NMR Studies of Active N-terminal Peptides of Stromal Cell-derived Factor-1
2000; Elsevier BV; Volume: 275; Issue: 35 Linguagem: Inglês
10.1016/s0021-9258(19)61446-8
ISSN1083-351X
AutoresElena Elisseeva, Carolyn M. Slupsky, Matthew P. Crump, Ian Clark‐Lewis, Brian D. Sykes,
Tópico(s)Immunotherapy and Immune Responses
ResumoStromal cell-derived factor 1 (SDF-1), a member of the CXC chemokine family, is the only chemokine to bind to the receptor CXCR4. This receptor is also a co-receptor for syncytia-inducing forms of HIV in CD4+ cells. In addition, SDF-1 is responsible for attracting mature lymphocytes to the bone marrow and can therefore contribute to host versusgraft rejection in bone marrow transplantation. Clearly, by manipulating SDF-1 activity, we could find a possible anti-viral AIDS treatment and aid in bone marrow transplantation. SDF-1 binds to CXCR4 primarily via the N terminus, which appears flexible in the recently determined three-dimensional structure of SDF-1. Strikingly, short N-terminal SDF-1 peptides have been shown to have significant SDF-1 activity. By using NMR, we have determined the major conformation of the N terminus of SDF-1 in a 17-mer (residues 1–17 of SDF-1) and a 9-mer dimer (residues 1–9 of SDF-1 linked by a disulfide bond at residue 9). Residues 5–8 and 11–14 form similar structures that can be characterized as a β-turn of the β-αR type. These structural motifs are likely to be interconverting with other states, but the major conformation may be important for recognition in receptor binding. These results suggest for the first time that there may be a link between structuring of short N-terminal chemokine peptides and their ability to activate their receptor. These studies will act as a starting point for synthesizing non-peptide analogs that act as CXCR4 antagonists. Stromal cell-derived factor 1 (SDF-1), a member of the CXC chemokine family, is the only chemokine to bind to the receptor CXCR4. This receptor is also a co-receptor for syncytia-inducing forms of HIV in CD4+ cells. In addition, SDF-1 is responsible for attracting mature lymphocytes to the bone marrow and can therefore contribute to host versusgraft rejection in bone marrow transplantation. Clearly, by manipulating SDF-1 activity, we could find a possible anti-viral AIDS treatment and aid in bone marrow transplantation. SDF-1 binds to CXCR4 primarily via the N terminus, which appears flexible in the recently determined three-dimensional structure of SDF-1. Strikingly, short N-terminal SDF-1 peptides have been shown to have significant SDF-1 activity. By using NMR, we have determined the major conformation of the N terminus of SDF-1 in a 17-mer (residues 1–17 of SDF-1) and a 9-mer dimer (residues 1–9 of SDF-1 linked by a disulfide bond at residue 9). Residues 5–8 and 11–14 form similar structures that can be characterized as a β-turn of the β-αR type. These structural motifs are likely to be interconverting with other states, but the major conformation may be important for recognition in receptor binding. These results suggest for the first time that there may be a link between structuring of short N-terminal chemokine peptides and their ability to activate their receptor. These studies will act as a starting point for synthesizing non-peptide analogs that act as CXCR4 antagonists. human immunodeficiency virus stromal cell derived factor 1 Nuclear Overhauser enhancement NOE spectroscopy rotating frame effect Overhauser enhancement parts per thousand high pressure liquid chromatography chemical shift deviation total correlation spectroscopy Chemokines are an important class of proteins in the immune system that act to recruit leukocytes to sites of inflammation and infection by interacting with specific receptors on the cell surface of their target cells (for reviews see Refs. 1Horuk R. Trends Pharmacol. Sci. 1994; 15: 159-165Abstract Full Text PDF PubMed Scopus (157) Google Scholar and 2Miller M.D. Krangel M.S. Crit. Rev. Immunol. 1992; 12: 17-46PubMed Google Scholar). In addition, chemokine receptors are coupled to several heterotrimeric G-proteins in natural killer cells that can recognize and kill transformed or infected cells (3Maghazachi A. Al-Aoukaty A. FASEB J. 1998; 12: 913-924Crossref PubMed Scopus (59) Google Scholar). Chemokines have been implicated in auto-immune diseases and allergic disorders (reviewed in Refs. 4Bacon K.B. Schall T.J. Int. Arch. Allergy Immunol. 1996; 109: 97-109Crossref PubMed Scopus (82) Google Scholar and 5Rollins B.J. Blood. 1997; 90: 909-928Crossref PubMed Google Scholar), and several chemokine receptors are necessary cofactors that permit HIV-11 cell entry (6Cocchi F. DeVico A.L. Garzino-Demo A. Arya S.K. Gallo R.C. Lusso P. Science. 1995; 270: 1811-1815Crossref PubMed Scopus (2633) Google Scholar, 7Bleul C. Farzan M. Choe H. Parolin C. Clark-Lewis I. Sodroski J. Springer T.A. Nature. 1996; 382: 829-832Crossref PubMed Scopus (1751) Google Scholar, 8Oberlin E. Amara A. Bachelerie F. Bessia C. Virelizier J. Arenzana-Seisdedos F. Schwartz O. Heard J. Clark-Lewis I. Legler D.F. Loetscher M. Baggiolini M. Moser B. Nature. 1996; 382: 833-835Crossref PubMed Scopus (1482) Google Scholar, 9Alkhatib G. Combadiere C. Broder C.C. Feng Y. Kennedy P.E. Murphy P.M. Berger E.A. Science. 1996; 272: 1955-1958Crossref PubMed Scopus (2445) Google Scholar, 10Deng H. Liu R. Ellmeier W. Choe S. Unutmax D. Burkhart M. Marzio P.D. Marmon S. Sutton R.E. Hill C.M. David C.B. Peiper S.C. Schall T.J. Littman D.R. Landau N.R. Nature. 1996; 381: 661-666Crossref PubMed Scopus (3198) Google Scholar, 11Dragic T. Litwin V. Allaway G.P. Martin S.R. Huang Y. Nagashima K.A. Cayanan C. Maddon P.J. Kroup R.A. Moore J.P. Paxton W.A. Nature. 1996; 381: 667-673Crossref PubMed Scopus (2815) Google Scholar, 12Doranz B.J. Rucher J. Yi Y. Xmyth R.J. Samson M. Peiper S.C. Parmentier M. Collman R.G. Doms R.W. Cell. 1996; 85: 1149-1158Abstract Full Text Full Text PDF PubMed Scopus (1685) Google Scholar, 13Choe H. Farzan M. Sun Y. Sullivan N. Rollins B. Ponath P.D. Wu L. Mackay C.R. LaRosa G. Newman W. Gerard N. Gerard C. Sodroski J. Cell. 1996; 85: 1135-1148Abstract Full Text Full Text PDF PubMed Scopus (2091) Google Scholar). Human chemokines are approximately 70–80 residues in length and share substantial sequence and structural similarity (14Clark-Lewis I. Kim K. Rajarathnam K. Gong J. Dewald B. Moser B. Baggiolini M. Sykes B.D. J. Leukocyte Biol. 1995; 57: 703-711Crossref PubMed Scopus (336) Google Scholar). There are two major classes of chemokines, the CC chemokines (RANTES, MCP-1, MIP-1α, and MIP-1β) and the CXC chemokines (IL-8, NAP-2, MGSA, and SDF-1), so named because of the spacing between the cysteine residues near the N terminus of these proteins. Stromal cell-derived factor-1 (SDF-1) is a member of the CXC chemokine family and is expressed constitutively in a broad range of tissues. SDF-1 has a fundamental role in trafficking, export, and homing of bone marrow cells (15Loetscher P. Gong J. Dewald B. Baggiolini M. Clark-Lewis I. J. Biol. Chem. 1998; 273: 22279-22283Abstract Full Text Full Text PDF PubMed Scopus (128) Google Scholar). Tissue distribution of SDF-1 suggests that it may have a role in immune surveillance rather than inflammation (16Bleul C.C. Fuhlbrigge R.C. Casasnovas J.M. Aiuti A. Springer T.A. J. Exp. Med. 1996; 184: 1101-1109Crossref PubMed Scopus (1284) Google Scholar). SDF-1 has exceptionally strong sequence conservation between species (17Shirouzu M. Nakano T. Inazawa J. Tashiro K. Tada H. Shinohara T. Honjo T. Genomics. 1995; 28: 495-500Crossref PubMed Scopus (536) Google Scholar) and is the only known natural ligand for the CXC chemokine receptor 4 (CXCR4) (8Oberlin E. Amara A. Bachelerie F. Bessia C. Virelizier J. Arenzana-Seisdedos F. Schwartz O. Heard J. Clark-Lewis I. Legler D.F. Loetscher M. Baggiolini M. Moser B. Nature. 1996; 382: 833-835Crossref PubMed Scopus (1482) Google Scholar). The recently solved structure of SDF-1 (18Crump M.P. Gong J. Loetscher P. Rajarathnam K. Amara A. Arenzana-Seisdedos F. Virelizier J. Baggiolini M. Sykes B.D. Clark-Lewis I. EMBO J. 1997; 16: 6996-7007Crossref PubMed Scopus (633) Google Scholar) reveals that it has a global fold similar to other chemokines with a flexible N-terminal region followed by a loop, three antiparallel β-strands, and one C-terminal α-helix. Interest in SDF-1 has grown since CXCR4 was identified as a co-receptor for syncytia-inducing forms of HIV in CD4+ T-cells. Through interaction with CXCR4, SDF-1 inhibits replication of the syncytia-inducing form (T-tropic) of HIV-1 (7Bleul C. Farzan M. Choe H. Parolin C. Clark-Lewis I. Sodroski J. Springer T.A. Nature. 1996; 382: 829-832Crossref PubMed Scopus (1751) Google Scholar, 8Oberlin E. Amara A. Bachelerie F. Bessia C. Virelizier J. Arenzana-Seisdedos F. Schwartz O. Heard J. Clark-Lewis I. Legler D.F. Loetscher M. Baggiolini M. Moser B. Nature. 1996; 382: 833-835Crossref PubMed Scopus (1482) Google Scholar). SDF-1 also appears to be important for attracting mature lymphocytes to the bone marrow. Antagonism of this function before the harvest of the bone marrow for transplantation could be clinically beneficial (16Bleul C.C. Fuhlbrigge R.C. Casasnovas J.M. Aiuti A. Springer T.A. J. Exp. Med. 1996; 184: 1101-1109Crossref PubMed Scopus (1284) Google Scholar). In addition, SDF-1 receptors are coupled to multiple G-proteins that may be important for initiating motility of natural killer cells (3Maghazachi A. Al-Aoukaty A. FASEB J. 1998; 12: 913-924Crossref PubMed Scopus (59) Google Scholar). Thus, SDF-1 plays an important role in mobilizing the immune system and may be important for the treatment of AIDS patients and in bone marrow transplantation. A low molecular weight antagonist for SDF-1 could provide a possible therapeutic in these areas. It has recently been observed in the CXC class of chemokines that important residues for receptor binding are at the N terminus and the loop region (RFFESH) following the two disulfide bridges (14Clark-Lewis I. Kim K. Rajarathnam K. Gong J. Dewald B. Moser B. Baggiolini M. Sykes B.D. J. Leukocyte Biol. 1995; 57: 703-711Crossref PubMed Scopus (336) Google Scholar, 15Loetscher P. Gong J. Dewald B. Baggiolini M. Clark-Lewis I. J. Biol. Chem. 1998; 273: 22279-22283Abstract Full Text Full Text PDF PubMed Scopus (128) Google Scholar,18Crump M.P. Gong J. Loetscher P. Rajarathnam K. Amara A. Arenzana-Seisdedos F. Virelizier J. Baggiolini M. Sykes B.D. Clark-Lewis I. EMBO J. 1997; 16: 6996-7007Crossref PubMed Scopus (633) Google Scholar, 19Clark-Lewis I. Schumacher C. Baggiolini M. Moser B. J. Biol. Chem. 1991; 266: 23128-23134Abstract Full Text PDF PubMed Google Scholar, 20Clark-Lewis I. Dewald B. Geiser T. Moser B. Baggiolini M. Proc. Natl. Acad. Sci. 1993; 90: 3574-3577Crossref PubMed Scopus (239) Google Scholar, 21Clark-Lewis I. Dewald B. Loetscher M. Moser B. Baggiolini M. J. Biol. Chem. 1994; 269: 16075-16081Abstract Full Text PDF PubMed Google Scholar, 22Williams G. Borkakoti N. Bottomley G.A. Cowan I. Fallowfield A.G. Jones P.S. Kirtland S.J. Price G.J. Price L. J. Biol. Chem. 1996; 271: 9579-9586Abstract Full Text Full Text PDF PubMed Scopus (64) Google Scholar), with the N terminus being the most critical receptor binding site (14Clark-Lewis I. Kim K. Rajarathnam K. Gong J. Dewald B. Moser B. Baggiolini M. Sykes B.D. J. Leukocyte Biol. 1995; 57: 703-711Crossref PubMed Scopus (336) Google Scholar). It is therefore tempting to suggest that the N terminus alone could be sufficient for binding activity. These two sites appear unstructured in the solution structure of SDF-1 (18Crump M.P. Gong J. Loetscher P. Rajarathnam K. Amara A. Arenzana-Seisdedos F. Virelizier J. Baggiolini M. Sykes B.D. Clark-Lewis I. EMBO J. 1997; 16: 6996-7007Crossref PubMed Scopus (633) Google Scholar). However, short N-terminal peptides of SDF-1 were found to have SDF-1 activity (15Loetscher P. Gong J. Dewald B. Baggiolini M. Clark-Lewis I. J. Biol. Chem. 1998; 273: 22279-22283Abstract Full Text Full Text PDF PubMed Scopus (128) Google Scholar). This is a very striking and important observation. Several sequences corresponding to residues 1–8, 1–9 monomer, 1–9 dimer, and 1–17 all bind to CXCR4 and induce intracellular calcium release and chemotaxis in T lymphocytes. It was determined that the 1–17 and 1–9 dimer peptides were similar in terms of receptor binding, whereas the 1–8 and 1–9 monomer peptides had significantly lower affinity. The 1–9 dimer had the greatest activity of all the peptides tested when compared with native SDF-1 with the 1–17, 1–9 monomer, and 1–8 peptides showing decreasing activity. The basis for the enhanced activity of the 1–9 dimer remains uncertain. Finally, P2G, an SDF 1–9 analog, attained binding similar to that of the 1–9 dimer yet acted as a receptor antagonist (15Loetscher P. Gong J. Dewald B. Baggiolini M. Clark-Lewis I. J. Biol. Chem. 1998; 273: 22279-22283Abstract Full Text Full Text PDF PubMed Scopus (128) Google Scholar). Although in the solution structure of SDF-1, the N-terminal region has significant flexibility (18Crump M.P. Gong J. Loetscher P. Rajarathnam K. Amara A. Arenzana-Seisdedos F. Virelizier J. Baggiolini M. Sykes B.D. Clark-Lewis I. EMBO J. 1997; 16: 6996-7007Crossref PubMed Scopus (633) Google Scholar), it is of interest to see if there is a significantly populated conformation of these peptides that might mimic the receptor bound conformation. We present here structural data on the 1–9 dimer as well as the 1–17-mer obtained using NMR spectroscopy at 8 °C. Although the 1–9 dimer and 1–17-mer peptides are conformationally flexible, analysis of the ensemble of structures calculated from the NMR data revealed a major family that consists of a β-turn structural motif. This motif was not detected in the structure of native SDF-1. These data support the structuring of the peptides into turns that may be important for recognition in receptor binding. By understanding the structural elements necessary for receptor binding, we hope to be able to develop therapeutics that are more cost-effective mimics of the peptide itself. The N-terminal fragments of SDF-1 1–9, KPVSLSYRC, and 1–17, KPVSLSYRCPCRFFESH, were synthesized by solid phase peptide synthesis and purified by reverse phase HPLC. The 1–9 dimer was prepared by oxidizing the 1–9 monomer peptide under dilute conditions in 100 mm ammonium bicarbonate buffer at pH 8.5. The solution was magnetically stirred for 24 h and then lyophilized. Verification of complete oxidation was indicated by reverse phase HPLC, electrospray mass spectroscopy, and NMR spectroscopy. The samples were prepared by dissolving each peptide in 500 μl of 90% H2O/10% D2O or 99.9% D2O, containing 20 mmCD3COO−Na+ and 1 mmNaN3, to a concentration of approximately 5 mm. 2,2-Dimethyl-2-sila-pentane sulfonate was added to a concentration of 1 mm as an internal chemical shift reference. The pH was subsequently adjusted to 5.0 using NaOH and HCl solutions (or NaOD for D2O samples, pH adjusted to 5.0 with no correction for isotope effects). 1H NMR spectra for the 1–9 monomer, 1–9 dimer, and 1–17 monomer peptides were acquired at 600 MHz using a Varian Unity 600 spectrometer. TOCSY, NOESY, and double quantum filtered COSY spectra acquired at 8 °C were used for1H resonance assignments. The WATERGATE pulse sequence (23Piotto M. Saudek V. Skelnar V. J. Biomol. NMR. 1992; 2: 661-665Crossref PubMed Scopus (3525) Google Scholar) was used for solvent suppression for spectra in H2O. Mixing times for NOESY experiments were set at 200, 300, 400, 500, and 600 ms to determine NOE build-up rates, which were found to be linear up to 500 ms. ROE spectroscopy data were collected for the 9-mer monomer with mixing times of 60, 120, and 150 ms. NOESY experiments with 500-ms mixing times were used for the integration of the NOE cross-peaks because build up rates were approximately linear up to this mixing time, and this spectrum gave best signal to noise ratio for our measurements. The integral volumes were converted into distance restraints using a reference distance of 2.5 Å between the ortho (δ) andmeta (ε) protons of the tyrosine ring. The NOE connectivities were classified as strong, medium, weak, and very weak, corresponding to upper distance restraints of 2.8, 3.5, 4.5, and 5.5 Å, respectively. Upper limits for nonstereospecifically assigned protons were corrected appropriately with center averaging. Structure calculations were performed on the 9-mer dimer and 17-mer using the simulated annealing method employing the SHAKE algorithm implemented in X-PLOR (24Brünger A.T. X-PLOR, version 3.1. Yale University Press, New Haven, CT1993Google Scholar) at an initial simulated annealing temperature of 800 K with 8000 high temperature and 6000 cooling steps. The initial structure was an extended chain, and the target function contained only potential terms for covalent geometry, experimental distance restraints, and a van der Waals' repulsion term for nonbonded contacts. The final structures, generated using the simulated annealing method, had no NOE violations >0.25 Å nor dihedral violations >5°. The 9-mer dimer was also subjected to the time-averaged distance restraint method as described in Ref. 25Torda A.E. Scheek R.M. Van Gunsteren W.F. J. Mol. Biol. 1990; 214: 223-235Crossref PubMed Scopus (273) Google Scholar. Families of structures were extracted by superimposing the backbone of residues 5–8 within one monomeric unit and utilizing the program NMRCLUST (26Kelly A.L. Gardner S.P. Sutcliffe M.J. Protein. Eng. 1996; 9: 1063-1065Crossref PubMed Scopus (417) Google Scholar). Three N-terminal fragments of SDF-1, which were found to have SDF-1 activity (15Loetscher P. Gong J. Dewald B. Baggiolini M. Clark-Lewis I. J. Biol. Chem. 1998; 273: 22279-22283Abstract Full Text Full Text PDF PubMed Scopus (128) Google Scholar), were studied. The sequence of the 9-mer monomer (to distinguish from the 9-mer dimer) corresponds to residues 1–9 of SDF-1. The 9-mer dimer represents two 9-mer monomers connected by a disulfide bond through Cys9. The sequence of the 17-mer corresponds to residues 1–17 of SDF-1. Two-dimensional homonuclear proton spectra, TOCSY, NOESY, and double quantum filtered COSY spectra were collected at 8 °C for all three peptides in aqueous solution. In addition, ROESY spectra were collected for the 9-mer monomer because the peptide was too small for suitable build-up of NOEs. Resonance assignments were made using standard two-dimensional methods (27Wüthrich K. NMR of Proteins and Nucleic Acids. John Wiley & Sons, New York1986Crossref Google Scholar). Chemical shift data were obtained for the 9-mer monomer, 9-mer dimer, and 17-mer of SDF-1. The 9-mer monomer and 9-mer dimer have identical chemical shifts (except for Cys9, which is in its reduced (δHα 4.40) and oxidized (δHα 4.66) forms, respectively), suggesting that both strands of the dimer have the same conformation but do not interact with each other. Chemical shifts of residues 1–7 of the 17-mer are identical to those of the 9-mer monomer and 9-mer dimer (data not shown). Vicinal proton coupling constants, 3JNH-CαH, were measured from well digitized one-dimensional proton spectra and, where measurements were not possible because of the resonance overlap, from double quantum filtered COSY spectra. For all three peptides, the 9-mer monomer, 9-mer dimer, and 17-mer, the coupling constants were all greater than 6 Hz and less than 8 Hz, indicating that these residues could potentially adopt any of a number of different φ angles. These data could therefore not be included in the structure calculations. The region of the NOESY spectrum of the 9-mer dimer containing d NN(i,i+1) NOEs and NOEs from amide protons to ring protons of Tyr7 is shown in Fig. 1 A. Regions of the 9-mer dimer NOESY spectrum including medium range NOEsd αN (i,i+2),d βN (i,i+2), and NOEs from α and side chain protons to the ring protons of Tyr7 are shown in Fig. 1 B. NOEs detected for the 9-mer dimer were assumed to arise from intra-monomeric contacts because there was no evidence of inter-monomer interactions. Moreover, this distinction was confirmed when the same NOEs were observed for residues 1–9 of the 17-mer and similar ROEs were observed for the 9-mer monomer (data not shown). Summaries of sequential and medium range NOE connectivities for the 9-mer dimer and 17-mer are illustrated in Fig.2.Figure 2Summary of sequential and medium range NOE connectivities, and amide proton temperature coefficients for the SDF-1 1–9 dimer (A) and the SDF-1 17-mer (B) as observed by NMR spectroscopy. Backbone NOE connectivities are indicated by horizontal lines between residues, with the line width indicating the relative magnitude for NOEs observed in the 500-ms NOESY spectrum.View Large Image Figure ViewerDownload (PPT) Observed NOEs suggest the presence of a β-turn conformation for the 9-mer dimer and 17-mer comprising residues Leu5, Ser6, Tyr7 and Arg8. The presence of a β-turn is usually indicated byd αN (2,3),d αN (3,4),d αN (2,4),d NN(2,3), and d NN(3,4) NOESY connectivities, where the numbering refers to the residue position in the turn. NOESY spectra for the 9-mer dimer and 17-mer peptide show a d αN (2Miller M.D. Krangel M.S. Crit. Rev. Immunol. 1992; 12: 17-46PubMed Google Scholar, 4Bacon K.B. Schall T.J. Int. Arch. Allergy Immunol. 1996; 109: 97-109Crossref PubMed Scopus (82) Google Scholar) cross-peak between Ser6 and Arg8 and ad NN (2Miller M.D. Krangel M.S. Crit. Rev. Immunol. 1992; 12: 17-46PubMed Google Scholar, 3Maghazachi A. Al-Aoukaty A. FASEB J. 1998; 12: 913-924Crossref PubMed Scopus (59) Google Scholar) cross-peak between Ser6and Tyr7. In addition, a dNN (1Horuk R. Trends Pharmacol. Sci. 1994; 15: 159-165Abstract Full Text PDF PubMed Scopus (157) Google Scholar, 2Miller M.D. Krangel M.S. Crit. Rev. Immunol. 1992; 12: 17-46PubMed Google Scholar) is observed between Leu5 and Ser6 as well as Arg8 and Cys9, and ad αN connectivity is observed between Tyr7 and Cys9 (Figs. 1 and 2), suggesting that Cys9 is also involved in the local structure. Apart from the backbone proton NOEs, NOEs to the Tyr7 ring protons were observed from the amide and α-protons of Cys9, Ser6, and Arg8, β and δ protons of Leu5, and γ protons of Val3 (Fig. 1, A and C). These connectivities show that the Tyr7 side chain is stabilized in the structure most likely by hydrophobic interactions with surrounding residues. There is evidence of a second local structure in the 17-mer involving residues Cys11, Arg12, and Phe13(data not shown). This is particularly interesting because the CRF portion of the CRFFESH sequence is a partial palindrome of the tail of the 9-mer sequence (KPASLSYRC) involved in the formation of the first β-turn. A d NN(i,i+1) NOESY connectivity was observed for residues Arg12 and Phe13 as well as ad αN (i,i+2) cross-peak between Cys11 and Phe13. In addition, NOEs were observed between Phe13 aromatic ring protons and the amide protons of residues Cys11, Arg12, and Phe14 as well as the Cα protons of Cys11 and Arg12. These data suggest that the Phe13 aromatic ring is stabilized in the structure, although it is less well defined than the Tyr7 aromatic ring. To further characterize the structuring of the 9-mer dimer and the 17-mer, we measured amide temperature coefficients (Δδ/ΔT). Although for a rigid structure, exposed NHs typically have gradients in the range of −6.0 to −8.5 ppb/°C, hydrogen-bonded or protected NHs apparently have Δδ/ΔT of −2.0 ± 1.4 ppb/°C (28Andersen N.H. Neidigh J.W. Harris S.M. Lee G.M. Liu Z. Tong H. J. Am. Chem. Soc. 1997; 119: 8547-8561Crossref Scopus (228) Google Scholar). For peptide fragments, however, Δδ/ΔT values may lie anywhere between −28 to +12 ppb/°C, resulting in a correlation between the gradient and structure that lies outside the rules mentioned above. Conformational averaging in peptides appears to be the major source of deviant values of Δδ/ΔT whereby temperature-induced changes in the population of the folded state are the major contributor to the observed NH shift temperature gradient for partially structured peptides (28Andersen N.H. Neidigh J.W. Harris S.M. Lee G.M. Liu Z. Tong H. J. Am. Chem. Soc. 1997; 119: 8547-8561Crossref Scopus (228) Google Scholar). A plot of Δδ/ΔT versus the chemical shift deviation (CSD) of the amide proton provides a better correlation with partial structuring of a peptide at lower temperatures. To measure Δδ/ΔT values for the 9-mer and 17-mer, TOCSY spectra were acquired at 5, 10, 15, 20, and 25 °C. Chemical shift deviations were derived from the lowest temperature set included (5 °C). Random coil chemical shifts (29Wishart D.S. Bigam C.G. Holm A. Hodges R.S. Sykes B.D. J. Biomol. NMR. 1995; 5: 67-81Crossref PubMed Scopus (1417) Google Scholar) were corrected to 5 °C according to Refs. 30Mertuka G. Dyson H.J. Wright P.E. J. Biomol. NMR. 1995; 5: 14-24Crossref PubMed Scopus (461) Google Scholar and 28Andersen N.H. Neidigh J.W. Harris S.M. Lee G.M. Liu Z. Tong H. J. Am. Chem. Soc. 1997; 119: 8547-8561Crossref Scopus (228) Google Scholar. Fig. 3shows a plot of the CSD versus Δδ/ΔT for the 9-mer dimer and 17-mer peptides. The dashed linerepresents the cutoff of Δδ/ΔT between exposed and sequestered NHs of proteins. Gradients above the dashed lineindicate exposed NHs, whereas those below indicate sequestered NHs. All of the residues in the 9-mer dimer and 17-mer are above thedashed line, indicating that these amides are somewhat exposed. However, according to Andersen et al. (28Andersen N.H. Neidigh J.W. Harris S.M. Lee G.M. Liu Z. Tong H. J. Am. Chem. Soc. 1997; 119: 8547-8561Crossref Scopus (228) Google Scholar), peptides that are structured at lower temperatures and become unstructured upon warming have a slope of the Δδ/ΔT versus NH-CSD graph in the −8 to −20 ppt/°C range. In addition, the gradient/CSD plot must display a correlation coefficient greater than 0.7 and significant NH and αH CSD values for reasonable assessment of NH sequestration. For the 9-mer dimer, the slope of the graph for residues 5–8 is −8 ppt/°C with anR 2 of 0.7 (for the unstructured residues, the slope was −8 ppt/°C with an R 2 of 0.3). For the 17-mer, residues 5–8 had a slope of −10 ppt/°C with anR 2 of 0.8; residues 11–14 had a slope of −4 ppt/°C with an R 2 of 0.8; for the unstructured residues, the slope was −5 ppt/°C with an R 2of 0.4. The NH and αH CSD values are shown in TableI. The greatest chemical shift deviations occur for residues 5–8 in the 9-mer dimer and 17-mer. In addition, residues 11–15 in the 17-mer show larger chemical shift deviations.Table INH and Hα 1 H NMR chemical shift deviationsResidue17-mer9-mer dimerNH CSDHα CSDNH CSDHα CSDLys1−0.24−0.23Pro20.120.13Val30.19−0.050.19−0.04Ser40.100.000.09−0.01Leu50.190.030.170.03Ser6−0.11−0.08−0.12−0.07Tyr7−0.27−0.04−0.280.00Arg8−0.29−0.10−0.16−0.07Cys90.08−0.180.03−0.05Pro10−0.05Cys110.13−0.16Arg120.09−0.11Phe13−0.20−0.07Phe14−0.23−0.09Glu15−0.19−0.14Ser16−0.03−0.13His170.28CSD for NHs were derived at 5 °C with appropriate random coil chemical shift correction as described in Refs. 28Andersen N.H. Neidigh J.W. Harris S.M. Lee G.M. Liu Z. Tong H. J. Am. Chem. Soc. 1997; 119: 8547-8561Crossref Scopus (228) Google Scholar and 29Wishart D.S. Bigam C.G. Holm A. Hodges R.S. Sykes B.D. J. Biomol. NMR. 1995; 5: 67-81Crossref PubMed Scopus (1417) Google Scholar. CSDs for Hα chemical shifts were calculated from peptide data at 8 °C, and random coil chemical shift data were calculated at 25 °C from Ref.29Wishart D.S. Bigam C.G. Holm A. Hodges R.S. Sykes B.D. J. Biomol. NMR. 1995; 5: 67-81Crossref PubMed Scopus (1417) Google Scholar. Open table in a new tab CSD for NHs were derived at 5 °C with appropriate random coil chemical shift correction as described in Refs. 28Andersen N.H. Neidigh J.W. Harris S.M. Lee G.M. Liu Z. Tong H. J. Am. Chem. Soc. 1997; 119: 8547-8561Crossref Scopus (228) Google Scholar and 29Wishart D.S. Bigam C.G. Holm A. Hodges R.S. Sykes B.D. J. Biomol. NMR. 1995; 5: 67-81Crossref PubMed Scopus (1417) Google Scholar. CSDs for Hα chemical shifts were calculated from peptide data at 8 °C, and random coil chemical shift data were calculated at 25 °C from Ref.29Wishart D.S. Bigam C.G. Holm A. Hodges R.S. Sykes B.D. J. Biomol. NMR. 1995; 5: 67-81Crossref PubMed Scopus (1417) Google Scholar. 49 inter-residue and 54 intra-residue NOEs were used to construct distance restraints for each monomeric unit of the 9-mer dimer. For structure calculations of the 17-mer, 96 inter-residue and 80 intra-residue distance restraints were used. No explicit dihedral or hydrogen bonding restraints were applied. Structure calculations were performed using a simulated annealing protocol (24Brünger A.T. X-PLOR, version 3.1. Yale University Press, New Haven, CT1993Google Scholar) for both the 9-mer dimer and 17-mer peptides. For both peptides, a family of 80 structures was calculated. Structures with the lowest energy and NOE violations of no more than 0.25 Å were selected from each group. Conformationally related subfamilies of structures were then extracted using the program NMRCLUST (26Kelly A.L. Gardner S.P. Sutcliffe M.J. Protein. Eng. 1996; 9: 1063-1065Crossref PubMed Scopus (417) Google Scholar) (TableII).Table IIStructural statistics and atomic root mean square differences17-mer9-mer dimerE TOT(kcal/mol)40.11 ± 12.0042.48 ± 5.85E NOE (kcal/mol)4.85 ± 2.087.40 ± 2.68E VDW(kcal/mol)6.86 ± 2.222.43 ± 0.60Deviations from idealized geometry Bonds (Å)0.0030 ± 0.00090.0032 ± 0.0003 Angles (°)0.5203 ± 0.14820.5488 ± 0.0247 Improper (°)0.3159 ± 0.08950.3760 ± 0.0247Atomic root mean square differences (Å) for backbone atoms Residues 5–80.69 ± 0.190.65 ± 0.23 Residues 12–141.03 ± 0.22E NOE was calculated with a square well potential with a force constant of 50 kcal/mol · Å2.E VDW was calculated with a force constant of 4 kcal/mol · Å−4 where final van der Waals' radii were set to 0.75 times their value in the CHARMM forcefield. Open table in a new tab E NOE was calcu
Referência(s)