Identification of HLA-DR9 (DRB1∗0901)-binding peptide motifs using a phage fUSE5 random peptide library

ELSEVIER

Identification of HLA-DRY (DRB l”O90 l)-Binding Peptide Motifs Using a Phage fUSE5 Random Peptide Library Shoji Fujisao, Sho Matsushita, Yasuharu Nishimura

Tohru Nishi, and

ABSTRACT: We identified HLA-DRB 1*0901-binding peptides by affinity-based selection of a phage random peptide library using the biotinylated DR9 complex. Analogue peptides with single amino acid residue substitutions of a DR9 binder revealed that two major anchors (WxxS, where x is any amino acid) play an essential role in binding to DR9. Determination of the binding affinity of synthetic wild-type-based analogue peptides showed that substituting W to F or L, and S to A, V, or F allow high affinity binding with DR9. Collectively, DR9binding peptide motifs identified in this study are char-

ABBREVIATIONS DR9BPl DRB 1*090 1 -binding

peptide

acteristic in that (a) only two anchors of the NH,-terminal half of binding peptides play important roles in binding, and (b) small neutral hydrophilic Ser is allowed as the second anchor for high-affinity binding, unlike the other DR-binding motifs heretofore reported. The implications of our results are discussed in light of the HLA-DR9associated susceptibility to juvenile-onset myasthenia gravis and systemic lupus erythematosus with antiphospholipid syndrome, in particular, T-cell responses to autoantigens.

1

INTRODUCTION HLA-DR9 (DRB1*0901) is the most frequent HLADRBl allele in the Japanese population and is more frequently observed in Orientals than in other ethnic groups (Japanese 30%, Caucasians 1%). Other studies have shown increased antigen frequencies of DR9 in some autoimmune diseases, including juvenile-onset myasthenia gravis {l] and systemic lupus erythematosus with antiphospholipid syndrome [H. Hashimoto and Y. Nishimura et al., manuscript in preparation). The p chain of the DR9 molecule has characteristic amino acid residues in “Lys, @l’Asp, P26Tyr, P28His, and P30Gly, which is distinct from other known DR alleles. Little is

known of the structural characteristics of DR9-binding peptides 121. In our previous studies [3] we identified the allele specificity of structural requirements for peptides bound to two DR4 subtypes (DRB1*0405 and DRBl*O406). These data predicted a T-cell epitope on the human insulin molecule in cases of insulin autoimwhich is strictly associated with mune syndrome, DRB1*0406 131. In the current study, we determined binding peptide motifs for HLA-DR9 using a phage random peptide library with 15-mer peptide inserts on amino-termini of PI11 minor coat protein molecules.

MATERIALS From the Division of Immunogenetics, Department of Neuroscience and Immunology, Kumumoto University Graduate Scboo( of Medical Sciences (S. F., S.M., Y.N.) and the Department of Neurosurgery (T.N.), Kumamoto University School of Medicine, Kumamoto, Japan. Address reprint requests to Dr. S. Matsushita or I’. Nisbimura, Division of Immunogenetics, Department of Neuroscience and Immunology, Kumamoto University Graduate School of Medical Sciences, Honjo 2-2-1, Kumamoto 860, Japan. Received June 2, 1995; accepted October 27, 1995. Human Immunology 45, 131-136 (1996) 0 American Society for Histocomparibility

and Immunogenetics,

1996

AND

METHODS

Construction of the IS-mer random peptide library. The library was in principle constructed as described [4], but with some modifications (manuscript in preparation). Bacteriophage fUSE5 vector and Escbwichia cofi KY1 and MC1061 were kindly provided by Dr. G. Smith of the University of Missouri. A mixture of oligonucleotides encoding

for possible

15-amino

acid

peptides

was syn-

019%88591961$15.00 SSDI 0198~8859(95)00169-7

132

S. Fujisao et al.

thesized with the sequence 5’-ACTCGGCCGACGGGGCT(NNK), ,GGGGCCGCTGGGGCCGAA3’(TN-l), in which N stands for an equal mixture of the deoxynucleotides G, A, T, and C, and K stands for a equal mixture of G and T. TN-1 single-strand degenerative oligonucleotides were converted to double-strand DNAs by polymerase chain reaction (PCR) amplification with 5’-biotinylated primers (TN-2: 5’ACTCGGCCGACGGGGC-3’, identical to the 5’ end of TN-1 and TN-3: 5’-TTCGGCCCCAGCGGCCC-3’, complementary to the 3’ end of TN-l). The doublestrand DNA fragments were digested with BglI and small fragments from both ends were absorbed by mixing with streptavidin-agarose beads. The sfi1 digested fUSE5 vector was ligated with purified BglI digested 63-bp inserts. Competent MC1061 cells were transformed with the ligated DNAs using electroporation and grown in culture medium containing tetracycline through approximately 10 doublings at 37°C to amplify the library. Phage from liquid culture were obtained by clearing the supernatant, and precipitating phage virions with polyethylene glycol, CsCI, gradient, and simple ultracentrifugations. The final phage pellet was dissolved in 1.5 ml of Tris-buffered saline containing 0.02% NaN,. This library consisted of 3.7 X 10’ independent phage particles displaying 15 random amino acid peptides flanked on the amino-terminal end by NH,ADGA- and on the carboxy terminal side by -GAAGA-. The flanking sequences are intended as structureless linkers to minimize the influence of pII1 on peptide conformation. Isolation and biotinylation of HLA-DR9.

HLA-DR9 complexes were isolated from the Epstein-Barr virus (EBV)transformed human B-lymphoblastoid cell line KT12 homozygous for the HLA-DR9 (DRB 1*090 l)-DR5 3 (DRB4*0101)-DQ3 haplotype. After preclearing the whole-cell lysate with protein-A Sepharose [3, 51, DR9 molecules were purified on affinity chromatography with monoclonal antibody (mAb) HU-4 (anti-DR IgG2a [3, 6]), which was previously shown to precipitate DR9 but not the DR53 complex 171. The solution was replaced with 0.25 M NaHC0,/0.2% NP-40 by repeated addition of the buffer and centrifugation in a Centricon(lo-kDa cut-off; Amicon, Beverly, MA, USA), and was then biotinylated. Unbound biotin was removed by five cycles of centrifugation on a Centricon with the addition of 50 mM Tris-Cl, pH 7.5, 150 mM NaCl, 1 mM EDTA, and 1 mM phenylmethylsulfonyl fluoride/O.2% NP-40.

complexes (10” molecules) for 24 hours at room temperature. The mixture was then applied to avidin-coated and BSA-blocked 35-mm polystyrene Petri dishes (Falcon) and the preparation incubated for 10 minutes at room temperature with continuous stirring. After extensive washing with 0.5% Tween 20/Tris-buffered saline, pH 7.5, the bound DR9-phage complexes were eluted with 400 1.11of 0.1 M glycine-HCl, pH 2.2. The neutralized eluate was concentrated, replaced with Trisbuffered saline, and used to infect E. co/i K9lkan cells, which were then expanded in the presence of tetracycline. Phage particles were harvested and purified three times with polyethylene glycol. The screening procedure was repeated six times. The eluate of the seventh round was used to isolate and sequence individual fUSE5 clones, using 5 ‘-TGAATTTTCTGTATGAGG-3 ’ as a sequencing primer. Peptide synthesis. All peptides

used for the binding assay were synthesized in a solid-phase simultaneous multiple peptide synthesizer PSSM-8 (Shimadzu Corp., Kyoto, Japan) based on Fmoc strategy and purified by Cl 8 reversephase high-pressure liquid chromatography (HPLC). Peptides DR9BPl (ANDGATGWVASMSSAY), some analogues and truncated peptides were sequenced, using a protein sequencer PPSQ-10 (Shimadzu) to confirm their structure. Class II HLA-peptide binding assays. The class II HLApeptide binding assay was performed as described elsewhere, but with minor modifications 23, S]. Purified HLA-DR9 molecules were incubated for 48 hours with 100 nM of radioiodinated DRB 1*090 l-binding peptide 1 (DRSBPl), with various doses of various unlabeled peptides at pH 7.0 in the presence of a protease inhibitor cocktail. HLA-peptide complexes were separated from free peptides by gel filtration and fractions were assayed for radioactivity. The fraction of peptide bound to HLADR9 molecules was calculated as the ratio of peptide in the void volume to the total peptide recovered. Peptide inhibitors were added to DR molecules simultaneously with the radioiodinated DR9BPl. Peptide inhibitors were tested at concentrations ranging from 100 pM to 50 nM, and percent inhibition of binding was determined at 5 PM, this being optimal to observe variability of competitive binding inhibition (not shown).

RESULTS Insert sequences of DRY-binding

pbages and direct binding bound to the DR9 complex were clonally expanded and peptide insert sequences of 27 clones were determined. Some identical insert sequences were observed in more than two clones, as phage bio-

assay. Phage particles Library screening. One hundred

billion phage particles (6 units) of the amplified fUSE5 peptide were incubated with 1 pg of biotinylated DR9

x lo9 transducing

library

HLA-DR9-Binding

panning ticles

Peptide

was performed that

bind

Motifs

seven

to DR9 with

times high

133

to select affinity,

phage

par-

and peptides

based on some representative clones with totally independent sequences. For radioiodination, Tyr was added to the COOH terminus of the peptides without Tyr, then radioiodinated peptides were tested for direct binding to the purified DR9 complex. All the four peptides tested showed positive binding to DR9; more than 5% of the radioiodinated peptides bound to DR9 in three of four peptides tested. Among these peptides, DR9BPl (lANDGATGWVASMSSAY16), where 16Y was inserted artificially, showed the highest binding (9.0%). were

synthesized

Identification of anchor residues and motifi. To identify

the anchor positions of DR9BPl (IANDGATGWVASMSSAY’“) for binding to the DR9 complex, we synthesized a series of DR9BPl analogue peptides carrying single residue substitutions to nonconservative amino acids 13). Thus we substituted hydrophobic residues with S, K, D, or T, and hydrophilic residues with A, V, or G. As shown in Fig. 1, nonconservative substitutions of ‘W to S, K, or D completely eliminated peptide binding to the DR9 complex. Furthermore, substitution of ‘lS with hydrophilic residues (K, D) completely eliminated the binding, whereas substitution with hydrophobic residues (A, V) did not. It is conceivable that A and V allowed for binding with DR9, as they are structurally close to S, regardless of polarity. Thus, the primary anchor positions were presumed to be ‘W and llS. Positive binding of a peptide with “Y to “G substitution suggests that “G flanked by the insert sequence on virions does not exert negative effects on binding to DR9. To determine whether the two residues (8W and llS) are sufficient for high-affinity binding to DR9, we synthesized a truncated peptide (KWVASMSSAY) and a polyalanine-based lo-mer analogue peptide (KWAASAAAAA), and examined their affinity for DR9. Lysine was inserted at the NH, terminus to increase the solubility, a procedure which should not create an artificial first anchor because the W to K substitution shown in Fig. 1 did not allow binding. As shown in Fig. 2, the same levels of affinity as the wild-type peptide were evident. These results confirmed that the WxxS motif plays a crucial role in binding of DR9BPl to DR9 molecules. To identify other structural motifs, peptides with substitutions of the 8W or “S residues of DR9BPl to representative amino acids with a distinct chemical nature (D, N, S, K, A, F, L, P, V, or W) were synthesized and evaluated by binding inhibition assay with DR9. As shown in Fig. 3, substitutions of both anchors led to a marked variation in binding affinity. Thus only hydrophobic residues, except Pro at the first anchor, showed

%Inhibition of peptide binding o

Unlabeled peptide

50

1Cfi

$NDGATG$jUl@dSSAP _______------__ =A-__--------__ A _____________ ___ S ____________

_________ ______ _______ _______ -___-t ______ _______ I n ______________ ______-i_______ ________ __--___

_____ wA__________

______

_______-----__ ______--i______ _____---______ _____---g______ __________ _____ __________________________ _____-----

B K

-mm_ --mm

_____

_____ S __

_______----______-----_________--__

__ y: mm

________---__ _____________ _____________

ii--

v

__ __

_______--------________-___ :-_ ________-----______________ _

F-

_______________ _______________ __________-____

ii :

FIGURE 1 Binding of single amino acid-substituted DR9BPl to the DRB1*0901 complexes. Peptides with single amino acid substitutions based on 16-mer peptide DR9BPl (ANDGATGWVASMSSAY) were synthesized and tested on a binding inhibition assay at 5 pM in the presence of 1251DR9BPl (100 nM). (-) indicates a residue without substitution. Larger percent inhibition indicates better binding to DR9. Substitutions yielding less than 50% inhibition and corresponding residues on the original peptide are boxed. binding to DR9. Substitutions of the second anchor S to small aliphatic (A or V) or F revealed high-affinity binding (lOO%, 97%, and 73% inhibition, respectively). Large neutral N or other hydrophobic residues exerted marginal binding, and none of charged residues (D or K) bound to DR9.

DISCUSSION Peptides lie within the peptide-binding groove of HLADR molecules with the peptide NH, and COOH termini projecting out of both ends of the binding site. A DRl-binding peptide was found to have a pronounced twist within the groove, and side chains project from the peptide backbone approximately every 130” in a confor-

S. Fujisao et al.

134

100

% Inhibition of peptide binding o

Unlabeled peptide

50

100

80

_____--F_____--a

0

1.

‘,.I

I

0.05

I

““‘I

0.2

1.0

”

5.0

Peptide concentration

(PM)

FIGURE 2 Competitive binding inhibition assay with truncated and polyalanine designed peptides. DRB1*0901 molecules were incubated with various concentrations of DR9BPl (closed sqz/are), KWVASMSSAY (open sqtiare), and KWAASAAAAA (open circle) in the presence of radioiodinated DR9BP1, and percent inhibition was determined.

mation

characteristic

Amino

acid residues

tides

of the type II polyproline of limited

are important

13, S-151.

Many

positions

for binding DR-binding

helix

in binding

to HLA-DR peptide

191. pep-

molecules

motifs

reported

to

date follow the 9-mer peptide pattern AxxBxCDxE, A, B, C, D, and E residues functioning as anchors.

with Dif-

ferent

pep-

tide

HLA-DR motifs.

vary among anchor

molecules

Thus DR

positions

alleles,

affinity

or 11 are influenced binding

affinity

and certain

manner.

in binding, (a) differences

bound

to DR 1, 4,

C [lo];

subgroups

do

at primary

differences

of the peptides DR4

positions

residues

For instance,

by the position

between

have various of these

reveal significant

in an allele-specific of the binding

apparently

the significance

(b) as to the and

peptides

131, the importance of the anchor positions is A, B, C, and D in that order, but position E does not count; and (c) differential DRBl*O406

binding of peptides to DRB1*0405 and are observed when specific amino acids are

located at positions A, B, and C [3]. Comparing binding peptide motifs

for

DR9

with

those for other DR molecules, distance between the first and second anchors and the polarity of these amino acid residues are conserved, except chor allows for a high-affinity conformations

that Ser at the second anbinding to DR9. Thus,

of HLA-DR9-binding

FIGURE 3 Peptides with single amino acid substitutions on the first and second anchor residues based on DR9BPl were synthesized and tested on binding inhibition assay at 5 pM in the presence of 100 nM ‘251-DR9BP1. (-) indicates a residue without substitution.

peptides

are con-

pears to be modulated by a GlyiVal dimorphism sition 686, and Trp is allowed for high-affinity only

when

drance DR9 with

Gly

is located

(P86Gly),

which

peptides

bearing

the crystallographic binding

at

13, IS]. Indeed,

f3, 12,

peptide,

allows

f386 because

of steric

residues

for high-affinity of DRl

binding In light

complexed

on HLA-DR9

interacting

paring

these

six

residues

forming

the

second

helices

hydrophilic

in this pocket

ap-

a

with

pocket

among DR9 and other alleles, DR9 has four hydrophilic residues (“‘Gin, P74G1u, P71Arg, and ““Arg), whereas DRBl*OlOl, DRB1*0401, and DRB1*0405 three, and DR15 has only two hydrophilic ones. this hydrophilic amino acid-rich circumstance,

Specificity

of

with

the second anchor should consist of the side chains of @l’Phe, “*Val, “*GIu, @‘lArg, and p70Arg, ““Gin, forming a smaller, shallower, and less hydrophobic pocket than that interacting with the first anchor. Com-

ceivably similar to peptides bound to other DR allelic products. The side chain of the first anchor is buried in the largest and most hydrophobic pocket formed by 01 of 01 and p chains.

hin-

this was also the case for

Trp at the first anchor. analysis

at pobinding

Ser might

be allowed,

thereby

being

have Due to neutral differ-

HLA-DR9-Binding

Peptide Motifs

ent from other DR-binding motifs. Unlike other DR molecules, residue selectivity at the third, fourth, and fifth anchors was not detected in the current study with DR9BPl. Thus, we made nonconservative substitutions in the residues at nonanchor positions, as shown in Fig. 1, but the binding affinity remained unchanged. All the 15-mer sequences obtained from 27 clones had first and second anchors that fit into the DR9-binding motif. Indeed, alignment of these sequences, taking the first and second anchors into account, revealed no residue selectivity at other positions; hence, the absence of third, fourth, and fifth anchor residues might be extrapolated to other DR9 binders. However, there is the possibility of the presence of an amino acid that causes a dominant negative effect [lb], as peptides with K (Lys) at the ninth position from the first anchor (P9) could not bind to HLA-DR9 although other motif-positive peptides derived from acetylcholine receptor 01 subunit bound to DR9 (not shown). Indeed, Fig. 1 shows that Lys at P9 reveals a moderate decrease of binding affinity, suggesting that Lys at P9 exerts a dominant negative effect on binding to DR9. It may be that the positive charge of 0176Arg is not offset by B57, as B57 of DR9 is not Asp but nonpolar Val, creating a repulsive interaction between Lys of P9 and “76Arg. Peptides without DR9-binding motifs such as MKRPSREKQDKKIFTEDIDIS or ISGKPGPPPMGFHSPLIK on the acetylcholine receptor (Y subunit did not bind as we expected (not shown). T-cell recognition of antigenic peptides in the context of HLA class II molecules expressed on antigenpresenting cells is the first step leading to cellular and humoral autoimmunity. There is little documentation with regard to antigenic peptide fragments recognized by human T cells in the context of DR9. A study on tetanus toxin-derived peptides reported by PaninaBordignon et al. [2] demonstrated that DR9-restricted T cells recognized FNNFTVSFWLRVEK, which contain DR9-binding motifs (FxxF and FxxS) delineated herein. DR9-binding self-peptides eluted and identified recently included KRKWEAAHAAEQQR (HLA-Al 1, 143156), GAKEKA&AQEAL (Neuropeptide Y-like receptor, 306-3 17), NKVSLTFSKQVALG (apolipoprotein ~-100, 1585-1597), &id EPKDEVYALNLTQTLNP (unknown), all of which contain the motifclarified in our current study [M. Katagiri et al., personal communication). Considering the pathogeneses of juvenile-onset myasthenia gravis or systemic lupus erythematosus with antiphospholipid syndrome, certain fragment(s) of disease-responsible autoantigens might bind with a higher affinity to HLA-DR9 than to other allelic products, if the development of these diseases are self-antigenspecific immune response gene phenomena f3f. In juvenile-onset myasthenia gravis, disease-responsible T cells may recognize acetylcholine receptor, (Y, B, y, and

135

6 subunits expressed in embryonic muscle [17). The putative DR9-binding motifs described herein are detected in 40, 41, 45, and 47 positions on the (Y, p, y, and 6 subunits, respectively, some of which might be recognized by disease-responsible autoreactive T cells. As for systemic lupus erythematosus with antiphospholipid syndrome, 13 peptide fragments with putative DR9binding motifs are found in the Pa-glycoprotein I, a candidate autoantigen for this syndrome [l&l. Interestingly, in these peptide fragments, there are 12, 12, 15, 10, and 4 positions in acetylcholine receptor 01, B, y, 6 and Pa-glycoprotein I, respectively, with Ser at the second anchor positions, and which fit peptide motifs characteristic of DR9. Isolation of T-cell clones specific to these antigens using peptides carrying DR9-binding peptide motifs is underway, and is expected to further our understanding of mechanisms of HLA-associated susceptibility to these diseases.

ACKNOWLEDGMENTS

We thank Dr. A. Wakisaka (Hokkaido University) for HU-4, Dr. N. Kashiwagi (Kitasato University) for KT12, and M. Ohara for reading the manuscript. This work was supported in part by Grants-in Aid 07670375, 06454222, 05278118, and 05272104 from the Ministry of Education, Science, Sports and Culture, Japan; a research grant for intractable diseases from the Ministry of Health and Welfare, Japan, the Japan Rheumatism Foundation, Kato Memorial Foundation; and the Mochida Memorial Foundation.

REFERENCES 1. Matsuki H, Soda Japanese 86:392,

K, Juji T, Tokunaga K, Takamizawa M, Maeda M, Nomura Y, Segawa M: HLA antigens in patients with myasthenia gravis. J Clin Invest 1990.

P, Tan A, Termijtelen A, Demotz S, 2. Panina-Bordignon Corradin G, Lanzavecchia A: Universally immunogenic T cell epitopes: promiscuous recognition by T cells. Eur J Immunol 19:2237, 1989. 3. Matsushita S, Takahashi K, Motoki M, Komoriya K, Ikagawa S, Nishimura Y: Allele specificity of structural requirement for peptides bound to HLA-DRB 1*0405 and -DRB1*0406 complexes: implication for the HLAassociated susceptibility to methimazole-induced insulin autoimmune syndrome. J Exp Med 180:873, 1994. 4. Scott JK, Smith GP: Searching for peptide epitope library. Science 249:386, 1990.

ligands with an

5. Buus S, Sette A, Colon SM, Jenis DM, Grey HM: Isolation and characterization of antigen-Ia complexes involved in T cell recognition. Cell 47:1071, 1986. F, Yoshida 6. Koide Y, Awashima Wakisaka A, Moriuchi J, Aizawa

TO, Takenouchi T, M: The role of three

136

S. Fujisao et al.

distinct Ia-like antigen molecules in human T cell proliferative responses: effect of monoclonal anti-Ia-like antibodies. J Immunol 129:1061, 1982. 7. Hirayama K, Nishimura Y, Tsukamoto K, Sasazuki T: Functional and molecular analysis of three distinct HLADR4 B-chains responsible for the MLR between HLADw4, Dwl5, and DKT2. J Immunol 137:924, 1986. 8. O’Sullivan D, Sidney J, Appella E, Walker L, Phillips L, Colon SM, Miles C, Chesnut RW, Sette, A: Characterization of the specificity of peptide binding to four DR haplotypes. J Immunol 145:1799, 1990. 9. Stern LJ, Brown JH, Jardetzky TS, Gorga JC, Urban RG, Strominger JL, Wiley DC: Crystal structure of the human class II MHC protein HLA-DRl complexed with an influenza virus peptide. Nature 368:215, 1994. 10. Hammer J, Valsasnini P, Tolba K, Bolin D, Higelin J, Takacs B, Sinigaglia F: Promiscuous and allele-specific anchors in HLA-DR-binding peptides. Cell 74:197, 1993. 11. O’Sullivan D, Arrhenius T, Sidney J, de1 Guercio M-F, Albertson M, Wall M, Oseroff C, Southwood S, Colon SM, Gaeta FCA, Sette A: On the interaction of promiscuous antigenic peptides with different DR allele: identification of common structural motifs. J Immunol 147: 2663, 1991. 12. Demotz S, Barbey C, Corradin G, Amoroso A, Lanzavecchia A: The set of naturally processed peptides displayed by DR molecules is tuned by polymorphism of residue 86. Eur J Immunol 23:425, 1993.

13. Krieger JI, Karr RW, Grey HM, Yu W-Y, O’Sullivan D, Batovsky L, Zheng Z-L, Colon SM, Gaeta FCA, Sidney J, Albertson M, de1 Guercio M-F, Chesnut RW, Sette A: Single amino acid changes in DR and antigen define residues critical for peptide-MHC binding and T cell recognition. J Immunol 146:2331, 1991. 14. Sidney J, Oseroff C, Southwood S, Wall M, Ishioka G, Koning F, Sette A: DRB1*0301 molecules recognize structural motif distinct from the one recognized by most DRBl alleles. J Immunol 149:2634, 1992. 15. Hammer J, Bono E, Gallazzi F, Belunis C, Nagy 2, Sinigaglia F: Precise prediction of major histocompatibility complex class II-peptide interaction based on peptide side chain scanning. J Exp Med 180:2353, 1994. 16. Boehncke W-H, Takeshita T, Pendleton CD, Houghten RA, Sadegh-Nasseri S, Racioppi L, Berzofsky JA, Germain RN: The importance of dominant negative effects of amino acid side chain substitution in peptide-MHC molecule interactions and T cell recognition. J Immunol 150: 331, 1993. 17. Moiola L, Protti MP, Manfredi AA, Yuen M-H, Howard Jr JF, Coti-Tronconi BM: T-helper epitopes on human nicotinic acetylcholine receptor in myasthenia gravis. Ann NY Acad Sci 681:198, 1993. 18. McNeil HP, Simpson RJ, Chesterman CN, Krilis SA: Anti-phospholipid antibodies are directed against a complex antigen that includes a lipid-binding inhibitor of coagulation: P,-glycoprotein I (apolipoprotein H). Proc Nat1 Acad Sci USA 87:4120, 1990.