fine specificity of the b-cell epitopes recognized in hiv-1 nef by human sera

VIROLOGY 187, 156-l 64 (1992)

Fine Specificity of the B-Cell Epitopes Recognized in HIV-I NEF by Human Sera

MARJA T;iHTINEN,**’ FRANK GOMBERT,t EIJA-RIll-TA HYYTINEN,* GUNTHER JUNG,t ANNAMARI RANKI,+ AND KAl J. E. KROHN”

*University of Tampere. Institute of Biomedical Sciences, Tampere, Finland; tEberhard-Karls-UniversiMt Tilbingen, lnstitut filr Organische Chemie, D-7400 TUbingen, Germany; and *Helsinki University Central Hospital, Department of Dermatology, Helsinki, Finland

Received July 5, 199 1; accepted November 8, 199 1

We have previously used partially overlapping synthetic nonapeptides to characterize the human natural antibody response against HIV-l negative regulatory factor (NEF), and identified nine 5 to 13 amino acid long regions that were recognized by sera of HIV-1 -infected individuals. In this report we define the minimal size of these epitopes with the use of shorter, from 3 to 8 amino acid long partially overlapping peptides covering the complete sequence of the previously identified reacting regions and the N- and C-terminal flanking sequences. We also introduce a new method for the analysis of the reactivities obtained with peptides of different lengths. In six of the antigenic regions the epitopes were found to be noncontiguous and to consist of multiple, down to three amino acid long separate reactive stretches (epitope 1: WSK, VGW, TVRERMRR; epitope 3A: PLRPM, SHFLK; epitope 3B: SQRRQD, DLW; epitope 3C: IYHT, QGYFPDWQN; epitope 4: SLL, VSL; epitope 5: EVLEWRFDSR, VAR). Three epitopes were clearly linear (epitope 2: CAWLE; epitope 3D: LTFGWC; epitope 6: PEYF). Interestingly, five of the minimized B-cell epitopes (1, 3A, 3C, 3D, 5) recognized by human sera overlap totally or partly with the previously identified T-cell epitopes in HIV-1 NEF. Also, only three of the epitopes (3C, 3D, 5) were in a computer-based homology search shown to contain strictly NEF-specific sequences. 0 1992 Academic Press, Inc.

INTRODUCTION

Negative regulatory factor (NEF) is one of the six nonstructural proteins coded by the genome of HIV-l, and in some circumstances it seems to suppress viral expression (Niedermann et al., 1989). More than 70% of HIV-l-positive persons have antibodies against NEF (Cheingsong-Popov et al., 1990; Sabatier et al,, 1989), and these antibodies sometimes appear before a sero- logical response toward the structural proteins can be detected (Ranki et al., 1987; Ameisen et a/., 1989). These early antibodies to NEF may therefore be a hall- mark of a latent HIV infection. However, the significance of the antibody response toward NEF is weak- ened by the fact that cross-reacting antibodies recog- nizing NEF can also be found among blood donors and dermatological patients with no known exposure to HIV (Ranki et al., 1990; Gombert et al., 1990). The problems caused by cross-reactivity could be over- come by identifying genuinely NEF-specific antigenic epitopes not recognized by sera from nonHIV-infected individuals.

Due to the fact that HIV regulatory proteins, such as NEF, are often synthesized even before the mRNA for the structural proteins appear (Krohn et a/., 1991) im-

’ To whom reprint requests should be addressed at the Institute of Biomedical Sciences, University of Tampere, P.O. Box 607. SF- 33101 Tampere, Finland.

0042-6822/92 $3.00 CopyrIght 0 1992 by Academic Press, Inc.

mune response against them might prevent the spreading of virus before virions containing structural proteins are synthesized. In the characterization of the NEF-specific immune response, both T- and B-cell epitopes have been searched for. Several helper T-cell and cytotoxic T-cell epitopes have been identified (Koenig et al., 1990; Cullman et a/., 1991; Wentworth, et a/., 1991). Also, the B-cell epitopes of NEF recognized by sera from HIV-infected individuals or from im- munized animals have been reported (Sabatier et a/., 1989; Ameisen et al., 1989; Gombert et a/., 1990; Schneider et al,, 1991).

We have previously used 9 amino acid long, partially overlapping peptides to characterize the sequences in NEF that are specifically recognized by sera from HIV- 1 -infected individuals as well as by sera from seronegative individuals with high-risk behavior and identified nine different B-cell epitopes (Gombert et al., 1990). The length of these regions in NEF varied from 5 to 13 amino acids. Also, antigenic regions in NEF identified by others have been rather long, from 14 to 34 amino acids (Sabatier et a/., 1989; Ameisen et al., 1989; Schneider et al., 1991). The average size of an anti- genie determinant is, however, generally thought to be only from 5 to 6 amino acids, and even 3 amino acid long reactive sequences have been identified with polyclonal sera (Geysen et a/., 1988). This would indicate that the reacting regions found so far in HIV-l NEF may actually be composed of shorter antigenic stretches.

156

All rights of reproduction in any form reserved

B-CELL EPITOPES IN HIV-1 NEF 157

In order to find out the fine, i.e., minimal, composi- tion of HIV-1 NEF B-cell epitopes, we have further studied the antigenic regions in a pin-Elisa system (Geysen et a/., 1987) with the use of synthetic peptides varying in size from three to eight amino acids. The results were gathered in table format and analyzed with a modified matrix analysis. In order to find out the most strictly NEF-specific sequences, we compared the identified NEF B-cell epitopes for homologies to known proteins, using the Swiss-Prot protein sequence bank. Also, we compared their sequences to known NEF- specific T-cell epitopes to identify regions in NEF which would be targets for both humoral- and cell-mediated immune response.

METHODS

Sera

Sera were obtained from HIV-l -infected patients, and 10 sera were selected for further analysis either due to a positive reaction with recombinant NEF in Western blot or due to reactivity against the NEF epitopes found earlier by us in a pin-ELISA assay. The pin-ELISA assay and Western blots were performed as described earlier (Ranki et a/., 1990; Gombert et al., 1990). As an antigen in Western blot, we used recombinant NEF protein representing the HIV-l BRU sequence (kindly provided by Dr. M.-P. Kieny and Dr. J.-P. Lecocq, Transgene, Strasbourg, France). Sera from four healthy HIV-1 -negative laboratory workers were used as controls. Also, in order to exclude the possibil- ity that the anti-NEF response seen in patient sera would be solely due to the polyclonal B-cell activation associated with HIV infection, we included one serum (control No. 5) from a patient with Yersinia enterocoli- tica type 3 infection, which is known to raise a strong polyclonal immune activation.

Peptide synthesis on polypropylene pins and ELISA assays

A total of more than 700 peptides covering the previously identified nine antigenic regions of NEF and the flanking sequences were synthesized on polypropylene pins with the Cambridge Research Biochemicals’ Epitope Scanning kit as described earlier (Gomber-t et al., 1990). The peptides represented the BRU isolate of HIV-l. All nine antigenic regions of NEF were synthesized as 8, 7, 6, 5, 4, or 3 amino acid long overlapping peptides as shown in Figs. 1 a and 1 b. The pins were incubated with diluted serum overnight, and bound antibodies were detected with peroxidase-conjugated

rabbit anti-human IgG in an enzymatic reaction with 2,2’-azino-bis(3-ethylbenzthiazoline-6-sulfonic acid) (ABTS) and H,O,. The photometric determination was carried out in ELlSAreader at 405 nm. Optimal serum dilution was from 1:500 to 1:800 depending on the reactivity of the serum.

Analysis of the epitopes

In order to define the minimal size of a reactive epitope, a matrix analysis of each serum against each epitope was performed. In the matrix, the x axis shows the sequence of the analyzed region and they axis the size (mer) of the studied peptides (Figs. 2a and 2b). Two separate matrix analysis were performed: the peptides were analyzed according to their first amino acid in the amino-terminal analysis and, correspondingly, the carboxy-terminal analysis was performed according to the last amino acid in a given peptide. This duplicate analysis was performed to ensure the detection of a minimal sequence which would react irrespec- tively of its situation in the amino or carboxy terminus of the peptide.

The matrix was filled in by giving the reaction with each peptide a positive or negative sign (+, -) as fol- lows: a reaction was considered positive if the absorbance given by a peptide was two times higher than the absorbance of background. The background for each serum against each epitope was calculated as the mean of the lowest 40% of ELISA values given by peptides from the area of the epitope in question. For a positive reaction, a + sign was placed into the matrix at the point which corresponds to the length and start point of the peptide (amino-terminal analysis), or at the point which corresponds the length and end point of the peptide (carboxy-terminal analysis). A - sign was correspondingly given to peptides yielding negative reactions. At the bottom of the matrix, arithmetical sum values of signs for each amino acid are given.

On the basis of these values, an epitope was deter- mined to be the region where the sum values ex- ceeded zero. This diminished the effect of an incidental positive reaction due to unspecific binding. As the shortest analyzed peptide was three amino acids long, an additional two amino acids at either the amino or the carboxy terminus had to be considered as part of the epitope, too. Finally, as an epitope should react irre- spectively of its position in a peptide, we deduced the final or consensus epitope to be the region which was found antigenic in both amino terminus and carboxy terminus analyses. This precondition excluded some sequences which were positive when situated in the carboxy terminus but not in the amino terminus of the

158 T;iHTINEN ET AL.

NTAATNAA

TAATNAAC AATNAACA

ATNAACAW

TNAACAWL

NAACAWLE

AACAWLEA

ACAWLEAQ

CAWLEAQE

AWLEAQEE

0.0 0.1 0.2 0.3 0.4

O.D.

NAACAW AACAWL ACAWLE CAWLEA AWLEAQ WLEAQE LEAQEE I I - I . I '

0.0 0.1 0.2 0.3 0.4

O.D.

NTAA TAAT AATN ATNA TNAA NAAC AACA ACAW CAWL AWLE WLEA LEAQ EAQE AQEE I ,

0.5

I * I. I. 8. I * I

0.0 0.1 0.2 0.3 0.4 0.5

O.D.

NTAATNA TAATNAA

!i AATNAAC

ACAWLEA

i CAWLEAQ AWLEAQE WLEAQEE 1

0.0 0.1 0.2 0.3 0.4 0.5 0.6

O.D.

NTAAT TAATN AATNA ATNAA TNAAC NAACA AACAW ACAWL CAWLE AWLEA WLEAQ LFAQE EAQEE

NTA TAA AAT ATN TNA NAA AAC ACA CAW AWL WLE LEA EAQ APE QEE I

1 I I

0.0 0.2 0.4 0.6 0.8

0.0 0.1 0.2 0.3 0.4 0.5 0.6

O.D.

O.D.

FIG. 1. (a) Absorbance results of serum No. 7 against peptides from epitope 2 showing a contiguous pattern of reactivity. White bars represent negative reactivities (absorbance below two times background) and black bars positive reactions. Due to the lack of space, some negative values in the amino terminus of the scanned area are left out of the figure. (b) Absorbance results of serum No. 7 against peptides from epitope 1 showing a noncontiguous reactivity.

peptide. It also prevented the edges of the matrix con- tope, the amount of mismatched amino acids was se- taining only a few values from giving false results. lected to allow up to 25% dishomology.

Homology search

In the homology search, the sequences of identified NEF epitopes were compared to the sequences found in the Swiss-Prot protein bank. For each stretch of epi-

Calculation of the variability index

A variability index for each amino acid inside the NEF epitopes was calculated according to the sequences found in the Human Retroviruses andAIDS 1990 book,


4

0.0 0.2 0.4 0.6 0.8

O.D.

WGWPT I VGWPTV I GWPTVR

WPTVRE

PTVRER

TVRERM

VRERMR

RERMRR 1 ’ I ’ I I I ’

0.0 0.2 0.4 0.6 0.8

O.D.

1.0

WGW VGWP GWPT WPTV PTVR TVRE VRER RERM ERMR RMRR

0.0 0.2 0.4 0.6 0.8

O.D.

VRERMRR

0.0 0.2 0.4 0.6 0.8

O.D.

WSKSS SKSSV KSSW SSWG SWGW WGWP VGWPT GWPTV WPTVR PTVRE TVRER VRERM REPMR

0.0 0.2 0.4 0.6 0.8

O.D.

WG VGW GWP

VRE RER ERM RMR MRR

0.0 0.2 0.4 0.6 0.8

O.D.

FIG. I-Continued

published by Los Alamos National Laboratory. Ten different isolates were compared to the BRU isolate used in our assay, and the amount of deviant amino acids was summed up to give the variability index.

RESULTS

To identify the minimal size of NEF epitopes recognized by man, we used HIV-l-positive sera, which were selected for this study due to their reactivity with recombinant NEF in Western blot or with the earlier defined NEF epitopes (Gombert et al., 1990) in a pin-

ELISA assay. All 10 sera could recognize at least four of the nine antigenic regions tested; however, the specific recognition pattern of some of the sera (e.g., Nos. 6 and 9) was clouded by their high background reactivity. When sera were tested against peptides varying in size from three to eight amino acids, two reaction pat- terns were seen. With epitopes 2, 3D, and 6 the reactive region clearly consisted of a linear sequence varying in length from four to six amino acids (Figs. 1 a, 2a, 3, 4, 5). In contrast, epitopes 1, 3A, 3B, 3C, 4, and 5 displayed a more complex reaction pattern (Figs. 1 b, 2b, 3, 4, 5). The positive reactivity often seemed to be

TAHTINEN ET AL.

Amino terminus analysis Semn87, epitcpz 2

Peptide Orer) HGAITSSNTAATNAACAWLEAQEE

8 --_-____ --ttttttt I -----------++++t++

6 ------------++++++-

5 -------------+++++--

4 --------------t+++---

3 ---.------------+++----

-6-6-6-6-6-6-6-6-6-6-4-2 Ot2+4+6+6+5-4-3-2-1 @it* NAACAWLE

Carboxy tezminus analysis Sem #7, epitqe 2

Peptide orer) HGAITSSNTAATNAACAWLEAQEE

8 -_- _---- --t+t+++t I ----m------t++++++

6 ------------++++++-

5 -------------+++++--

4 --------------+t++---

3 ---------------tt+----

-l-2-3-4-5-6-6-6-6-6-6-6-6-6-6+6c6+6+4+2 O-2 Epitope CAWLEAQE

Consensus epitope CAWLE

Amino tenninw analysis Semm X7, epitope 1

mid 6~21) WSKSSVVGWPTVRERMRR

8 -tt --tttttt I --+t--tttttt 6 m-o++- -++++++ 5 +--- tt-ttttttt 4 +---- tttttttttt 3 +----- ttt-tttttt

O-4-2-2-2 0+2+6+6+4t6+5+4+3+2+1 Wss- WSK VVGWPTVRERMRR

Carboxy terminus analysis Smnn X7, epitqze 1

peptie (nerd WSKSSVVGWPTVRERMRR

8 -tt--tttttt 7 --tt--tttttt 6 -- -tt--tttttt 5 +-.-- tt-ttttttt 4 +---- tttttttttt 3 +--- --ttt-tttttt

tl 0-1-4-5-6+6t6-2-2+6t6+6+6t6+6 WSKS VGWP

TVRERMRR

consensus WSK VGWP epitome TVRERYRR

FIG. 2. (a) Matrix analysis of serum No. 7 against peptides from epitope 2 showing a typical contiguous reactivity. A consensus epitope deduced from separate amino- and carboxy-terminal analyses is shown below. (b) Matrix analysis of serum No. 7 against peptides from epitope 1 showing a typical noncontiguous reactivity.

concentrated into several small, down to three amino acid long regions, and the reactivity was even depen- dent on the localization (Iv- or C-terminus) of these regions within the peptide tested. Such epitopes also often showed variability between different sera. We therefore developed the matrix analysis method to de- termine which parts of the antigenic region were man- datory for a positive reaction.

All sera were tested against each epitope and the results of their matrix analysis are exemplified in Figs, 2a and 2b and summarized in Figs. 3, 4, and 5. The results of their homology search are shown in Table 1. Epitope 1 (Figs. 1 b, 2b, and 3) contains three different reactive stretches (WSK, VGW, TVRERMRR), the last of which gives the most intense color reaction. However, this sequence also showed reactivity with one of the control sera and was not shown to be clearly NEF-specific in the homology search (Table 1). In epitope 2, all HIV-l-positive sera recognized a uniform linear sequence (CAWLE) (Figs. 1 a, 2a, and 3), but the reactivities were quite low. Also, an exactly identical sequence

could be found, for example, in human, rat, and chicken estrogen-receptor proteins. Epitope 3A is again noncontiguous (Fig. 3), consisting of two reactive stretches (PLRPM, SHFLK), the latter of which was more often recognized by HIV-l-infected sera and which also gave stronger positive reactions. Neither of these stretches is clearly NEF-specific as similar sequences can be found in other proteins, even when only one of the five amino acids is allowed to vary (Ta- ble I).

The reactivities of the sera against epitopes 3B, 3C, and 3D are shown in Fig. 4. All three of these epitopes derived from the middle part of the NEF gave a strong reaction in ELISA. Epitope 3B (Fig. 4) contains two separate reactive stretches (SQRRQD, DLW), the longer of which is more specific for NEF (Table 1). However, similar sequences differing in only one amino acid can be found in four other proteins, interestingly including the p55v-fos transforming protein of mouse. Epitope 3C (Fig. 4) contains three different reactive stretches (IYHT, QGYFPDWQN, GVR), two of which are adjacent


Variation 000538500255000323

Epitope 2 E$dtcpe 3A HGAITS PI 101110011134017011101155 0000001050501000000100

EIV+ sera #l #2

#3

z WSK #6 #-i WSK #8

#9 K% #lO WSK

Im HFL SHFL

nKK PIBFM Hm

- HFL HFL

PIR SHETI SHET

HFLKEKG SHFLK

Control sera #l m - -

ENR

ti - - - - - -

#4 - - - #s - - -

FIG. 3. Reactivity of the sera against epitopes 1,2, and 3A. The scanned area for each epitope and the sequences recognized by the 10 patient sera and five control sera are shown in one letter amino acid code. A straight line indicates no reactivity. Below the scanned area, a variation index for each amino acid is also shown.

to each other, but still distinguishable with the matrix analysis. Of these three stretches, the middle one seems to be highly NEF-specific, because even when variability is allowed in two amino acid positions, no similar sequences can be found (Table 1). Epitope 3D (LTFGWC) (Fig. 4) is contiguous and also highly NEF- specific as no similar sequences can be found even when variability in one amino acid is allowed (Table 1).

Figure 5 shows the reactivity of the sera against epitopes 4, 5, and 6. Epitope 4 was previously shown to

be HIV MAL-isolate-specific and, in fact, only one of our sera was able to recognize it (Fig. 5). It also seems to contain two separate stretches, but further studies are needed to confirm the exact sequence and length of the epitope. Epitope 5, again, contains two antigenic regions (EVLEWRFDSR, VAR), the longer of which gives more intense color reactions. It was also found to be highly NEF-specific even when a 20% variability (2 amino acids of 10) was allowed. Epitope 6, instead, consists of only one linear stretch (PEYF) and gives

Epitope 3B Epitape 3C Epitope 3D ~IHsQRIp;;D~~ IY PLTFGcfKLvFvEi?

Variation 02002815811610111 813111311111111101071 002000082000250

EIV+ sera #l #2

Control sera

ii - - - - - -

#3 - - -

r5 - - - - - -

FIG. 4. Reactivity of the sera against epitopes 38, 3C, and 3D

162 TAHTINEN ET AL.

Variation

z;sw4 Epitope 5 Z-&

11200009341 3200805021400350900 9001500039071

EIV+ sera #l rit%EmmS~ VAR PEX - #2 IEM x3 SC VSL -m wa-

iz Fc!VLz Em -

- - -

; -

I.ismkG VAR -

- - 118 Im - -

EO - - r.23R6z -F&EK

Control sera #l - - -

E - - - - - -

iz - - - - - -

FIG. 5. Reactivity of the sera against epitopes 4, 5, and 6.

relatively high absorbancies especially when situated at the carboxy terminus. As a matter of fact, in this position it was also recognized by four additional sera (Nos. 6, 7, 8, and 9) but since we defined a sequence to be an epitope only if it shows reactivity regardless of its position in a peptide, the reactions of these four

TABLE 1

COMPUTER-BASEDHOMOLOGYSEARCHFORTHEB-CELLEPITOPES FOUNDIN HIV-1 NEF

Allowed Number of mismatch

NEF Amino acid homological (number of epitope sequence proteins found amino acids)

1 WSK 129 0 VGW 88 0 TVRERMRR 0 1

1 2 2 CAWLE 5 0 3A PLRPM 0 0

53 1 SHFLK 1 0

61 1 38 SQRRQD 0 0

4 1 DLW 80 0

3c IYHT 2 0 QGYFPDWQN 0 1

0 2 GVR 321 0

3D LTFGWC 0 0 0 1

4 Not done 5 EVLEWRFDSR 0 1

0 2 VAR 367 0

6 PEYF 7 0

other sera were not considered positive. This carboxy terminal reactivity may be caused by the extra alanine residue, which is included in every pin as the starting point of the peptide synthesis, but which may make the overall peptide a target for unspecific binding of antibodies.

Figures 3, 4, and 5 also show a variability index for each amino acid in the studied regions. Interestingly, some epitopes (e.g., 1, ?, 3A, 3C, 3D) are situated in an area with low variation. Also, while the actual antigenic regions of some nonlinear epitopes (e.g., 1, 3A) are conserved, the intervening sequences have rather high variability. However, the pattern of a variable, nonreacting stretch around a constant reactive epitope is not a uniform phenomenon, as constant but still nonreacting sequences can also be found between reactive areas (e.g., 3C, 4). Reactive epitopes may themselves also be located in a rather variable region (e.g., 3B, 5, 6).

DISCUSSION

The characterization of the host immune response toward NEF is of importance from several points of view. First, NEF seems to be expressed early in the infected cells and thus a cellular cytotoxic immune response toward NEF-expressing cells may destroy infected cells before the release of viral infectious parti- cles and, consequently, NEF could be of importance in HIV vaccine development. Second, antibodies toward NEF are seen in HIV-infected individuals and this NEF- specific humoral response can be used in improving diagnostic methods for HIV infection. Finally, as antibodies are often targeted against functionally active sites of a protein, the characterization of the defined


specificity of host immune response may be used for the characterization of the structure of NEF.

We have previously used both recombinant whole NEF protein (Ranki et al., 1990) and synthetic peptides (Gombert et al., 1990) to characterize the extent and nature of NEF-specific immune response in HIV-infected individuals. NEF-specific antibodies, recognized by Western blotting, were seen in 54% of infected individuals. Anti-NEF antibodies were also sometimes observed in HIV-exposed persons before seroconversion toward structural viral proteins, and this NEF-specific immune response could be taken as an indication of silent infection with low or lacking virus expression (Ranki et al., 1987). These findings were hampered by the observation that cross-reacting NEF antibodies were sometimes observed in non-HIV-exposed persons having some other infectious or au- toimmune disease, and even in healthy blood donors (Ranki eta/,, 1990). The immune response toward NEF thus seems to resemble that seen against HIV p24 core protein, toward which antibodies are found in ca. 1% of the normal population (Biberfeld et al., 1986).

In order to characterize which part of NEF the antibodies were targeted toward and which of the recognized NEF epitopes might be specific for HIV infection, we first used nine amino acid long synthetic peptides to characterize the antibody response toward NEF (Gombert et al., 1990). Nine reacting regions (1, 2, 3A, 3B, 3C, 3D, 4, 5, and 6) were observed and some of these regions were also recognized by sera from seronegative risk-group individuals. In order to define more specifically the sequences within HIV NEF that were recognized strictly by antibodies elicited by HIV infection, we undertook the present work where three to eight amino acid long synthetic peptides covering the previously identified sequences and the adjacent regions were analyzed for immunogenic reactivity.

The results obtained could be categorized into two kinds of reactivities. First, within some antigenic regions a clearly defined linear epitope could be identified. Second, with most regions the reactive pattern was irregular, variation being seen with different sera tested but also depending on the length of the peptide used. We therefore developed a modified matrix method where the reactivity of the peptide was evalu- ated in relation to its N- and C-terminal amino acid so that the obtained results with peptides of different sizes, but starting from or ending at the same amino acid, were all taken into account when the reactive region was defined. With this method, we could show that several of the previously identified longer antigenic regions contain, in fact, short intervening stretches of reactive and nonreactive areas. The reactive areas could sometimes represent only three amino acid long

TABLE 2

COMPARISONOF B-CELL,TH-CELL,ANDCTL EPITOPES RECOGNIZED IN HIV-1 NEF

B-cell epitope 1

B-Epitope

Th-Epitope

WSK...VGWP**** TVRERMM****

WSAIRERMRRA*

B-cell epitope 3A

B-epitope CTL-epitope CTL-epitope CTL-epitope Th-epitope

PLRPM....... SHFLK**** AAVDLSHFLKEK***

QVPLRPMTYK*** QVPLRPMTYK**

EVGFPVRPQVPLRP*

B-ceil epitope 3C

B-epitope

CTL-epitope CTL-epitope

IYHT**** QGYFPDWQN ..GVR****

YHTQGYFPDWQ*** TQGYFPDWQNYT***

NYTPGVRYPLT***

B-cell epitope 30

B-epitope CTL-epitope

LTFGWC**** GVRYPLTFGWCYKLVP***

B-ceil epitope 5

B-epitope Th-epitope

* Wentworth et al. ** Koenig et a/.

*** Cullmann et a/. **** Tahtinen et al.

EVLEWRFDSR.....VAR**** AEKEVLVWRFDSKL*

sequences, which were then separated from the next reactive sequence by equally short nonreactive areas.

The finding of noncontiguous linear epitopes could be due to at least three reasons. First, the closely situated reactive regions could, in fact, represent several contiguous independent epitopes recognized by different antibodies. Second, some of these closely situated regions could actually be formed by one longer epitope and the nonreactivity seen in the assays might be due to the fact that the sequences in these areas contain high variation. The nonreactivity would thus mean that the serum tested would not recognize the sequence in the BRU isolate, although it could still react strongly with a sequence corresponding to the individual’s au- tologous virus. Finally, the finding of noncontiguous epitopes could represent conformational epitopes.

In this respect, the finding that many of the B-cell epitopes within NEF overlap, either partially or totally, with the identified helper T-cell or cytotoxic T-cell epitopes of NEF (Table 2), is of special interest. Helper T-cell epitopes represent peptides that can bind simul- taneously to the MHC II molecule and to the immuno-

164 T;iHTINEN ET AL

logically specific T-cell receptor (TCR), and thus the amino acids reacting with TCR may be located nonlin- early, like on one side of an amphipathic a-helix. It would theoretically be possible that an antibody molecule would also recognize the noncontiguously located amino acids in a fashion similar to that of the T-cell receptor. However, finding such conformational epitopes with linear synthetic peptides is extremely dif- ficult and other studies are required to confirm that hypothesis. Also; in view of the recent findings regard- ing the groove of the MHC I molecule and the motifs of octa- and nonapeptides representing CTL epitopes, the assumption of a-helical oligopeptides bound to MHC I is not very likely (Falk et al., 1991).

The identification of the minimal size of the epitopes recognized by human sera of HIV-infected individuals allows us to study further the epitopes within NEF that might be suitable for diagnostic assays. In view of the fact that cross-reacting antibodies toward NEF were seen in noninfected individuals, we looked at sequence homologies between the newly identified epitopes in NEF and known protein sequences. With some epitopes, a variety of similar sequences could be identified (Table 1) and, interestingly, many of these were associated with regulation of cell proliferation and/or differentiation. On the other hand, several of the epitopes (i.e., the middle part of the epitope 3C, 3D, and the beginning of epitope 5) seem to be genuinely NEF-specific and we are currently testing whether these could be used in the development of an HIV-1 infection-specific NEF peptide ELISA.

ACKNOWLEDGMENTS

The study was supported by the Academy of Finland (M.T., A.R., and K.K.), the American Foundation for AIDS Research (AmFAR) (A.R. and K.K.), EC Grant MR4*/0188/SF, and the German Bundes- ministerium f[lr Forschung und Technologie (Project G. Jung, FKZ II-01 2-86).

REFERENCES

AMEISEN, J.-C., GUY, B., CHAMARET, S., LOCHE, M., MOUTON, Y., NEY- RINCK, J.-L., KHALIFE, J., LEPREVOST, C., BEAUCAIRE, G., BOUTILLON, C., GRAS-MASSE, H., MANIEZ, M., KIENY, M.-P., LAUSTRIAT, D., BER- THIER, A., MACH, B., MONTAGNIER, L., LECOCQ. J.-P., and CAPRON, A. (1989). Antibodies to the NEF protein and to NEF peptides in HIV-1 infected seronegative individuals. AlDS Res. Hum. Retroviruses 5, 279-291.

BIBERFELD, G., BRENBERG-RADEN, U., B~TTLIGER, B., PUTKONEN, P.-O., BLOMBERG, I., JUTO, P., and WADELL, G. (1986). Blood donor sera with false-positive Western blot reactions to human immunodeficiency virus. Lancef ii, 289-290.

CHEINGSONG-POPOV, R., PANAGIOTIDI, C., Au, M., BOWCOCK, S., WAT- KINS, P., ARONSTAM, A., WASSEF, M., and WEBER, J. (1990). Antibod- ies to HIV-1 nef (p27): Prevalence, significance and relationship to seroconversion. AlDS Res. Hum. Retroviruses 6, 1099-l 105.

CULLMAN, B., GOMARD, E., KIENY, M.-P., GUY, B., DREYFUS, F., SAIMOT, A.-G., SERENI, D., SICARD, D., and LEW, J.-P. (1991). Six epitopes reacting with human cytotoxic CD8+T cells in the central region of the HIV-1 NEF protein. /. Immunol. 146, 1560-l 565.

FALK, K., R~STZSCHKE, O., STEVANOVIC, S., JUNG, G., and RAMMENSEE, H.-G. (1991). Nature 351, 290-293.

GEYSEN, M., RODDA, J. S., MASON, J. T., TRIBBICK, G., and SHOOFS, P. G. (1987). Strategies for epitope analysis using peptide synthesis. J. Immunol. Methods 102, 259-274.

GEYSEN, M. H., MASON, T. J., and RODDA, S. J. (1988). Cognitive fea- tures of continuous antigenic determinants. J. Mol. Rec. 1, 32-39.

GOMBERT, F. O., BLECHA, W., TZ~HTINEN, M., RANKI, A., PFEIFER, S., TR~GER, W., BRAUN, R., MOLLER-LANTZSCH, N., JUNG, G., ROBSAMEN- WAIGMANN, H., and KROHN, K. (1990). Antigenic epitopes of NEF proteins from different HIV-1 strains as recognized by sera from patients with manifest and latent HIV infection. Virology 176,458- 466.

KOENIG, S., FUERST, T. R., WOOD, L. V., WOODS, R. M., SUZCH, J. A., JONES, G. M., DE LA CRUZ, V., DAVEY, R. T., VENKATESAN, S., Moss, B., BIDDISON, W. E., and FAUCI, A. S. (1990). Mapping of the fine specificity of a cytolytic T cell response to HIV-1 nef protein. J. Immunol. 14.5, 127-135.

KROHN, J. E., OVOD, V., LAGERSTEDT, A., SAARIALHO-KERE, U., RANKI, A., GOMBERT, F. O., JUNG, G., and MULLING, K. (1991). Expression kinetics and cellular localization of HIV TAT and NEF proteins in relation to the expression of viral structural mRNA and viral parti- cles. In “Vaccines-91” (R. A. Chanock, H. S. Ginsberg, F. Brown, and R. A. Lerner, Eds.), pp. 97-l 01. Cold Spring Harbor Labora- tory, Cold Spring Harbor, NY.

NIEDERMANN, T. M. J., THIELAN, B. J., and RATNER, L. (1989). Human immunodeficiency virus type 1 negative factor is a transcriptional silencer. Proc. Nat/. Acad. Sci. USA 86, 1 128-l 132.

RANKI, A., VALLE, S.-L., KROHN, M., ANTONEN, J., ALLAIN, J.-P., LEUTHER, M., FRANCHINI, G., and KROHN, K. (1987). Long latency precedes overt seroconversion in sexually transmitted human immunodeficiency virus infection. Lancer ii, 589-593.

RANKI, A., JARVINEN, K., VALLE, S.-L., NURMILAAKSO, P., and KROHN, K. (1990). Antibodies to recombinant HIV-1 NEF protein detected in HIV-1 infection as well as in nonrisk individuals. J. Acquired Im- mune Deft. Syndr. 3, 348-355.

SABATIER, J.-M., CLERGET-RASLAIN, B., FONTAN, G., FENOUILLET, E., Ro- CHAT, H., GARNIER, C., GLUCKMAN, J.-C., VAN RIETSCHOTEN, J., MON- TAGNIER, L., and BAHRAOUI, E. (1989). Use of synthetic peptides for the detection of antibodies against the NEF regulating protein in sera of HIV-infected patients. AlDS 3, 2 15-220.

SCHNEIDER, T., HARTHUS, H.-P., HILDEBFUINDT, P., NIEDRIG, M., BROKER, M., WEIGELT, W., BECK, A., and PAULI, G. (1991). Epitopes of the HIV-1 negative factor (nef) reactive with murine monoclonal antibodies and human HIV-1 positive sera. AIDS Res. Hum. Retro- viruses 7, 37-44.

WENTWORTH, P. A., ABRIGNANI, S.. BATHURST, I. C., FEUCHT, P. H., DOE, B., LEE, D. J., and STEIMER, K. S. (1991). Human HIV-l Nef- specific CD4+ T cell clones. In “Vaccines-91” (R. A. Chanock, H. S. Ginsberg, F. Brown, and R. A. Lerner, Eds.), pp. 65-69. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY.

fine specificity of the b-cell epitopes recognized in hiv-1 nef by human sera

Documents