Conformational Analysis of Peptide T and of Its C-Pentapeptide Fragment

ANDREA MO'ITA, Istituto Chimica MIB del CNR, via Toiano 6, Arco Felice, Napoli; DELIA PICONE, PIER0 A. TEMUSSI,

Dipartimnto di Chimica, Uniwrsitd di Napoli, via Mezzocannone 4, 80134 Napoli; MAURO MARASTON1 and ROBERTO TOMATIS,

Dipartimento di Scienze Farmaceutiche, Universitri di Ferrara, uia Scandianu, Ferrara, Italy

Synopsis

The synthetic peptide of sequence H-Ala-Ser-Thr-Thr-Thr-Am-Tyr-Thr-OH, termed peptide T, a competitor of the Human Immunodeficiency Virus in the binding to human T cells, and its C - t d a l pentapeptide fragment, were studied by 'H-nmr in DMSO solution to determine conformational preferences. The observation of nuclear Overhauser enhancements (NO&) for both peptides, and unusual finding for small linear peptides, allowed complete sequence-specific r-ance assignments. Long-range NO& ring-current s h i h and the very small temperature coefficient of the Thr* NH chemical shifi suggest, for the zwitterionic form of peptide T, the preaence in solution of a involving Thr', Am6, Tyr' and Thr'. This conformational feature is consistent with previous structure-activity relationship studies indicating the invariance of the same reaidues in several potent pentapeptide analogues. The studied pentapeptide fragment, although lees structured, shows some tendency to fold even in a polar solvent such as DMSO.

Preliminary chemotaxis data on some pentapeptide analogues are consistent with our struc- tural model.

INTRODUCTION

Peptide T, a synthetic peptide of sequence Ala'-Ser2-Thr3-Thr4-Thr5-Asn6- Tyr7-Thr8, which corresponds to a region of the external envelope glycopro- b i n (gp120) of the human immunodeficiency virus (HIV), was found to inhibit the binding of the '%abeled gp 120 to brain membranes.' Since peptide T is very active in antagonizing the in vitro infection of human T cells by HIV, it has been proposed that this sequence of the gp 120 protein is responsible for the HIV virus attachment to its receptor(s) on T cells.'

We have recently reported a conformational study of peptide T in DMSO solution by means of 'H-nmr Although small linear peptides are seldom chara- by well-de6ned s t r u h in such a polar medium, we found clear indications of the presence of a fairly rigid conformation, consisting of a &turn involving the four C-terminal residues. This structure is stabilized by the salt bridge between the charged ends of the peptide, as suggested by the absence of a similar structure in the cationic form3 of peptide T. Interestingly, the proposed structure for the zwitterionic peptide ascribes a fundamental role to residues 4-8, whose crucial function in the biological activity of peptide T was previously assessed from classical structure-activity

Biopolymers, Vol. 28,479-486 (1989) Q 1989 John Wdey & Sane, Inc. CCC oaoS-3525/89/010479-~.~

480 MO'ITA ET AL.

relationship (SAR) data4 It has been shown that the pentapeptide of se- quence Thr4-Thr5-Asn6-Tyr7-Thr8 (peptide T [4-81, this numbering of residues will be used throughout the paper to allow a direct comparison with the parent octapeptide) is active in the in oitro human monocyte chemotaxis assay, as the whole peptide T, both with an EC, in the pM range.4 The above amino acidic sequence is present, with minor replacements, in several viruses not obviously related to HIV, as Epstein-Barr (Thr-Glu-Am-"yr-Thr) and polio (Thr-Ile-Am-Tyr-Thr), as well as in the vamactive intestinal polypeptide (VIP; Thr-A~p-Asn-Tyr-Thr)~ and in the isolated nine AIDS- related virus4 (ARV). Synthetic peptides corresponding to all these sequences are all active in the human monocyte he mot ax is.^ By comparing the ARV sequences and that of peptide T [l-81 (and its [4-81 C-terminal fragment) the fundamental structural requirements of these peptides became evident. In fact, the major changes are the replacement of Ser and Thr, two closely related amino acids, whereas Tyr is an invariant f e a t ~ r e . ~ . ~ The carboxyl- terminal amino acid also seems to play a significant role, since the [D-Ala', D-Thr8]-peptide T-NH, has a lower chemotactic a~t iv i ty .~

The present study aims at the understanding of the conformation of the peptide T [4-8] in DMSO solution. The structural data, compared with those of the octapeptide, are expected to be very useful as a guideline in designing analogues endowed with more potent and/or long-lasting biological activity.

RESULTS AND DISCUSSION Figure 1 compares the low-field region of the one-dimensional (1D) spectra

of both the octa- (spectrum A) and the pentapeptide (spectrum B) in DMSO-d, at 500 MHz. Some backbone NH resonances (Ser2 and Thr*) are broad in peptide T, while the ones stemming from all pentapeptide residues appear as well-resolved doublets. Overall, the NH resonances are well separated in both sp t ra , even those originating from the four Thrs in the octa and from the three Thrs in the penta.

Resonance identification was achieved by means of double quantum filtered (DQF) correlated spectroscopy for both peptides, since chemical shifts of homologous residues show no obvious correspondence in the two peptides. The full assignment for the octapeptide was achieved on the basis of nuclear Overhauser enhancement spectroscopy (NOESY)8vg experiments. Fig- ure 2 reports a NOESY spectrum (500 ms mixing time) of the octapeptide at 500 MHz. Besides the NHraCHi-, sequential cross peaks (all labeled in Fig. 2), we were able to observe diagnostic NH-NH effects near the diagonal. These data, combined with the very small temperature coefficient of Thr8 NH and the high-field shifts undergone by the resonances of Thr8 (Table I) imply the presence of a /Mum3 involving Thr5-Asn6-Tyr7-Thr8. Figure 3 shows a schematic model of this conformation, emphasizing the role of the salt bridge. The moiety of the peptide comprising the terminal four residues was recog- nized to be of primary importance in previous SAR s tudie~ .~

Sequence-specific resonance assignment in the pentapeptide was achieved by rotating f r m e NOESY (ROESY)'**'' experiments. Figure 4 shows the low-field region of the contour plot of a ROESY spectrum (200 ms d g time), which contains cross peaks from amidic to its own and previous aCH

CONFORMATION OF PEPTIDE T 481

9 . 0 8 , s 8 . 0 7 . 5 7 . 0 6 . 5 w n

Fig. 1. Comparison of the low-field ‘H-nmr spectral regions of peptide T (A) and its C-termi- nal pentapeptide fragment (B) in DMSO-dB at 500 MHz. Standard one-letter code for amino acids is used to indicate the assignments for NH and CH aromatic protons, as obtained from two-dimensional (2D) experiments (see text).

protons, but completely lacks those effects diagnostic of ordered structures, such as NH-NH cross peaks.12 Table I compares the assignments of all protons, together with the temperature coefficients of the amidic protons, both for peptide T (indicated with T) and its C-terminal [4-81 fragment (indicated with P). As it can be seen, no anomalous values of chemical shifts or very low-temperature coefficients (i.e., < 2.4 ppb/K) are observed for the penta, including Thr’. All these nmr data point to the absence of a single ordered structure in the pentapeptide fragment. This would be consistent with our studies on the octapeptide. In fact, in the case of the peptide T we have suggested that the electrostatic attraction between the charged ends of the peptide plays a crucial role in favoring a B-turn involving the four C-terminal residues.

The same residues are present in the C-terminal pentapeptide fragment, but in this case the charged ends cannot play any “catalytic” role, probably owing to both the length of the peptide and the increased flexibility. It must be considered, however, that the ROESY experiment has some disadvantages if compared with the NOE experiment, one of them being the smaller maximum theoretical NOES that may be obtained.’O

482 MOTTA ET AL.

7:O PPM

Fig. 2. Low-field region of a 500 MHz NOESY spectrum of peptide T in DMSO showing cross peaks from amidic and aromatic protons t o NH, a-CH, and 8-CH. Mising time = 500 ms, T=298K.

In fact a NOESY experiment at 400 MHz, with a 500 ms mixing time, shows, besides the effects detected in the ROESY spectrum of Fig. 4, two small but diagnostically valuable NH-NH cross peaks between Thr5-Asn6 and Asn6-Tyr7 (Fig. 5), which are consistent with a sizeable contribution from folded structures.

At high magnetic fields NOE effects are seldom detectable in medium-sized linear peptides,13 i.e., molecules of weights between 500 and 2000 daltons approximately, owing to the unfavorable value of we, to the flexibility of the molecules, and/or to the small fractional population of folded conformem. The observation of all the sequential effects in the pentapeptide NOESY experiment may indicate the presence, in DMSO solution, of several quasi- rigid folded conformers, in equilibrium with extended forms. The large num- ber of OH groups in this peptide may play a role in favoring some preferential conformation since their facile solvation by DMSO molecules may partially shield the NH groups from hydrogen-bond solvation by bulk DMSO. It can be concluded that, although the pentapeptide does not show a clear preference for a /I-turn conformation involving the last four residues, it tends to fold consistently with the &turn found for the parent peptide.

CONFORMATION OF PEPTIDE T 483

TABLE I Chemical Shib [a, Referred to Internal Tetramethylsilane (TMS)] and Amidic Temperature

Coefficients of Peptide T 0 and Its C-Terminal Pentapeptide Fragment (P) in DMSO-d6

3.81 4.42 4.31 4.30

1.32 3.66 4.09 4.07

8.69 7.90 7.85

OH 5.16 CH3 1.05 CH, 1.05

- 5.4 - 4.0 - 2.8

8.06 7.69

3.87 4.25

3.77 4.03

1.17 CH, 1.02

- 1.5 - 2.3

8.50 8.02

4.34 4.51

3.99 2.36 2.56 2.35 2.53 2.74 2.95 2.67 2.95 3.96

1.04 NH, 6.97

7.40 6.95 7.39

Ar 6.64 7.03 6.62 7.06

CH3 0.94

- 4.0 - 3.3

8.10 4.55 - 4.0

8.00 4.37 - 3.3

7.94 4.49 - 3.0

- 0.65 7.54 3.96

7.99 4.16 4.13 1.06 - 6.0

(Thr’, 6

2 Ser

Fig. 3. Schematic model of the conformation proposed for peptide T in the zwitterionic form. The skeleton of the /hum is shown with stick and balls, whereas all side chains and the first four residues are simply indicated by the three-letter code for amino acids.

484 MO'M'A ET AL.

9 V." I ' ' ' ' I " r T' ' ' ' I

9.5 8.5 7.5 6.5

In

3n

5.0

7.0

m.0

PPM

Fig. 4. ROESY spectrum (400 MHz) of 3 mM peptide T(4-8) fragment in DMSO-d, at 298 K. The shown spectral region contains cross peaks from amide and aromatic protons to (r-CH and 8-protons.

T5

7.8

8,2

8.6

-1. 8.6 8.2 7.8

P P M

Fig. 5. Low-field region of a 400-MHz NOESY spectrum of 3 mM peptide T(4-8) fragment in DMSO-d, showing cross peaks among amidic protons. Mixing time = 500 ms, T = 298 K.

CONFORMATION OF PEPTIDE T 485

These conformational tendencies were tested by designing several ana- logues, both linear and cyclic. The biological testing is still in progress but it is possible to anticipate some relevant results (R. Tomatis, personal communica- tion).

According to the proposed &turn structure, systematic substitution of the Thr residues shows that Thm 5 and 8, which are inside the putative turn, are essential, whereas Thr4 can be substituted by Abu without affecting the biological activity. In fact, [Abu4]-Tf4-8] has a chemotactic behavior almoet identical to that of the parent pentapeptide, while [Abu6]-Tf4-8] and [Abu'l- 1'14-83 have no activity.

EXPERIMENTAL

Peptide T and its C-terminal[4-8] fragment were synthesized by classical solution methods and purified to homogeneity by silica gel column chromatog- raphy, eluted with n-butanol-acetic acid-H,O (6 : 2 : 2) and/or reverse-phase high performance liquid chromatography (HPLC) on c18, eluted with a hear gradient of acetonitrile in 0.1% trifluoroacetic acid.

Thin-layer chromatography, amino acid and elemental analysis and analyti- cal HPLC indicated that the peptides employed in this study are > 97% homogeneous. All nmr experiments were run on 2 mM (peptide T) or 3 mM (C-terminal

[4-81 fragment) solutions in 99.8% deuterated DMSO-d, (C. Erba, Milan, Italy).

H spectra were recorded on a Bruker WM-500 spectrometer interfaced to an Aspect 2000 computer or on a Bruker AM-400 spectrometer interfaced to an Aspect 3000 computer, and referenced to internal TMS. 1D spectra were typically acquired in quadrature detection. 2D experiments were performed by using the standard Bruker microprograms, and quadrature detection was used in both dimensions, with the carrier in the center of the spectrum. For phase-aemitive (DQF)-COSY,%' ROESY,'O* l1 and NOESY8*' experiments, the time-proportional phase incrementation scheme was used.'' Before Fourier transformation, the time domain data matrix was multiplied by a Lorentz-te Gauss function in both dimemions.

1

References 1. Pert,C.B.,Hill,J.M.,Ruff,M.R,Berman,RM.,Robey,W.G.,Arthur,LO.,Ruscetti,

2. Picone, D., TmuSei, P. A, Mamatmi, M., Tomatis, R & Motta, A. (1968) FEBS Lett., 231,

3. Motta, A, Picone, D., MarastWi, M., Tomatis, R. & TemusPi, P. A. (1966) J . Am. Chem.

F. W. & F ~ M , W. L. (1986) A-oc. N ~ t l . A d . Sci. 23,9264-9258.

169- 163.

Soc., submitted. 4. Ruff, M. R., Wdd, S. M & Pert, C. B. (1986) Peptides 6, 107-111. 5. Ruff, M. R., Martin, B. M., Ginns, E. J., Farrar, W. L & Pert, C. B. (1987) FEBS Lett. 211,

6. Shaka, A. J. & Freeman, R (1983) J . Magn. Reson. 61,169-173. 7. Rance, M., Sorenaen, 0. W., Bodenhauaen, G., Wagner, G., Emst, R. €2. & Wuethrich, K.

8. Macur4 S. & Emet, €2. R. (1979) Md. Phys. 41.95-101.

17-22.

(1984) Biochem. Biophys. Res. Comm~ur. 117,479-485.

486 MOTTA ET AL.

9. Bodenhausen, G., Kogler, H. & Emst, R. R. (1984) J. Mogn. Reson. 5,370-388. 10. Bother-By, A. A., Stephens, R. L., Lee, J., Warren, C. D. & Jeanloz, R. V. (1984) J. Am.

11. Bax, A. &Davis, D. G. (1985) J. M e . Reson. 63,207-211. 12. Wuethrich, K. (1986) NMR of Proteins and Nucleic Acids, J. Wiley & Sons, New York. 13. Motta, A., Picone, D., Tancrd, T. & Temussi, P. A. (1987) J. Mogn. Reson. 75, 364-370. 14. Marion, D. & Wuethrich, K. (1983) Biochem. Bwphys. Res. Commun. 113,967-974.

Chem. SOC. 106,811-813.

Received July 5,1988 Accepted August 4,1988