CLB-19.7 and MAb CLB-MG reacted also with peptide 1 (Fig. lb; Fig. 2a). Although we cannot exclude the possibility that both peptides reacted with the MAbs because they are part of a single discontinuous epitope, a more likely explanation is that both peptides react due to their sequence homology; the amino acid sequence of reactive peptide 1 is PIVQNIQGQ whereas the shared amino acids of the reactive peptides 89 and 90 are PIAPGQMR. This partial sequence homology may account for the crossreactivity of the MAbs. The rabbit antisera reacted with the turns between strands B and C, strands D and E, strands H and I and ahelix and strand D, the N-terminal part of the a-helix, Cterminal part of the puff, strand E and the N-terminal part of strand D (Fig. 2b; Fig. 3b). The sheep antisera reacted with turns between strands B and C, a-helix and strand D, strands D and E, strands G and H, the C-terminal part of strand B, a-helix, Cterminal part of the puff and a major part of strands D and E (Fig. 2b). The human sera reacted with the turns between strands B and C, a-helix and strand D, strands D and E, strands F and G, the C-terminal part of the puff, the centre of the a-helix and strand D (Fig. 2c).

(c) OY37A

0

.................................................. ~,,~.,,.,,,;.,h,~,,,,,,,,ll,.,,.,.,l~ ............~,L ..................................................................................................................

20

40

60

80

100

120

140

160

180

200

220

(d) OT38B

o ............. ~...................................................................... Ill.................................................................................................................................... i

,

~

20

i

i

*

i

,

,

i

,

i

i

,

,

i

,

i

,

I

~

i

40

60

80

100

120

140

160

180

200

220

40

60

80

100

120

140

160

180

200

220

60 80 100 120 140 Amino acid number

160

180

(e) OT39A

20 ( f ) Sheep 1

E,,rrt,JJ,i 20

40

Fig. 1. Pepscan results obtained with the overlapping nonapeptides of p24 of HIV-I (strain HTLV-IIIB) (Rattler et a L , 1985), five MAbs (a, CLB-21; b, CLB-19.7; c, OT37A; d, OT38B; e, OT39A) and one polyclonal sheep antiserum (f, sheep 1) raised with HIV-1/H9 (strain IIIB) viral lysate. The number on the horizontal axis corresponds with the N-terminal amino acid of the nonapeptide. Absorbances at 405 nm obtained with each peptide in an ELISA are plotted vertically.

The reactivity o~"the polyclonal antibodies is broader than the reactivities of the MAbs taken together; the polyclonal antibodies reacted not only with amino acid sequences on the exposed loops but also with amino acid sequences that seem inaccessible in the model, i.e. strands E or D (Fig. 3). As the best possible alignment of HIV p24 and picornaviral VP2 is achieved by introducing some deletions and insertions, some parts of the proposed model may differ from the unknown three-dimensional structure of p24.

J. P. M. Langedijk and others

(a)

CLB-21

i

CLB-59

i

CLB-19.7

i

i I

I

I

I

I !

CLB-MG

,

I

OT37A

,

I

OT38B

i

OT39A Summary

I I

(b) Rabbit l Rabbit2

,

2611

,m* m

I

II

i

i

m

I

•

,

, ~

, m

I

~

,

, ,

Sheep 1 , ~

~

,

Sheep 2

~,

,

,~.

I

,

~

I

'

,

,

,

m ~

I

'

I

'

,

,

~'

,

,

m

m

,

~.

,

'

~"

,

I

'

I

'

,

,

,

,

,

,

,

m

,

,

,

,

,

~-

,

I

'

,

,

I

I

,

'~

,

,

,

m

,

.m,

,

,

,

.m,

,

,

,

(c)

I

.'~

I I |

7 8 9 10 11 12 Summary

m

' I

p24 Sequence

0

' I

I

20

'B I

I

40

C ' I

I

60

Helix' !

I

80

'D I

m

I~

I

!

100

Puff

I

I

120

I

140

' I

I

F

'G

/

160

I

180

H' I

I

200

I ' I

I

220

Fig. 2. (a) Location of the antigenic siteson the primary sequence of the major core protein p 2 4 o f H I V - 1 for different MAbs. The bars indicate reactive amino acid sequences giving absorbance readings threefold higher than the background. Site 1 (amino acids 1 to 9) and 2 (amino acids 32 to 40) are not part of the model. The reactive sequences are summarized on the bottom line. The n u m b e r s on the bottom line correspond with the p24 sequence of HIV-1. (b) As (a) for rabbit and sheep sera. (c) As (a) for HIV-l-infected h u m a n sera. Summarized are the reactive sequences with an absorbance reading threefold higher than the background and which occurred in three or more of the 12 h u m a n sera. On the bottom line, the secondary structures of p24 of H IV-1 proposed by Argos (1989) are shown. The flstrands are marked B to I according to the nomenclature of R o s s m a n n et al. (1985) for the rhinoviral coat protein fold. In (c) above the p24 sequence, the position of//-strands B to I are represented by a bar ( ); the or-helix is represented by an open box ( r - " l ) .

Secondary structure predictions were made using the Garnier algorithms (Garnier et al., 1978) for those parts which correspond with two major 'insertions' in the primary sequence of p24 when compared to the primary sequence of VP2 (Argos, 1989). The secondary structure predictions for the puff region and the region between strands C and D are shown in Fig. 4. Unlike the ribbon drawing of Fig. 3, the puff region has a low turn and coil propensity (Fig. 4a) and high sheet propensity. Except for the C-terminal part, the region is quite hydrophobic,

Table 1. Amino acid sequences which react with p24-specific MAbs

1 2 3 4 5 6

Site

Sequence

Reactive M A b s

1-9 32-40 85-91 89-97 111-115 205-213

PIVQNIQGQ FSPEVIPMF PVHAGPI GPIAPGQMR LQEQI LGPAATLEE

CLB-19.7, C L B - M G CLB-21 OT39A CLB-19.7, C L B - M G , OT38B OT37A CLB-59

Epitopesof the p24 protein of HIV

2612

(a)

::~ \~ II

Exterior

(b)

\\ II

// // It"~'