Portal de Periódicos da CAPES

Crystal Structures of HIV-1 Tat-derived Nonapeptides Tat-(1–9) and Trp2-Tat-(1–9) Bound to the Active Site of Dipeptidyl-peptidase IV (CD26)

2005; Elsevier BV; Volume: 280; Issue: 15 Linguagem: Inglês

10.1074/jbc.m413400200

1083-351X

Wilhelm A. Weihofen, Jianguo Liu, Werner Reutter, Wolfram Saenger, Hua Fan,

vaccines and immunoinformatics approaches

CD26 or dipeptidyl-peptidase IV (DPPIV) is engaged in immune functions by co-stimulatory effects on activation and proliferation of T lymphocytes, binding to adenosine deaminase, and regulation of various chemokines and cytokines. DPPIV peptidase activity is inhibited by both Tat protein from human immunodeficiency virus (HIV)-1 and its N-terminal nonapeptide Tat-(1–9) with amino acid sequence MDPVDPNIE, suggesting that DPPIV mediates immunosuppressive effects of Tat protein. The 2.0- and 3.15-Å resolution crystal structures of the binary complex between human DPPIV and nonapeptide Tat-(1–9) and the ternary complex between the variant MWPVDPNIE, called Trp2-Tat-(1–9), and DPPIV bound to adenosine deaminase show that Tat-(1–9) and Trp2-Tat-(1–9) are located in the active site of DPPIV. The interaction pattern of DPPIV with Trp2-Tat-(1–9) is tighter than that with Tat-(1–9), in agreement with inhibition constants (Ki) of 2 × 10–6 and 250 × 10–6 m, respectively. Both peptides cannot be cleaved by DPPIV because the binding pockets of the N-terminal 2 residues are interchanged compared with natural substrates: the N-terminal methionine occupies the hydrophobic S1 pocket of DPPIV that normally accounts for substrate specificity by binding the penultimate residue. Because the N-terminal sequence of the thromboxane A2 receptor resembles the Trp2-Tat-(1–9) peptide, a possible interaction with DPPIV is postulated. CD26 or dipeptidyl-peptidase IV (DPPIV) is engaged in immune functions by co-stimulatory effects on activation and proliferation of T lymphocytes, binding to adenosine deaminase, and regulation of various chemokines and cytokines. DPPIV peptidase activity is inhibited by both Tat protein from human immunodeficiency virus (HIV)-1 and its N-terminal nonapeptide Tat-(1–9) with amino acid sequence MDPVDPNIE, suggesting that DPPIV mediates immunosuppressive effects of Tat protein. The 2.0- and 3.15-Å resolution crystal structures of the binary complex between human DPPIV and nonapeptide Tat-(1–9) and the ternary complex between the variant MWPVDPNIE, called Trp2-Tat-(1–9), and DPPIV bound to adenosine deaminase show that Tat-(1–9) and Trp2-Tat-(1–9) are located in the active site of DPPIV. The interaction pattern of DPPIV with Trp2-Tat-(1–9) is tighter than that with Tat-(1–9), in agreement with inhibition constants (Ki) of 2 × 10–6 and 250 × 10–6 m, respectively. Both peptides cannot be cleaved by DPPIV because the binding pockets of the N-terminal 2 residues are interchanged compared with natural substrates: the N-terminal methionine occupies the hydrophobic S1 pocket of DPPIV that normally accounts for substrate specificity by binding the penultimate residue. Because the N-terminal sequence of the thromboxane A2 receptor resembles the Trp2-Tat-(1–9) peptide, a possible interaction with DPPIV is postulated. CD26 or dipeptidyl-peptidase IV (DPPIV) 1The abbreviations used are: DPPIV, dipeptidyl-peptidase IV; HIV, human immunodeficiency virus; bADA, bovine adenosine deaminase; hDPPIV, human DPPIV; SDF, stromally derived factor; PEG, polyethylene glycol. (EC 3.4.14.5) is a ubiquitous, multifunctional integral type II membrane glycoprotein located on the surface of a variety of epithelial, endothelial, and lymphoid cells. As an exopeptidase, it cleaves N-terminal dipeptides from polypeptides with proline or alanine in the penultimate position, thereby regulating the activity of a variety of biologically important peptides (2De Meester I. Korom S. Van Damme J. Scharpe S. Immunol. Today. 1999; 20: 367-375Abstract Full Text Full Text PDF PubMed Scopus (407) Google Scholar, 3Mentlein R. Regul. Pept. 1999; 85: 9-24Crossref PubMed Scopus (1151) Google Scholar). Some of these peptides are closely related to immune function, such as GLP-1 (glucagon-like peptide), GIP (glucose-dependent insulinotropic polypeptide) (4Marguet D. Baggio L. Kobayashi T. Bernard A.M. Pierres M. Nielsen P.F. Ribel U. Watanabe T. Drucker D.J. Wagtmann N. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 6874-6879Crossref PubMed Scopus (489) Google Scholar), RANTES (regulated on activation, normal T cell expressed and secreted) (5Oravecz T. Pall M. Roderiquez G. Gorrell M.D. Ditto M. Nguyen N.Y. Boykins R. Unsworth E. Norcross M.A. J. Exp. Med. 1997; 186: 1865-1872Crossref PubMed Scopus (318) Google Scholar), MDC (monocyte-derived chemokine) (6Proost P. Struyf S. Schols D. Opdenakker G. Sozzani S. Allavena P. Mantovani A. Augustyns K. Bal G. Haemers A. Lambeir A.M. Scharpe S. Van Damme J. De Meester I. J. Biol. Chem. 1999; 274: 3988-3993Abstract Full Text Full Text PDF PubMed Scopus (140) Google Scholar), SDF-1α and SDF-1β (stromally derived factors) (7Shioda T. Kato H. Ohnishi Y. Tashiro K. Ikegawa M. Nakayama E.E. Hu H. Kato A. Sakai Y. Liu H. Honjo T. Nomoto A. Iwamoto A. Morimoto C. Nagai Y. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 6331-6336Crossref PubMed Scopus (160) Google Scholar), eotaxin (8Struyf S. Proost P. Schols D. De Clercq E. Opdenakker G. Lenaerts J.P. Detheux M. Parmentier M. De Meester I. Scharpe S. Van Damme J. J. Immunol. 1999; 162: 4903-4909Crossref PubMed Google Scholar), and LD78β (9Proost P. Menten P. Struyf S. Schutyser E. De Meester I. Van Damme J. Blood. 2000; 96: 1674-1680Crossref PubMed Google Scholar). The activities of eotaxin, MDC, SDF-1α, and SFD-1β were abolished after truncation by DPPIV, whereas the chemotactic and anti-human immunodeficiency virus (HIV)-1 activities of RANTES and LD78β were increased significantly after processing by DPPIV (10Ohtsuki T. Tsuda H. Morimoto C. J. Dermatol. Sci. 2000; 22: 152-160Abstract Full Text Full Text PDF PubMed Scopus (33) Google Scholar, 11Struyf S. Proost P. Van Damme J. Adv. Immunol. 2003; 81: 1-44Crossref PubMed Scopus (103) Google Scholar). Inhibition of DPPIV activity by synthetic inhibitors resulted in suppression of T-cell proliferation in vitro and a decrease of antibody production in mice immunized with bovine serum albumin in vivo (12Kubota T. Flentke G.R. Bachovchin W.W. Stollar B.D. Clin. Exp. Immunol. 1992; 89: 192-197Crossref PubMed Scopus (101) Google Scholar, 13Augustyns K. Bal G. Thonus G. Belyaev A. Zhang X.M. Bollaert W. Lambeir A.M. Durinx C. Goossens F. Haemers A. Curr. Med. Chem. 1999; 6: 311-327PubMed Google Scholar). The HIV-1 Tat protein is a transactivator regulating the transcription of HIV-1 genes and is essential for viral replication in vitro (14Sodroski J. Rosen C. Wong-Staal F. Salahuddin S.Z. Popovic M. Arya S. Gallo R.C. Haseltine W.A. Science. 1985; 227: 171-173Crossref PubMed Scopus (307) Google Scholar, 15Dayton A.I. Sodroski J.G. Rosen C.A. Goh W.C. Haseltine W.A. Cell. 1986; 44: 941-947Abstract Full Text PDF PubMed Scopus (411) Google Scholar). HIV-1-infected cells release Tat into extracellular space, where Tat suppresses antigen- and mitogen-induced activation of human T cells. Full-length Tat inhibits DPPIV with high affinity (Ki = 0.02–11 nm) (16Gutheil W.G. Subramanyam M. Flentke G.R. Sanford D.G. Munoz E. Huber B.T. Bachovchin W.W. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 6594-6598Crossref PubMed Scopus (131) Google Scholar), whereas the N-terminal nonapeptide Tat-(1–9) with sequence MDPVDPNIE acts as weak inhibitor of DPPIV activity at a Ki of 250 μm (17Wrenger S. Reinhold D. Hoffmann T. Kraft M. Frank R. Faust J. Neubert K. Ansorge S. FEBS Lett. 1996; 383: 145-149Crossref PubMed Scopus (26) Google Scholar), and full-length Tat protein that was N-terminally modified with rhodamine lacks any inhibitory potential. Much improved inhibition was observed with a peptide analogue of Tat-(1–9) featuring Trp in the second position, Trp2-Tat-(1–9), with a Ki of 2 μm (18Lorey S. Stockel-Maschek A. Faust J. Brandt W. Stiebitz B. Gorrell M.D. Kahne T. Mrestani-Klaus C. Wrenger S. Reinhold D. Ansorge S. Neubert K. Eur. J. Biochem. 2003; 270: 2147-2156Crossref PubMed Scopus (50) Google Scholar). Interestingly, the N-terminal sequence Met-Trp-Pro is found in the G-protein-coupled thromboxane A2 receptor located on the surface of monocytes known as antigen-presenting cells. Because monocytes interact strongly with T cells carrying DPPIV, it was suggested that, similar to Tat protein, an interaction between the N terminus of thromboxane A2 receptor and DPPIV might modulate T-cell activation by inhibiting DPPIV proteolytic activity (19Wrenger S. Faust J. Mrestani-Klaus C. Fengler A. Stockel-Maschek A. Lorey S. Kahne T. Brandt W. Neubert K. Ansorge S. Reinhold D. J. Biol. Chem. 2000; 275: 22180-22186Abstract Full Text Full Text PDF PubMed Scopus (34) Google Scholar). Crystal structures of DPPIV free and in complexes with different inhibitors (20Rasmussen H.B. Branner S. Wiberg F.C. Wagtmann N. Nat. Struct. Biol. 2003; 10: 19-25Crossref PubMed Scopus (362) Google Scholar, 21Thoma R. Loffler B. Stihle M. Huber W. Ruf A. Hennig M. Structure (Lond.). 2003; 11: 947-959Abstract Full Text Full Text PDF PubMed Scopus (210) Google Scholar, 22Engel M. Hoffmann T. Wagner L. Wermann M. Heiser U. Kiefersauer R. Huber R. Bode W. Demuth H.U. Brandstetter H. Proc. Natl. Acad. Sci. U. S. A. 2003; 100: 5063-5068Crossref PubMed Scopus (287) Google Scholar, 23Gorrell M.D. Gysbers V. McCaughan G.W. Scand. J. Immunol. 2001; 54: 249-264Crossref PubMed Scopus (335) Google Scholar) and with a decapeptide substrate (24Aertgeerts K. Ye S. Tennant M.G. Kraus M.L. Rogers J. Sang B.C. Skene R.J. Webb D.R. Prasad G.S. Protein Sci. 2004; 13: 412-421Crossref PubMed Scopus (155) Google Scholar) showed that DPPIV occurs as homodimer. The DPPIV monomer is composed of a 6-residue-long N-terminal cytoplasmic tail followed by a 22-residue-long transmembrane α-helix, and the extracellular ectodomain is divided into an eight-bladed β-propeller domain (Arg54-Asn497) and a C-terminal α/β-hydrolase domain (Gln508-Pro766) harboring the catalytic triad (Asp708, His740, and Ser630) that is required for peptidase activity. DPPIV binds specifically to adenosine deaminase (23Gorrell M.D. Gysbers V. McCaughan G.W. Scand. J. Immunol. 2001; 54: 249-264Crossref PubMed Scopus (335) Google Scholar, 25Wilson D.K. Rudolph F.B. Quiocho F.A. Science. 1991; 252: 1278-1284Crossref PubMed Scopus (441) Google Scholar). Binding of adenosine deaminase to DPPIV is involved in regulation of the extracellular local concentration of adenosine (26Kameoka J. Tanaka T. Nojima Y. Schlossman S.F. Morimoto C. Science. 1993; 261: 466-469Crossref PubMed Scopus (455) Google Scholar, 27Schrader W.P. West C.A. Miczek A.D. Norton E.K. J. Biol. Chem. 1990; 265: 19312-19318Abstract Full Text PDF PubMed Google Scholar) that is important for the activation and proliferation of T lymphocytes. Recently, we determined the structure of the complex between human DPPIV (hDPPIV) and bovine adenosine deaminase (bADA) at 3.0-Å resolution (28Weihofen W.A. Liu J. Reutter W. Saenger W. Fan H. J. Biol. Chem. 2004; 279: 43330-43335Abstract Full Text Full Text PDF PubMed Scopus (81) Google Scholar) and 22-Å resolution (29Ludwig K. Fan H. Dobers J. Berger M. Reutter W. Bottcher C. Biochem. Biophys. Res. Commun. 2004; 313: 223-229Crossref PubMed Scopus (17) Google Scholar). In the complex with [hDPPIV·bADA]2 stoichiometry, the active sites of hDPPIV and bADA are fully accessible for substrates, and in hDPPIV, the possible alternative paths of substrate through the central tunnel of the β-propeller or through the side opening to the active site of hDPPIV are not blocked by bound bADA (Fig. 1). We report here the three-dimensional structures of complexes formed between hDPPIV and Tat-(1–9) with sequence MDPVDPNIE and between [hDPPIV·bADA]2 and the nonapeptide MWPVDPNIE that differs from Tat-(1–9) only in the replacement D2W, referred to henceforth as Trp2-Tat-(1–9). Our intention was to elucidate the binding site of these peptides and reveal their mechanism of DPPIV inhibition. Enzyme Purification—Full-length human DPPIV was produced and purified as described previously (30Dobers J. Zimmermann-Kordmann M. Leddermann M. Schewe T. Reutter W. Fan H. Protein Expression Purif. 2002; 25: 527-532Crossref PubMed Scopus (26) Google Scholar). Because the N-terminal 28 residues were uniformly truncated during purification as confirmed by N-terminal sequencing, hDPPIV was soluble and further purified by gel filtration on Superdex200 using buffer containing 20 mm Tris, pH 8.0, and 150 mm NaCl. Peptide Synthesis—Tat-(1–9) with sequence MDPVDPNIE and Trp2-Tat-(1–9) (MWPVDPNIE) were custom synthesized and high pressure liquid chromatography purified by Dr. Volkmar Engert (Charité Berlin, Germany). Crystallization of Tat-(1–9)·hDPPIV—hDPPIV was concentrated to 5 mg/ml and crystallized by sparse matrix screening (Hampton Research) in sitting drop wells. Small crystals grew in conditions containing PEG 3350 and PEG 4000, but just a single crystal of hDPPIV grew to a size suitable for x-ray analysis from reservoir solution containing 20% PEG 3350, 100 mm KCl, and 100 mm Tris, pH 8.5. This crystal was soaked for 2 h in reservoir solution supplemented with 1 mm Tat-(1–9) and 20% glycerol for cryo-cooling. Crystallization of Trp2-Tat-(1–9)·hDPPIV·bADA—hDPPIV·bADA was produced and purified as described previously (30Dobers J. Zimmermann-Kordmann M. Leddermann M. Schewe T. Reutter W. Fan H. Protein Expression Purif. 2002; 25: 527-532Crossref PubMed Scopus (26) Google Scholar) and crystallized with microseeding from a reservoir solution containing 20–22% PEG 3350, 200 mm NaCl, and 100 mm Tris-HCl, pH 8.0 (28Weihofen W.A. Liu J. Reutter W. Saenger W. Fan H. J. Biol. Chem. 2004; 279: 43330-43335Abstract Full Text Full Text PDF PubMed Scopus (81) Google Scholar). Crystals were incubated for 2 h in reservoir solution containing Trp2-Tat-(1–9) peptide at 100 μm concentration. Subsequently, crystals were soaked in 25% PEG 3350, 200 mm NaCl, and 100 mm Tris, pH 8.0, supplemented with 20% glycerol and cryo-cooled. X-ray diffraction power and quality of these crystals were improved by annealing crystals to room temperature two times for 3 s. Structure Determination—X-ray data were collected at beamline ID14-2, European Synchrotron Radiation Facility (Grenoble, France) and processed with DENZO and SCALEPACK (31Otwinowski Z. Minor W. Carter Jr., C.W. Sweet R.M. Macromolecular Crystallography Part A. 276. Academic Press, New York1997: 307-326Google Scholar) (see Table I). Because the crystals of the Tat-(1–9) and Trp2-Tat-(1–9) complexes were isomorphous to (hDPPIV)2 with one dimer in the asymmetric unit (Protein Data Bank code 1N1M) and [hDPPIV·bADA]2 (Protein Data Bank code 1W1I) with two heterotetramers in the asymmetric unit, respectively, the initial structures used glycoside-depleted hDPPIV and were adjusted for small packing differences by rigid body refinements with the program REFMAC5.Table ICrystallographic data and refinement statisticsTat-(1-9)·DPPIVTrp2-Tat-(1-9)·DPPIV·ADASpace groupP212121C2Unit cell constants (Å)a = 158.1, b = 168.5, c = 238a = 158.1, b = 168.5, c = 238.8, β = 100.5°Resolution range (Å)30.0 to 2.030.0 to 3.15No. of observations548,144243,240No. of unique reflections128,89897,782Completeness (%)aValues in parentheses refer to the outer resolution shell.92.2 (87.7)89.3 (85.9)〈I/σ(I)〉aValues in parentheses refer to the outer resolution shell.19.8 (3.1)9.1 (2.6)Rsym (%)aValues in parentheses refer to the outer resolution shell.,bRsym = (Σ|Ihkl — 〈I〉|)/(Σ Ihkl), where Ihkl is the observed intensity, and 〈I〉 is the average intensity obtained from multiple observations of symmetry-related reflections.6.8 (37)10.6 (41)Refinement statisticsNo. of residues/atoms1,538/14,4544,440/35,904Rwork16.022.4Rfree20.324.6Root mean square deviationscComputed with PROCHECK.Bond lengths (Å)0.0120.01Bond angles (o)1.51.5a Values in parentheses refer to the outer resolution shell.b Rsym = (Σ|Ihkl — 〈I〉|)/(Σ Ihkl), where Ihkl is the observed intensity, and 〈I〉 is the average intensity obtained from multiple observations of symmetry-related reflections.c Computed with PROCHECK. Open table in a new tab The crystal structures were determined by difference Fourier techniques. The (Fo – Fc) electron density maps showed the bound peptides Tat-(1–9) and Trp2-Tat-(1–9) in the active sites of DPPIV. The three N-terminal residues of Tat-(1–9) and glycosides bound to each DPPIV, but not the remaining 6 residues of Tat-(1–9), could be modeled manually into the electron density. Similarly, for Trp2-Tat-(1–9), only the 6 N-terminal residues and the sugars could be modeled. Subsequent restrained refinement applied 2-fold non-crystallographic symmetry to the protein chains throughout in [Tat-(1–9)·hDPPIV]2 and 4-fold (two times 2-fold) non-crystallographic symmetry in [Trp2-Tat-(1–9)·hDPPIV·bADA]2, but the peptides were refined independently. Further TLS refinement implemented in REFMAC5 yielded the R and Rfree values provided in Table I together with other refinement statistics. Figures of molecules were prepared with MOLSCRIPT (32Kraulis P.J. J. Appl. Crystallogr. 1991; 24: 946-950Crossref Google Scholar) and RASTER3D (33Merritt E.A. Bacon D.J. Methods Enzymol. 1997; 277: 505-524Crossref PubMed Scopus (3875) Google Scholar). Atomic coordinates were deposited in the Protein Data Bank under accession codes 2BGR and 2BGN for Tat-(1–9)·DPPIV and Trp2-Tat-(1–9)·hDPPIV·bADA, respectively. Overall Structure of [Trp2-Tat-(1–9)·hDPPIV·bADA]2—Because crystals of [hDPPIV·bADA]2 were incubated with Trp2-Tat-(1–9), the stoichiometry was maintained, and neither hDPPIV nor bADA suffered significant structural changes upon binding of the peptide, as shown by a root mean square deviation of 0.3 Å between protein backbone positions of the complex with the peptide and of free [hDPPIV·bADA]2. The crystal asymmetric unit contains two structurally isomorphous complexes [Trp2-Tat-(1–9)·hDPPIV·bADA]2. Each hDPPIV binds one bADA at the periphery of the β-propeller domain (Fig. 1), and one Trp2-Tat-(1–9) is located in the active site of DPPIV between the α/β-hydrolase domain and the β-propeller domain. The N-terminal 4 residues of Trp2-Tat-(1–9) are clearly defined in the electron density and permitted unambiguous modeling of this part of the peptide. They form intermolecular interactions with DPPIV as shown in Figs. 2A and 3 that are of van der Waals type. In addition, the N-terminal ammonium group forms salt bridges to Glu205 and Glu206 carboxylates and a hydrogen bond to Tyr662Oη, and the Trp2 indole Nϵ hydrogen bonds to Glu206O. The main chains of Asp5-Pro6 are also well defined in the electron density, but only poor density is observed for the side chains because they do not form contacts to DPPIV. These side chains were consequently placed with the most likely conformation determined by NMR spectroscopy of peptide Trp2-Tat-(1–9) in aqueous solution, which indicated that the main chain conformation of this nonapeptide is determined by Pro3 (17Wrenger S. Reinhold D. Hoffmann T. Kraft M. Frank R. Faust J. Neubert K. Ansorge S. FEBS Lett. 1996; 383: 145-149Crossref PubMed Scopus (26) Google Scholar, 34Wrenger S. Hoffmann T. Faust J. Mrestani-Klaus C. Brandt W. Neubert K. Kraft M. Olek S. Frank R. Ansorge S. Reinhold D. J. Biol. Chem. 1997; 272: 30283-30288Abstract Full Text Full Text PDF PubMed Scopus (57) Google Scholar, 35Wrenger S. Reinhold D. Faust J. Mrestani-Klaus C. Brandt W. Fengler A. Neubert K. Ansorge S. Adv. Exp. Med. Biol. 2000; 477: 161-165Crossref PubMed Google Scholar). It is notable that the conformation of Trp2-Tat-(1–9) in the NMR study is comparable to that found here for the N-terminal 6 residues of Trp2-Tat-(1–9). The C-terminal 3 residues of Trp2-Tat-(1–9) are not seen in the electron density due to solvent exposure and possible disorder.Fig. 3Stereo representation of surface area around the active site of hDPPIV with the backbone of bound six N-terminal residues MWPVDP of Trp2-Tat-(1–9) shown in green; the remaining 3 residues are not seen in the electron density. Residues of hDPPIV engaged in peptide binding are indicated in gray. Oxygen atoms are red, nitrogen is blue, and sulfur is yellow. Hydrogen bonds are indicated by black dashed lines, and important amino acids are labeled.View Large Image Figure ViewerDownload Hi-res image Download (PPT) The backbones of the independently refined peptides superimpose with root mean square deviations of about 0.4 Å, indicating that they adopt comparable conformations. Upon binding to the active site, 850 Å2 of surface area are buried for each peptide. A second binding site for the inhibitor or a binding site outside the active center as suggested from kinetic studies for full-length Tat protein (16Gutheil W.G. Subramanyam M. Flentke G.R. Sanford D.G. Munoz E. Huber B.T. Bachovchin W.W. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 6594-6598Crossref PubMed Scopus (131) Google Scholar, 18Lorey S. Stockel-Maschek A. Faust J. Brandt W. Stiebitz B. Gorrell M.D. Kahne T. Mrestani-Klaus C. Wrenger S. Reinhold D. Ansorge S. Neubert K. Eur. J. Biochem. 2003; 270: 2147-2156Crossref PubMed Scopus (50) Google Scholar) was not observed in the electron density. Overall Structure of [Tat-(1–9)·hDPPIV]2—The 2.0-Å resolution crystal structure of hDPPIV complexed with Tat-(1–9) was refined to R and Rfree values of 16.0 and 20.3%, respectively (Table I). The asymmetric unit contains one homodimeric [Tat-(1–9)·hDPPIV]2 with comparable geometry as shown in Fig. 1 (with bADA omitted). An Fo – Fc electron density peak above 10.0 σ in the active site of hDPPIV served as a landmark to place the sulfur of the N-terminal methionine side chain containing the only sulfur atom in Tat-(1–9). Given the position of the methionine side chain, the N terminus of Tat-(1–9) could be modeled. The N-terminal ammonium group forms salt bridges and a hydrogen bond similar to that described for hDPPIV bound to Trp2-Tat-(1–9) and to the substrate Ile-Pro-Ile (21Thoma R. Loffler B. Stihle M. Huber W. Ruf A. Hennig M. Structure (Lond.). 2003; 11: 947-959Abstract Full Text Full Text PDF PubMed Scopus (210) Google Scholar), see Fig. 2, A, C, and E. Difference in electron density allowed unambiguous modeling of the conformation of Asp2 of Tat-(1–9). Only sparse density was observed for the main chain of Pro3 (Fig. 2C), and the rest of the peptide lacks any difference electron density and could not be modeled. This is surprising in comparison to crystal structures of hDPPIV in complex with Trp2-Tat-(1–9) presented above and with neuropeptide Y (24Aertgeerts K. Ye S. Tennant M.G. Kraus M.L. Rogers J. Sang B.C. Skene R.J. Webb D.R. Prasad G.S. Protein Sci. 2004; 13: 412-421Crossref PubMed Scopus (155) Google Scholar) because both structures show electron density for at least 6 residues in the active site. The peptides in each half site of the DPPIV dimer were independently refined. Their backbones show comparable conformations and superimpose with root mean square deviations of about 0.5 Å. Upon binding to the active site, 600 Å2 of surface area are buried for each peptide. As noticed for Trp2-Tat-(1–9) (see above) a second binding site for the inhibitor or a binding site outside the active site was not observed. Substrate Access to the Binding Site—As shown by crystal structures of DPPIV in complex with different inhibitors (20Rasmussen H.B. Branner S. Wiberg F.C. Wagtmann N. Nat. Struct. Biol. 2003; 10: 19-25Crossref PubMed Scopus (362) Google Scholar, 21Thoma R. Loffler B. Stihle M. Huber W. Ruf A. Hennig M. Structure (Lond.). 2003; 11: 947-959Abstract Full Text Full Text PDF PubMed Scopus (210) Google Scholar, 22Engel M. Hoffmann T. Wagner L. Wermann M. Heiser U. Kiefersauer R. Huber R. Bode W. Demuth H.U. Brandstetter H. Proc. Natl. Acad. Sci. U. S. A. 2003; 100: 5063-5068Crossref PubMed Scopus (287) Google Scholar, 23Gorrell M.D. Gysbers V. McCaughan G.W. Scand. J. Immunol. 2001; 54: 249-264Crossref PubMed Scopus (335) Google Scholar) and by the complex of DPPIV and a decapeptide substrate (24Aertgeerts K. Ye S. Tennant M.G. Kraus M.L. Rogers J. Sang B.C. Skene R.J. Webb D.R. Prasad G.S. Protein Sci. 2004; 13: 412-421Crossref PubMed Scopus (155) Google Scholar), there are two possible entrances for substrates to the active site of DPPIV. The two routes are both fully accessible in the complex [hDPPIV·bADA]2 as shown in the present study and in our earlier study (28Weihofen W.A. Liu J. Reutter W. Saenger W. Fan H. J. Biol. Chem. 2004; 279: 43330-43335Abstract Full Text Full Text PDF PubMed Scopus (81) Google Scholar). The side opening formed between the β-propeller domain and the α/β-hydrolase domain and the central tunnel of the eight-bladed β-propeller domain measures 20 and 15 Å in diameter, respectively. The bound Trp2-Tat-(1–9) points with the N terminus to the active site of hDPPIV and with the C terminus pointing to the side opening, suggesting that this is the entrance to the active site utilized by the peptide (Fig. 1). Tat-(1–9) and Trp2-Tat-(1–9) Do Not Adopt the Same Conformation as a Substrate in the Active Site of DPPIV—As serine exopeptidase, DPPIV cleaves the N-terminal dipeptide of substrates with Pro or Ala in the penultimate position called P1 in the protease nomenclature introduced by Berger and Schechter (1Berger A. Schechter I. Philos. Trans. R Soc. Lond. B Biol. Sci. 1970; 257: 249-264Crossref PubMed Scopus (378) Google Scholar), see Scheme 1. Substrate specificity is associated with the narrow hydrophobic pocket S1 (Fig. 2E) that accommodates the side chain of residue P1 of a substrate and restricts possible residues fitting to this pocket in DPPIV to be unpolar and as large or smaller than Pro, consequently also allowing for Ala, as actually observed for the Pro2/Ala2 dimorphism of DPPIV substrates. The side chain of the upstream N-terminal residue P2 points to the large cavity S2 that gives rise to low specificity of DPPIV for this position. The catalytically active Ser630 attacks the scissile carbonyl group between P1 and P1′ to from a tetrahedral intermediate as observed in the structures of DPPIV in complex with two peptide inhibitors, the tripeptide Ile-Pro-Ile (Fig. 2E) (21Thoma R. Loffler B. Stihle M. Huber W. Ruf A. Hennig M. Structure (Lond.). 2003; 11: 947-959Abstract Full Text Full Text PDF PubMed Scopus (210) Google Scholar) and the N-terminal decapeptide of neuropeptide Y (24Aertgeerts K. Ye S. Tennant M.G. Kraus M.L. Rogers J. Sang B.C. Skene R.J. Webb D.R. Prasad G.S. Protein Sci. 2004; 13: 412-421Crossref PubMed Scopus (155) Google Scholar). Compared with a natural substrate, Trp2-Tat-(1–9) exhibits a different backbone conformation when bound to the active site of DPPIV. Except for the N-terminal ammonium group that is anchored to the active site in the same way as found for natural substrates, the side chains of P1 and P2 interchange their supposed binding pockets, with P1 binding to S2 and P2 binding to S1 (Scheme 1; Fig. 2, A, B, and E). The side chain of Met1 (P2) mimics the conformation of proline in the P1 position of a feasible substrate and fills pocket S1, which is supposed to assure substrate specificity and is lined by Ser630, Tyr631, Val656, Tyr662, and Tyr666 (Fig. 2, A and B). The subsequent Trp2 of Trp2-Tat-(1–9) points to and is located in cavity S2, which is supposed to nonspecifically accommodate the side chain of P2. The bulky indole group of Trp2 (P1) forms a hydrogen bond to the backbone carbonyl of Glu206 and packs against Arg125, Phe357, and Tyr547 (Fig. 2, A and B). The side chain arrangements of Met1 and Trp2 force the backbone of Trp2-Tat-(1–9) into a conformation that shows the peptide plane of the scissile bond (Trp2-Pro3) rotated by 120° so that the carbonyl carbon of Trp2 is displaced from Ser630Oγ by 5 Å, and this geometry prevents formation of a tetrahedral intermediate and turnover (see Figs. 2, A and B, and 3; compare with Fig. 2E). Pro3 in position P1′ and Val4 in position P2′ form hydrophobic interactions with Tyr547 and Trp629, respectively (Figs. 2, A and B, and 3). Asp5 (P3′) and Pro6 (P4′) could be located from the electron density but are more than ∼4 Å distant to residues of DPPIV so that only the N-terminal 4 residues of the peptide interact tightly with DPPIV. Tat-(1–9) occupies the active site of DPPIV with the N-terminal Met1 (P2) bound to the S1 pocket (Fig. 2, C and D), and the N-terminal ammonium group interacts with the active site as described above for the Trp2-Tat-(1–9) complex. However, the subsequent residues Asp2 and Pro3 adopt a different position and orientation because Asp2 is rotated by ∼180° compared with Trp2 in Trp2-Tat-(1–9) to form a hydrogen bond (Asp2Oδ–Ser630Oγ) to the catalytically active Ser630 and is placed between Ser630 and the scissile bond of the peptide, thereby preventing turnover. Difference electron density only allowed modeling the backbone of Pro3, and no density was observed for the side chain of Pro3 (Fig. 2C) that was modeled based on the geometry of the main chain. In the final model, Pro3 is displaced by >4 Å from any DPPIV residue, and all interactions are restricted to Met1 and Asp2 (Fig. 2, C and D). Considering the less dense interaction pattern of Tat-(1–9) bound to the DPPIV active site and the buried surface area of only 600 Å2 upon complex formation, the reported lower binding affinity (see below) of Tat-(1–9) to the active site of DPPIV relative to Trp2-Tat-(1–9) is in good agreement with our structural data. Structural Data Suggest a Competitive Type of Inhibition Contrasting Kinetic Data—Extensive kinetic studies were performed regarding the inhibition potentials of the original Tat-(1–9) peptide and Tat-(1–9) derivatives on DPPIV activity. Some results of these studies are given in Table II. Depending on the experimental setup, Ki values and type of inhibition were determined for Tat-(1–9)-derived peptides with some of the N-terminal amino acids exchanged. Here we only discuss the results for Tat-(1–9) and Trp2-Tat-(1–9).Table IIKinetic constants of Tat-(1–9)-derived peptides for inhibition of DPPIV, taken from Ref.18Lorey S. Stockel-Maschek A. Faust J. Brandt W. Stiebitz B. Gorrell M.D. Kahne T. Mrestani-Klaus C. Wrenger S. Reinhold D. Ansorge S. Neubert K. Eur. J. Biochem. 2003; 270: 2147-2156Crossref PubMed Scopus (50) Google ScholarPeptideSequenceKiType of inhibitionTat-(1-9)MDPVDPNIE2.5 × 10-4Parabolic mixed typeTrp2-Tat-(1-9)MWPVDPNIE1.9 × 10-6Linear mixed typeMet-Trp-ProMWP2.45 × 10-5CompetitiveThromboxane A2-receptor-(1-9)MWPNGSSLG5.02 × 10-5CompetitiveGly3-Tat-(1-9)MDGVDPNIE4.9 × 10-4Parabolic mixed typeIle3-Tat-(1-9)MDIVDPNIE1.7 × 10-3Parabolic mixed typeTrp2,Ile3-Tat-(1-9)MWIVDPNIE4.4 × 10-5Competitive Open table in a new tab As shown by the data in Table II, none of the investigated single replacements among the 4 N-terminal residues of Tat-(1–9) fully inhibited DPPIV. They slightly altered Ki, and the type of inhibition changed in some cases to a competitive or parabolic/linear mixed type. Compared with Tat-(1–9), which inhibits DPPIV weakly with a Ki of about 250 μm, the best inhibitor is Trp2-Tat-(1–9) with Ki of 1.9 μm. The types of inhibition observed in these two cases (parabolic mixed and linear mixed, respectively) suggest Tat-(1–9) and Trp2-Tat-(1–9) to be located not only in the active site but also in a noncompetitive binding site on DPPIV that is distant from the active site (18Lorey S. Stockel-Maschek A. Faust J. Brandt W. Stiebitz B. Gorrell M.D. Kahne T. Mrestani-Klaus C. Wrenger S. Reinhold D. Ansorge S. Neubert K. Eur. J. Biochem. 2003; 270: 2147-2156Crossref PubMed Scopus (50) Google Scholar). This contradicts our findings that suggest purely competitive inhibition and demands a new interpretation of the kinetic data. One could argue, however, that we find Tat-(1–9) and Trp2-Tat-(1–9) only in the active site and not in a second site because the complexes were generated by soaking of the [hDPPIV]2 and [hDPPIV·bADA]2 crystals and that crystal packing contacts and binding of ADA to DPPIV had obscured secondary binding sites of the ligands. This, however, can be ruled out because [hDPPIV]2 and [hDPPIV·bADA]2 crystallized in two different crystal forms with different crystal contacts and a second binding site was not observed in either case. Structures Partly Explain the Kinetic Behavior of Single Mutants of Tat-(1–9)—The structural data of Tat-(1–9) and Trp2-Tat-(1–9) bound to the active site of DPPIV suggest a new interpretation of kinetic data determined for Tat-(1–9) and derivatives. For this we presuppose that peptides with inhibitory potential are bound with their N-terminal ammonium group to Glu205 and Glu206 in the active site of DPPIV (see above). On this basis, we propose that peptides comprising a hydrophobic N-terminal residue show the same binding pattern to DPPIV as observed for Tat-(1–9) and Trp2-Tat-(1–9). The hydrophobic side chain would likely replace the Tat-(1–9) methionine side chain in the S1 pocket of DPPIV. This view is supported by a kinetic study showing that the peptides FAPAG, FHPKR, and IKPEA have an inhibitory potential comparable to peptides featuring an N-terminal Met (17Wrenger S. Reinhold D. Hoffmann T. Kraft M. Frank R. Faust J. Neubert K. Ansorge S. FEBS Lett. 1996; 383: 145-149Crossref PubMed Scopus (26) Google Scholar). Peptides with Ala, Gly, or Thr in the N-terminal position lacked the ability to inhibit DPPIV and indicated that an N-terminal hydrophobic residue is a prerequisite for a peptide to act as a DPPIV inhibitor. The influence of single amino acid exchanges at position 2 of Tat-(1–9) on the ability to inhibit DPPIV was investigated in another study (35Wrenger S. Reinhold D. Faust J. Mrestani-Klaus C. Brandt W. Fengler A. Neubert K. Ansorge S. Adv. Exp. Med. Biol. 2000; 477: 161-165Crossref PubMed Google Scholar). Only relative inhibitory potentials were determined, and the ranking of residues with increasing influence on the binding affinity was found to be D < P ≪ S < G < K < F < A ≪ W. Tat-(1–9) comprising Asp2 shows no prominent role, but Ala2 features the second best inhibition potential and suggests that neither extensive hydrophobic contacts nor polar interactions of residue 2 are important determinants for strong DPPIV binding affinity. How much the presence of Pro in position 3 accounts for binding affinity to DPPIV was investigated previously (18Lorey S. Stockel-Maschek A. Faust J. Brandt W. Stiebitz B. Gorrell M.D. Kahne T. Mrestani-Klaus C. Wrenger S. Reinhold D. Ansorge S. Neubert K. Eur. J. Biochem. 2003; 270: 2147-2156Crossref PubMed Scopus (50) Google Scholar). The results in Table II indicate that the inhibition constant of Trp2-Tat-(1–9) is about 20-fold lower compared with that of Trp2, Ile3-Tat-(1–9). Pro3 stabilizes the conformation of the peptide, which reduces loss of entropy upon binding and results in stronger binding affinity (34Wrenger S. Hoffmann T. Faust J. Mrestani-Klaus C. Brandt W. Neubert K. Kraft M. Olek S. Frank R. Ansorge S. Reinhold D. J. Biol. Chem. 1997; 272: 30283-30288Abstract Full Text Full Text PDF PubMed Scopus (57) Google Scholar). Trp2-Tat-(1–9) bound to the active site of DPPIV shows Pro3 stacked with Tyr547, but in the case of Tat-(1–9), no contacts were found between Pro3 and any DPPIV residues (see above). To what extent Pro3 contributes to binding affinity can hardly be derived from our structural data because the two peptides show different backbone conformations. Kinetic studies of Tat-(1–9) peptides with replacements in position 5 or 6 showed similar inhibitory potential compared with wild type Tat-(1–9) (34Wrenger S. Hoffmann T. Faust J. Mrestani-Klaus C. Brandt W. Neubert K. Kraft M. Olek S. Frank R. Ansorge S. Reinhold D. J. Biol. Chem. 1997; 272: 30283-30288Abstract Full Text Full Text PDF PubMed Scopus (57) Google Scholar). This is consistent with our structure and the crystal structure of DPPIV in complex with the N-terminal decapeptide of neuropeptide Y (24Aertgeerts K. Ye S. Tennant M.G. Kraus M.L. Rogers J. Sang B.C. Skene R.J. Webb D.R. Prasad G.S. Protein Sci. 2004; 13: 412-421Crossref PubMed Scopus (155) Google Scholar), both showing that substrate recognition does not extend beyond P2′ (Scheme 1). Implications for Tat Binding—Tat protein was reported to bind to DPPIV at physiological salt concentrations without inhibiting the protease activity of DPPIV against small chromogenic substrates used to assay enzymatic activity (16Gutheil W.G. Subramanyam M. Flentke G.R. Sanford D.G. Munoz E. Huber B.T. Bachovchin W.W. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 6594-6598Crossref PubMed Scopus (131) Google Scholar). The kinetics of inhibition indicate that the affinity of Tat for DPPIV varies from 20 pm to 11 nm, and the activity of DPPIV varies from 13% to 100% depending on NaCl concentration in the range between 0 and 140 mm. Another study reported that the N-terminal residues of Tat protein are essential for the inhibition of DPPIV-catalyzed degradation of interleukin-2 (1Berger A. Schechter I. Philos. Trans. R Soc. Lond. B Biol. Sci. 1970; 257: 249-264Crossref PubMed Scopus (378) Google Scholar, 2De Meester I. Korom S. Van Damme J. Scharpe S. Immunol. Today. 1999; 20: 367-375Abstract Full Text Full Text PDF PubMed Scopus (407) Google Scholar, 3Mentlein R. Regul. Pept. 1999; 85: 9-24Crossref PubMed Scopus (1151) Google Scholar, 4Marguet D. Baggio L. Kobayashi T. Bernard A.M. Pierres M. Nielsen P.F. Ribel U. Watanabe T. Drucker D.J. Wagtmann N. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 6874-6879Crossref PubMed Scopus (489) Google Scholar, 5Oravecz T. Pall M. Roderiquez G. Gorrell M.D. Ditto M. Nguyen N.Y. Boykins R. Unsworth E. Norcross M.A. J. Exp. Med. 1997; 186: 1865-1872Crossref PubMed Scopus (318) Google Scholar, 6Proost P. Struyf S. Schols D. Opdenakker G. Sozzani S. Allavena P. Mantovani A. Augustyns K. Bal G. Haemers A. Lambeir A.M. Scharpe S. Van Damme J. De Meester I. J. Biol. Chem. 1999; 274: 3988-3993Abstract Full Text Full Text PDF PubMed Scopus (140) Google Scholar, 7Shioda T. Kato H. Ohnishi Y. Tashiro K. Ikegawa M. Nakayama E.E. Hu H. Kato A. Sakai Y. Liu H. Honjo T. Nomoto A. Iwamoto A. Morimoto C. Nagai Y. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 6331-6336Crossref PubMed Scopus (160) Google Scholar, 8Struyf S. Proost P. Schols D. De Clercq E. Opdenakker G. Lenaerts J.P. Detheux M. Parmentier M. De Meester I. Scharpe S. Van Damme J. J. Immunol. 1999; 162: 4903-4909Crossref PubMed Google Scholar, 9Proost P. Menten P. Struyf S. Schutyser E. De Meester I. Van Damme J. Blood. 2000; 96: 1674-1680Crossref PubMed Google Scholar, 10Ohtsuki T. Tsuda H. Morimoto C. J. Dermatol. Sci. 2000; 22: 152-160Abstract Full Text Full Text PDF PubMed Scopus (33) Google Scholar, 11Struyf S. Proost P. Van Damme J. Adv. Immunol. 2003; 81: 1-44Crossref PubMed Scopus (103) Google Scholar, 12Kubota T. Flentke G.R. Bachovchin W.W. Stollar B.D. Clin. Exp. Immunol. 1992; 89: 192-197Crossref PubMed Scopus (101) Google Scholar) (17Wrenger S. Reinhold D. Hoffmann T. Kraft M. Frank R. Faust J. Neubert K. Ansorge S. FEBS Lett. 1996; 383: 145-149Crossref PubMed Scopus (26) Google Scholar). Modification of the N termini of Tat-(1–72) and Tat-(1–9) with rhodamine abolished inhibition in both cases. These observations and the reported low Ki values for wild-type Tat binding in comparison to moderate binding of Tat-(1–9) to DPPIV suggest that the N terminus of Tat is not the only determinant for binding and that other intermolecular interactions might contribute as well. DPPIV has been shown to play a crucial role in T-cell activation and in several functions of the immune system (2De Meester I. Korom S. Van Damme J. Scharpe S. Immunol. Today. 1999; 20: 367-375Abstract Full Text Full Text PDF PubMed Scopus (407) Google Scholar, 36Morimoto C. Schlossman S.F. Immunol. Rev. 1998; 161: 55-70Crossref PubMed Scopus (366) Google Scholar, 37Franco R. Valenzuela A. Lluis C. Blanco J. Immunol. Rev. 1998; 161: 27-42Crossref PubMed Scopus (157) Google Scholar, 38von Bonin A. Huhn J. Fleischer B. Immunol. Rev. 1998; 161: 43-53Crossref PubMed Scopus (132) Google Scholar). Previous evidence implicated DPPIV in antigen-specific T-cell activation events and Tat protein in suppression of antigen-induced, but not mitogen-induced, T-cell proliferation. In the present work, we show that the N-terminal residues of Tat-(1–9) bind to the active site of DPPIV and provide evidence that Tat-(1–9) might inhibit DPPIV activity. This suggests that the immunosuppressive effect of Tat is mediated, at least partly, by competitive inhibition of DPPIV activity. In addition, the inhibition of DPPIV activity influences the processing of chemokines SDF-1α, SDF-1β, RANTES, and LD78β, thereby modulating their anti-HIV activities. We are grateful to Dr. Jörg Dobers for initial work and to Melanie Leddermann and Sabine Stehling for excellent technical assistance.