_Archives
Vi rology
Arch Virol (1992) 126:12%146
© Springer-Verlag 1992 Printed in Austria
Comparison and fine mapping of both high and low neutralizing monoclonal antibodies against the principal neutralization domain of HIV-1 J. P. M. Langedijk 1, Nicole K. T. Back 2, Elaine Kinney-Thomas 3, Claudine Bruck 4, Myriam Francotte 4, J. Goudsmit 2, and R. H. Meloen 1 1Central Veterinary Institute, Lelystad 2Human Retrovirus Laboratory, Academic Medical Center, Amsterdam, The Netherlands 30ncogen, Seattle, Washington, U.S.A. 4 Smith Kline Biologicals, Rixensart, Belgium
Accepted February 3, 1992
Summary. Monoclonal antibodies raised against viral lysate of HIV-1 (strain LAV-1) and against recombinant gp 160 of HIV-1 (strain HTLV IIIB) which neutralized HIV- 1 in a type specific manner were mapped with the aid ofpeptides (Pepscan analysis). Each of these monoclonal antibodies bound to peptides located on the principal neutralizing domain (PND) of HIV-1. We found that the antigenic sites of the MAbs described in this paper are represented by linear peptides of at least 10 amino acids long. The affinity of the MAbs is high for these peptides and in the same order of magnitude as for native gp 160. The fine mapping of the epitopes may reflect structural features of the PND, for instance which amino acid side chains are exposed and which are buried in the protein. Furthermore the fine mapping of the epitopes explained the HIV typespecifc neutralizing activity of the MAbs. Antibodies that bound to the tip of the loop (amino acids QRGPGRAF) have a higher neutralizing activity than antibodies that bound to amino acids towards the N-terminal side of the loop (amino acids KSIRI). Furthermore, MAbs that bound to virtually the same amino acids on the tip of the loop (amino acids IQRGPGRAF and RGPGRAFV) had different neutralizing activities due to different affinities for native gp 160. These data reveal that neutralizing activity not only is determined by the affinity of an antibody to the neutralizing site but also by its fine binding specificities to the V 3 loop of gp 120.
Introduction The human immunodeficiency virus (HIV) is the etiological agent of the acquired immunodeficiency syndrome (AIDS) [7]. The most important targets for HIV
130
J.P.M. Langedijk etal.
infection appear to be the monocytes and T-lymphocytes that express the primary cellular receptor for the virus, the CD4 glycoprotein. The infection is mediated by binding of the external envelope glycoprotein (gp 120) with CD 4 [26, 12, 42], followed by the fusion of the virus particle with the lymphocyte. Antibodies against gp 120 of HIV-1 may play an important role in preventing infection by the virus 1,17, 39, 37, 40]. Isolate-restricted neutralizing antibodies bind to a highly variable part of gp 120 (amino acids 296-331) within the third variable region (V 3) [41, 22, 35]. This 35 amino acid sequence is contained in a loop between two cysteines which form a disulfide bridge 1,29]. A conserved secondary structure in this hypervariable region is the predicted type 2 reversed turn 1-28] on the tip of the loop, which is mostly constituted of the amino acid sequence GPGR. This immunodominant principal neutralization domain (PND) induces type-specific neutralizing and cell fusion inhibiting antibodies 1-41]. The neutralization mechanism of these antibodies takes place after gp 120 binds to CD 4 [32, 46, 30, 4, 5]. The PND appears to have an unknown but pivotal functional role in infection and cell tropism which is blocked by the neutralizing antibodies against this region. Because of the critical function of the PND during HIV-1 infection [56, 53, 24] and because antibodies induced by the PND peptide may be of major importance to prevent infection with HIV-I, we studied in detail the epitopes of a set of MAbs with different neutralizing activities to understand the criteria for a high neutralizing antibody and to get more insight in the surface structure of the PND. Three different Pepscan analysis approaches were performed to define antibody specificity at the amino acid level: firstly by using peptides in which each amino acid was deleted one at a time, secondly by using overlapping peptides of multiple length and thirdly, by using peptide analogues in which each amino acid position was substituted by every naturally occurring amino acid. The affinity of each MAb for both the PND peptide and for native gp 160 was determined. The results show which amino acids are exposed at the PND surface and that MAbs with high affinity for the tip of the PND on gp 160 have the highest neutralizing activity; MAbs which are directed to amino acids which flank the tip of the PND have lower neutralizing activity even if they bind to the PND with high affinity. Material and methods Monoctonal antibodies
MAbs 110.3, 110.4, 110.5, and 110.6 were raised against viral lysate of the LAV-1 isolate, which has the same amino acid sequence in the PND as HIV-1, type BH 10. The MAbs reacted against gp 120 of HIV-1, type LAV-1 as described [25]. MAb 178.1 was raised with yeast recombinant gp 160 of the HIV-1, type BH 10 isolate as described [54]. MAb 5023A was raised against a PND peptide with the amino acid sequence RIQRGPGRAFVTIGK (HIV-I type BH 10) coupled to chicken ovalbumin with glutaraldehyde as described [16].
Mapping of HIV- 1 neutralizing antibodies
131
Recombinant protein HIV-1 gp 160, used for affinity measurements, was expressed in mammalian cells by using the vaccinia vector system as described [54]. The recombinant virus (vaccinia-gp 160) contained the complete coding sequence of the HIV-1 env gene of molecular clone BH 10 (nucleotides 5802-8487 [38]).
Peptide synthesis Peptides with the amino acid sequence RIQRGPGRAFVTIGK and CKSIRIQRGPGRAFVTC, used for affinity measurements and Ab titer measurements, were synthesized using a slightly modified solidphase synthesis method [6, 43].
Pepscan Peptides were synthesized on polyethelene rods and tested for their reactivity with MAbs in an enzyme linked immunosorbent assay (ELISA) according to established procedures [20, 21]. The following peptides were synthesized. All 536 overlapping 9-amino acid long peptides of HIV-1 gp 120; deletion peptides for R I Q R G P G R A F V T I G K and KSIRIQRGPGRA in which each amino acid was consecutively deleted were synthesized and tested; overlapping 2 to 18 amino acid long peptides from the region between amino acids 296-331 and a set of analogues for nonapeptides KSIRIQRGP, I Q R G P G R A F and RGPGRAFVT in which each amino acid was consecutively replaced by the other 19 natural occurring amino acids. Amino acid sequence was derived from the nucleotide sequence of HTLVIIIB-BH 10 [38]. The purity of the peptides covalently linked to the rods is not tested standardly. However, possible impurites like deletions and truncations of peptides might in many cases not react but will not disturb the reaction with the prototype peptide which is also present. When peptides with the same amino acid sequences are synthesized in a classical fashion, the reactivities closely resemble those of the peptides used in Pepscan analysis.
Neutralization assay Syncytium inhibition assays using HIV- 1 types HX 10 and HXB 3 were performed according to Back et al. [4] using virus producing H 9 cells and C 8166 as target cells. The neutralization titers of the panel of MAbs were determined in the same assay on the same day using a defined TCIDs0. The neutralization titers are the average of three independent assays.
Serology The Ab titers were determined in a indirect ELISA with mirotiter plates coated with an excess of the peptide CKSIRIQRGPGRAFVTC (1 ~tg/well), as described E44, 52].
Affinity measurements The affinity constants (Karl) for the PND peptides and gp 160 were determined using an ELISA as described by Friguet et al. [19], with adaptations to bivalent binding as described by Stevens [50]. This involved the determination of the concentration of free antibody at equilibrium between the MAb and gp 160, CKSIRIQRGPGRAFVTC or RIQRGPGRAFVTIGK. This method produces accurate affinity constants provided that there is a linear correlation
132
J . P . M . Langedijk etal.
between the absorbance and the range of antibody concentrations and, moreover, that readjustment of the equilibrium is negligible in the liquid phase during incubation of the mixture in the coated wells. This was verified using established procedures described by Friguet et al. [19]. Both conditions were satisfied in the experiments. The following equation is used to deduce the dissociation constant: A0/(A0- A) = 1 + Kd/a0, where A0 is the absorbance without blocking peptide; A, absorbance with blocking peptide; a0, peptide or protein concentration; Kd, dissociation constant; Ka~ = 1/Kd, affinity constant.
Modelling Sybyl version 5.41 was used for molecular modelling (Tripos ass., St. Louis, Mo., U.S.A.). We have used the secondary structure predictions described by LaRosa et al. [28]. A J3hairpin loop of REI Immunoglobulin which has a sequence similarity with the PND of gp 120 [3t] was used as a starting conformation rather than an ab initio conformational search algorithm. The Sybyl force force field was used; the minimum energy conformation was 200 kJ/mol.
Results
Locating epitopes with overlappingnonapeptides M A b s 110.3, 110.4, 110.5, 110.6, a n d 178.1 were tested by P e p s c a n for their reactivity with all 536 o v e r l a p p i n g n o n a p e p t i d e s o f H T L V - I I I - B gp 120 (Fig. 1). M A b s 110.3 a n d 110.4 s h o w e d the same specificities as M A b 110.5 in all experiments, therefore only the results o f M A b 110.5 are s h o w n in this report. M A b 110.5 b o u n d to a peptide with the a m i n o acid sequence I Q R G P G R A F ( a m i n o acid 309-317); M A b 110.6 b o u n d to a peptide with the a m i n o acid IQRGPGRAF 2
M A b 110.5
'
' 3"00
'
'
' 350' '
QRGPGRAFV 0.51
M A b 110.6
Fig. 1. Pepscan with overlapping peptides. Plot of 8 ] Illl]ltlullllUlf1lt1tl1tl11111111 ltlJltllllLIllli$11hml ttlttl ELISA-reactivity of a section of the individual o
2
'
]
'36o
M A b 178.1
. . . .
a o*
KSIRIQRGP
o 0
'a6o
. . . .
a, o*
overlapping nonapeptides with three different monoclonal antibodies: A 110.5, B 110.6, and C 178.1. The other nonapeptides did not show any reactivity with the MAbs. The ascites fluids were tested at a 1:250 dilution. The numbers on the horizontal axis correspond to the N-terminal amino acid of the nonapeptide. Absorbances at 405 nm were plotted vertically
Mapping of HIV-1 neutralizing antibodies
133
sequence QRGPGRAFV (amino acid 310-318) and MAb 178.1 bound to a peptide with amino acid sequence KSIRIQRGP (amino acid 305-313). The precise length of the antigenic site does not directly follow from these data. For instance the epitope of MAb 178.1 may be contained within the 4 amino acids which are shared by the 5 successive reactive nonapeptides (i.e., KSIRI), the epitope could also be longer (i.e., NTRKSIRIQRGP) and theoretically the epitope could contain an additional part located elsewhere in the primary structure. To estimate the size of the part of the epitope located on the PND we determined the reactivity of the MAbs with peptides in which amino acids were systematically deleted (deletion peptides) and overlapping peptides of multiple length [36, 1].
Determination of core sequences of epitopes The binding of MAb 110.5 to the deletion peptides is shown in Fig. 2. The deletion of amino acid R (311), G (312), P (313), G (314), R (315), A (316) and F (317) resulted in a loss of binding, and the deletion of I (309), Q (310) and K (322) resulted in a reduction of the binding by more than 50%. Binding of MAb 110.6 with the PND peptides was completely lost when amino acid R (311), G (312), P (313), G (314), R (315), A (316) and F (317) were deleted and the binding was reduced by more than 50% when V (318) and K (322) were deleted. For MAb 178.1 the deletion of K (305), S (306), I (307), R (308), I (309) resulted in loss of binding and the deletion of G (3 t2) and R (313) reduced the binding by more than 50%, while the rest of the deletions had no influence on the binding. Thus the core sequence for MAb 110.5 and 110.6 is RGPGRAF (amino acid 311-317), and for MAb 178.1 it is KSIRI (amino acid 305-309).
Binding contribution determined at single amino acid level MAbs 110.5, 110.6, and 178.1 were tested for their reactivity with overlapping peptides of the PND (amino acid 296-331) varying in length from 2 to 18 amino acids to determine the core (shortest reactive peptide) and the extent of the epitope recognized by each MAb. Representative results obtained for MAb 110.5 are shown in Fig. 3. When the reactivity of antibodies for each pair of peptides differing in length by one amino acid were compared with each other it was possible to estimate the contribution of a specific amino acid to the binding. For instance, in Fig. 3 the absorbance for the nonapeptide (309) IQRGPGRAF is 3.5 times higher than the absorbance for (310) QRGPGRAF, therefore we assume that the I (309) in the peptide sequence (309) IQRGPGRAF contributes to binding. The binding contribution of each amino acid flanking the core sequence was calculated and is expressed as arbitrary units in Fig. 4. The binding contribution was calculated as follows: For An > A~n- 1~, B = [~;(An/A(n_ 1)I/N; for An < A(n_ 1) a negative contribution is calculated as follows:
134
J . P . M . Langedijk et al. MAb 110,5 800 700 600 500 400 300 200 100 0 noR
8
I
1o°ooooooi
Q R G P G R A
F V T
I
G K
omitted amino acids MAb 110.6 1400 1200 1000 800
8
600 400 200 no R b
i,ooooooo
I Q R G P G R A F V T
I
G K
omi~edaminoacids MAb 178.1
1200 1000 800 8
600 400 200 0
C
P G R A n o K S I R I Q R G omiRed amino acids
Fig. 2. Pepscan with deletion peptides and MAb 110.5 (a), 110.6 (b), and 178.1 (e). Antibody binding activity for the deletion peptides are proportional to the vertical bars. • Reactivity of the original sequence without any deletions. The amino acid deleted from the original sequence is indicated on the horizontal axis. The set of deletion analogues were derived from RIQRGPGRAFVTIGK for MAbs 110.5 and 110.6 and KSIRIQRGPGRA for MAb 178.1
Mapping of HIV-1 neutralizing antibodies
135
LL < rr (..9
°]
LL < Cr L9 m (5 rr 0
1
o
0
i
lll}lhIMIMhl}lhllll!llIll ~llit~liHilllllhlillllilllIllll lilillllllllh,lh!lllllhilllIlt llllhmllllhMktlhi~lhh rlllhl,lhlh!lllhhuilllIl illllllllMh,hhlllElllll ~,ll111h*l,llihlhMllll ,lhn,l,ill i*ililltll,ii 2
3
5
4
6
7
8
9
Vi rology
Arch Virol (1992) 126:12%146
© Springer-Verlag 1992 Printed in Austria
Comparison and fine mapping of both high and low neutralizing monoclonal antibodies against the principal neutralization domain of HIV-1 J. P. M. Langedijk 1, Nicole K. T. Back 2, Elaine Kinney-Thomas 3, Claudine Bruck 4, Myriam Francotte 4, J. Goudsmit 2, and R. H. Meloen 1 1Central Veterinary Institute, Lelystad 2Human Retrovirus Laboratory, Academic Medical Center, Amsterdam, The Netherlands 30ncogen, Seattle, Washington, U.S.A. 4 Smith Kline Biologicals, Rixensart, Belgium
Accepted February 3, 1992
Summary. Monoclonal antibodies raised against viral lysate of HIV-1 (strain LAV-1) and against recombinant gp 160 of HIV-1 (strain HTLV IIIB) which neutralized HIV- 1 in a type specific manner were mapped with the aid ofpeptides (Pepscan analysis). Each of these monoclonal antibodies bound to peptides located on the principal neutralizing domain (PND) of HIV-1. We found that the antigenic sites of the MAbs described in this paper are represented by linear peptides of at least 10 amino acids long. The affinity of the MAbs is high for these peptides and in the same order of magnitude as for native gp 160. The fine mapping of the epitopes may reflect structural features of the PND, for instance which amino acid side chains are exposed and which are buried in the protein. Furthermore the fine mapping of the epitopes explained the HIV typespecifc neutralizing activity of the MAbs. Antibodies that bound to the tip of the loop (amino acids QRGPGRAF) have a higher neutralizing activity than antibodies that bound to amino acids towards the N-terminal side of the loop (amino acids KSIRI). Furthermore, MAbs that bound to virtually the same amino acids on the tip of the loop (amino acids IQRGPGRAF and RGPGRAFV) had different neutralizing activities due to different affinities for native gp 160. These data reveal that neutralizing activity not only is determined by the affinity of an antibody to the neutralizing site but also by its fine binding specificities to the V 3 loop of gp 120.
Introduction The human immunodeficiency virus (HIV) is the etiological agent of the acquired immunodeficiency syndrome (AIDS) [7]. The most important targets for HIV
130
J.P.M. Langedijk etal.
infection appear to be the monocytes and T-lymphocytes that express the primary cellular receptor for the virus, the CD4 glycoprotein. The infection is mediated by binding of the external envelope glycoprotein (gp 120) with CD 4 [26, 12, 42], followed by the fusion of the virus particle with the lymphocyte. Antibodies against gp 120 of HIV-1 may play an important role in preventing infection by the virus 1,17, 39, 37, 40]. Isolate-restricted neutralizing antibodies bind to a highly variable part of gp 120 (amino acids 296-331) within the third variable region (V 3) [41, 22, 35]. This 35 amino acid sequence is contained in a loop between two cysteines which form a disulfide bridge 1,29]. A conserved secondary structure in this hypervariable region is the predicted type 2 reversed turn 1-28] on the tip of the loop, which is mostly constituted of the amino acid sequence GPGR. This immunodominant principal neutralization domain (PND) induces type-specific neutralizing and cell fusion inhibiting antibodies 1-41]. The neutralization mechanism of these antibodies takes place after gp 120 binds to CD 4 [32, 46, 30, 4, 5]. The PND appears to have an unknown but pivotal functional role in infection and cell tropism which is blocked by the neutralizing antibodies against this region. Because of the critical function of the PND during HIV-1 infection [56, 53, 24] and because antibodies induced by the PND peptide may be of major importance to prevent infection with HIV-I, we studied in detail the epitopes of a set of MAbs with different neutralizing activities to understand the criteria for a high neutralizing antibody and to get more insight in the surface structure of the PND. Three different Pepscan analysis approaches were performed to define antibody specificity at the amino acid level: firstly by using peptides in which each amino acid was deleted one at a time, secondly by using overlapping peptides of multiple length and thirdly, by using peptide analogues in which each amino acid position was substituted by every naturally occurring amino acid. The affinity of each MAb for both the PND peptide and for native gp 160 was determined. The results show which amino acids are exposed at the PND surface and that MAbs with high affinity for the tip of the PND on gp 160 have the highest neutralizing activity; MAbs which are directed to amino acids which flank the tip of the PND have lower neutralizing activity even if they bind to the PND with high affinity. Material and methods Monoctonal antibodies
MAbs 110.3, 110.4, 110.5, and 110.6 were raised against viral lysate of the LAV-1 isolate, which has the same amino acid sequence in the PND as HIV-1, type BH 10. The MAbs reacted against gp 120 of HIV-1, type LAV-1 as described [25]. MAb 178.1 was raised with yeast recombinant gp 160 of the HIV-1, type BH 10 isolate as described [54]. MAb 5023A was raised against a PND peptide with the amino acid sequence RIQRGPGRAFVTIGK (HIV-I type BH 10) coupled to chicken ovalbumin with glutaraldehyde as described [16].
Mapping of HIV- 1 neutralizing antibodies
131
Recombinant protein HIV-1 gp 160, used for affinity measurements, was expressed in mammalian cells by using the vaccinia vector system as described [54]. The recombinant virus (vaccinia-gp 160) contained the complete coding sequence of the HIV-1 env gene of molecular clone BH 10 (nucleotides 5802-8487 [38]).
Peptide synthesis Peptides with the amino acid sequence RIQRGPGRAFVTIGK and CKSIRIQRGPGRAFVTC, used for affinity measurements and Ab titer measurements, were synthesized using a slightly modified solidphase synthesis method [6, 43].
Pepscan Peptides were synthesized on polyethelene rods and tested for their reactivity with MAbs in an enzyme linked immunosorbent assay (ELISA) according to established procedures [20, 21]. The following peptides were synthesized. All 536 overlapping 9-amino acid long peptides of HIV-1 gp 120; deletion peptides for R I Q R G P G R A F V T I G K and KSIRIQRGPGRA in which each amino acid was consecutively deleted were synthesized and tested; overlapping 2 to 18 amino acid long peptides from the region between amino acids 296-331 and a set of analogues for nonapeptides KSIRIQRGP, I Q R G P G R A F and RGPGRAFVT in which each amino acid was consecutively replaced by the other 19 natural occurring amino acids. Amino acid sequence was derived from the nucleotide sequence of HTLVIIIB-BH 10 [38]. The purity of the peptides covalently linked to the rods is not tested standardly. However, possible impurites like deletions and truncations of peptides might in many cases not react but will not disturb the reaction with the prototype peptide which is also present. When peptides with the same amino acid sequences are synthesized in a classical fashion, the reactivities closely resemble those of the peptides used in Pepscan analysis.
Neutralization assay Syncytium inhibition assays using HIV- 1 types HX 10 and HXB 3 were performed according to Back et al. [4] using virus producing H 9 cells and C 8166 as target cells. The neutralization titers of the panel of MAbs were determined in the same assay on the same day using a defined TCIDs0. The neutralization titers are the average of three independent assays.
Serology The Ab titers were determined in a indirect ELISA with mirotiter plates coated with an excess of the peptide CKSIRIQRGPGRAFVTC (1 ~tg/well), as described E44, 52].
Affinity measurements The affinity constants (Karl) for the PND peptides and gp 160 were determined using an ELISA as described by Friguet et al. [19], with adaptations to bivalent binding as described by Stevens [50]. This involved the determination of the concentration of free antibody at equilibrium between the MAb and gp 160, CKSIRIQRGPGRAFVTC or RIQRGPGRAFVTIGK. This method produces accurate affinity constants provided that there is a linear correlation
132
J . P . M . Langedijk etal.
between the absorbance and the range of antibody concentrations and, moreover, that readjustment of the equilibrium is negligible in the liquid phase during incubation of the mixture in the coated wells. This was verified using established procedures described by Friguet et al. [19]. Both conditions were satisfied in the experiments. The following equation is used to deduce the dissociation constant: A0/(A0- A) = 1 + Kd/a0, where A0 is the absorbance without blocking peptide; A, absorbance with blocking peptide; a0, peptide or protein concentration; Kd, dissociation constant; Ka~ = 1/Kd, affinity constant.
Modelling Sybyl version 5.41 was used for molecular modelling (Tripos ass., St. Louis, Mo., U.S.A.). We have used the secondary structure predictions described by LaRosa et al. [28]. A J3hairpin loop of REI Immunoglobulin which has a sequence similarity with the PND of gp 120 [3t] was used as a starting conformation rather than an ab initio conformational search algorithm. The Sybyl force force field was used; the minimum energy conformation was 200 kJ/mol.
Results
Locating epitopes with overlappingnonapeptides M A b s 110.3, 110.4, 110.5, 110.6, a n d 178.1 were tested by P e p s c a n for their reactivity with all 536 o v e r l a p p i n g n o n a p e p t i d e s o f H T L V - I I I - B gp 120 (Fig. 1). M A b s 110.3 a n d 110.4 s h o w e d the same specificities as M A b 110.5 in all experiments, therefore only the results o f M A b 110.5 are s h o w n in this report. M A b 110.5 b o u n d to a peptide with the a m i n o acid sequence I Q R G P G R A F ( a m i n o acid 309-317); M A b 110.6 b o u n d to a peptide with the a m i n o acid IQRGPGRAF 2
M A b 110.5
'
' 3"00
'
'
' 350' '
QRGPGRAFV 0.51
M A b 110.6
Fig. 1. Pepscan with overlapping peptides. Plot of 8 ] Illl]ltlullllUlf1lt1tl1tl11111111 ltlJltllllLIllli$11hml ttlttl ELISA-reactivity of a section of the individual o
2
'
]
'36o
M A b 178.1
. . . .
a o*
KSIRIQRGP
o 0
'a6o
. . . .
a, o*
overlapping nonapeptides with three different monoclonal antibodies: A 110.5, B 110.6, and C 178.1. The other nonapeptides did not show any reactivity with the MAbs. The ascites fluids were tested at a 1:250 dilution. The numbers on the horizontal axis correspond to the N-terminal amino acid of the nonapeptide. Absorbances at 405 nm were plotted vertically
Mapping of HIV-1 neutralizing antibodies
133
sequence QRGPGRAFV (amino acid 310-318) and MAb 178.1 bound to a peptide with amino acid sequence KSIRIQRGP (amino acid 305-313). The precise length of the antigenic site does not directly follow from these data. For instance the epitope of MAb 178.1 may be contained within the 4 amino acids which are shared by the 5 successive reactive nonapeptides (i.e., KSIRI), the epitope could also be longer (i.e., NTRKSIRIQRGP) and theoretically the epitope could contain an additional part located elsewhere in the primary structure. To estimate the size of the part of the epitope located on the PND we determined the reactivity of the MAbs with peptides in which amino acids were systematically deleted (deletion peptides) and overlapping peptides of multiple length [36, 1].
Determination of core sequences of epitopes The binding of MAb 110.5 to the deletion peptides is shown in Fig. 2. The deletion of amino acid R (311), G (312), P (313), G (314), R (315), A (316) and F (317) resulted in a loss of binding, and the deletion of I (309), Q (310) and K (322) resulted in a reduction of the binding by more than 50%. Binding of MAb 110.6 with the PND peptides was completely lost when amino acid R (311), G (312), P (313), G (314), R (315), A (316) and F (317) were deleted and the binding was reduced by more than 50% when V (318) and K (322) were deleted. For MAb 178.1 the deletion of K (305), S (306), I (307), R (308), I (309) resulted in loss of binding and the deletion of G (3 t2) and R (313) reduced the binding by more than 50%, while the rest of the deletions had no influence on the binding. Thus the core sequence for MAb 110.5 and 110.6 is RGPGRAF (amino acid 311-317), and for MAb 178.1 it is KSIRI (amino acid 305-309).
Binding contribution determined at single amino acid level MAbs 110.5, 110.6, and 178.1 were tested for their reactivity with overlapping peptides of the PND (amino acid 296-331) varying in length from 2 to 18 amino acids to determine the core (shortest reactive peptide) and the extent of the epitope recognized by each MAb. Representative results obtained for MAb 110.5 are shown in Fig. 3. When the reactivity of antibodies for each pair of peptides differing in length by one amino acid were compared with each other it was possible to estimate the contribution of a specific amino acid to the binding. For instance, in Fig. 3 the absorbance for the nonapeptide (309) IQRGPGRAF is 3.5 times higher than the absorbance for (310) QRGPGRAF, therefore we assume that the I (309) in the peptide sequence (309) IQRGPGRAF contributes to binding. The binding contribution of each amino acid flanking the core sequence was calculated and is expressed as arbitrary units in Fig. 4. The binding contribution was calculated as follows: For An > A~n- 1~, B = [~;(An/A(n_ 1)I/N; for An < A(n_ 1) a negative contribution is calculated as follows:
134
J . P . M . Langedijk et al. MAb 110,5 800 700 600 500 400 300 200 100 0 noR
8
I
1o°ooooooi
Q R G P G R A
F V T
I
G K
omitted amino acids MAb 110.6 1400 1200 1000 800
8
600 400 200 no R b
i,ooooooo
I Q R G P G R A F V T
I
G K
omi~edaminoacids MAb 178.1
1200 1000 800 8
600 400 200 0
C
P G R A n o K S I R I Q R G omiRed amino acids
Fig. 2. Pepscan with deletion peptides and MAb 110.5 (a), 110.6 (b), and 178.1 (e). Antibody binding activity for the deletion peptides are proportional to the vertical bars. • Reactivity of the original sequence without any deletions. The amino acid deleted from the original sequence is indicated on the horizontal axis. The set of deletion analogues were derived from RIQRGPGRAFVTIGK for MAbs 110.5 and 110.6 and KSIRIQRGPGRA for MAb 178.1
Mapping of HIV-1 neutralizing antibodies
135
LL < rr (..9
°]
LL < Cr L9 m (5 rr 0
1
o
0
i
lll}lhIMIMhl}lhllll!llIll ~llit~liHilllllhlillllilllIllll lilillllllllh,lh!lllllhilllIlt llllhmllllhMktlhi~lhh rlllhl,lhlh!lllhhuilllIl illllllllMh,hhlllElllll ~,ll111h*l,llihlhMllll ,lhn,l,ill i*ililltll,ii 2
3
5
4
6
7
8
9
