Isoleucine, Asparagine, Glutamine, Proline, Leucine, and Glycine Which Bears a Carboxamide Group

Home , Asparagine, Asparagine (data page), Cysteine (data page), Glycine (data page), Isoleucine (data page), Leucine (data page)

OXYTOCIN AND NEUROHYPOPHYSEAL PEPTIDES: SPECTRAL ASSIGNMENT AND CONFORMA TIONAL ANAL YSIS BY 220 M1Hz NUCLEAR MAGNETIC RESONANCE*,t BY LEROY F. JOHNSON, I. L. SCHWARTZ, AND RODERICH WALTERt VARIAN ASSOCIATES, ANALYTICAL INSTRUMENT DIVISION, PALO ALTO, CALIF.; DEPARTMENT OF PHYSIOLOGY, MOUNT SINAI MEDICAL AND GRADUATE SCHOOLS OF THE CITY UNIVERSITY OF NEW YORK; AND MEDICAL RESEARCH CENTER, BROOKHAVEN NATIONAL LABORATORY, UPTON, NEW YORK Communicated by Maurice Goldhaber, May 26, 1969 Abstract. 1\Jagnetic resonance peaks have been assigned to individual protons of the constituent amino acids in the neurohypophyseal hormone, oxytocin, and in related peptides. The assignments were made possible by operation at 220 M\Hz with the use of variable temperature studies, proton homonuclear spin- decoupling, and comparison of spectra of oxytocin analogs. Some of the observed chemical shifts, and NH-CHa coupling constants were studied in relation to the conformation of the hormone.

Earlier we investigated the conformation of neurohypophyseal hormones by means of partition chromatography1 and circular dichroism.2 In continuation of these studies we turned to 220 1\IHz proton nuclear magnetic resonance (NM'\1R)- a technique which offers great promise in revealing information about helix-- coil transitions and mobility of side chains as well as intra- and intermolecular interactions in peptides and proteins. In this paper' we wish to report on the analysis of NM'R spectra, in deuterated dimethylsulfoxide, of oxytocin a cyclic peptide hormone composed of eight different amino acids, viz., cystine, tyrosine, isoleucine, asparagine, glutamine, proline, leucine, and glycine which bears a carboxamide group. Materials and Methods.-Spectra were recorded using a Varian Associates HR- 220 spectrometer. The sample concentrations were 4 to 6 weight per cent and the internal reference was tetramethylsilane. The temperature in the sample zone was controlled and known within + 20. Proton spin-decoupling was done by the field sweep method using a side band generated by an HP 4204-A oscillator. Results and Discussion.-One prerequisite for a successful conformational in- vestigation of a peptide or protein by NM\IR is the identification of the resonance pattern of individual protons of the constituent amino acids. To assign resonance peaks to specific residues in the neurohypophyseal peptides we used: (a) examination of NMIR spectra at different temperature levels; (b) comparison of spectra of intermediates of oxytocin and of analogs with selected structural modi- fications; and (c) homonuclear proton spin decoupling. In the course of the assignment, the variable temperature studies were particularly helpful in re- solving overlapping resonance signals, while the spectra of intermediates and analogs give an indication of the chemical shifts and the resonance pattern to be expected for the individual amino acid residues in the hormone molecule. The decoupling experiments were vital for deciphering which NH, C', and C" protons belong to a particular amino acid residue. When maximal fingerprinting 1269 Downloaded by guest on October 1, 2021 1270 BIOCHEMISTRY: JOHNSON ET AL. PROC. N. A. S.

was required, i.e., for determining which amide proton -peaks were coupled with which Ca-proton peaks, the samples were dissolved in deuterated dimethylsulfoxide in order to prevent proton exchange of the NH moiety; when we were interested in determining which C' and Ca protons were coupled, it was occasion- ally advantageous to record the spectra in a solvent system which promotes exchange (deuterated dimethylsulfoxide containing 5 per cent deuterium oxide). Considering the structure of oxytocin4 or deamino-oxytocin,5 it is evident that two amino acid residues, glycine and proline, will exhibit resonance patterns qualitatively different from those of the other residues in the molecule: (a) the amide proton of the glycine residue ought to be split by the two Ca-protons into a triplet, while all the optically active amino acid residues should ideally exhibit NH doublets. IThe one-proton triplet of the glycine residue is readily recognized in the spectrum of oxytocin at 7.97 ppm (Fig. 1). Upon irradiation 988 Hz upfield the triplet collapses indicating the resonances ofthe glycine methylene group to be located at 3.47 ppm;6 (b) the proline residue lacks the amide proton, and hence the C' proton ought only be coupled with protons appearing further upfield. This criterion is met in the spectrum of oxytocin by the doublet -actually two unresolved, closely spaced doublets-centered at 4.40 ppm which collapses to a broad singlet when irradiated 515 Hz upfield. By virtue of their special proton patterns, the glycine and proline residue were readily detected. More difficulties may be expected in the assignment of the remaining amino acid residues, because even after decoupling experiments reveal which NH, C' and Ca protons are linked, a criterion must still be found for associating these proton resonances with a particular amino acid residue in the hormonal molecule. Since the NH doublets per se never reveal the particular

9 8 7 6 5 ppm

FiG. l.-Oxytocin. Bridges indicate resonance patterns connected by decoupling. Downloaded by guest on October 1, 2021 VOL. 64, 1969 BIOCHEMISTRY: JOHNSON ET AL. 1271

optically active amino acid residue with which they are associated and since the complex Ca-proton patterns will yield such information only in rare in- stances, it is the Ca-proton resonance pattern which carries the major burden for providing the clue to a more specific assignment. For example, the observed unsymmetrical two-proton triplet at 1.47 ppm in the spectrum of deamino- oxytocin (see Fig. 3) can only be attributed to leucine, which possesses a C- methylene moiety flanked by two single protons (decoupling experiment of the leucine residue is shown in expansion of Fig. 2). The correctness of the assignment was attested with deamino-8-alanine-oxytocin, an analog in which the leucine residue is replaced by an alanine residue.8 The spectrum of this analog lacks the signals assigned to the C"-methylene moiety of leucine and, although the Ca- proton resonances of the alanine residue exhibit approximately the same chemical shift as those of the replaced leucine residue the splitting pattern differs. To complete the determination of the NMR pattern of the C-terminal tri- peptide sequence of oxytocin, we examined the spectrum of Z-Pro-Leu-GlyNH2. The two amide protons of the glycine amide moiety are centered at 7.00 ppm, therefore, they are totally hidden in the spectrum of oxytocin (Fig. 1) and par- tially buried in the spectrum of deamino-oxytocin (Fig. 2) by the low field doublet of the tyrosine residue. From this spectrum we also obtained the chemical shifts and splitting patterns of the isopropyl group of the leucine and the fl-, ay-, and 5-methylene groups of the proline residue, respectively; these data were directly applicable to the resonance assignment of these groups in the oxytocin and deamino-oxytocin spectra. The next target was the assignment of the amino acid residues which com-

... , . .

DEAMINO-OXYTOCIN tWo:) 220 MHz A a.- o._ IN DMSO-da AT 30 C IM

AT &C 1125HHi.2

Ir U) _J _J 5~~~~~05z - 4-

OH 836 H

I~~~~~~~ 730 Hz ~ ~ IMlXI X XU

95Hz r_ Ii[Ii II Ii 865 Hz

FIG.~~~~~~~~9Demn-xtcn>- (~Exanio onteletilsresdcuinofheL11ppmppm poonf theprolin reides11111111exasoH~~on th86rihH lutae tedculn f h mdC n C~~~~~~~~~3proonof th leucin reides

I ~~~~~~AA!XXTYR 4HzI P z ------I AI - 10 9 8 7 6 5 4 3 2 0 ppm

FIG. 2.-Deamino-oxytocin. Expansion on the left illustrates decoupling of the C" proton of the proline residues; expansion on the right illustrates the decoupling of the amide, C' and Co protons of the leucine residues. Downloaded by guest on October 1, 2021 1272 BIOCHEMISTRY: JOHNSON ET AL. PROC. N. A. S.

IA t) v636A

, I -_ -, -, -, t A. I i '- -, i 9 8 7 6 5 4 3 2 0 PPM

FIG. 3.-Protected C-terminal pentapeptide of oxytocin. Expansion shows the decoupling of the NH proton of the cysteine, leucine and glycine residues.

prise the ring component of the hormonal molecule. Although the spectrum of the C-terminal pentapeptide of oxytocin (Fig. 3) exhibits the characteristic proton resonance associated with the proline, leucine, and glycine residues, the problem of assigning the corresponding resonances for the cysteine and asparagine residues persists. Therefore, we studied the spectrum of a pentapeptide derivative in which the asparagine residue had been replaced by a valine residue (Fig. 4) -for our purpose a particularly advantageous replacement. In the course of this study, we had noted that the chemical shift of the Ca-proton resonances of the proline residue remained remarkably constant when the peptide chain was lengthened, when the hydrogen ion concentration or the temperature of the sample was changed, and, hence, that the Ca-proton resonances of this residue may serve in first approximation as a reference, at least in the case of neurohypophyseal hormones and their peptide derivatives. The Ca-proton resonances of amino acid residues which bear an electron withdrawing group on the fl-carbon such as SH, CONH2, or p-hydroxyphenyl appear downfield, while the Ca-proton resonances of amino acid residues bearing hydrogen, methylene, or methyl groups on the ,8-carbon appear upfield from those of the C' proton of proline. However, this rule has to be applied with caution because any significant change in the magnetic anisotropies (e.g., as a result of change in conformation or solva- tion) of the three groups attached to a CH' group would of course change its chemical shift. Returning to Figure 4, it can be seen that the leucyl and valyl CH' resonances appear further upfield when compared with those of prolyl, while the signal of the cysteine CH' residue is downfield at 4.68 ppm. Therefore, the one-proton signal at 4.65 ppm in Figure 3 is associated with the CH' of the cysteine residue and the one-proton signal at 4.36 ppm with that of the asparagine residue. The Downloaded by guest on October 1, 2021 VOL. 64, 1969 BIOCHEMISTRY: JOHNSON ET AL. 1273

Ppem FIG. 4.-Protected C-terminal pentapeptide derivative of oxytocin. Expansion shows the decoupling of the CO-methylene protons of the cysteine residue.

elongation of the pentapeptide of oxytocin to the hexapeptide resulted in the appearance of a new CHa-proton resonance at 3.97 ppm associated with the glutamine residue just introduced; according to prediction, this resonance appears upfield from that corresponding to the proline residue. Next, we turned to the spectrum of oxytocin (Fig. 1). As anticipated, the Ca- proton resonance of the tyrosine residue appears downfield from that of the proline residue and, more precisely, is located between those of the cysteine and the asparagine- residues. This designation was confirmed by the spectrum (not shown) of deamino-2-alanine-oxytocin.9 The new one-proton resonance at 3.90 ppm in the oxytocin spectrum just upfield of that corresponding to the glutaminyl CH' is due to the C' proton of the isoleucine residue. In this case, decoupling experiments of the CH' were the basis for the assignment because only the CH' of the isoleucine residue in oxytocin is expected to go to a doublet upon decoupling. The three carboxamide groups of oxytocin gave rise to interesting two-line patterns. While the two amide proton resonances of the glycine amide moiety were separated at room temperature by 10 to 15 cps depending on the peptide (Figs. 1, 2, and 4), the primary carboxamide protons of the asparagine residue were separated by 95 to 100 cps (Figs. 1, 2, and 3), and those of the glutamine residue by 105 to 110 cps (Figs. 1 and 2). As the temperature is raised from 400 to 550, the two peaks centered at 7.00 ppm, which are due to -CONH2 of the glycine amide moiety, coalesce (this reversible temperature-dependent effect was studied with deamino-2-alanine-oxytocin to simplify the spectrum by avoiding interference by the p-hydroxyphenyl protons of the tyrosine residue present in oxytocin), while the peaks at 6.65 and 7.15 ppm broaden. An increase of the temperature to 60° and 650 further sharpens the glycine amide peak while broadening the amide absorptions of the asparagine and glutamine residues even Downloaded by guest on October 1, 2021 1274 BIOCHEMISTRY: JOHNSON ET AL. PROC. N. A. S.

more and, at 820, virtually one peak is obtained. The amide group is planar and the partial double-bond character of the C-N bond gives rise to a high energy barrier to rotation around this bond resulting in separated resonances. This restricted rotation is overcome as the system is supplied with energy. It has been shown that the N-proton cis to the carbonyl is more shielded than the trans,'0 i.e., the signal with the smaller chemical shift from tetramethylsilane belonging to the two-line pattern of a primary carboxamide is attributable to the cis proton. The chemical shift differences within the amide protons of the glycine amide, asparagine, and glutamine residues appear to result from conformational features of the oxytocin molecule as well as inherent characteristics of the amino acid amides. "I A comparison of the spectra of oxytocin and deamino-oxytocin reveals no striking differences. This and the reversibility of temperature effects are both reminiscent of results obtain in our circular dichroism studies.2 On comparing the two spectra, the most readily detectable discrepancy is the phenomenon associated with the temperature-dependent proton resonance at 9.13 ppm which is due to the tyrosine hydroxyl proton. While deamino-oxytocin exhibits a sharp signal at room temperature oxytocin shows a broad band which is sometimes not detectable at all-indicating that the rate of exchange of the phenolic proton must be considerably greater in the natural hormone. Recently Stern, Gibbons, and Craig,'2 while studying the conformation of the antibiotic gramicidin S-A by NMR, have shown that the magnitude of the coupling constants between the N-Ca protons (JNC) yields information about the backbone stereochemistry, i.e., rotation around the N-C bonds. If their argu- ment, viz., that a small value of JNC corresponds to a large NC dihedral angle, whereas a large value of JNC corresponds to a small dihedral angle, is applied to the spectra of oxytocin and deamino-oxytocin, in which the JNC's are in the order of 6 to 7.5 cps (except for Ile -- 4 Hz), it suggests that the average dihedral angle of the N-C bonds of the optically active amino acid residues are largely trans in oxytocin and deaminooxytocin. One of the most intriguing regions in the oxytocin and deamino-oxytocin spectra is that between 2.5 and 3.5 ppm. In this region one finds, along with two protons associated with the proline residue, the signals of the C~-methylene moieties of the cysteine, tyrosine, and asparagine residues. The Ca protons of the cysteine and tyrosine residues exhibit different chemical shifts. Since the Ca-C" coupling constants (Jaw) should yield information about the average rotation of the CalCa bonds and thus about the stereochemistry of the amino acid side chains, we are presently studying this problem in more detail. Recently, McDonald and Phillips'3 observed, while comparing the computed spectrum of oxytocin with that recorded in neutral aqueous solution, that the actual spectrum departs significantly from expectation at the cystine A-hydrogen position. We found that the Cal protons of cysteine (e.g., Fig. 4) remain constant as the peptide chain is elongated; however, as the number of the possible conformers is drastic- ally reduced by ring closure of the key intermediate to form the hormonal peptide, the position of the Cal protons of the cysteine residue is shifted. Since this cysteine residue in position 6 of the hormonal molecule does not occupy an N- Downloaded by guest on October 1, 2021 VOL. 64, 1969 BIOCHEMISTRY: JOHNSON ET AL. 1275

terminal position, the observed shift must be associated with its restricted envi- ronment in the cyclic molecule. In conclusion, assignments have been made for proton magnetic resonances in the spectra of oxytocin and neurohypophyseal peptide analogs as a further step in our program of studies directed at the elucidation of the conformation of the posterior pituitary hormones. We are indebted to Dr. R. T. Havran, Miss Patricia Evans, and Mr. Fred Davidson for preparing oxytocin and deamino-oxytocin, and Lewis Cary for his excellent technical assistance recording the spectra. * This work was supported, in part, by U.S. Public Health Service grant AM 10080 of the National Institute of Arthritis and Metabolic Diseases and by the U.S. Atomic Energy Com- mission. t Neurohypophyseal hormones are denoted in accordance with the IUPAC-IUB Tentative Rules (Biochemistry, 6, 362 (1967)) and standard abbreviations are used for amino acid residues (Biochemistry, 6, 2485 (1966)). The amino acids (except glycine) are of the Lconfiguration. To whom reprint requests should be addressed. Gordon, W., R. T. Havran, I. L. Schwartz, and R. Walter, these PROCEEDINGS, 60, 1353 (1968). 'Walter, R., F. Quadrifoglio, and D. W. Urry, 164th National Meeting of the American Chemical Society, C-174 (1967); D. W. Urry, F. Quadrifoglio, R. Walter, and I. L. Schwartz, these PROCEEDINGS, 60, 967 (1968); Walter, R., W. Gordon, I. L. Schwartz, F. Quadrifoglio, and D. W. Urry, in Peptides, Proceedings of the 9th European Peptide Symposium, ed. E. Bricas (North-Holland Publ. Co., 1968), p. 50. ' Walter, R., and L. F. Johnson, 13th Annual Biophysical Society Meeting, Biophys. J., 9, A-159 (1969), where part of this work was presented. 4du Vigneaud, V., C. Ressler, J. M. Swan, C. W. Roberts, P. G. Katsoyannis, and S. Gordon, J. Am. Chem. Soc., 75, 4879 (1953). 5Ferrier, B. M., D. Jarvis, and V. du Vigneaud, J. Biol. Chem., 240, 4264 (1965). 6 Upon close examination, the methylene signals of the glycine amide were found as two separate groups with chemical shift differences of 0.10 ppm. Similar proton inequivalences were reported for peptidyl glycine zwitterions.7 Mandel, M., J. Biol. Chem., 240, 1586 (1965); van Gorkom, M., Tetr. Letters, 5433 (1966); Beecham, A. F., and N. S. Ham, Tetrahedron, 24, 2773 (1968). 8 Walter, R., and V. du Vigneaud, Biochemistry, 5, 3720 (1966). Walter, R., and I. L. Schwartz, Life Sciences, 7, 545 (1968). '0Anet, F. A. L., and A. J. R. Bourn, J. Am. Chem. Soc., 87, 5350 (1965). 11Walter, R., and L. F. Johnson, to be published. 12Stern, A., W. A. Gibbons, and L. C. Craig, these PROCEEDINGS, 61, 734 (1968). 13 McDonald, C. C., and W. D. Phillips, J. Am. Chem. Soc., 91, 1513 (1969). Downloaded by guest on October 1, 2021