HLA-DQA1*0102/DQB1*0602 (DQ0602) is observed at a decreased frequency in insulin-dependent diabetes mellitus in different ethnic groups, suggesting a protective role for DQ0602. Analysis of overlapping peptides from human insulin found that insulin B(1–15) bound well to DQ0602 and exhibited a high degree of allelic specificity. Truncation analysis of insulin B(1–15) identified insulin B(5–15) as the minimal peptide for DQ0602 binding. Insulin B(5–15) bound to DQ0602 with an apparent KD of 0.7 to 1.0 μM and peptide binding reached equilibrium at 96 h. Single arginine substitutions at each position of the insulin B(5–15) peptide identified amino acids 6, 8, 9, 11, and 14 (relative positions P1, P3, P4, P6, and P9) as important for binding. Extensive substitutions for each of these amino acids revealed that amino acids 11 and 14 (P6 and P9) exhibited the highest specificity. Amino acid 11 (P6) preferred large aliphatic amino acids, while amino acid 14 (P9) preferred smaller aliphatic and hydroxyl amino acids. Binding of an overlapping series of peptides from a randomly chosen protein, the herpes simplex virus-2 tegument protein UL49, correlated completely with the presence or absence of the DQ0602 peptide binding motif. Peptides 11 amino acids long were selected from GAD65, IA-2, and proinsulin, that contained the DQ0602 peptide binding motif. Of these, 79% (19 of 24) were able to bind DQ0602. This study identifies a peptide binding motif for DQ0602 and peptides from insulin-dependent diabetes mellitus autoantigens that bind DQ0602 in vitro.

The MHC on human chromosome 6p21 is the major susceptibility locus in insulin-dependent diabetes mellitus (IDDM)3 (1). The strongest genetic association on the MHC is with the highly polymorphic HLA-DR and -DQ genes. Susceptibility is associated primarily with DRB1*04-DQB1*0302 and DRB1*0301-DQB1*0201 haplotypes, whereas protection is associated with the DR2 haplotype (2). Protection on the DR2 haplotype associates most closely with the DQB1*0602 gene which has been observed at a decreased frequency in different ethnic groups, including Caucasian, Black, and Japanese (3). Protection attributed to DQB1*0602 is dominant to the susceptibility associated with other DQ alleles but is not absolute given that IDDM patients have been identified with the DQB1*0602 allele (4, 5, 6, 7).

The MHC class II molecules, HLA-DR and -DQ, are heterodimeric proteins that function as peptide receptors for presentation of antigenic peptide to T lymphocytes (8). The rules that govern peptide interaction with MHC class II molecules have been described for some HLA-DR and -DQ alleles through binding studies with synthetic peptides, biochemical isolation of naturally associated MHC class II peptides, and x-ray crystallography. The x-ray crystal structure of HLA-DRA/DRB1*0101 complexed with the influenza virus hemagglutinin peptide, HA(306–318), has provided a foundation for current knowledge on MHC class II/peptide interaction (9). The more recent elucidation of the x-ray crystal structure for HLA-DRA/DRB1*0301 complexed with the class II-associated invariant chain peptide, CLIP(81–104), demonstrated that CLIP(81–104) interaction with HLA-DRA/DRB1*0301 was almost identical with HA(306–318) interaction with HLA-DRA/DRB1*0101 (10), suggesting universal rules for peptide binding to MHC class II.

Peptide binding motifs have been described for many HLA-DR molecules and a few HLA-DQ molecules (8). These include peptide binding motifs for molecules encoded by HLA-DRA/DRB1*0401 (11), HLA-DRA/DRB1*0402 (12), HLA-DRA/DRB1*0405 (13), HLA-DRA/DRB1*0301 (14), HLA-DQA1*0301/DQB1*0302 (15), and HLA-DQA1*0501/DQB1*0201 (16, 17), MHC class II molecules found on haplotypes associated with susceptibility to IDDM. A peptide binding motif for HLA-DRA/DRB1*1501 and HLA-DRA/DRB5*0101 found on the DR2 protective haplotype in the Caucasian population has also been described (18, 19). However, a peptide binding motif for the MHC class II molecule most closely associated with IDDM protection, HLA-DQA1*0102/DQB1*0602 (DQ0602), has not been determined.

More than a dozen putative autoantigens for IDDM have been identified, including GAD65, IA-2, and (pro)insulin (20). Of these, proinsulin is unique in its tissue distribution, being expressed primarily in the pancreas, and at very low levels in the fetal and postnatal thymus (21, 22). In the thymus, a higher level of insulin mRNA expression correlates with the class III variable number of tandem repeats polymorphism found in the insulin gene promoter, an allele associated with protection in IDDM. The correlation of high thymic insulin expression with protection has raised the hypothesis that higher concentrations of (pro)insulin expression result in negative selection of (pro)insulin-specific T lymphocytes. In nonobese diabetic (NOD) mice, it was shown that mice made transgenic for proinsulin expression in thymus were protected from diabetes (23).

A number of additional lines of evidence are supportive of proinsulin as an important autoantigen in IDDM. In humans, insulin and proinsulin autoantibodies have been associated with an increased risk for IDDM development (24, 25). Insulin-specific T lymphocyte proliferation has been demonstrated in peripheral blood derived from prediabetics and diabetics (26), including low levels of proliferation to the proinsulin peptide B24–C36 (27). In NOD mice, insulin-specific T lymphocytes represent a predominant component of islet infiltrates (28). Adoptive transfer experiments have shown that insulin-specific T cell clones are capable of accelerating IDDM in NOD mice (29). A direct role for the insulin B(9–23) peptide in the development of IDDM in NOD mice has been suggested based on the observation that s.c. and intranasal administration of insulin B(9–23) resulted in a marked delay in the onset and a decrease in the incidence of diabetes relative to mice given the control peptide, tetanus toxin-(830–843) (30).

The potential role of HLA-DQ in presenting peptides to disease-mediating T lymphocytes in IDDM led us to examine binding of peptides derived from IDDM autoantigens to HLA-DQ. We report the identification of insulin B(5–15) as a peptide that binds well and with allelic specificity to DQ0602. A peptide binding motif for DQ0602 was elucidated and used to identify additional peptides from IDDM autoantigens that bind DQ0602 in vitro.

Homozygous EBV-transformed B-lymphoblastoid cell lines (B-LCL) from the Tenth International Histocompatibility Workshop include MGAR (DQA1*0102/DQB1*0602), AMAI (DQA1*0102/DQB1*0602), HOM-2 (DQA1*0101/DQB1*0501), KT3 (DQA1*0301/DQB1*0401), AMALA (DQA1*0501/DQB1*0301), JVM (DQA1*0501/DQB1*0301), DEU (DQA1*0301/DQB1*0301), BSM (DQA1*0301/DQB1*0302), and COX (DQA1*0501/DQB1*0201) (31). Other EBV-transformed B-LCLs used in this study include LG2 (DQA1*0101/DQB1*0501), HAS-15 (DQA1*0301/DQB1*0401), PF97387 (DQA1*0301/DQB1*0301), PRIESS (DQA1*0301/DQB1*0302), and MAT (DQA1*0501/DQB1*0201), and they were HLA typed by high resolution oligonucleotide typing (Puget Sound Blood Center, Seattle, WA). HLA class II-deficient BLS-1 was a gift from Dr. Janet Lee (32). Cells were grown in Iscove’s modified Dulbecco’s medium with l-glutamine and 25 mM HEPES buffer (Life Technologies, Gaithersburg, MD) supplemented with 10% FCS, 1 mM sodium pyruvate, 50 U/ml penicillin, and 50 μg/ml streptomycin.

Peptides were synthesized with an Applied Biosystems 432 Peptide Synthesizer (Foster City, CA) or purchased from GeneMed Synthesis, Inc. (South San Francisco, CA). Peptides were biotinylated as described (33). The m.w. of each peptide was analyzed by mass spectrometry. Mass spectrometry was performed by Anaspec, Inc. (San Jose, CA), GeneMed Synthesis, Inc., and the Protein and Carbohydrate Structure Facility at the University of Michigan (Ann Arbor, MI). The amino acid sequence of peptides that are not given elsewhere are: insulin A(1–15), GIVEQCCTSICSLYQ; insulin A(7–21), CTSICSLYQLENYCN; insulin B(1–15), FVNQHLCGSHLVEAL; insulin B(9–23), SHLVEALYLVCGERG; insulin B(16–30), YLVCGERGFFYTPKT.

SPVL3 (anti-DQ) hybridoma cells were kindly provided by Dr. Hans Yssel of DNAX Research Institute of Molecular and Cellular Biology, Palo Alto, CA. L243 (anti-DR) hybridoma cells were purchased from American Type Culture Collection (Rockville, MD). Mouse IgG was purchased from Sigma BioSciences (St. Louis, MO). SPVL3 and L243 ascites were prepared at the University of Washington (Seattle, WA). SPVL3 and L243 were purified from ascites using protein A-Sepharose chromatography. Mouse IgG-, L243-, and SPVL3-Sepharose columns were prepared by coupling 20 mg of purified Ab with 5 ml of cyanogen bromide-activated Sepharose 4B (Sigma).

DQ0602 was purified from 1010 MGAR cells. All manipulations occurred at 4°C. Cells were lysed in 100 ml of 20 mM Tris-HCl (pH 7.5), 100 mM NaCl, 1% Nonidet P-40, 5 mM EDTA, 1 mM PMSF, 1 mM iodoacetamide by mixing for 1 h. The lysate was centrifuged at 100,000 × g for 60 min. The supernatant was precleared by sequential incubations on a rotator with mouse IgG-Sepharose (5 ml) for 1 h and L243-Sepharose (5 ml) for 3 h. The supernatant was than incubated with SPVL3-Sepharose (5 ml) overnight at 4°C. The SPVL3-Sepharose was poured into a column and the flowthrough was collected. The resin was than washed with Buffer 1 (10 mM Tris-HCl, pH 7.5, 0.1% deoxycholate), Buffer 2 (10 mM Tris-HCl, pH 7.5, 1 M NaCl, 1% n-octyl-β-d-glucopyranoside), and Buffer 3 (10 mM Tris-HCl, pH 7.5, 1% n-octyl-β-d-glucopyranoside). DQ0602 was eluted with 100 mM Tris (pH 11.2), 1% n-octyl-β-d-glucopyranoside and neutralized immediately with concentrated acetic acid. Purified DQ0602 was stored at −80°C.

EBV-transformed B-LCLs (1.5 × 106 cells) were washed with HBSS and than incubated for 20 min in 1% paraformaldehyde. Fixed cells were washed with Iscove’s complete medium followed by PBS. Cells were resuspended in 200 μl of 150 mM citrate-phosphate (pH 5.4), 5 mM EDTA, 1 mM iodoacetamide, 1 mM benzamidine, 1 mM PMSF. Biotinylated peptide was added to the cells in 4 μl of a DMSO:β-mercaptoethanol solution (1 part DMSO and 1 part β-mercaptoethanol diluted 1:10 in whole cell peptide binding buffer) to a final concentration of 10 μM and incubated for 18 h at 37°C. Cells were washed with HBSS and lysed by resuspending in 100 μl of 50 mM Tris-HCl (pH 7.5), 5 mM EDTA, 0.15 M NaCl, 1% Nonidet P-40, 1 mM PMSF, 1 μg/ml leupeptin, 1 μg/ml pepstatin A for 1 h on ice. The lysates were centrifuged at 20,000 × g for 10 min, and the supernatants were transferred to a 96-well microtiter plate coated with 10 μg/ml SPVL3, neutralized with 100 μl of 50 mM Tris-HCl (pH 8.0), 0.02% n-dodecyl-β-d-maltoside, and incubated overnight at 4°C. The plate was washed with PBS containing 0.05% Tween-20. DELFIA europium-labeled streptavidin (Wallac, Turku, Finland) diluted 1:1000 in DELFIA assay buffer (Wallac) was added to the wells and incubated for 4 h at room temperature. The plate was washed with PBS containing 0.05% Tween-20. DELFIA enhancement solution (Wallac) was added to the wells and incubated for 1 h at room temperature. Fluorescence was measured using a DELFIA 1232 fluorometer (Wallac).

The reaction mixture consisted of 48 μl of affinity-purified DQ0602 in 150 mM citrate-phosphate (pH 5.4), 1 mM PMSF, 0.02% n-dodecyl-β-d-maltoside at a final concentration of 25 nM and 2 μl of biotinylated peptide in a DMSO:β-mercaptoethanol solution (1 part DMSO and 1 part β-mercaptoethanol diluted 1:20 in purified peptide binding buffer) at a final concentration of 0.001 to 30 μM. Nonspecific binding was determined by omitting DQ0602 from the reaction mixture. DQ0602 was incubated with peptide for 48 h at 37°C. The reaction mixture was transferred to a 96-well microtiter plate coated with 10 μg/ml SPVL3, neutralized with 50 μl of 50 mM Tris-HCl (pH 8.0), 0.02% n-dodecyl-β-d-maltoside, and incubated overnight at 4°C. The detection of bound biotinylated peptide was conducted as described above for the whole cell peptide binding assay. For the peptide saturation curve, nonspecific binding was determined by the addition of 200 μM nonbiotinylated peptide to the reaction mixture and was conducted under equilibrium conditions with a 96-h incubation time. For the time course experiment, the reaction was started by the addition of 10 μM biotinylated peptide and was stopped by the addition of 200 μM nonbiotinylated peptide. The reaction time varied from 15 min to 120 h.

The reaction mixture consisted of 46 μl of affinity-purified DQ0602 in 150 mM citrate-phosphate (pH 5.4), 1 mM PMSF, 0.02% n-dodecyl-β-d-maltoside at a final concentration of 25 nM, 2 μl of nonbiotinylated competitor peptide in DMSO:β-mercaptoethanol solution (1 part DMSO and 1 part β-mercaptoethanol diluted 1:20 in purified peptide binding buffer), and 2 μl of biotinylated insulin B(5–15) in DMSO:β-mercaptoethanol solution at a final concentration of 0.25 μM. The competitor peptide was added first at 0.1, 0.3, 1.0, 3.0, 10, 30, and 100 μM final concentrations to the assay. The incubation was conducted at 37°C for 48 h. The reaction mixture was transferred to a microtiter plate coated with 10 μg/ml SPVL3, and the remainder of the assay was conducted as described above for the purified peptide binding assay. The concentration at which 50% inhibition occurs (IC50) was determined by plotting a curve for each peptide examined and extrapolating from the curve the concentration at which 50% inhibition occurs. Relative binding values were calculated by dividing the IC50 for insulin B(5–15) by the IC50 for the analogue peptide.

Binding of the insulin peptides to HLA-DQ alleles was examined in an assay that uses paraformaldehyde-fixed EBV-transformed B-LCLs derived from individuals who are homozygous for different HLA-DQ alleles. The insulin peptides chosen, insulin A(1–15), insulin A(7–21), insulin B(1–15), insulin B(9–23), and insulin B(16–30), represent the primary structure of insulin and were used previously for mapping insulin T cell epitopes in NOD mice (29). Figure 1 shows the results of insulin peptide binding to DQA1*0101/DQB1*0501, DQA1*0102/DQB1*0602, DQA1*0501/DQB1*0201, DQA1*0501/DQB1*0301, DQA1*0301/DQB1*0301, DQA1*0301/DQB1*0302, and DQA1*0301/DQB1*0401. These molecules are representative of HLA-DQ serologic specificities and are also commonly found in the Caucasian population with the exception of DQA1*0301/DQB1*0401, which is rare in Caucasians but is prevalent in Japanese (3). Figure 1 shows that insulin B(1–15) bound the best to DQ0602, whereas the other peptides showed lower levels of binding to the HLA-DQ alleles examined. In addition, binding of insulin B(1–15) to DQ0602 occurred with high allelic specificity. Similar results were obtained with a second panel of EBV-transformed B-LCLs (data not shown).

FIGURE 1.

Binding of insulin A(1-15) (A), insulin A(7–21) (B), insulin B(1–15) (C), insulin B(9–23) (D), and insulin B(16–30) (E) to HLA-DQ alleles on B-LCLs. Biotinylated insulin peptides (10 μM) were incubated with 1.5 × 106 paraformaldehyde-fixed B-LCLs in whole cell peptide binding buffer. Cells were washed to remove unbound peptide and lysed. HLA-DQ was bound to a microtiter plate coated with SPVL3. Bound biotinylated peptide was detected by fluorescence using a europium-labeled streptavidin system. Data are the means ± SD of triplicate determinations. The HLA-DQ genotype of the homozygous B-LCL is indicated on the y-axis and was represented by BLS-1 (none), LG2 (0101/0501), AMAI (0102/0602), COX (0501/0201), JVM (0501/0301), DEU (0301/0301), BSM (0301/0302), and HAS-15 (0301/0401).

FIGURE 1.

Binding of insulin A(1-15) (A), insulin A(7–21) (B), insulin B(1–15) (C), insulin B(9–23) (D), and insulin B(16–30) (E) to HLA-DQ alleles on B-LCLs. Biotinylated insulin peptides (10 μM) were incubated with 1.5 × 106 paraformaldehyde-fixed B-LCLs in whole cell peptide binding buffer. Cells were washed to remove unbound peptide and lysed. HLA-DQ was bound to a microtiter plate coated with SPVL3. Bound biotinylated peptide was detected by fluorescence using a europium-labeled streptavidin system. Data are the means ± SD of triplicate determinations. The HLA-DQ genotype of the homozygous B-LCL is indicated on the y-axis and was represented by BLS-1 (none), LG2 (0101/0501), AMAI (0102/0602), COX (0501/0201), JVM (0501/0301), DEU (0301/0301), BSM (0301/0302), and HAS-15 (0301/0401).

Close modal

The minimal insulin B(1–15) peptide required for binding to DQ0602 was determined by measuring binding of truncated biotinylated insulin peptides to purified DQ0602. Figure 2 shows that insulin B(1–15) and insulin B(5–15) bound to DQ0602 in a similar fashion. Further truncations at the amino- and carboxyl-terminal end of insulin B(1–15) resulted in a decrease in binding, suggesting that the minimal peptide for maximal binding was insulin B(5–15). Insulin B(6–15) and insulin B(1–14) were efficient binders compared with insulin B(7–15) and insulin B(1–13) but bound less well than insulin B(1–15), suggesting that the minimal required epitope for binding is insulin B(6–14).

FIGURE 2.

Binding of truncated insulin B(1–15) peptides to DQA1*0102/DQB1*0602. Biotinylated insulin B(1–15), insulin B(5–15), insulin B(6–15), insulin B(7–15), insulin B(1–14), and insulin B(1–13) at concentrations from 0.001 to 10 μM were incubated with 25 nM purified DQA1*0102/DQB1*0602 in purified peptide binding buffer for 48 h at 37°C. HLA-DQ was bound to a microtiter plate coated with SPVL3, and samples were washed to remove unbound peptide. Bound biotinylated peptide was detected by fluorescence using a europium-labeled streptavidin system. Data are the means ± SD of triplicate determinations.

FIGURE 2.

Binding of truncated insulin B(1–15) peptides to DQA1*0102/DQB1*0602. Biotinylated insulin B(1–15), insulin B(5–15), insulin B(6–15), insulin B(7–15), insulin B(1–14), and insulin B(1–13) at concentrations from 0.001 to 10 μM were incubated with 25 nM purified DQA1*0102/DQB1*0602 in purified peptide binding buffer for 48 h at 37°C. HLA-DQ was bound to a microtiter plate coated with SPVL3, and samples were washed to remove unbound peptide. Bound biotinylated peptide was detected by fluorescence using a europium-labeled streptavidin system. Data are the means ± SD of triplicate determinations.

Close modal

The interaction of insulin B(5–15) with DQ0602 was characterized in a peptide saturation curve and time course experiment. A linear saturation isotherm for insulin B(5–15) binding to purified DQA1*0102/DQB1*0602 is shown in Figure 3,A. An apparent KD calculated by Scatchard analysis for insulin B(5–15) binding to DQ0602 was 0.7 to 1.0 μM. An apparent KD of 0.01 to 3 μM has been reported for other MHC class II molecules (34, 35, 36). Figure 3 B shows the time course for binding of insulin B(5–15) to purified DQA1*0102/DQB1*0602. A slow association time was exhibited, with 96 h being required to reach equilibrium. The association of peptides with purified MHC class II molecules has been shown to occur slowly, ranging between 1 and 5 days for various HLA-DR, HLA-DQ, IA, and IE alleles (34, 35, 36, 37). These results suggest that insulin B(5–15) interaction with DQ0602 is occurring in a manner consistent with conventional MHC class II peptide interaction.

FIGURE 3.

Peptide saturation curve (A) and time course (B) of insulin B(5–15) binding to DQA1*0102/DQB1*0602. A, Biotinylated insulin B(5–15) (0.001–30 μM) was incubated with 25 nM purified DQA1*0102/DQB1*0602 in purified peptide binding buffer for 96 h at 37°C. Nonspecific binding was determined by the addition of 200 μM nonbiotinylated insulin B(5–15). B, Biotinylated insulin B(5–15) (10 μM) was incubated with 25 nM purified DQA1*0102/DQB1*0602 in purified peptide binding buffer for 15 min to 120 h at 37°C. The reaction was stopped with 200 μM nonbiotinylated insulin B(5–15). Nonspecific binding was determined by omitting HLA-DQ from the reaction mixture. Bound peptide was determined in both A and B by transferring the samples to a microtiter plate coated with SPVL3 and washing to remove unbound peptide. Bound biotinylated peptide was detected by fluorescence using a europium-labeled streptavidin system. Data are the means ± SD of triplicate determinations.

FIGURE 3.

Peptide saturation curve (A) and time course (B) of insulin B(5–15) binding to DQA1*0102/DQB1*0602. A, Biotinylated insulin B(5–15) (0.001–30 μM) was incubated with 25 nM purified DQA1*0102/DQB1*0602 in purified peptide binding buffer for 96 h at 37°C. Nonspecific binding was determined by the addition of 200 μM nonbiotinylated insulin B(5–15). B, Biotinylated insulin B(5–15) (10 μM) was incubated with 25 nM purified DQA1*0102/DQB1*0602 in purified peptide binding buffer for 15 min to 120 h at 37°C. The reaction was stopped with 200 μM nonbiotinylated insulin B(5–15). Nonspecific binding was determined by omitting HLA-DQ from the reaction mixture. Bound peptide was determined in both A and B by transferring the samples to a microtiter plate coated with SPVL3 and washing to remove unbound peptide. Bound biotinylated peptide was detected by fluorescence using a europium-labeled streptavidin system. Data are the means ± SD of triplicate determinations.

Close modal

The peptide binding motif for DQ0602 was defined by examining the effect of single amino acid substitutions in insulin B(5–15) on binding to DQ0602. Arginine substitutions were chosen to map the primary anchors, even though its effect could be pleiotropic, because of the effectiveness of using positively charged substitutions to determine peptide contact sites for HLA-DQ (15, 17, 38). Alanine substitutions, which typically have been used to define motifs for HLA-DR alleles, often have little effect on binding to HLA-DQ alleles (17, 38). Figure 4 shows the effect of single arginine (R) substitutions in biotinylated insulin B(5–15) on binding to DQ0602 on AMAI B-LCLs. Binding of 6R, 8R, 9R, 11R, and 14R insulin B(5–15) peptides to DQ0602 on AMAI B-LCLs was greatly reduced, whereas 5R, 7R, 10R, 12R, 13R, and 15R insulin B(5–15) peptides bound as well as the unsubstituted peptide. Comparable results were obtained with MGAR B-LCL (data not shown). Table I shows the results of analyzing nonbiotinylated arginine (R)-substituted insulin B(5–15) peptides in a competition assay with biotinylated insulin B(5–15) and purified DQ0602. In this assay, the binding of 6R, 8R, 9R, 11R, and 14R insulin B(5–15) peptides to DQ0602 was also greatly reduced.

FIGURE 4.

Binding of single Arg-substituted insulin B(5–15) peptides to DQA1*0102/DQB1*0602. Biotinylated Arg analogue insulin B(5–15) peptides (10 μM) were incubated with 1.5 × 106 paraformaldehyde-fixed B-LCLs in whole cell peptide binding buffer for 18 h at 37°C. Cells were washed to remove unbound peptide. Cells were lysed and DQA1*0102/DQB1*0602 was bound to a microtiter plate coated with SPVL3. Bound biotinylated peptide was detected by fluorescence using a europium-labeled streptavidin system. The HLA-DQ genotype of the homozygous B-LCLs is BLS-1 (none), AMAI (0102/0602). The fluorescence units for insulin B(5–15) binding are: BLS-1, 8,324 ± 822; AMAI, 576,001 ± 8,608. Data are means ± SD of triplicate determinations.

FIGURE 4.

Binding of single Arg-substituted insulin B(5–15) peptides to DQA1*0102/DQB1*0602. Biotinylated Arg analogue insulin B(5–15) peptides (10 μM) were incubated with 1.5 × 106 paraformaldehyde-fixed B-LCLs in whole cell peptide binding buffer for 18 h at 37°C. Cells were washed to remove unbound peptide. Cells were lysed and DQA1*0102/DQB1*0602 was bound to a microtiter plate coated with SPVL3. Bound biotinylated peptide was detected by fluorescence using a europium-labeled streptavidin system. The HLA-DQ genotype of the homozygous B-LCLs is BLS-1 (none), AMAI (0102/0602). The fluorescence units for insulin B(5–15) binding are: BLS-1, 8,324 ± 822; AMAI, 576,001 ± 8,608. Data are means ± SD of triplicate determinations.

Close modal
Table I.

Binding capacity of single Arg insulin B(5–15) analogues for DQA1*0102/DQB1*0602a

PeptideSequenceIC50 (μM)
Insulin B(5–15) HLCGSHLVEAL 1.6 
Insulin B(5–15)5R RLCGSHLVEAL 1.5 
Insulin B(5–15)6R HRCGSHLVEAL >100 
Insulin B(5–15)7R HLRGSHLVEAL 2.5 
Insulin B(5–15)8R HLCRSHLVEAL 25 
Insulin B(5–15)9R HLCGRHLVEAL >100 
Insulin B(5–15)10R HLCGSRLVEAL 2.2 
Insulin B(5–15)11R HLCGSHRVEAL >100 
Insulin B(5–15)12R HLCGSHLREAL 2.0 
Insulin B(5–15)13R HLCGSHLVRAL 3.0 
Insulin B(5–15)14R HLCGSHLVERL 52 
Insulin B(5–15)15R HLCGSHLVEAR 1.4 
PeptideSequenceIC50 (μM)
Insulin B(5–15) HLCGSHLVEAL 1.6 
Insulin B(5–15)5R RLCGSHLVEAL 1.5 
Insulin B(5–15)6R HRCGSHLVEAL >100 
Insulin B(5–15)7R HLRGSHLVEAL 2.5 
Insulin B(5–15)8R HLCRSHLVEAL 25 
Insulin B(5–15)9R HLCGRHLVEAL >100 
Insulin B(5–15)10R HLCGSRLVEAL 2.2 
Insulin B(5–15)11R HLCGSHRVEAL >100 
Insulin B(5–15)12R HLCGSHLREAL 2.0 
Insulin B(5–15)13R HLCGSHLVRAL 3.0 
Insulin B(5–15)14R HLCGSHLVERL 52 
Insulin B(5–15)15R HLCGSHLVEAR 1.4 
a

Binding was measured as described for the DQ0602 competition assay in Materials and Methods.

The importance of the nature of the amino acids at amino acids 6, 8, 9, 11, and 14 (P1, P3, P4, P6, and P9) for binding was further investigated by making additional single amino acid-substituted peptides. Peptides were chosen to represent the general classes of amino acid side chains: glycine (G), aliphatic (A, V, L, I), cyclic imino acid (P), hydroxyl (S, T), acidic (D, E), amide (N, Q), basic (K, R), aromatic (F, Y, W), and sulfur-containing (C, M) (39). Histidine (H), although having unique properties, was treated as a basic amino acid since the pH of the peptide binding assay is 5.4. The effect of these substitutions in insulin B(5–15) is shown in Figure 5 and is expressed as relative binding. The relative binding value was determined by dividing the IC50 for insulin B(5–15) by the IC50 for each insulin B(5–15) analogue peptide. At amino acids 6, 8, and 9 (P1, P3, and P4), Arg (R)- and Pro (P)-substituted peptides had a binding capacity of <0.1 (Fig. 5, AC). In addition, a Gly (G)-substituted peptide was not tolerated at amino acid 6; Cys (C)- and Asp (D)-substituted peptides were not tolerated at amino acid 9. At amino acid 11 (P6), Arg (R)-, Cys (C)-, Phe (F)-, Asp (D)-, Asn (N)-, and Gly (G)-substituted peptides had a relative binding capacity of <0.1 and Thr (T)-, Pro (P)-, and Ala (A)-substituted peptides had a relative binding capacity between 0.1 and 0.3 (Fig. 5,D). Only the highly conservative Val- and Ile-substituted peptides bound well to DQ0602 with relative binding capacities of 0.63 and 1.0, respectively (data not shown), suggesting that a large aliphatic amino acid at amino acid 11 (P6) is important for binding. At amino acid 14 (P9), Arg (R)-, Cys (C)-, Asp (D)-, Asn (N)-, and Phe (F)-substituted peptides had a relative binding capacity of 0.1 or less, whereas Pro (P)- and Leu (L)-substituted peptides exhibited relative binding capacities between 0.1 and 0.3, and Gly (G)-, Thr (T)-, and Ser (S)-substituted peptides exhibited relative binding capacities between 0.3 and 0.5 (Fig. 5,E). Insulin B(5–15) contains an Ala at amino acid 14, thus suggesting a general trend toward smaller amino acids, which possess aliphatic and hydroxyl side chains at this position. The deduced DQ0602 peptide binding motif is shown in Figure 6.

FIGURE 5.

Relative binding capacity of insulin B(5–15) analogues for DQA1*0102/DQB1*0602. A, amino acid 6-substituted insulin B(5–15) analogues. B, amino acid 8-substituted insulin B(5–15) analogues. C, amino acid 9-substituted insulin B(5–15) analogues. D, amino acid 11-substituted insulin B(5–15) analogues. E, amino acid 14-substituted insulin B(5–15) analogues. Insulin B(5–15) analogues (0.1–100 μM) were incubated with 25 nM purified DQA1*0102/DQB1*0602 and biotinylated insulin B(5–15) (0.25 μM) in purified peptide binding buffer for 48 h at 37°C. The reaction mixture was transferred to a microtiter plate coated with SPVL3, and samples were washed to remove unbound peptide. Bound biotinylated peptide was detected by fluorescence using a europium-labeled streptavidin system. Relative binding values were calculated by dividing the IC50 for insulin B(5–15) by the IC50 for the analogue peptide.

FIGURE 5.

Relative binding capacity of insulin B(5–15) analogues for DQA1*0102/DQB1*0602. A, amino acid 6-substituted insulin B(5–15) analogues. B, amino acid 8-substituted insulin B(5–15) analogues. C, amino acid 9-substituted insulin B(5–15) analogues. D, amino acid 11-substituted insulin B(5–15) analogues. E, amino acid 14-substituted insulin B(5–15) analogues. Insulin B(5–15) analogues (0.1–100 μM) were incubated with 25 nM purified DQA1*0102/DQB1*0602 and biotinylated insulin B(5–15) (0.25 μM) in purified peptide binding buffer for 48 h at 37°C. The reaction mixture was transferred to a microtiter plate coated with SPVL3, and samples were washed to remove unbound peptide. Bound biotinylated peptide was detected by fluorescence using a europium-labeled streptavidin system. Relative binding values were calculated by dividing the IC50 for insulin B(5–15) by the IC50 for the analogue peptide.

Close modal
FIGURE 6.

DQA1*0102/DQB1*0602 peptide binding motif. The arrows indicate the amino acids critical for binding to DQA1*0102/DQB1*0602 and the designated relative position. Residues in parentheses are weakly tolerated.

FIGURE 6.

DQA1*0102/DQB1*0602 peptide binding motif. The arrows indicate the amino acids critical for binding to DQA1*0102/DQB1*0602 and the designated relative position. Residues in parentheses are weakly tolerated.

Close modal

The predictive power of the deduced motif for DQ0602 was tested by examining the binding capacity of a randomly chosen set of seven overlapping peptides to DQ0602 in the DQ0602 competition assay. Binding of overlapping peptides from a region of herpes simplex virus-2 (HSV-2) UL49, amino acids 105 to 190, is shown in Table II. The correlation of motif with binding was 100%. The four peptides that did not contain the motif bound to DQ0602 with an IC50 of >100 μM. The three peptides that did contain the motif bound to DQ0602 with IC50 values between 8.6 and 27 μM. Each of the peptides that contained the motif had one weakly tolerated anchor amino acid (relative binding between 0.1 and 0.3), whereas the four other anchor positions were tolerated (relative binding between 0.3 and 1.0). The DQ0602 motif in UL49(135–156) also lacked carboxyl-terminal flanking residues which are predicted to increase binding (Fig. 2). UL49(145–166), which contains the same motif found in UL49(135–156), but located centrally within the peptide, bound 2.9-fold better to DQ0602. These results suggest that the DQ0602 motif identified is typical because the presence of the motif correlates with binders and the absence of the motif correlates with nonbinders.

Table II.

Binding capacity of overlapping peptides from HSV-2 UL49(105–190) to DQA1*0102/DQB1*0602a

PeptideSequenceMotifIC50 (μM)b
UL49(105–126) GGPVGAGGRSHAPPARTPKMTR − >100 
UL49(115–136) HAPPARTPKMTRGAPKASATPA − >100 
UL49(125–146) TRGAPKASATPATDPARGRRPA − >100 
UL49(135–156) PATDPARGRRPAQADSAVLLDAc 25 
UL49(145–166) PAQADSAVLLDAPAPTASGRTK 8.6 
UL49(161–176) ASGRTKTPAQGLAKKLd − >100 
UL49(165–190) TKTPAQGLAKKLHFSTAPPSPTAPWT 27 
PeptideSequenceMotifIC50 (μM)b
UL49(105–126) GGPVGAGGRSHAPPARTPKMTR − >100 
UL49(115–136) HAPPARTPKMTRGAPKASATPA − >100 
UL49(125–146) TRGAPKASATPATDPARGRRPA − >100 
UL49(135–156) PATDPARGRRPAQADSAVLLDAc 25 
UL49(145–166) PAQADSAVLLDAPAPTASGRTK 8.6 
UL49(161–176) ASGRTKTPAQGLAKKLd − >100 
UL49(165–190) TKTPAQGLAKKLHFSTAPPSPTAPWT 27 
a

Binding was measured as described for the DQ0602 competition assay in Materials and Methods.

b

The IC50 for insulin B(5–15) in this experiment was 1.7 μM.

c

Region of peptide containing the DQA1*0102/DQB1*0602 peptide binding motif is underlined.

d

UL49(161–176) is shorter than the other peptides in this series due to difficulties in extending the synthesis amino terminal to residue 161.

In addition, the predictive power of the deduced DQ0602 peptide binding motif was examined by identifying all of the peptides in GAD65, IA-2, and proinsulin (40, 41, 42) that contain the DQ0602 peptide binding motif and testing them for binding to DQ0602. GAD65, IA-2, and proinsulin were selected because of their putative role as autoantigens in IDDM, and thus the potential importance of the identification of peptides derived from these Ags that bind to DQ0602. A total of 24 peptides were identified that contained tolerated anchors in all 5 positions. These were synthesized including the amino acid at the −1 position and the +10 position and were examined in the DQ0602 competition assay for their capacity to bind DQ0602. GAD65(503–513), GAD65(526–536), and insulin B(11–21) (proinsulin (11–21)), which are found within sequences identified as immunodominant epitopes in NOD mice, were among the peptides identified to contain the DQ0602 peptide binding motif (30, 43). Table III shows the IC50 for competition with biotinylated insulin B(5–15) for binding to DQ0602. Of these peptides, 79% (19 of 24) bound to DQ0602 (IC50 >100 μM). The IC50 value of the binders ranged from 0.7 to 90 μM, with insulin B(5–15) binding with an IC50 of 1.7 μM. The range of binding may be explained, in part, by a loose correlation between binding and the number of highly preferred amino acids at anchor positions. These results suggest that the DQ0602 motif, deduced within the context of the insulin B(5–15) peptide, is able to select peptides which bind DQ0602. However, it is also apparent that there are factors that have not been elucidated by this study that will modify the DQ0602 peptide binding motif.

Table III.

Binding capacity of GAD65, proinsulin, and IA-2 peptides containing the deduced DQA1*0102/DQB1*0602 peptide binding motif for DQA1*0102/DQB1*0602a

PeptideSequencebIC50 (μM)
IA-2(586–596) GVAGLLVALAV 0.70 
IA-2(499–509) MSSGSFINISV 0.94 
Insulin B(5–15) HLCGSHLVEAL 1.7 
GAD65(334–344) ATAGTTVYGAF 1.7 
GAD65(91–101) FLHATDLLPAC 1.9 
GAD65(396–406) KMMGVPLQCSA 3.0 
Proinsulin(11–21) LVEALYLVCGE 5.8 
IA-2(576–586) SVLLTLVALAG 6.6 
GAD65(116–126) NILLQYVVKSF 12 
IA-2(504–514) FINISVVGPAL 13 
IA-2(543-553) AQTGLQILQTG 13 
IA-2(544–554) QTGLQILQTGV 23 
IA-2(584–594) LAGVAGLLVAL 31 
Proinsulin(36–46) DLQVGQVELGG 40 
IA-2(379–389) VNVGADIKKTM 48 
GAD65(503–513) NVCFWYIPPSL 58 
IA-2(229–239) MVSVGPLPKAE 62 
IA-2(530–540) VTQQAGLVKSE 64 
GAD65(365–375) AAWGGGLLMSR 70 
GAD65 (86–96) DVNYAFLHATD 90 
IA-2(361–371) LTLLQLLPKGA >100 
GAD65(378–388) KWKLSGVERAN >100 
GAD65(526–536) LSKVAPVIKAR >100 
IA-2(373–383) RNPGGVVNVGA >100 
IA-2(335–345) LQRLAAVLAGY >100 
PeptideSequencebIC50 (μM)
IA-2(586–596) GVAGLLVALAV 0.70 
IA-2(499–509) MSSGSFINISV 0.94 
Insulin B(5–15) HLCGSHLVEAL 1.7 
GAD65(334–344) ATAGTTVYGAF 1.7 
GAD65(91–101) FLHATDLLPAC 1.9 
GAD65(396–406) KMMGVPLQCSA 3.0 
Proinsulin(11–21) LVEALYLVCGE 5.8 
IA-2(576–586) SVLLTLVALAG 6.6 
GAD65(116–126) NILLQYVVKSF 12 
IA-2(504–514) FINISVVGPAL 13 
IA-2(543-553) AQTGLQILQTG 13 
IA-2(544–554) QTGLQILQTGV 23 
IA-2(584–594) LAGVAGLLVAL 31 
Proinsulin(36–46) DLQVGQVELGG 40 
IA-2(379–389) VNVGADIKKTM 48 
GAD65(503–513) NVCFWYIPPSL 58 
IA-2(229–239) MVSVGPLPKAE 62 
IA-2(530–540) VTQQAGLVKSE 64 
GAD65(365–375) AAWGGGLLMSR 70 
GAD65 (86–96) DVNYAFLHATD 90 
IA-2(361–371) LTLLQLLPKGA >100 
GAD65(378–388) KWKLSGVERAN >100 
GAD65(526–536) LSKVAPVIKAR >100 
IA-2(373–383) RNPGGVVNVGA >100 
IA-2(335–345) LQRLAAVLAGY >100 
a

Binding was measured as described for the DQ0602 competition assay in Materials and Methods.

b

The putative DQ0602 peptide binding motifs are shown in bold letters.

HLA-DR and -DQ molecules have been associated with susceptibility and protection in IDDM (2). The HLA molecule most strongly associated with protection in IDDM is DQ0602. Very little is known regarding the biochemical properties of DQ0602 and peptide ligands for DQ0602. Only one peptide, the p21 ras oncogene peptide derived from mutated ras genes found in human cancer, has previously been reported to bind DQ0602 (44). A peptide binding motif for DQ0602 was not determined.

The approach utilized in this report to identify a DQ0602 peptide ligand was to screen synthetic peptide ligands derived from human insulin, a putative autoantigen in IDDM, for binding to HLA-DQ on EBV-transformed B-LCLs. Insulin is considered to be an important autoantigen in IDDM because detection of insulin autoantibodies along with GAD65 and IA-2 contributes to accurate prediction of IDDM (20). Insulin is of particular interest as an autoantigen in IDDM because the expression of its precursor, proinsulin, occurs primarily in the pancreas along with low levels in thymus. Recently, a correlation between higher insulin expression level in thymus and the presence of the class III variable number of tandem repeats allele, associated with protection in IDDM, has been drawn (21, 22).

The peptide identified as a model peptide ligand for DQ0602 was insulin B(5–15). Insulin B(5–15) was shown to bind DQ0602 following the general rules that govern most MHC class II/peptide interactions and not in a fashion that would suggest peptide groove-independent interaction as was recently reported for insulin B(10–30) and HLA-DRA/DRB1*0101 (45). Interaction of insulin B(5–15) with the peptide binding groove of DQ0602 is supported by the allelic specificity of binding. A peptide binding motif suggestive of pockets at P1, P4, P6, and P9 is consistent with the x-ray crystal structure deduced for HLA-DRA/DRB1*0101 complexed with HA(306–318) and HLA-DRA/DRB1*0301 complexed with CLIP(81–104) (9, 10). In addition, the time course of insulin B(5–15) binding with purified DQ0602 requires days to reach equilibrium and not hours as was seen for insulin B(10–30) and HLA-DRA/DRB1*0101.

The peptide binding motif for DQ0602 deduced using insulin B(5–15) as a model peptide suggests that large aliphatic amino acids in relative position 6 and small aliphatic or hydroxyl amino acids in relative position 9 are most important for binding to DQ0602. A critical role for position 6 in allele-specific binding was previously demonstrated by Hammer et al. (46) by selection of peptides from a M13 phage display library with HLA-DRA/DRB1*0101, HLA-DRA/DRB1*0401, and HLA-DRA/DRB1*1101. Sequence analysis revealed peptide binding motifs for the three HLA-DR molecules that shared anchor residues at relative positions 1 and 4 while having an allele-specific anchor at position 6. This suggests that pocket 6 plays a critical role in peptide binding; however, the mechanism by which position 6 confers allelic specificity is not clear.

The preference for a small aliphatic or hydroxyl residue in position 9 of DQ0602-binding peptides is consistent with previous observations that β57-Asp containing MHC class II alleles bind peptides with Ala in position 9. These studies showed that peptides with acidic residues at position 9 bind well to non-Asp-containing alleles but do not bind well to Asp-containing alleles (47, 48). By changing the acidic residue at position 9 to Ala, peptides reverse their binding pattern, binding well to Asp-containing alleles and not binding well to non-Asp-containing alleles. This phenomenon has been attributed to a salt bridge that forms between β57-Asp in P9 and Arg76 of the α-chain in HLA-DRA/DRB1*0101. In HLA-DQ molecules, an analogous salt bridge has been proposed to form between β57-Asp in P9 and Arg79 of the α-chain.

Naturally processed DQ0602 peptides have not been described, and very few DQ0602 binders are known. Therefore, examination of these peptides for the DQ0602 peptide binding motif could not be used as a method for validating the motif determined herein. In note, the p21 ras oncogene peptide (VVGAAGVGKSA) previously identified to bind DQ0602 (44) does contain the deduced DQ0602 peptide binding motif. As a result, the approach that was taken to address the validity of the motif was to correlate the presence of the motif with binding to overlapping peptides from HSV-2 UL49 and insulin. As is shown in Table II, the binding of overlapping peptides from HSV-2 UL49 correlated completely (7 of 7) with the presence or absence of the motif. For insulin, the whole cell peptide binding data (Fig. 1) suggested that insulin A(1–15), insulin A(7–21), and insulin B(9–23) bound moderately to DQ0602 whereas insulin B(16–30) bound poorly. The presence or absence of the motif correlated with binding for insulin A(1–15), insulin B(9–23), and insulin B(16–30), with insulin A(7–21) being the exception. In addition, the DQ0602 peptide binding motif was used to identify peptides from GAD65, proinsulin, and IA-2. Of the peptides identified as containing the motif, 79% (19 of 24) competed for binding with insulin B(5–15) to DQ0602. The correlation of binding with motif in overlapping peptides from randomly selected Ags as well as the ability to select peptides that bind to DQ0602 provides evidence that the preferences determined with the insulin B(5–15) peptide define, at least in part, the requirements of conventional DQ0602 binding peptides.

The inability of the deduced DQ0602 peptide binding motif to correctly predict all peptides that will bind to DQ0602 suggests that there are additional factors involved in determining motif. This has been observed in other studies determining peptide binding motifs for MHC class II molecules (11, 14). One explanation for the inadequacy of the motif may be the overall amino acid composition of the peptide which was not addressed by single amino acid substitutions in the insulin B(5–15) peptide. The amino acid composition may affect the ability of the peptide to interact with the peptide binding groove or simply its solubility in aqueous solution. This point is exemplified by three of the IDDM autoantigen peptides that did not bind and were somewhat unusual in their amino acid composition; IA-2(361–371) had five leucine amino acids, and GAD65(378–388) and GAD65(526–536) both had three positively charged residues. Another factor that was not addressed by the single amino acid substitutions in insulin B(5–15) is the contribution of each individual anchor to the overall motif. The wide range (0.7–90 μM) of IC50 values obtained for the peptides containing a DQ0602 peptide binding motif may partially result from different additive effects between the anchors.

The DQ0602 binding peptides identified from GAD65 (86–96, 91–101, 116–126, 334–344, 365–375, 396–406, 503–513), proinsulin (5–15, 11–21, 36–46, 66–80, 72–86), and IA-2 (229–239, 379–389, 499–509, 504–514, 530–540, 543–553, 544–554, 576–586, 584–594, 586–596) have not previously been implicated in the context of DQ0602. However, autoantibodies to GAD65, insulin, and IA-2 have been identified in DQ0602 individuals who are first-degree relatives of type I diabetics (49). Of the peptides found within regions identified as immunodominant T cell epitopes in NOD mice, insulin B(11–21) bound moderately well, GAD65(503–513) bound poorly, and GAD65(526–536) did not bind to DQ0602.

Insulin B(5–15) was found to bind well and with specificity to DQ0602 in vitro, but whether insulin B(5–15) will be presented by DQ0602 on APC in vivo remains to be determined. The only human insulin-specific T cell clones to be reported were restricted by HLA-DRA/DRB1*0406 and came from healthy donors and insulin autoimmune syndrome patients (50). The epitope specificity of these T cell clones was not determined. Jensen (51) determined that reduction of disulfide bonds is both necessary and sufficient for presentation of insulin to a major population of class II-restricted T cells from H-2d and H-2b mice. Insulin A(1–13), existing in an extended conformation with its A-loop disulfides reduced, was characterized as the major immunogenic determinant presented by I-Ad (52). However, the effect of different MHC class II molecules and APC on insulin processing and presentation is not known.

In conclusion, we have identified peptides that bind DQ0602 in vitro and determined a peptide binding motif for DQ0602. These results provide a stepping stone for understanding the biochemical properties of DQ0602 and its ligands in IDDM. However, these results do not address the physiologic relevance of the peptides being identified. This will be the objective of further study.

We thank Patricia Byers for peptide synthesis and Gerald T. Nepom, Helena Reijonen, and Susan Masewicz for critically reading the manuscript.

1

This work was supported by Grants DK02319 and DK40964 from the National Institutes of Health.

3

Abbreviations used in this paper: IDDM, insulin-dependent diabetes mellitus; DQ0602, HLA-DQA1*0102/DQB1*0602; B-LCL, B-lymphoblastoid cell line; NOD, nonobese diabetic; IC50, concentration at which 50% inhibition occurs; HSV-2, herpes simplex virus 2.

1
Todd, J. A., M. Farrall.
1996
. Panning for gold: genome-wide scanning for linkage in type I diabetes.
Hum. Mol. Genet.
5
:
1443
2
Reijonen, H., and G. T. Nepom. 1997. Role of HLA susceptibility in predisposing to insulin-dependent diabetes mellitus. In Frontiers in Hormone Research. Vol. 22: Molecular Pathogenesis of Diabetes Mellitus, R. D. G. Leslie, ed. Basel, S. Karger AG, p. 46.
3
Rønningen, K. S., A. Spurkland, B. D. Tait, B. Drummond, C. Lopez-Larrea, F. S. Baranda, M. J. Menendez-Diaz, S. Caillat-Zucman, G. Beaurain, H.-J. Garchon, J. Ilonen, H. Reijonen, M. Knip, B. O. Boehm, C. Rosak, C. Löliger, P. Kuhnl, T. Ottenhoff, L. Contu, C. Carcassi, M. Savi, P. Zanelli, T. M. Neri, K. Hamaguchi, A. Kimura, R. P. Dong, N. Chikuba, S. Nagataki, C. Gorodezky, H. Debaz, C. Robles, H. B. Coimbra, A. Martinho, M. A. Ruas, J. A. Sachs, M. Garcia-Pachedo, A. Biro, A. Nikaein, L. Dombrausky, T. Gonwa, C. Zmijewski, D. Monos, M. Kamoun, Z. Layrisse, M. C. Magli, P. Balducci, E. Thorsby.
1992
. HLA class II associations in insulin-dependent diabetes mellitus among Blacks, Caucasoids, and Japanese. K. Tsuji, and M. Aizawa, and T. Sasazuki, eds.
HLA 1991
713
Oxford University Press, New York.
4
Baisch, J. M., T. Weeks, R. Giles, M. Hoover, P. Stastny, J. D. Capra.
1990
. Analysis of HLA-DQ genotypes and susceptibility in insulin-dependent diabetes mellitus.
N. Engl. J. Med.
322
:
1836
5
Zeliszewski, D., J.-M. Tiercy, C. Boitard, X.-F. Gu, M. Loche, R. Krishnamoorthy, N. Simonney, J. Elion, J.-F. Bach, B. Mach, G. Sterkers.
1992
. Extensive study of DRB, DQA, and DQB gene polymorphism in 23 DR2-positive, insulin-dependent diabetes mellitus patients.
Hum. Immunol.
33
:
140
6
Reijonen, H., J. Ilonen, H. K. Å kerblom, M. Knip, H.-M. Dosch.
1994
. Multi-locus analysis of HLA class II genes in DR2-positive IDDM haplotypes in Finland.
Tissue Antigens
43
:
1
7
Sanjeevi, C. B., T. P. Lybrand, M. Landin-Olsson, I. Kockum, G. Dahlquist, W. A. Hagopian, J. P. Palmer, Å. Lernmark.
1994
. Analysis of antibody markers, DRB1, DRB5, DQA1, and DQB1 genes and modeling of DR2 molecules in DR2-positive patients with insulin-dependent diabetes mellitus.
Tissue Antigens
44
:
110
8
Rammensee, H.-G., T. Friede, S. Stevanović.
1995
. MHC ligands and peptide motifs: first listing.
Immunogenetics
41
:
178
9
Stern, L. J., J. H. Brown, T. S. Jardetzky, J. C. Gorga, R. G. Urban, J. L. Strominger, D. C. Wiley.
1994
. Crystal structure of the human class II MHC protein HLA-DR1 complexed with an influenza virus peptide.
Nature
368
:
215
10
Ghosh, P., M. Amaya, E. Mellins, D. C. Wiley.
1995
. The structure of an intermediate in class II MHC maturation: CLIP bound to HLA-DR3.
Nature
378
:
457
11
Sette, A., J. Sidney, C. Oseroff, M.-F. del Guercio, S. Southwood, T. Arrhenius, M. F. Powell, S. M. Colón, F. C. A. Gaeta, H. M. Grey.
1993
. HLA DR4w4-binding motifs illustrate the biochemical basis of degeneracy and specificity in peptide-DR interactions.
J. Immunol.
151
:
3163
12
Hammer, J., F. Gallazzi, E. Bono, R. W. Karr, J. Guenot, P. Valsasnini, Z. A. Nagy, F. Sinigaglia.
1995
. Peptide binding specificity of HLA-DR4 molecules: correlation with rheumatoid arthritis association.
J. Exp. Med.
181
:
1847
13
Matsushita, S., K. Takahashi, M. Motoki, K. Komoriya, S. Ikagawa, Y. Nishimura.
1994
. Allele specificity of structural requirement for peptides bound to HLA-DRB1*0405 and -DRB1*0406 complexes: implication for the HLA-associated susceptibility to methimazole-induced insulin autoimmune syndrome.
J. Exp. Med.
180
:
873
14
Geluk, A., K. E. van Meijgaarden, S. Southwood, C. Oseroff, J. Wouter Drijfhout, R. R. P. de Vries, T. H. M. Ottenhoff, A. Sette.
1994
. HLA-DR3 molecules can bind peptides carrying two alternative specific submotifs.
J. Immunol.
152
:
5742
15
Kwok, W. W., M. E. Domeier, F. C. Raymond, P. Byers, G. T. Nepom.
1996
. Allele-specific motifs characterize HLA-DQ interactions with a diabetes-associated peptide derived from glutamic acid decarboxylase.
J. Immunol.
156
:
2171
16
Johansen, B. H., F. Vartdal, J. A. Eriksen, E. Thorsby, L. M. Sollid.
1996
. Identification of a putative motif for binding of peptides to HLA-DQ2.
Int. Immunol.
8
:
177
17
van de Wal, Y., Y. M. C. Kooy, J. W. Drijfhout, R. Amons, F. Koning.
1996
. Peptide binding characteristics of the coeliac disease-associated DQ(α1*0501, β1*0201) molecule.
Immunogenetics
44
:
246
18
Wucherpfennig, K. W., A. Sette, S. Southwood, C. Oseroff, M. Matsui, J. L. Strominger, D. A. Hafler.
1994
. Structural requirements for binding of an immunodominant myelin basic protein peptide to DR2 isotypes and its recognition by human T cell clones.
J. Exp. Med.
179
:
279
19
Vogt, A. B., H. Kropshofer, H. Kalbacher, M. Kalbus, H.-G. Rammensee, J. E. Coligan, R. Martin.
1994
. Ligand motifs of HLA-DRB5*0101 and DRB1*1501 molecules delineated from self-peptides.
J. Immunol.
153
:
1665
20
Christie, M. R..
1996
. Islet cell autoantigens in type I diabetes.
Eur. J. Clin. Invest.
26
:
827
21
Vafiadis, P., S. T. Bennett, J. A. Todd, J. Nadeau, R. Grabs, C. G. Goodyer, S. Wickramasinghe, E. Colle, C. Polychronakos.
1997
. Insulin expression in human thymus is modulated by INS VNTR alleles at the IDDM2 locus.
Nat. Genet.
15
:
289
22
Pugliese, A., M. Zeller, A. Fernandez, Jr, L. J. Zalcberg, R. J. Bartlett, C. Ricordi, M. Pietropaolo, G. S. Eisenbarth, S. T. Bennett, D. D. Patel.
1997
. The insulin gene is transcribed in the human thymus and transcription levels correlate with allelic variation at the INS VNTR-IDDM2 susceptibility locus for type 1 diabetes.
Nat. Genet.
15
:
293
23
French, M. B., J. Allison, D. S. Cram, H. E. Thomas, M. Dempsey-Collier, A. Silva, H. M. Georgiou, T. W. Kay, L. C. Harrison, A. M. Lew.
1997
. Transgenic expression of mouse proinsulin II prevents diabetes in nonobese diabetic mice.
Diabetes
46
:
34
24
Kuglin, B., H. Kolb, C. Greenbaum, N. K. Maclaren, Å. Lernmark, J. P. Palmer.
1990
. The Fourth International Workshop on the Standardisation of Insulin Autoantibody.
Diabetologia
33
:
638
25
Bohmer, K., H. Keilacker, B. Kuglin, A. Hubinger, J. Bertrams, F. A. Gries, H. Kolb.
1991
. Proinsulin autoantibodies are more closely associated with type 1 (insulin-dependent) diabetes mellitus than insulin autoantibodies.
Diabetologia
34
:
830
26
Roep, B. O..
1996
. T-cell responses to autoantigens in IDDM.
Diabetes
45
:
1147
27
Rudy, G., N. Stone, L. C. Harrison, P. G. Colman, P. McNair, V. Brusic, M. B. French, M. C. Honeyman, B. Tait, A. M. Lew.
1995
. Similar peptides from two β cell autoantigens, proinsulin and glutamic acid decarboxylase, stimulate T cells of individuals at risk for insulin-dependent diabetes.
Mol. Med.
1
:
625
28
Wegmann, D. R., M. Norbury-Glaser, D. Daniel.
1994
. Insulin-specific T cells are a predominant component of islet infiltrates in pre-diabetic NOD mice.
Eur. J. Immunol.
24
:
1853
29
Daniel, D., R. G. Gill, N. Schloot, D. Wegmann.
1995
. Epitope specificity, cytokine production profile and diabetogenic activity of insulin-specific T cell clones isolated from NOD mice.
Eur. J. Immunol.
25
:
1056
30
Daniel, D., D. R. Wegmann.
1996
. Protection of nonobese diabetic mice from diabetes by intranasal or subcutaneous administration of insulin peptide B-(9–23).
Proc. Natl. Acad. Sci. USA
93
:
956
31
Kimura, A., R.-P. Dong, H. Harada, T. Sasazuki.
1992
. DNA typing of HLA class II genes in B-lymphoblastoid cell lines homozygous for HLA.
Tissue Antigens
40
:
5
32
Hume, C. R., N. L. A. Shookster, R. Collins, J. O’Reilly, S. Lee..
1989
. Bare lymphocyte syndrome: altered HLA class II expression in B cell lines derived from two patients.
Hum. Immunol.
25
:
1
33
Kwok, W. W., G. T. Nepom, F. C. Raymond.
1995
. HLA-DQ polymorphisms are highly selective for peptide binding interactions.
J. Immunol.
155
:
2468
34
Buus, S., A. Sette, S. M. Colon, D. M. Jenis, H. M. Grey.
1986
. Isolation and characterization of antigen-Ia complexes involved in T cell recognition.
Cell
47
:
1071
35
Roche, P. A., P. Cresswell.
1990
. High-affinity binding of an influenza hemagglutinin-derived peptide to purified HLA-DR.
J. Immunol.
144
:
1849
36
Jensen, P. E..
1991
. Enhanced binding of peptide antigen to purified class II major histocompatibility glycoproteins at acidic pH.
J. Exp. Med.
174
:
1111
37
Johansen, B. H., S. Buus, F. Vartdal, H. Viken, J. A. Eriksen, E. Thorsby, L. M. Sollid.
1994
. Binding of peptides to HLA-DQ molecules: peptide binding properties of the disease-associated HLA-DQ(α1*0501, β1*0201) molecule.
Int. Immunol.
6
:
453
38
Raddrizzani, L., T. Sturniolo, J. Guenot, E. Bono, F. Gallazzi, Z. A. Nagy, F. Sinigaglia, J. Hammer.
1997
. Different modes of peptide interaction enable HLA-DQ and HLA-DR molecules to bind diverse peptide repertoires.
J. Immunol.
159
:
703
39
Creighton, T. E..
1984
. Chemical nature of polypeptides.
Proteins Structures and Molecular Principles
1
W. H. Freeman and Co., New York.
40
Karlsen, A. E., W. A. Hagopian, C. E. Grubin, S. Dube, C. M. Disteche, D. A. Adler, H. Bärmeier, S. Mathewes, F. J. Grant, D. Foster, Å. Lernmark.
1991
. Cloning and primary structure of a human islet isoform of glutamic acid decarboxylase from chromosome 10.
Proc. Natl. Acad. Sci. USA
88
:
8337
41
Lan, M. S., J. Lu, Y. Goto, A. L. Notkins.
1994
. Molecular cloning and identification of a receptor-type protein tyrosine phosphatase, IA-2, from human insulinoma.
DNA Cell Biol.
13
:
505
42
Sures, I., D. V. Goeddel, A. Gray, A. Ullrich.
1980
. Nucleotide sequence of human preproinsulin complementary DNA.
Science
208
:
57
43
Kaufman, D. L., M. Clare-Salzler, J. Tian, T. Forsthuber, G. S. P. Ting, P. Robinson, M. A. Atkinson, E. E. Sercarz, A. J. Tobin, P. V. Lehmann.
1993
. Spontaneous loss of T-cell tolerance to glutamic acid decarboxylase in murine insulin-dependent diabetes.
Nature
366
:
69
44
Johansen, B. H., T. Gedde-Dahl, III, L. M. Sollid, F. Vartdal, E. Thorsby, G. Gaudernack.
1994
. Binding of ras oncogene peptides to purified HLA-DQ(α1*0102,β1*0602) and -DR(α,β1*0101) molecules.
Scand. J. Immunol.
39
:
607
45
Tompkins, S. M., J. C. Moore, P. E. Jensen.
1996
. An insulin peptide that binds an alternative site in class II major histocompatibility complex.
J. Exp. Med.
183
:
857
46
Hammer, J., P. Valsasnini, K. Tolba, D. Bolin, J. Higelin, B. Takacs, F. Sinigaglia.
1993
. Promiscuous and allele-specific anchors in HLA-DR-binding peptides.
Cell
74
:
197
47
Kwok, W. W., M. E. Domeier, M. L. Johnson, G. T. Nepom, D. M. Koelle.
1996
. HLA-DQB1 codon 57 is critical for peptide binding and recognition.
J. Exp. Med.
183
:
1253
48
Nepom, B. S., G. T. Nepom, M. Coleman, W. W. Kwok.
1996
. Critical contribution of β chain residue 57 in peptide binding ability of both HLA-DR and -DQ molecules.
Proc. Natl. Acad. Sci. USA
93
:
7202
49
Gianani, R., C. F. Verge, R. I. Moromisato-Gianani, L. Yu, Y. J. Zhang, A. Pugliese, G. S. Eisenbarth.
1996
. Limited loss of tolerance to islet autoantigens in ICA+ first degree relatives of patients with type I diabetes expressing the HLA DQB1*0602 allele.
J. Autoimmun.
9
:
423
50
Ito, Y., M. Nieda, Y. Uchigata, M. Nishimura, K. Tokunaga, S. Kuwata, F. Obata, K. Tadokoro, Y. Hirata, Y. Omori, T. Juji.
1993
. Recognition of human insulin in the context of HLA-DRB1*0406 products by T cells of insulin autoimmune syndrome patients and healthy donors.
J. Immunol.
151
:
5770
51
Jensen, P. E..
1991
. Reduction of disulfide bonds during antigen processing: evidence from a thiol-dependent insulin determinant.
J. Exp. Med.
174
:
1121
52
Williams, D. B., J. Ferguson, J. Gariepy, D. McKay, Y.-T. Teng, S. Iwasaki, N. Hozumi.
1993
. Characterization of the insulin A-chain major immunogenic determinant presented by MHC class II I-Ad molecules.
J. Immunol.
151
:
3627