Synthetic Peptides Reproducing Tissue Transglutaminase–Gliadin Complex Neo-epitopes as Probes for Antibody Detection in Celiac Disease Patients’ Sera

Article pubs.acs.org/jmc

Synthetic Peptides Reproducing Tissue Transglutaminase−Gliadin Complex Neo-epitopes as Probes for Antibody Detection in Celiac Disease Patients’ Sera Margherita Di Pisa,†,‡,▽ Simona Pascarella,†,§ Mario Scrima,∥ Giuseppina Sabatino,† Feliciana Real-Fernández,†,§ Mario Chelli,† Daniela Renzi,⊥ Antonio Calabrò,⊥ Anna Maria D’Ursi,∥ Anna Maria Papini,†,‡ and Paolo Rovero*,†,§ †

Laboratory of Peptide and Protein Chemistry and Biology, University of Florence, I-50019 Sesto Fiorentino, Italy Department of Chemistry “Ugo Schiff”, University of Florence, Via della Lastruccia 3/13, I-50019 Sesto Fiorentino, Italy § Department of Neurosciences, Psychology, Drug Research and Child Health, Section of Pharmaceutical Sciences and Nutraceutics, University of Florence, Via Ugo Schiff 6, I-50019 Sesto Fiorentino, Italy ∥ Department of Pharmacy, University of Salerno, Via Giovanni Paolo II 132, I-84084 Fisciano, Italy ⊥ Department of Experimental and Clinical Biomedical Sciences, Gastroenterology Unit, University of Florence, Viale Morgagni 50, 50139 Florence, Italy ‡

S Supporting Information *

ABSTRACT: Celiac disease (CD) patients usually present high levels of circulating IgA antibodies directed to different antigens, in particular tissue transglutaminase (tTG), gliadin (Glia), and endomysium. A series of synthetic peptide constructs containing cross-linked tTG and Glia deamidated peptides have been synthesized. Peptides were tested in enzyme-linked immunosorbent assays against celiac disease patients’ sera versus normal blood donors, and their conformational features were evaluated by molecular modeling techniques. Four peptides were recognized as epitopes by autoantibodies (IgG class) circulating in CD patients’ sera before gluten-free diet. The peptide II, containing Ac-tTG(553−564)-NH2 sequence cross-linked with deamidated Ac-α2Glia(63−71)-NH2, was able to identify specific disease antibodies with a sensitivity of 50% and a specificity of 94.4%. Structural conformations of the linear fragments Ac-tTG(553−564)-NH2 and Ac-α2-Glia(63−71)-NH2 and the corresponding cross-linked peptide II were calculated by molecular modeling. Results showed that cross-linking is determinant to assume conformations, which are not accessible to the linear fragments.

■

diagnostic difficulties.11−13 These facts highlight the need for novel/improved noninvasive serological markers in CD, which assist and support clinicians in diagnosis, evaluation, and prognosis of the disease. In clinical chemistry, the recognition of antibodies circulating in patients’ sera as disease biomarkers has been strongly pursued, aiming at the development of noninvasive reliable diagnostic/prognostic tools.14,15 Directly damaged patients’ tissues or native proteins containing triggering agents have been exploited to identify specific antibodies endowed with diagnostic value. However, the native antigen recognized by these antibodies may be elusive, while univocally characterized peptides, mimicking the relevant epitopes of the native antigens, can display enhanced recognition specificity eliminating or minimizing potential cross-reactivity between structurally

INTRODUCTION Celiac disease (CD) diagnosis requires heterogeneous serological, genetic, histological, and clinical evaluation due to the large variety of associated symptoms.1−4 Up to now, the gold standard for the diagnosis of disease remains the histological analysis of duodenal biopsy samples even though serological tests are in rapid development in terms of both specificity and sensitivity.5 From a serological point of view, CD patients usually present high levels of circulating immunoglobulin A (IgA) antibodies directed to different antigens, in particular tissue transglutaminase (tTG), gliadin (Glia), and endomysium (EMA), whose presence is correlated with gluten dietary intake.6−9 Unfortunately, emerged symptoms upon gluten intake varied amply from days to months and did not always occur in parallel to serum antibody or histological changes.10 Moreover, levels of anti-tTG antibodies are detected in other autoimmune diseases, such as primary biliary cirrhosis and type I diabetes, increasing © XXXX American Chemical Society

Received: November 4, 2014

A

DOI: 10.1021/jm5017126 J. Med. Chem. XXXX, XXX, XXX−XXX

Article

Journal of Medicinal Chemistry Table 1. Sequences and Analytical Data of Synthesized Peptides

a Analytical HPLC on Alliance Chromatography (Waters) with a Phenomenex Kinetex C-18 column, 2.6 μm (100 × 3.0 mm), working at 0.6 mL/ min. Solvent systems are A (0.1% TFA in H2O) and B (0.1% TFA in CH3CN). bPeptides purified on a Phenomenex Jupiter C-18 (250 × 4.6 mm) column using a Waters instrument (separation module 2695, detector diode array 2996) working at 4 mL/min. Solvent systems are A (0.1% TFA in H2O) and B (0.1% TFA in CH3CN). cAnalytical gradient 30−70% B in 5 min. dAnalytical gradient 20−60% B in 5 min. eAnalytical gradient 25−65% B in 5 min. f Semipreparative gradient 20−70% B in 30 min. gSemipreparative gradient 25−70% B in 30 min.

homologous protein epitopes.16−18 For this reason, replacing extracted or recombinant antigens with synthetic peptide probes may enhance sensitivity and specificity of serological tests, a result that assumes particular relevance when applied to CD, which is an underdiagnosed and misdiagnosed disease.19 Indeed, the production of synthetic peptides bearing posttranslational modifications and/or reproducing neo-epitopes is a quite easy task from a synthetic point of view, in comparison to recombinant techniques. In fact, a new generation of CD diagnostic tools proposes synthetic gliadin peptides bearing deamidated sequences reproducing in vivo modifications introduced by tTG enzyme.20,21 These deamidated gliadin peptides (DGP) showed more specific recognition of antibodies than native sequences of α-gliadin for both IgA and IgG isotypes, leading to increased sensitivity, particularly interesting in the case of IgA-deficient patients.22 Molecular chemistry can explain these observations, since deamidation of glutamine side chain to glutamic acid catalyzed by tTG increases the negative charge of the sequences.23 These charges modified the binding of gluten-derived peptides to DQ2 and DQ8 HLA allotypes, creating novel T cell epitopes.24 A recently proposed second generation of CD diagnostic tools proposes an immunoenzymatic assay based on the unique

antigen human tTG cross-linked with deamidated gliadin peptides.25 Following the hypothesis that this covalently linked complex is a neo-epitope in CD, formed under in vivo physiological conditions, a diagnostic kit containing as antigen the recombinant tTG cross-linked with specific peptides of gliadin has been developed, thus improving IgA and IgG antibody detection.26 Building on this hypothesis, we propose herein synthetic antigenic peptides from tTG covalently linked to well-known gliadin epitopes as probes to recognize specific autoantibodies in CD patients’ sera by enzyme-linked immunosorbent assay (ELISA) experiments. Synthetic peptides as neo-epitopes from tTG−gliadin adducts should be relatively stable and could replace protein extracts, thus giving rise to a new class of even more sensitive and specific molecular diagnostic tools. To this purpose, a series of constructs containing tTG peptides crosslinked with deamidated gliadin fragments were designed, synthesized, and tested in ELISA.

■

RESULTS AND DISCUSSION Peptide Design. The goal of this research was to obtain a designed peptide reproducing the linkage of tTG protein crosslinked with gliadin fragments and presenting the chemical and B

DOI: 10.1021/jm5017126 J. Med. Chem. XXXX, XXX, XXX−XXX

Article

Journal of Medicinal Chemistry structural features necessary for detection of celiac-diseasespecific autoantibodies. Fleckenstein et al.27 identified gliadin peptides covalently linked to tTG either via thioester bond to the active-site Cys or via isopeptide bond to specific Lys residues of the enzyme. In this experiment, six Lys residues inside tTG were identified as participating in isopeptide linkage formation to the targeted Gln residues on gliadin, giving important insights on how tTG and gliadin are linked. Among these lysine residues, we selected those showing higher frequency in giving the isopeptide bonds: Lys-590, which is located in C-terminal domain 4, and Lys-562, placed in domain 2. Interestingly, searching inside the tTG protein, Lys-677 is located at the interface of domain 2 and 4, hidden in the core of the protein rather than exposed at the surface.27,28 Lys-562 is partly buried, whereas the other lysine residues are clearly exposed to the surface of tTG. Indeed, all these lysine residues are uncovered when the tTG protein assumes an open conformation.29 Gliadin fragments were selected from the identified gliadin peptide, which contains the known 9-mer core region, either intact or truncated.30,31 The so-called “33-mer” α2-gliadin (57− 89) fragment (LQLQPFPQPQLPYPQPQLPYPQPQLPYPQPQPF), derived from proteolytic digestion of gliadin from food sources such as wheat, barley, and rye, has been identified as an antigen leading to activation of the T-cellmediated response in genetically predisposed subjects.32 Three distinct 9-mer epitopes were previously identified within 33mer: PFPQPQLPY, PQPQLPYPQ (three copies), and PYPQPQLPY (two copies). With all these considerations in mind, we synthesized the covalently cross-linked tTG−gliadin complex peptides listed in Table 1. Peptide Synthesis. Linear peptide fragments ActTG(553−564)-NH 2 , Ac-tTG(581−592)-NH 2 , and ActTG(675−680)-NH2 and α2-gliadin peptides Ac-α2-Glia(61− 69)-NH2, Ac-α2-Glia(63−71)-NH2, and Ac-α2-Glia(68−76)NH2 were prepared by a 9-fluorenylmethoxycarbonyl/tert-butyl (Fmoc/tBu) solid-phase peptide synthesis (SPPS) strategy. An automatic peptide synthesizer was employed according to the general procedure for SPPS, using commercially available Fmoc-protected amino acids and TBTU/DIPEA activation. Fragment Ac-tTG(553−564)-NH2 was synthesized manually due to unsuccessful results from automatic synthesis. After different experiments to overcome synthetic problems, fragment Ac-tTG(553−564)-NH2 was finally obtained by use of a Rink Amide Chem matrix resin and changing protecting groups of Ser side chain. The target Gln residues in gliadin fragments were replaced with Glu, for synthetic requirements to generate the new crosslinked peptides. Amino- and carboxy-termini of each peptide were acetylated and amidated, respectively, in order to remove free terminal charges, which are not present in the native protein sequence and may interfere with antibody recognition.33 Gliadin fragments were cleaved, purified by SLFC, and then coupled in heterogeneous phase to the tTG fragments still anchored on the resin (Scheme 1). After cross-linking reaction, all peptides were cleaved, further purified by semipreparative RP-HPLC to obtain a final purity ≥95%, and characterized by electrospray ionization mass spectrometry (ESI-MS). Analytical HPLC methods, retention times, and observed mass peaks for each peptide are summarized in Table 1. Enzyme-Linked Immunosorbent Assay Experiments. The nine cross-linked peptides I−IX were then tested for their

Scheme 1. Deprotection of Lysine Side Chain, CrossLinking, and Peptide Cleavage Reactionsa

a

Reagents and conditions: (a) 2% hydrazine in DMF; (b) AcPQPELPYPQ-NH2 (1.5 equiv), HATU (1.5 equiv), DIPEA (5 equiv), DMF/NMP (1:1); (c) TFA/H2O/TIS (95:2.5:2.5).

ability to detect diagnostically relevant IgG or/and IgA autoantibodies in celiac disease patients’ sera. Cross-linked peptides I−IX were separately coated as antigens in microplates and incubated with 24 sera from biopsy-confirmed celiac patients (age range 2.2−47.9 years, 20 females and 4 males), and 18 sera from normal blood donors (NBDs; age range 6.1− 36.4 years, 18 males and 4 females). For each antigen, IgG and IgA measurements and statistical analysis were performed separately. Data distribution of IgG antibody responses to peptides I−IX are summarized in Figure 1. The two-tailed Mann−Whitney test revealed significant differences between patients and healthy controls in the case of peptides I, II, V, and VII (P value 0.3303 >0.4213 >0.344

29.2 50 20.8 25

16.9−44.1 35.2−64.8 7.1−42.1 13.6−39.6

94.4 94.4 94.4 94.4

81.3−99.3 81.3−99.3 72.7−99.8 81.3−99.3

5.25 9.00 3.75 4.50

a

Values are given for area under the curve, P value, established cutoff and the corresponding sensitivity, specificity, and likelihood ratio. 95% CI = 95% confidence interval.

models in two different molecular modeling procedures. (i) Molecular dynamic simulations in explicit water (500 ns) (Figure 4, step 2) provided information on the conformational preferences of free Ac-α2-Glia(63−71)-NH2 and Ac-tTG(553− 564)-NH2 (ii) Molecular docking calculations (Figure 4, step 3) were used to build the peptide II structural model. Figure 5 reports the results of DSSP analysis for Ac-α2Glia(63−71)-NH2 and Ac-tTG(553−564)-NH2 during 500 ns of MD runs. DSSP analysis indicates that only a minor amount of regular conformations (i.e., the initial type I β-turn and γturn) are sampled during the molecular dynamics simulation. On the contrary, due to conformational flexibility, these conformers change their structural arrangement to assume

conformations corresponding to Leu5−Tyr7 and Tyr7−Gln9 segments, with the Pro residues inducing “cycle-like” structures, stabilized by hydrogen bonds between the backbone carbonyl oxygen of Gln9 and the backbone amide NH of Gln2. An experimental procedure analogous to that previously mentioned for Ac-α2-Glia(63−71)-NH2, was followed for ActTG(553−564)-NH2: accordingly, a structural model was generated characterized by two N- and C-terminal strands connected by a Ser6−Ile9 type I β-turn, and stabilized by hydrogen bonds between Thr4−Ile9 and Cys2−Val11. According to the protocol shown in Figure 4, the secondary structures previously described for Ac-tTG(553−564)-NH2 and deamidated Ac-α2-Glia(63−71)-NH2 were used as starting E

DOI: 10.1021/jm5017126 J. Med. Chem. XXXX, XXX, XXX−XXX

Article

Journal of Medicinal Chemistry

calculation. The best complex, judged according to the scoring function value,38 was selected as the starting conformation in the subsequent 500 ns MD simulation carried out for determination of the peptide II structural model (Figure 4, step 4). All docking runs were realized by use of the HADDOCK Web server.38 DSSP analysis shoes the secondary structure features of peptide II. (Figure 5) The peptide II conformers sampled during the molecular dynamics simulation were clustered by use of a cutoff of 2.7 Å. Figure 6 shows the representative conformers for the two most populated clusters. In the most representative peptide II conformer (Figure 6A), the structure corresponding to Ac-tTG(553−564)-NH2 is very similar to that of the free peptide, characterized by two β-strands connected by a type I β-turn of the Thr4−Asn7 residues. The structural features corresponding to Ac-α2-Glia(63−71)-NH2 are instead modified in comparison to the free peptide. A careful inspection of the model (Figure 6A) suggests that this conformational modification is reasonably due to the formation of intermolecular H-bonds between Ac-α2-Glia(63−71)-NH2 and Ac-tTG(553−564)-NH2; H bonds specifically involve the side chains of Glu4 and Glu9 in Ac-α2-Glia(63−71)-NH2 and Leu8 and Asn7 in Ac-tTG(553−564)-NH2. Figure 6B shows a less representative structural model of peptide II. In this case the strand−turn−strand arrangement of Ac-tTG(553−564)NH2 is perturbed owing to the interposition of Ac-α2-Glia(63− 71)-NH2 strand; as a consequence, the Leu3−Glu5 segment of Ac-tTG(553−564)-NH2 is free to assume a short turn−helix conformation. In an attempt to correlate the immunogenic activity data reported for the free Ac-α2-Glia(63−71)-NH2 and ActTG(553−564)-NH2 peptides and for the cross-linked peptide II, we compared the structural models calculated for Ac-α2Glia(63−71)-NH2 and Ac-tTG(553−564)-NH2 free peptides with that of cross-linked peptide II. This operation showed that the cross-linking in both Ac-α2-Glia(63−71)-NH2 and ActTG(553−564)-NH2 stabilizes structural features that may be essential to generate epitopes in peptide II for immunogenic activity.

Figure 3. Data distribution of IgG antibody responses to peptide II and the mix of linear peptides Ac-tTG(553−564)-NH2 and Ac-α2Glia(63−71)-NH2 in sera of celiac disease patients (CD) and normal blood donors (NBD) determined by ELISA. Data are reported as absorbance at 405 nm of sera diluted 1:100. Lines represents the mean values. (Significance level *p < 0.05, two-tailed Mann−Whitney nonparametric test).

variable, noncanonical secondary structures. At variance, DSSP analysis evidence the stability of the secondary structure predicted by the Robetta34 server for Ac-tTG(553−564)NH2, with the most sampled Ac-tTG(553−564)-NH2 conformations very similar to the previously described starting model. As previously anticipated, to build the conformational model of peptide II, the initial Ac-α2-Glia(63−71)-NH2 and ActTG(553−564)-NH2 conformers generated by Robetta (Figure 4, step 1) were subjected to a preliminary step of molecular docking with HADDOCK software (Figure 4, step 3).38 This step was necessary to calculate the relative positioning of Acα2-Glia(63−71)-NH2 and Ac-tTG(553−564)-NH2 in the construction of the cross-linked peptide II. The calculation was carried out by imposing intermolecular contact between Glu4 of Ac-α2-Glia(63−71)-NH2 and Lys10 of Ac-tTG(553− 564)-NH2, while the single peptides were considered as rigid. Ten complexes of Ac-α2-Glia(63−71)-NH2 with Ac-tTG(553− 564)-NH2 were generated at the end of the molecular docking

Figure 4. Computational approach followed for molecular modeling of peptide II. F

DOI: 10.1021/jm5017126 J. Med. Chem. XXXX, XXX, XXX−XXX

Article

Journal of Medicinal Chemistry

Figure 5. DSSP analysis for Ac-tTG(553−564)-NH2, deamidated Ac-α2-Glia(63−71)-NH2, and peptide II in 500 ns of MD runs.

Figure 6. Representative structural models of peptide II obtained by GROMOS clusterization method with 2.7 Å as cutoff, on 500 ns of MD runs. (A) Representative model of the most populated cluster; (B) representative model of the second most populated cluster. Ac-tTG(553−564)-NH2 and Ac-α2-Glia(63−71)-NH2 fragments are distinguished by yellow and red ribbons, respectively. Red dotted lines indicate H-bonds.

■

CONCLUSIONS In conclusion, we synthesized a series of synthetic peptide constructs containing cross-linked tTG and gliadin fragments, we tested these peptides in ELISA against sera of CD patients versus normal blood donors, and finally we evaluated their conformational features by molecular modeling techniques. Four peptides were recognized as epitopes by autoantibodies (IgG class) circulating in CD patients’ sera before gluten-free diet. Peptide II, containing Ac-tTG(553−564)-NH2 fragment cross-linked with deamidated Ac-α2-Glia(63−71)-NH2, is able to identify specific disease antibodies with a sensitivity of 50% and a specificity of 94.4%. Although this sensitivity value can appear uncompetitive in comparison with commercial tests, a

careful analysis of patients’ subgroups established a possible clinical correlation not detected by established tests. This study provides the basis for further investigation of anti-peptide II antibodies and their clinical relevance. Structure conformations of the linear fragments ActTG(553−564)-NH2 and Ac-α2-Glia(63−71)-NH2 and the corresponding cross-linked peptide II were calculated by molecular modeling. Comparison of calculated conformations of the free peptides with that of the cross-linked peptide II showed that the cross-linking is determinant to assume conformations that are not accessible to the isolated fragments and that may be specifically recognized by autoantibodies, as G

DOI: 10.1021/jm5017126 J. Med. Chem. XXXX, XXX, XXX−XXX

Article

Journal of Medicinal Chemistry

Cleavage from Resin. Gliadin fragments were cleaved from the resin, with contemporary deprotection of the amino acid side chains, by treatment with a cleavage mixture consisting of TFA/H2O/TlS (95:2.5:2.5). The resin was treated with the reaction mixture for 3 h at room temperature, then filtered off, and the solution was concentrated by flushing with N2. The peptides were precipitated from cold Et2O, centrifuged, and lyophilized. Gliadin fragment crude products were purified by SLFC and characterized by RP-HPLC/ESI-MS. Tissue Transglutaminase Fragments. Synthesis. Ac-tTG(675− 680)-NH2 and Ac-tTG(581−592)-NH2 were synthesized automatically employing Fmoc/tBu strategy and starting from Fmoc-Rink Amide (AM)-PS resin (0.35 g, 0.63 mmol/g). Ac-tTG(553−564)-NH2 was synthesized manually according to Fmoc/tBu chemistry, starting from Fmoc Rink-Amide Chem matrix (0.1 g, 0,56 mmol/g) and inserting a Fmoc-Ser(Trt)-OH residue. Fmoc-Lys(Dde)-OH residues were opportunely inserted during chain elongation in the three tTG peptides. Coupling cycle steps were carried out as previously described for gliadin fragments. After Fmoc protecting group removal from the α-N-terminus, the peptides were acetylated according to the procedure described for gliadin fragments. Chain elongation was monitored through cleavage reactions executed on a small amount of resin every four coupling cycles. Dde Protecting Group Removal. After N-terminal removal of Fmoc protecting group and acetylation, Dde removal on lysine residues was achieved by treating the resin, properly swollen in DCM (40 min), with a solution of 2% hydrazine in DMF (3 × 3 min) at room temperature. The resin was then washed with DCM (3 × 2 min) and dried. tTG fragments were left anchored to the resin for further cross-coupling. Cross-Linking. Reactions. Resins anchoring tTG fragments were swollen for 40 min in DMF. Gliadin fragment (1.5 equiv) was dissolved in DMF/NMP (1:1), and then the coupling reagents HATU (1.5 equiv) and DIPEA (3 equiv) were added to the gliadin fragments solution and the resin was treated with the resulting reaction mixture for 1 h at room temperature. After the coupling reaction, the resin was washed with DMF (3 × 2 min) and DCM (2 × 2 min) and dried. Cross-linked peptides were then cleaved from the resin as described for gliadin fragments. Purification and Characterization of Cross-Linked Fragments. Cross-linked peptides were purified through semipreparative RPHPLC and characterized through RP-HPLC ESI-MS. Purification and analytical methods and observed and calculated masses are summarized in Table 1. Solid-Phase Not-Competitive Indirect Enzyme-Linked Immunosorbent Assays. After informed consent was obtained, serum samples were stored at −20 °C until use. Sera from 48 CD patients were collected at the diagnosis before gluten-free diet (39 females, 9 males; age range 2.2−52 years). A total of 6/48 (12.5%) of the CD samples were potential celiac patients presenting a susceptible genetic profile, anti-tTG and anti-EMA IgA antibodies, and normal small bowel mucosa biopsy (Marsh 0−2). Two of 48 (4%) CD patients presented IgA deficiency. CD diagnosed patients’ sera and healthy donors’ sera were screened in parallel for each ELISA plate. Antibody responses were determined in SP-ELISA. All buffers were brought to room temperature prior to use and filtered through 0.22 μm filters. Activated polystyrene 96-well ELISA plates were coated with each cross-linked peptide, 50 μL/well of 1:100 diluted peptide in pure carbonate buffer 0.05 M (pH 9.6). After 3 h of incubation at room temperature, nonspecific binding sites were blocked with buffer 5 (5% FBS in washing buffer, consisting of 0.9% NaCl and 0.05% Tween 20 in water) 50 μL/well, at room temperature for 1 h. After two washes, plates were incubated with diluted patients’ sera and healthy controls’ sera (50 μL/well, diluted 1:100 in blocking buffer 2.5, consisting of 2.5% FBS in washing buffer) at 4 °C overnight. After four washes, plates were treated with 50 μL/well alkaline phosphatase-conjugated antibody (anti-human IgG or anti-human IgA), appropriately diluted in blocking buffer 2.5 (1:3000 for anti-human IgG, 1:1500 for antihuman IgA). After 3 h of incubation at room temperature and four washes, 50 μL of substrate solution was added to each well. Substrate

confirmed by the lower activity of the mixture of two fragments versus their covalently cross-linked analogue, peptide II. These observations support the hypothesis that a neoepitope may be formed in CD patients’ sera under in vivo physiological conditions, by a covalent cross-link between tTG and deamidated gliadin peptides, and this neo-antigen may be specifically recognized by autoantibodies. Accordingly, a fully synthetic molecule, such as the peptide II designed within this study, appears to be an efficient antigenic probe to be used in a new-generation CD diagnostic assay.

■

EXPERIMENTAL SECTION

Materials and Methods. Peptide-synthesis-grade N,N-dimethylformamide (DMF) was from Scharlau (Barcelona, Spain). HPLCgrade CH3CN was purchased from Carlo Erba (Milano, Italy). Protected amino acids and resins were obtained from Iris Biotech AG (Marktredwitz, Germany). Coupling reagents O-(benzotriazol-1-yl)N,N,N′,N′-tetramethyluronium tetrafluoroborate, 1-[bis(dimethylamino)methylene]-1H-1,2,3-triazolo[4,5-b]pyridinium-3oxide hexafluorophosphate, N-hydroxybenzotriazole, and N,N′-diisopropylcarbodiimide were purchased from Advanced Biotech Italia (Milano, Italy). Gliadin fragments were synthesized on the multiple Apex 396 synthesizer (Advanced ChemTech, Louisville, KY) equipped with an 8-well reaction block. tTG-derived peptides were synthesized on a manual batch synthesizer PLS 4 × 4 (Advanced ChemTech, Louisville, KY). Liquid chromatography flash (SLCF) was performed by use of a Li-Chroprep C-18 column on an Armen Instrument (VWR, Milano, Italy) working at 20 mL/min, with solvent system A (0.1% TFA in H2O) and B (0.1% TFA in CH3CN). Semipreparative RP-HPLC was performed on a Phenomenex Jupiter C-18 (250 × 4.6 mm) column at 28 °C with a Waters separation module 2695 and diode-array detector 2996 (Waters, Milano, Italy), working at 4 mL/ min with solvent system A (0.1% TFA in H2O) and B (0.1% TFA in CH3CN). Analytical HPLC system Alliance Chromatography with a Phenomenex Kinetex C-18 column 2.6 μm (100 × 3.0 mm) working at 0.6 mL/min, with the indicated linear gradients, coupled to a single quadrupole ESI-MS (Micromass ZQ) were purchased from Waters (Milano, Italy). The solvent systems used were A (0.1% TFA in H2O) and B (0.1% TFA in CH3CN). Peptides were lyophilized by use of an Edward Modulyo lyophilizer (5pascal, Milano, Italy). SP-ELISA assays were performed on 96-well plates, Nunc Maxisorp (Sigma−Aldrich, Milano, Italy). Washings steps were executed with automatic Hydroflex microplate washer (Tecan Italia, Milano, Italy). Antihuman IgA−alkaline phosphatase and anti-human IgG−alkaline phosphatase conjugates were purchased from Sigma−Aldrich (Milano, Italy). p-Nitrophenyl phosphate was purchased from Fluka (Milano, Italy). Absorbance values were measured on a Sunrise Tecan ELISA plate reader purchased by Tecan (Tecan Italia, Milano, Italy). Gliadin Fragments. Synthesis. Peptides Ac-α2-Glia(61−69)-NH2, Ac-α2-Glia(63−71)-NH2, and Ac-α2-Glia(68−76)-NH2 selected from gliadin were automatically synthesized by a Fmoc/tBu protection strategy, starting from Fmoc-Rink Amide (AM)-PS resin (0.35 g, 0.63 mmol/g). A robotic needlelike arm allows the distribution of dissolved reagents (0.5 M amino acids and coupling reagent, TBTU, in DMF, 2 M DIPEA in NMP). During peptide synthesis steps, the resins were treated with 1 mL of proper solution per 100 mg of resin. The resin was swollen for 40 min in DMF. Each amino acid cycle was characterized by the following four steps: (1) Fmoc deprotection with a solution of 20% piperidine in DMF (1 × 5 min + 1 × 15 min); (2) washing with DMF (3 × 2 min); (3) coupling reaction (1 × 15 min); and (4) washings with DMF (3 × 2 min) and DCM (1 × 2 min). α-N-Acetylation. Gliadin linear peptides were acetylated at the α-Nterminus, after removal of the Fmoc protecting group on the α-amino function of the last residue anchored to the resin. After 20 min of swelling in DCM, the resin was stirred in a solution of Ac2O (20 equiv) and NMM (20 equiv) in DCM at room temperature, and the solution was refreshed after 1 h (2 × 1 h). Resin was then washed with DCM (3 × 2 min) and dried. H

DOI: 10.1021/jm5017126 J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

■

pNPP was previously dissolved (1 mg/mL) in alkaline substrate buffer (pH 9.8), consisting of 10% diethanolamine and 0.1% MgCl2 in water. After 45 min of incubation at room temperature (1 h for IgA detection), the absorbance was read in a multichannel ELISA reader at 405 nm (ref 620 nm). ELISA plates, coating conditions, reagent dilutions, buffers, and incubation times were tested in preliminary experiments. The antibody levels are expressed as absorbance in arbitrary units at 405 nm. Molecular Modeling. The sequences of Ac-tTG(553−564)-NH2 (DCLTESNLIKVR) and Ac-α2-Glia(63−71)-NH2 (PQPELPYPQ) of peptide II were respectively submitted to a preliminary full-chain protein structure prediction by use of the server Robetta.35 The quality of the structural models generated was evaluated by the Q-mean server39 and PROMOTIF36 and PROCHECK37 software. The starting structures were selected according to the low value of the scoring function. The structural interactions between Ac-tTG(553−564)-NH2 and Ac-α2-Glia(63−71)-NH2 were computed by molecular docking on the HADDOCK web server.38 Molecular docking calculations were performed by forcing intermolecular contacts between Lys10 of ActTG(553−564)-NH2 and Glu4 of Ac-α2-Glia(63−71)-NH2. Among the 10 Ac-tTG(553−564)-NH2/Ac-α2-Glia(63−71)-NH2 complexes generated, the best complex judged according to the HADDOCK scoring function was selected as starting model for further MD simulations. Molecular dynamics (MD) simulations were independently carried out on (1) Ac-α2-Glia(63−71)-NH2, (2) Ac-tTG(553−564)-NH2, and (3) peptide II by use of GROMACS 4.6.5 MD simulations package.40 Each peptide system was built to include the peptides embedded in explicit water environment. TIP3P was used as water model. Peptides were parametrized with ff99SBildn41 and solvated by use of the GROMACS genbox tool. All systems were prepared by 5000 steps of conjugate gradient energy minimization, 100 ps of NVT molecular dynamics with positional restraints on the peptide at 300 K, and 100 ps of NPT MD at 300 K and 1.01325 bar, with positional restraints. At the end of equilibration procedure, the final size of the cubic solvent box was 54.83 Å. All systems were sampled for 500 ns of NPT MD simulations. Electrostatic interactions were computed by the particle mesh Ewald method42,43 using a fourth order spline for interpolation. van der Waals interactions were computed with a cutoff of 10 Å. Time-step simulations of 2 fs was used, with production runs conducted for 500 ns total simulation time. During the simulation periods, Parrinello− Rahmann barostat44 and Nosé−Hoover thermostat were used.45,46 Coordinates were saved every 20 ps for subsequent analysis to determine changes in the secondary structure of the peptides. Hydrogen bonds were measured with the g_hbond GROMACS tool. To analyze changes in secondary structure of the peptides in each simulation, the secondary structure assigning mechanism within the dictionary of secondary structure program (DSSP) was implemented in GROMACS in combination with the PROMOTIF program.36,47 Plots of the secondary structure content per residue over time were then generated for each simulation. The backbone φ and ψ dihedral angles values were also used to construct Ramachandran plots for each residue of the peptides. Clustering of the conformations sampled by the peptide in each simulation was done with the g_cluster analysis tool in GROMACS. The peptides in each simulation were clustered with a 2.7 Å cutoff and GROMOS method48, based on the mutual root-mean-square deviation (RMSD) between all conformations sampled during the MD simulation.

■

Article

AUTHOR INFORMATION

Corresponding Author

*Phone +39 055 4573724; fax +39 055 4573584; e-mail paolo. rovero@unifi.it. Present Address ▽

(For M.D.P.) Laboratoire des BioMolécules, UMR 7203, Université Pierre et Marie Curie 7500, Paris Cedex 05, France. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS

We gratefully acknowledge Ente Cassa di Risparmio di Firenze for financial support.

■

ABBREVIATIONS USED AUC, area under the curve; CD, celiac disease; DCM, dichloromethane; Dde, N-[1-(4,4-dimethyl-2,6dioxocyclohexylidene)ethyl]; DIPEA, diisopropylethylamine; DMF, N,N-dimethylformamide; DSSP, Dictionary of Secondary Structure Program; ELISA, enzyme-linked immunosorbent assay; EMA, endomysium; Glia, gliadin; HATU, 2-(1H-7azabenzotriazol-1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphate methanaminium; HLA, human leukocyte antigen; HPLC, high-pressure liquid chromatography; IgA, immunoglobulin A; IgG, immunoglobulin G; MD, molecular dynamics; NBD, normal blood donors; NMM, N-methylmorpholine; NMP, N-methyl-2-pyrrolidone; pNPP, p-nitrophenyl phosphate; ROC, receiver operating characteristic; RP-HPLC, reverse-phase high-preassure liquid chromatography; SLFC, spot liquid flash chromatography; SPPS, solid-phase peptide synthesis; TBTU, 2-(1H-benzotriazol-1-yl)-1,1,3,3-tetramethyluronium tetrafluoroborate; TFA, trifluoroacetic acid; TIS, triisopropylsilane; tTG, tissue transglutaminase

■

REFERENCES

(1) Murray, J. A. The widening spectrum of celiac disease. Am. J. Clin. Nutr. 1999, 69, 354−365. (2) Shaoul, R.; Lerner, A. Associated autoantibodies in celiac disease. Autoimmun. Rev. 2007, 6, 559−565. (3) Wieser, H. Chemistry of gluten proteins. Food Microbiol. 2007, 24, 115−159. (4) Schuppan, D.; Dennis, M. D.; Kelly, C. P. Celiac disease: Epidemiology, pathogenesis, diagnosis, and nutritional management. Nutr. Clin. Care 2005, 8, 54−69. (5) Rubio-Tapia, A.; Hill, I. D.; Kelly, C. P.; Calderwood, A. H.; Murray, J. A. ACG clinical guidelines: Diagnosis and management of celiac disease. Am. J. Gastroenterol. 2013, 108, 656−676 (quiz 677).. (6) Dieterich, W.; Laag, E.; Schöpper, H.; Volta, U.; Ferguson, A.; Gillett, H.; Riecken, E. O.; Schuppan, D. Autoantibodies to tissue transglutaminase as predictors of celiac disease. Gastroenterology 1998, 115, 1317−1321. (7) Troncone, R.; Maurano, F.; Rossi, M.; Micillo, M.; Greco, L.; Auricchio, R.; Salerno, G.; Salvatore, F.; Sacchetti, L. IgA antibodies to tissue transglutaminase: An effective diagnostic test for celiac disease. J. Pediatr. 1999, 134 (2), 166−171. (8) Reif, S.; Lerner, A. Tissue transglutaminase - the key player in celiac disease: a review. Autoimmun. Rev. 2004, 3, 40−45. (9) Husby, S.; Koletzko, S.; Korponay-Szabó, I. R.; Mearin, M. L.; Phillips, A.; Shamir, R.; Troncone, R.; Giersiepen, K.; Branski, D.; Catassi, C.; Lelgeman, M.; Mäki, M.; Ribes-Koninckx, C.; Ventura, A.; Zimmer, K. P. European Society for Pediatric Gastroenterology, Hepatology, and Nutrition guidelines for the diagnosis of coeliac disease. J. Pediatr. Gastroenterol. Nutr. 2012, 54 (1), 136−160.

ASSOCIATED CONTENT

* Supporting Information S

Two figures showing IgA data distribution and Ramachandran plots. This material is available free of charge via the Internet at http://pubs.acs.org. I

DOI: 10.1021/jm5017126 J. Med. Chem. XXXX, XXX, XXX−XXX

Article

Journal of Medicinal Chemistry (10) Bruins, M. J. The clinical response to gluten challenge: A review of the literature. Nutrients 2013, 5, 4614−4641. (11) Bizzaro, N.; Tampoia, M.; Villalta, D.; Platzgummer, S.; Liguori, M.; Tozzoli, R.; Tonutti, E. Low specificity of anti-tissue transglutaminase antibodies in patients with primary biliary cirrhosis. J. Clin. Lab. Anal. 2006, 20, 184−189. (12) Bakker, S. F.; Tushuizen, M. E.; Stokvis-Brantsma, W. H.; Aanstoot, H. J.; Winterdijk, P.; van Setten, P. A.; von Blomberg, B. M.; Mulder, C. J.; Simsek, S. Frequent delay of coeliac disease diagnosis in symptomatic patients with type 1 diabetes mellitus: Clinical and genetic characteristics. Eur. J. Intern. Med. 2013, 24, 456−460. (13) Scaramuzza, A. E.; Mantegazza, C.; Bosetti, A.; Zuccotti, G. V. Type 1 diabetes and celiac disease: The effects of gluten free diet on metabolic control. World J. Diabetes 2013, 4, 130−134. (14) Rolan, P.; Atkinson, A. J.; Lesko, L. J.; Committee, S. O.; Committee, C. R. Use of biomarkers from drug discovery through clinical practice: Report of the Ninth European Federation of Pharmaceutical Sciences Conference on Optimizing Drug Development. Clin. Pharmacol. Ther. 2003, 73, 284−291. (15) Baker, M. In biomarkers we trust? Nat. Biotechnol. 2005, 23, 297−304. (16) Meloen, R. H.; Puijk, W. C.; Langeveld, J. P.; Langedijk, J. P.; Timmerman, P. Design of synthetic peptides for diagnostics. Curr. Protein Pept. Sci. 2003, 4, 253−260. (17) Papini, A. M. The use of post-translationally modified peptides for detection of biomarkers of immune-mediated diseases. J. Pept. Sci. 2009, 15, 621−628. (18) Ballew, J. T.; Murray, J. A.; Collin, P.; Mäki, M.; Kagnoff, M. F.; Kaukinen, K.; Daugherty, P. S. Antibody biomarker discovery through in vitro directed evolution of consensus recognition epitopes. Proc. Natl. Acad. Sci. U.S.A. 2013, 110, 19330−19335. (19) Nenna, R.; Tiberti, C.; Petrarca, L.; Lucantoni, F.; Mennini, M.; Luparia, R. P.; Panimolle, F.; Mastrogiorgio, G.; Pietropaoli, N.; Magliocca, F. M.; Bonamico, M. The celiac iceberg: Characterization of the disease in primary schoolchildren. J. Pediatr. Gastroenterol. Nutr. 2013, 56, 416−421. (20) Aleanzi, M.; Demonte, A. M.; Esper, C.; Garcilazo, S.; Waggener, M. Celiac disease: Antibody recognition against native and selectively deamidated gliadin peptides. Clin. Chem. 2001, 47, 2023−2028. (21) Mothes, T. Deamidated gliadin peptides as targets for celiac disease-specific antibodies. Adv. Clin. Chem. 2007, 44, 35−63. (22) Schwertz, E.; Kahlenberg, F.; Sack, U.; Richter, T.; Stern, M.; Conrad, K.; Zimmer, K. P.; Mothes, T. Serologic assay based on gliadin-related nonapeptides as a highly sensitive and specific diagnostic aid in celiac disease. Clin. Chem. 2004, 50, 2370−2375. (23) Qiao, S. W.; Sollid, L. M.; Blumberg, R. S. Antigen presentation in celiac disease. Curr. Opin. Immunol. 2009, 21, 111−117. (24) Tollefsen, S.; Arentz-Hansen, H.; Fleckenstein, B.; Molberg, O.; Ráki, M.; Kwok, W. W.; Jung, G.; Lundin, K. E.; Sollid, L. M. HLADQ2 and -DQ8 signatures of gluten T cell epitopes in celiac disease. J. Clin. Invest. 2006, 116, 2226−2236. (25) Matthias, T.; Neidhöfer, S.; Pfeiffer, S.; Prager, K.; Reuter, S.; Gershwin, M. E. Novel trends in celiac disease. Cell. Mol. Immunol. 2011, 8, 121−125. (26) Matthias, T.; Pfeiffer, S.; Selmi, C.; Gershwin, M. E. Diagnostic challenges in celiac disease and the role of the tissue transglutaminaseneo-epitope. Clin. Rev. Allergy Immunol. 2010, 38, 298−301. (27) Fleckenstein, B.; Qiao, S. W.; Larsen, M. R.; Jung, G.; Roepstorff, P.; Sollid, L. M. Molecular characterization of covalent complexes between tissue transglutaminase and gliadin peptides. J. Biol. Chem. 2004, 279, 17607−17616. (28) Liu, S.; Cerione, R. A.; Clardy, J. Structural basis for the guanine nucleotide-binding activity of tissue transglutaminase and its regulation of transamidation activity. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 2743− 2747. (29) Pinkas, D. M.; Strop, P.; Brunger, A. T.; Khosla, C. Transglutaminase 2 undergoes a large conformational change upon activation. PLoS Biol. 2007, 5, No. e327.

(30) Miller, M. L.; Johnson, G. V. Rapid, single-step procedure for the identification of transglutaminase-mediated isopeptide crosslinks in amino acid digests. J. Chromatogr. B: Biomed. Sci. Appl. 1999, 732, 65− 72. (31) Dørum, S.; Arntzen, M.; Qiao, S. W.; Holm, A.; Koehler, C. J.; Thiede, B.; Sollid, L. M.; Fleckenstein, B. The preferred substrates for transglutaminase 2 in a complex wheat gluten digest are peptide fragments harboring celiac disease T-cell epitopes. PLoS One 2010, 5, No. e14056. (32) Shan, L.; Molberg, Ø.; Parrot, I.; Hausch, F.; Filiz, F.; Gray, G. M.; Sollid, L. M.; Khosla, C. Structural basis for gluten intolerance in celiac sprue. Science 2002, 297, 2275−2279. (33) Van Regenmortel, M. H. Antigenicity and immunogenicity of synthetic peptides. Biologicals 2001, 29, 209−13. (34) Habbema, J. D.; Eijkemans, R.; Krijnen, J.; Knottnerus, J. E. Analysis of data on the accuracy of diagnostic tests. In: The Evidence Base of Clinical Diagnosis; Knottnerus, J. E., Ed.; BMJ Books: London, 2002; pp 117−143. (35) Kim, D. E.; Chivian, D.; Baker, D. Protein structure prediction and analysis using the Robetta server. Nucleic Acids Res. 2004, 32, W526−531. (36) Hutchinson, E. G.; Thornton, J. M. PROMOTIF–a program to identify and analyze structural motifs in proteins. Protein Sci. 1996, 5, 212−220. (37) Laskowski, R. A.; MacArthur, M. W.; Moss, D. S.; Thornton, J. M. PROCHECK: a program to check the stereochemical quality of protein structures. J. Appl. Crystallogr. 1993, 26, 283−291. (38) De Vries, S. J.; van Dijk, M.; Bonvin, A. M. The HADDOCK web server for data-driven biomolecular docking. Nat. Protoc 2010, 5, 883−897. (39) Benkert, P.; Biasini, M.; Schwede, T. Toward the estimation of the absolute quality of individual protein structure models. Bioinformatics 2011, 27, 343−350. (40) Pronk, S.; Páll, S.; Schulz, R.; Larsson, P.; Bjelkmar, P.; Apostolov, R.; Shirts, M. R.; Smith, J. C.; Kasson, P. M.; van der Spoel, D. GROMACS 4.5: a high-throughput and highly parallel open source molecular simulation toolkit. Bioinformatics 2013, 29, 845−854. (41) Lindorff-Larsen, K.; Piana, S.; Palmo, K.; Maragakis, P.; Klepeis, J. L.; Dror, R. O.; Shaw, D. E. Improved side-chain torsion potentials for the Amber ff99SB protein force field. Proteins: Struct., Funct., Bioinf. 2010, 78, 1950−1958. (42) Darden, T.; York, D.; Pedersen, L. Particle mesh Ewald: An N· log(N) method for Ewald sums in large systems. J. Chem. Phys. 1993, 98, 10089−10092. (43) Essmann, U.; Perera, L.; Berkowitz, M. L.; Darden, T.; Lee, H.; Pedersen, L. G. A smooth particle mesh Ewald method. J. Chem. Phys. 1995, 103, 8577−8593. (44) Parrinello, M.; Rahman, A. Polymorphic transitions in single crystals: A new molecular dynamics method. J. Appl. Phys. 1981, 52, 7182−7190. (45) Nosé, S. A molecular dynamics method for simulations in the canonical ensemble. Mol. Phys. 1984, 52, 255−268. (46) Hoover, W. G. Canonical dynamics: eEquilibrium phase-space distributions. Phys. Rev. A 1985, 31, 1695. (47) Kabsch, W.; Sander, C. Dictionary of protein secondary structure: Pattern recognition of hydrogen-bonded and geometrical features. Biopolymers 1983, 22, 2577−2637. (48) Daura, X.; Gademann, K.; Jaun, B.; Seebach, D.; van Gunsteren, W. F.; Mark, A. E. Peptide folding: When simulation meets experiment. Angew. Chem., Int. Ed. 1999, 38, 236−240.

J

DOI: 10.1021/jm5017126 J. Med. Chem. XXXX, XXX, XXX−XXX