Engineering the Polyproline II Propensity of a Class II Major

Sep 3, 2013 - diabetes and celiac disease are associated with HLA-DQ2 and. HLA-DQ8.2 In all cases, autoimmunity is triggered when a disease relevant ...
0 downloads 0 Views 838KB Size
Letters pubs.acs.org/acschemicalbiology

Engineering the Polyproline II Propensity of a Class II Major Histocompatibility Complex Ligand Peptide Sathya Dev Unudurthi, Kinya Hotta, and Chu-Young Kim* Department of Biological Sciences, National University of Singapore, 14 Science Drive 4, 117543 Singapore S Supporting Information *

ABSTRACT: Our immune system constantly samples peptides found inside the body as a means to detect foreign pathogens, infected cells, and tumorous cells. T cells, which carry out the critical task of distinguishing self from nonself peptides, can only survey peptides that are presented by the major histocompatibility complex protein. We investigated how the secondary structure of a peptide, namely, the polyproline II helix content, influences major histocompatibility complex binding. We synthesized 12 analogues of the wheat gluten derived α-I-gliadin peptide and tested their binding to the celiac disease associated HLA-DQ2 protein. Our analogue library represents a broad spectrum of polyproline II propensities, ranging from random coil structure to high polyproline II helix content. Overall, there was no noticeable correlation between the peptide polyproline II helix content and HLA-DQ2 binding. One analogue peptide, which has low polyproline II helix content, showed a 4.5-fold superior binding compared to native α-I-gliadin.

M

hydrophobic side chain preferred), P4 (negatively charged or aliphatic side chain preferred), P6 (Pro or negatively charged side chain preferred), P7 (negatively charged side chain preferred), and P9 (bulky hydrophobic side chain preferred) is important.5 Another key factor is the enzyme HLA-DM. Newly synthesized class II MHC proteins contain a segment of the invariant chain called CLIP (class II-associated Ii peptide) in the peptide binding groove.6 HLA-DM removes CLIP from the nascent MHC and catalyzes loading of the incoming antigenic peptide. A much less well understood aspect of peptide−MHC complex formation is the influence of the epitope’s secondary structure. Because class II MHC ligand peptides are relatively short (typically 12−15 amino acids in length), they generally lack a definitive backbone conformation. However, celiac disease causing α-I-gliadin and other gluten derived peptides adopt a polyproline II (PPII) helix conformation in solution because of their high Pro and Gln content.7,8 PPII helix has backbone dihedral angles of roughly −75° (φ) and +145° (ψ), lacks intramolecular backbone hydrogen bonds, and contains approximately three residues per turn.9 PPII helix is most frequently observed in peptides and proteins that are rich in Pro and Gln.10 Peptides bound to class II MHC proteins also adopt the PPII helix conformation because of the layout of the peptide binding groove in the MHC (Figure 1).11 We have previously reported the crystal structure of HLA-DQ2 in complex with α-I-gliadin (QLQPFPQPELPY), a gluten-derived peptide associated with celiac disease.12 α-I-gliadin has a PPII

ajor histocompatibility complex (MHC) protein, or human leukocyte antigen (HLA) in humans, is a glycoprotein found on the surface of antigen presenting cells. An individual may express as many as six unique class II MHC proteins; two HLA-DQ, two HLA-DR, and two HLA-DP proteins. Such a small set of MHC proteins is able to cover an almost infinite peptide sequence space of potential antigens because each MHC can bind and present multiple peptide sequences. CD4+ T cells constantly probe peptide−MHC complexes displayed on the antigen presenting cell surface and carry out the critical job of discriminating self-peptides from foreign peptides. When T cells encounter a foreign peptide, in the context of a peptide−MHC complex, a cascade of immune reaction ensues that eradicates the pathogen or pathogen infected host cells.1 Certain class II MHC proteins are also associated with autoimmune disorders. For example, rheumatoid arthritis is associated with HLA-DR4, multiple sclerosis is associated with HLA-DR2 and HLA-DQ6, and both type 1 diabetes and celiac disease are associated with HLA-DQ2 and HLA-DQ8.2 In all cases, autoimmunity is triggered when a disease relevant peptide−MHC complex is detected by the T cell. Therefore, formation of the peptide−MHC complex is a critical and necessary event in autoimmune disease pathogenesis. There are multiple factors that influence the kinetics and thermodynamics of peptide−MHC complex formation. The main criterion for MHC association is the peptide amino acid sequence. Although peptides of varying lengths associate with class II MHC molecules, the presence of a consensus sequence is necessary for MHC binding.3 In the case of HLA-DQ2, 432 natural ligands representing 155 distinct sequences have been identified so far.4 Analysis of HLA-DQ2 ligands showed that the nature of the amino acid side chain at positions P1 (bulky © 2013 American Chemical Society

Received: June 19, 2013 Accepted: September 3, 2013 Published: September 3, 2013 2383

dx.doi.org/10.1021/cb400594q | ACS Chem. Biol. 2013, 8, 2383−2387

ACS Chemical Biology

Letters

We have synthesized the native α-I-gliadin peptide and 12 αI-gliadin analogues using standard solid phase peptide synthesis (Table 1). The secondary structure content of each peptide was determined using circular dichroism spectroscopy at pH 5.5 and 37 °C (Figure 2). This buffer condition mimics the native

Figure 1. X-ray crystal structure of α-I-gliadin (PFPQPELPY) in complex with HLA-DQ2 (PDB ID: 1S9V). α-I-gliadin adopts a polyproline II helix conformation in the peptide binding groove of HLA-DQ2. Side chain of amino acids at positions P1, P4, P6, P7, and P9 contact HLA-DQ2, whereas the side chain of amino acids at positions P2, P3, P5, and P8 are solvent exposed (C, yellow; N, blue; O, red).

helix conformation when it is in solution as well as when it is bound to HLA-DQ2, which suggests that α-I-gliadin incurs a minimal entropic penalty when binding to HLA-DQ2. Therefore the intrinsic PPII propensity of gluten peptides may contribute to their high immunogenicity. To test this hypothesis, we have systematically modulated the PPII propensity of α-I-gliadin and studied its effect on HLA-DQ2 binding. Our goal was to create a series of α-I-gliadin analogues that have the same MHC interacting residues but have different backbone conformations in solution. The α-I-gliadin−HLADQ2 crystal structure shows that P2 (Phe), P3 (Pro), P5 (Pro), and P8 (Pro) side chains of α-I-gliadin do not make contact with HLA-DQ2 (Figure 1).12 Therefore, amino acid substitution at these positions is unlikely to alter the enthalpy of peptide−MHC formation. However, because different amino acids have different α-helix, β-sheet, or PPII helix promoting tendencies,13−19 backbone conformation of the altered peptide may be different from that of the original peptide. We substituted the P3, P5, and P8 residues of α-I-gliadin with either (2S,4R)-4-hydroxyproline (Hyp), which is known to promote PPII helix formation, or with N-methyl-L-alanine (MeA), which is known to disfavor PPII helix formation.20,21 We did not modify the P2 position because the main chain C O and N−H groups of P2 (Phe) form a bidentate hydrogen bond with Asn-β82 of HLA-DQ2, which would be lost upon Hyp or MeA substitution.

Figure 2. Circular dichroism spectra of α-I-gliadin and analogues at pH 5.5 and 37 °C. Each trace represents an average of 10 scans with no additional smoothing. The presence of ellipticity maxima near 223−225 nm signifies polyproline II helix content in the peptide.

environment of the endosome, the cellular compartment where peptide−MHC complex formation takes place. As expected, all Hyp containing analogues displayed greater PPII helix content

Table 1. Amino Acid Sequence of α-I-Gliadin and Its Analoguesa α-I-gliadin Hyp-3 Hyp-5 Hyp-8 Hyp-3-5 Hyp-3-8 Hyp-5-8 Hyp-3-5-8 MeA-3 MeA-5 MeA-3-5 MeA-3-8 MeA-5-8 a

P-3

P-2

P-1

P1

P2

P3

P4

P5

P6

P7

P8

P9

Gln Gln Gln Gln Gln Gln Gln Gln Gln Gln Gln Gln Gln

Leu Leu Leu Leu Leu Leu Leu Leu Leu Leu Leu Leu Leu

Gln Gln Gln Gln Gln Gln Gln Gln Gln Gln Gln Gln Gln

Pro Pro Pro Pro Pro Pro Pro Pro Pro Pro Pro Pro Pro

Phe Phe Phe Phe Phe Phe Phe Phe Phe Phe Phe Phe Phe

Pro Hyp Pro Pro Hyp Hyp Pro Hyp MeA Pro MeA MeA Pro

Gln Gln Gln Gln Gln Gln Gln Gln Gln Gln Gln Gln Gln

Pro Pro Hyp Pro Hyp Pro Hyp Hyp Pro MeA MeA Pro MeA

Glu Glu Glu Glu Glu Glu Glu Glu Glu Glu Glu Glu Glu

Leu Leu Leu Leu Leu Leu Leu Leu Leu Leu Leu Leu Leu

Pro Pro Pro Hyp Pro Hyp Hyp Hyp Pro Pro Pro MeA MeA

Tyr Tyr Tyr Tyr Tyr Tyr Tyr Tyr Tyr Tyr Tyr Tyr Tyr

Hyp = (2S,4R)-4-hydroxyproline; MeA = N-methyl-L-alanine. 2384

dx.doi.org/10.1021/cb400594q | ACS Chem. Biol. 2013, 8, 2383−2387

ACS Chemical Biology

Letters

than α-I-gliadin, as shown by the increase in ellipticity maxima near 223−225 nm.20 All MeA containing analogues displayed reduced PPII helix content compared to α-I-gliadin, as seen by lower or lack of ellipticity maxima near 223−225 nm. In general, the effect of Hyp and MeA substitution on the PPII helix content of the parent peptide was additive. Multiple Hyp substitutions resulted in greater PPII helix content than single substitutions, and multiple MeA substitutions led to lower PPII helix content than single substitutions. Overall, we have prepared a series of α-I-gliadin analogues with low to high PPII helix circular dichroism signature (Figure 2). High-performance liquid chromatography (HPLC)-based peptide exchange assay was carried out to assess the extent of peptide−HLA-DQ2 complex formation (Figure 3). In this

Hyp-5 (18%) showed approximately 2-fold improvement in complex formation. Interestingly, all MeA containing analogue peptides showed higher degree of complex formation compared to α-I-gliadin; MeA-3-5 (15%), MeA-5 (20%), MeA-5-8 (28%), MeA-3 (33%), and MeA-3-8 (37%). Analogue peptides containing Hyp at P3 (Hyp-3, Hyp-3-5, Hyp-3-8, and Hyp-35-8) failed to form the complex, likely due to a steric clash between the hydroxyl group attached to the gamma carbon of P3 Hyp and the pocket wall of HLA-DQ2. In this study, we modulated the solution PPII propensity of α-I-gliadin by substituting up to three non-MHC-interacting residues with Hyp or MeA. This strategy has led to the discovery of MeA-3-8, which showed a 4.5-fold improvement in HLA-DQ2 complex formation. Interestingly, no direct correlation was observed between the peptide PPII helix propensity and MHC binding. Some analogues with higher PPII content than α-I-gliadin as well as those with lower PPII content formed a complex with HLA-DQ2 more readily than αI-gliadin. Hyp-3-8, which has the highest PPII helix content as assessed by circular dichroism spectroscopy, was not the best binder. Instead, MeA-3-8, which has a relatively low PPII helix content, formed the greatest amount of peptide−MHC complex. This unexpected result is likely due to the transient nature of the PPII helix. Unlike the α-helix or β-sheet, PPII helix does not contain backbone-mediated hydrogen bonds that stabilize the overall secondary structure. As a result, the pitch and translation of a PPII helix may vary along the length of the peptide and also over time. Because such local or temporal variations in a PPII helix cannot be detected by circular dichroism spectroscopy, it is not surprising that the rate of peptide−MHC complex formation is not a simple function of the ellipticity maximum determined by circular dichroism spectroscopy. When binding to HLA-DQ2, the α-I-gliadin backbone does not adopt an ideal polyproline II helix conformation (φ = −75° and ψ = 145°) (Figure 1). The φ angle varies from −66° to −143° (average φ = −100°), and ψ angle varies from −40° to 169° (average ψ = 116°) along the length of the peptide. Furthermore, the PPII helix formed by the α-I-gliadin backbone is not straight but has a gentle overall curvature. We hypothesize that, among the 12 analogues, the solution backbone conformation of MeA-3-8 most closely matches the backbone conformation of α-I-gliadin bound to HLA-DQ2, thus resulting in the most efficient MHC binding. With respect to celiac disease, our result indicates that the solution PPII helix conformation of α-I-gliadin is not critical for HLA-DQ2 binding. This is in agreement with the previous finding that HLA-DQ2 can bind a large number of natural peptides that completely lack Pro in their sequence.4 However, it should be noted that the high Pro content of gluten peptides, in addition to promoting PPII helix backbone conformation, confers resistance against proteolytic degradation since human digestive enzymes are generally incapable of cleaving peptide bonds adjacent to Pro. Such proteolytic resistance has the undesired effect of raising the concentration of incompletely digested gluten peptides in blood, thus increasing the likelihood of triggering an autoimmune reaction in celiac disease patients. In fact, one of the current celiac disease therapy development efforts involves administration of an exogenous prolyl endopeptidase to break down gluten peptides into fragments too small to bind to the MHC.23,24

Figure 3. Association of α-I-gliadin and its analogues with HLA-DQ2 at 37 °C and pH 5.5. Percentage of bound peptide is calculated by measuring the fluorescence signal corresponding to the peptide−HLADQ2 complex. After 120 min, 8.1% of α-I-gliadin formed a complex with HLA-DQ2, while 36.6% of MeA-3-8 associated with HLA-DQ2. Hyp-3, Hyp-3-5, Hyp-3-8, and Hyp-3-5-8 did not show any detectable binding (data not shown).

assay, we used an engineered HLA-DQ2 protein that contains a covalently attached CLIP peptide (VSKMRMATPLLMQAL). This ensures that the peptide binding pocket of all HLA-DQ molecules are fully occupied by the CLIP peptide, akin to newly synthesized class II MHC proteins found in vivo. CLIP is attached to the N-terminal end of the HLA-DQ2 β-chain via a Gly-rich flexible linker that contains a thrombin cleavage site. Thrombin is added to the protein solution immediately prior to the experiment so that the HLA-DQ2 bound CLIP can be exchanged with an incoming peptide. The assay is initiated by adding fluorescently tagged α-I-gliadin or an analogue peptide to the thrombin treated recombinant CLIP−HLA-DQ2. An aliquot of the reaction mixture is removed at various time points and analyzed using an HPLC equipped with an analytical gel filtration column and a fluorescence detector. Percent peptide−MHC complex formed for a given peptide is determined by calculating the area of the peak corresponding to the newly formed peptide−MHC complex in the HPLC fluorescence spectrum. For native α-I-gliadin, 8% of the total peptide in the reaction mixture formed a complex with HLADQ2 after 120 min incubation, which is consistent with previous reports.22 Hyp 5-8 (7%) and Hyp-8 (10%) showed similar levels of complex formation as native α-I-gliadin, while 2385

dx.doi.org/10.1021/cb400594q | ACS Chem. Biol. 2013, 8, 2383−2387

ACS Chemical Biology



METHODS



ASSOCIATED CONTENT



Peptide Synthesis and Purification. Peptides used in this study were prepared using Fmoc/HBTU-based solid phase peptide synthesis with PS-Wang resin (Novabiochem, 100−200 mesh, and 0.48 mmol gm−1) as solid support.25 All amino acids used were procured as FmocL-α-amino acid (Novabiochem). Peptides were labeled at the Nterminus with 5-(and-6)-carboxyfluorescein (2.5-fold molar excess) using N,N′-diisopropylcarbodiimide and 1-hydroxybenzotriazole in 1:1 dimethyl formamide and dichloromethane as the solvent. Peptides were cleaved from the resin by applying trifluoroaceticacid/ triisopropylsilane/water (TFA/TIS/H2O 95:2.5:2.5, v/v/v) for 4 h and were precipitated in ice cold methyl tert-butyl ether. Crude peptides were lyophilized and were purified by reverse-phase HPLC on a semipreparative column (Luna C18, 250 × 10 mm, 10 μm) using a water−acetonitrile gradient in 0.05% (v/v) TFA, to purity greater than 95% and stored at −20 °C. Purity and identity of each peptide was confirmed using liquid chromatography coupled electrospray mass spectrometry (LC−MS). Circular Dichroism Spectroscopy. Circular dichroism (CD) spectra were measured on a Jasco J810 spectropolarimeter. Peptides were dissolved in a 1:1 mixture of phosphate buffered saline and McIlvaine’s citrate-phosphate buffer, pH 5.5, to a final peptide concentration of 35 μM. Spectra were recorded with a quartz cell of 0.1 cm path length. Scanning was performed at a resolution of 0.1 nm, from 200 to 240 nm. Temperature was maintained at 37 °C using a Peltier temperature control device. CD spectra in Figure 2 represent an average of 10 accumulations, which were corrected by subtracting the buffer signal. No additional smoothing or filtering was performed. Peptide Exchange Assay. Peptide exchange experiments were carried out as described previously.22 Briefly, HLA-DQ2 covalently attached to CLIP peptide through a thrombin labile linker was treated with thrombin [40:1 (w/w)] for 90 min at 0 °C, after which fluorescein-conjugated peptide was added to a final concentration of 5.0 μM HLA-DQ2 and 0.2 μM peptide in 1:1 mixture of phosphate buffered saline and McIlvaine’s citrate-phosphate buffer at pH 5.5. HLA-DM was not used in this assay. The mixture was incubated for 2 h at 37 °C and complex formation was monitored every 30 min using high performance size exclusion chromatography. Two microliters of the reaction mixture was injected into a Shodex RW 803 size exclusion column using an auto injector and eluted with PBS buffer at 0.7 mL/ min flow rate. Complex formation was monitored by quantifying the area under the HLA-DQ2−peptide complex peak, which eluted ∼2.5 min earlier than the unbound peptide. Fluorescence signal was recorded using an in-line Shimadzu RF 10-AxL fluorescent detector with excitation wavelength set to 495 nm and detection wavelength set to 517 nm.

REFERENCES

(1) Rose, N. R. and MacKay, I. R. (2006) The Autoimmune Diseases, 4th ed., Academic Press, Boston, MA. (2) Thorsby, E. (1997) Invited anniversary review: HLA associated diseases. Hum. Immunol. 53, 1−11. (3) Rudensky, A., Preston-Hurlburt, P., al-Ramadi, B. K., Rothbard, J., and Janeway, C. A., Jr. (1992) Truncation variants of peptides isolated from MHC class II molecules suggest sequence motifs. Nature 359, 429−431. (4) Stepniak, D., Wiesner, M., de Ru, A. H., Moustakas, A. K., Drijfhout, J. W., Papadopoulos, G. K., van Veelen, P. A., and Koning, F. (2008) Large-scale characterization of natural ligands explains the unique gluten-binding properties of HLA-DQ2. J. Immunol. 180, 3268−3278. (5) Vartdal, F., Johansen, B. H., Friede, T., Thorpe, C. J., Stevanovic, S., Eriksen, J. E., Sletten, K., Thorsby, E., Rammensee, H. G., and Sollid, L. M. (1996) The peptide binding motif of the disease associated HLA-DQ (alpha 1* 0501, beta 1* 0201) molecule. Eur. J. Immunol. 26, 2764−2772. (6) Jones, P. P., Murphy, D. B., Hewgill, D., and McDevitt, H. O. (1979) Detection of a common polypeptide chain in I···A and I···E sub-region immunoprecipitates. Mol. Immunol. 16, 51−60. (7) Parrot, I., Huang, P. C., and Khosla, C. (2002) Circular dichroism and nuclear magnetic resonance spectroscopic analysis of immunogenic gluten peptides and their analogs. J. Biol. Chem. 277, 45572− 45578. (8) Herrera, M. G., Zamarreno, F., Costabel, M., Ritacco, H., Hutten, A., Sewald, N., and Dodero, V. I. (2013) Circular dichroism and electron microscopy studies in vitro of 33-mer gliadin peptide revealed secondary structure transition and supramolecular organization. Biopolymers, DOI: 10.1002/bip.22288. (9) Creamer, T. P., Campbell, M. N., and George, D. R. (2002) Determinants of the polyproline II helix from modeling studies. Adv. Protein Chem. 62, 263−282. (10) Stapley, B. J., and Creamer, T. P. (1999) A survey of left-handed polyproline II helices. Protein Sci. 8, 587−595. (11) Jardetzky, T. S., Brown, J. H., Gorga, J. C., Stern, L. J., Urban, R. G., Strominger, J. L., and Wiley, D. C. (1996) Crystallographic analysis of endogenous peptides associated with HLA-DR1 suggests a common, polyproline II-like conformation for bound peptides. Proc. Natl. Acad. Sci. U.S.A. 93, 734−738. (12) Kim, C. Y., Quarsten, H., Bergseng, E., Khosla, C., and Sollid, L. M. (2004) Structural basis for HLA-DQ2-mediated presentation of gluten epitopes in celiac disease. Proc. Natl. Acad. Sci. U.S.A. 101, 4175−4179. (13) Lyu, P. C., Liff, M. I., Marky, L. A., and Kallenbach, N. R. (1990) Side chain contributions to the stability of alpha-helical structure in peptides. Science 250, 669−673. (14) Chakrabartty, A., Kortemme, T., and Baldwin, R. L. (1994) Helix propensities of the amino acids measured in alanine-based peptides without helix-stabilizing side-chain interactions. Protein Sci. 3, 843−852. (15) Smith, C. K., Withka, J. M., and Regan, L. (1994) A thermodynamic scale for the beta-sheet forming tendencies of the amino acids. Biochemistry 33, 5510−5517. (16) Kelly, M. A., Chellgren, B. W., Rucker, A. L., Troutman, J. M., Fried, M. G., Miller, A. F., and Creamer, T. P. (2001) Host−guest study of left-handed polyproline II helix formation. Biochemistry 40, 14376−14383. (17) Rucker, A. L., Pager, C. T., Campbell, M. N., Qualls, J. E., and Creamer, T. P. (2003) Host−guest study of left-handed polyproline II helix formation. Proteins 53, 68−75. (18) Brown, A. M., and Zondlo, N. J. (2012) A propensity scale for type II polyproline helices (PPII): aromatic amino acids in proline-rich sequences strongly disfavor PPII due to proline−aromatic interactions. Biochemistry 51, 5041−5051. (19) Shi, Z., Chen, K., Liu, Z., Ng, A., Bracken, W. C., and Kallenbach, N. R. (2005) Polyproline II propensities from GGXGG

S Supporting Information *

LC−ESI−MS profile of all peptides used in this study. This material is available free of charge via the Internet at http:// pubs.acs.org.



Letters

AUTHOR INFORMATION

Corresponding Author

*(C.-Y.K.) E-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS We thank L. Sollid (University of Oslo) for the kind provision of the recombinant HLA-DQ2 protein. This research was supported by the Singapore Ministry of Health’s National Medical Research Council (grant: NMRC/NIG/0013/2007). 2386

dx.doi.org/10.1021/cb400594q | ACS Chem. Biol. 2013, 8, 2383−2387

ACS Chemical Biology

Letters

peptides reveal an anticorrelation with beta-sheet scales. Proc. Natl. Acad. Sci. U.S.A. 102, 17964−17968. (20) Horng, J. C., and Raines, R. T. (2006) Stereoelectronic effects on polyproline conformation. Protein Sci. 15, 74−83. (21) Zhang, S., Prabpai, S., Kongsaeree, P., and Arvidsson, P. I. (2006) Poly-N-methylated alpha-peptides: synthesis and X-ray structure determination of beta-strand forming foldamers. Chem. Commun. 5, 497−499. (22) Xia, J., Sollid, L. M., and Khosla, C. (2005) Equilibrium and kinetic analysis of the unusual binding behavior of a highly immunogenic gluten peptide to HLA-DQ2. Biochemistry 44, 4442− 4449. (23) Mitea, C., Havenaar, R., Drijfhout, J. W., Edens, L., Dekking, L., and Koning, F. (2008) Efficient degradation of gluten by a prolyl endoprotease in a gastrointestinal model: implications for coeliac disease. Gut 57, 25−32. (24) Bethune, M. T., and Khosla, C. (2012) Oral enzyme therapy for celiac sprue. Methods Enzymol. 502, 241−271. (25) Chan, W. C. and White, P. D. (2000) Fmoc Solid Phase Peptide Synthesis: A Practical Approach, Oxford University Press, New York.

2387

dx.doi.org/10.1021/cb400594q | ACS Chem. Biol. 2013, 8, 2383−2387