Ion Binding of Cyclolinopeptide A: An NMR and CD Conformational Study TEODORICO TANCREDI,',* ETTORE BENEDETT1,2 MARIA GRIMALDI,' CARL0 PEDONE,' FILOMENA ROSSI,' MICHELE SAVIANO,' PIER0 ANDREA TEMUSSI,' and CIANCARLO ZANOTTI

'Istituto Chirnica MIB del CNR, Via Toiano 6, 80072 Arc0 Felice, 'Universita di Napoli Federico II, Dipartirnento di Chirnica, Napoli, and 3Centrodi Studi Chirnica del Farrnaco del CNR, Rorna, Italy

SYNOPSIS

CD and nmr techniques have been used to study, in acetonitrile solution, the ion-complexing capability of cyclolinopeptide A (CLA) , a cyclic nonapeptide of sequence cyclo- ( Pro-Pro-Phe-Phe-Leu-Ile-Ile-Leu-Val) endowed with remarkable cytoprotective ability in vitro, and the conformation of the Ba2'/ CLA complex. At room temperature, CLA in acetonitrile shows a proton nmr spectrum characteristic of the coexistence of many different conformers in intermediate exchange. The backbone contains a cis Pro-Pro bond, with all other peptide bonds in the trans conformation. CLA binds Ba2+more tightly than the other cations studied, namely K', N a + , Mg2+, and Ca2+;CD data are indicative of the presence of both 1 : 2 (sandwich) and 1 : 1 (equimolar) type complexes, depending on the Ba2+ion concentration, whereas nmr data are consistent with an equimolar form. The relevant conformational features of the equimolar Ba2+/CLAcomplex are that the backbone contains all trans peptide bonds, a type I 6 -P 3 @-turnand a 3 -P 1 y-turn (or a distorted 3 + 9 @-turn).The global shape of the complexed peptide can be described as a bowl, with the concave (polar) side hosting Ba2+and the convex side predominantly apolar.

INTRODUCTION Cyclolinopeptide A CLA peptide of sequence

, a natural cyclic nona-

cyclo- ( Pro-Pro-Phe-Phe-Leu-Ile-Ile-Leu-Val )

isolated from linseed by Kaufmann and Tobshirbel' in 1959, and later synthesized by Prox and Weygand' by classical solution methods, has recently been shown to be endowed with remarkable cytoprotective ability in vit1-0,~ as measured by the inhibition of the transport system responsible for the uptake of cholate into hepatocytes. This behavBiopolymers, Vol. 31, 761-767 (1991) 0 1991 John Wiley & Sons, Inc.

CCC 0006-3525/91/060761-07$04.00

' To whom correspondence should be addressed.

ior is shared by other natural cyclic peptides like antamanide, somatostatin, and synthetic analogues, all characterized by sequential homologies with CLA. Many attempts have been made to solve the structure of CLA both in solution and in the solid state, by theoretical5 and spectroscopic m e t h ~ d s . ~ - l ~ Earlier studies did not provide definite conformational conclusions, due to the limited spectroscopic data and t o the experimental conditions chosen for the solution studies; in fact, the flexibility of CLA and the use of polar solvents always prevented the observation of a single conformer. Only recently has it been possible to solve the solid state structure of CLA,1°-12and by the combined use of apolar solvent ( CHC13) and low temperature (214 K ) , the solution one, by nmr methods'','2: in these conditions, they are practically identical.

76 1

762

TANCREDI ET AL.

In order to shed light on the mechanism of action of CLA, it may be important to establish its conformational preferences in different environments. Only preliminary data on the ion-binding properties of CLA here we present a CD and nmr study on the ion-binding characteristics of CLA and the conformational features of the CLA/Ba2+complex in acetonitrile at room temperature.

EXPERIMENTAL

ones, have been acquired in the phase-sensitive mode, by use of the time proportional phase increment (TPPI).I5All the spectra of the complexed peptide have been run at 298 K. The package INSIGHT/DISCOVER (Biosym Technologies, San Diego, CA) was used for energy minimization of the models emerging from the nmr data. The steepest descent algorithm was used in the early stages of refinement, whereas the quasiNewton-Raphson algorithm was consistently employed in the final stages.

Materials

RESULTS Spectrophotometric grade acetonitrile (99%, Jansen), refluxed and distilled in the presence of calcium hydride, was used for CD measurements. Hexahydrate calcium perchlorate, magnesium, barium and sodium perchlorates, and potassium chloride were supplied by Aldrich (Milwaukee, WI) . Five-millimeter tubes (Wilmad, Buena, N J ) and deuterated acetonitrile (99.98% isotopical purity, Aldrich, Milwaukee, WI) were used for nmr spectra.

Circular Dichroism

CD experiments show a significant change in the molar ellipticity of CLA spectrum by addition of increasing amounts of Ca2+,Ba2+,and Mg2+.No change has been detected in the spectrum of CLA in the case of addition of K + or Na' ions. Figure 1 shows the CD spectra of CLA in its free form and in the presence of an excess of Ba2+cation. Figure

Methods

CD spectra were recorded with a Jasco J-500A dichrograph, using a cylindrical silica cell with path lengths of 0.1 and 1.0 cm at room temperature. The molar ellipticity of the peptide in solution ( [ 61, in deg cm2/d mol) has been measured in the range of 190-260 nm as a function of the added cation solution R ( R = [Mltot/[Pltot, where [MItotand [PItot are the total cation and peptide concentrations, respectively). The different cation titrations have been compared by plotting, as a function of R ,the values N = ( Oo - Oi)/ ( Bo - 8, ) , where O0 is the molar ellipticity of the free peptide, 8i that at the Ri value, and 8, is the one at R -P co. Nuclear magnetic resonance spectra were recorded on Bruker spectrometers (AM 400 and WM 500). The peptide concentration was 6 m M in deuterated acetonitrile. All chemical shifts in part per millions (ppm), in the proton spectra, are referred to internal tetramethylsilane ( T M S ) ; in the I3C spectra, they are referred to the resonance of the methyl carbon of acetonitrile, at 1.5 ppm. Proton one-dimensional ( 1D) spectra have been acquired using typically 16-32 scans with 16K data size; 32K and 5000-15000 scans have been used for 13C experiments. For the two-dimensional (2D ) experiments, both homo- and heterocorrelated pulse programs of the standard Bruker software library were used. All 2D experiments except the heterocorrelated

-5

u I '

200

220

240

:

.l(nm)

Figure 1. CD spectra of CLA in acetonitrile at 298 K. The continuous curve refers to the free peptide; the dashed one to the Ba'+ complex.

763

ION BINDING OF CYCLOLINOPEPTIDE A

A

I

0

0.5

1

2

Da2+/CLA

-

0-

0.5

1

2

Mg2+/CLA

Figure 2. CD saturation plots of CLA with Ba2+,Ca2+,and Mg2+ ions.

2 shows the saturation curves of CLA with Ba2+ Ca2' , and Mg2+ions, respectively. From these plots it is possible to extract, in the range of concentration examined ( 10p4-10-5 M ) , a cation/CLA coordination ratio of 0.5, in agreement with the following global equilibrium:

2P

+ M = P2M

sically high degree of mobility both of the side chains and of the peptide skeleton, l 2 at room and high temperature. Only in chloroform a t very low temperature it has been possible to freeze a definite conformation of CLA, whose main structural features consist of five strong transannular hydrogen bonds, with the formation of one Ci7 two Clo, one C13, and one C I i ring structures; the P'-P2 peptide bond is cis,

The equilibrium constant

K

=

[P2M1/[Pl2[M1

was then calculated using the method reported by Grimaldi et a1.16 The values of K found for the different cation/ CLA complexes are the following: R=2

K

~= ~ 3 x ~1 0 9+~ - 2

Kca2+= 3 X 108M-2 R=O

KMgz+= 5 X 1OiMP2 70

Nuclear Magnetic Resonance

We have recently shown t h a t CLA is characterized, in a wide spectrum of solvent systems, by a n intrin-

60

50

40

30

20

10

PPm

Figure 3. Proton 1D nmr spectra (500 MHz) of CLA in acetonitrile in absence ( A ) and in excess ( B ) of Ba2* ions.

764

TANCREDI ET AL.

R=2

R=0.5

R=0.3

f

5.0

lllM'

R=O

4.0

3.0

PPm

Figure 4. a-CH region of 500-MHz proton 1D nmr spectra of CLA in acetonitrile at various Ba2+/CLAratios.

with all remaining peptide bonds in the trans conformation. Figure 3A shows the room temperature spectrum of CLA in acetonitrile: it is clearly indicative of the

coexistence of many different conformers in intermediate exchange. In particular, the resonances of the amide protons are so broad to prevent almost completely sequential assignment; also notably broad are the CY proton resonances of the two Phe residues, a t 4.90 and 4.45 ppm. It has been possible, however, to identify almost all spin systems belonging to all residue types, except one NH resonance of one of the two Phe residues, by the combined use of double quantum filtered correlated spectroscopy (DQF COSY), l 7 homonuclear Hartman-Hahn (HOHAHA) ,I8 and H-C J correlation" experiments. Very few sequential information can be extracted by rotating frame nuclear Overhauser enhancement spectroscopy ( ROESY ) experiment, which show practically no correlation in the NHCH region; only the sequence V '-PI has been identified, by means of a strong da6correlation, assessing the trans conformation of the Val-Pro linkage. The resonances of the two CY protons of the Pro residues are almost coincident, so that a d,,, typical of a cis peptide bond, cannot be resolved notwithstanding, the lack of any Pro'-Pro2 da6effect speaks against the hypothesis of a trans peptide bond between these amino acids. A final answer is provided by the proton-carbon correlation spectrum: in fact, the p and y carbons of P' resonate a t 29.6 and 25.5 ppm, and those of P 2a t 32.1 and 22.5 ppm, respectively; the Asa, value of 4.4 ppm for P1 is characteristic of a

*'

Table I Proton and Carbon (Lower Part) Chemical Shifts, in ppm from TMS, of the Equimolar Ba2+/CLA Complex in Acetonitrile

NH aCH PCH PCHz -YCH yCH2 -YC& 6CHz

P'

P2

F3

F4

L5

I6

I7

La

V9

4.36

4.10

6.87 4.20

7.71 4.52

6.71 4.47

7.16 4.06 2.51

7.55 4.25 1.94

6.53 4.08

6.87 4.99 2.06

2.03 2.28

0.82 1.65

3.29 3.83

3.11 3.27

1.65

2.00 2.16

1.53 1.65

3.61 4.00

3.23 3.42

1.65

0.94

6CH3

aC PC $2 6C 6CH3 6CH3

1.54 1.71 1.54

66.0 27.3 26.2 49.2

63.1 29.3 26.2 48.3

8.1 35.6

58.7 37.4

2.6 41.7 25.8

21.2

1.11 1.36 1.01

1.33 1.54 0.99

0.72

0.95

60.0 34.1 25.1

61.4 37.4 26.6

16.1

16.6

10.1

12.4

0.93

0.88 0.99 54.7 42.0 25.8

55.5 32.7

23.5 23.8 17.5 20.4

ION BINDING OF CYCLOLINOPEPTIDE A

765

*

1

*

I

I

I

60

IIIJ

50

J

1

20

30

40

10

PPm

Figure 5. Aliphatic portion of the proton-carbon J correlated spectrum of the Ba2+/ CLA complex.

trans Xxx-Pro peptide bondz1;that of 9.6 ppm for P2 is, on the other hand, indicative of a cis conformation of that bond. We can then say that the backbone conformation of the Val-Pro-Pro moiety of CLA in acetonitrile, as far as the peptide bonds are concerned, is identical t o that found for CLA in chloroform at 214 K." All spectra of CLA in acetonitrile, a t room temperature, do not show the presence of a less populated all-trans isomer. The nmr study of the ion binding of CLA has been restricted to the interaction with Ca2+and Baz+ ions, whose association constants (vide supra) exhibit the higher values. T h e spectrum of CLA shows many perturbations upon addition of hexahydrate calcium perclorate, but even a t a molar ratio ion/ peptide higher than 2, there is no indication of the stabilization of a complex of definite stoichiometry and/or conformation; the spectrum, in fact, shows different sets of resonances for each proton species, which is indicative of the presence of different complexed forms in slow exchange in the nmr time scale. On the other hand, addition of barium ions to the CLA/acetonitrile solution results in the appearance of only one set of new resonances, attributable to the complexed form of CLA. The room temperature spectrum of the equimolar Ba2+/CLA solution is shown in Figure 3B. Figure 4 shows the behavior of the CH,, resonances a t different ion/peptide ratios; it is possible to monitor the increasing area of the resonances of the complex, coexisting with the decreasing ones of the free form of the peptide; no shifting or clear broadening of these resonances are

observed throughout the titration; i.e., the free and complexed forms of the peptide are in the slow exchange case. All spectral perturbations end a t a ion/ peptide ratio of 1.0, indicating a complexing stoichiometry of 1 : 1. The observation of only one set of resonances for the complexed form of CLA, during the entire barium titration, with constant halfheight width, means that if a sandwich-type complex is formed a t low Ba2+ concentration, a s suggested by the CD data, it must be in very fast exchange with the equimolar one. All useful data for the conformational study of the Ba2+/CLA complex in acetonitrile have been extracted from nmr spectra of a 2 : 1 Ba2+/CLA solution: the 1D spectrum in this condition does not differ from that of the 1: 1solution. The use of DQF COSY and HOHAHA spectra allowed the identifiTable I1 Coupling Constants, Corresponding 9 Angles, and Temperature Coefficients of the Amidic Protons of the Ba2+/CLAComplex JNHCH

cp

-As/AT

Residue

(Hz)

(degrees)

(PPb/K)

F3 F4 L5

9.6 5.6 9.3 8.2 5.0 5.8 9.6

-105 to -130,60 -170, -70, 25,95 -100 to -140 -160, -90,60 -170, -65, 20, 105 -170, -70, 25,95 -105 to -130,60

0.5 2.4 1.8 0.3 2.5 3.0 1.9

16

LR v9

766

TANCREDI ET AL.

cation of all proton subspectra belonging to each type of amino acid; the sequential assignment has been achieved by means of a ROESY spectrum. Table I shows the chemical shift values, in ppm from TMS, of all protons of the equimolar Ba"/CLA complex; in the lower part of the table are reported the carbon chemical shift values, determined by the H-C correlation spectrum shown in Figure 5. T h e chemical shifts of the 0 and y carbons of the two Pro residues are characterized by values different from those of the free CLA form; their absolute and Assr values are typical, in both cases, of trans XxxPro peptide bonds2'; the all-trans arrangement of the peptide bonds is, moreover, confirmed by a do& Val-Pro and a do, Pro-Pro ROE effects. Table I1 shows the NHCH coupling constant values corrected for electronegativity, the corresponding allowed cp ranges, 23 and the N H temperature coefficients. The Phe3 and Ile6 amide protons exhibit very low coefficients (-0.3 and -0.5 parts per billion/K respectively), so that they can be confidently considered a s intramolecularly hydrogen bonded, possibly through the formation of turns; the hypothesis of a type I 6 --* 3 P-turn, with the N H proton of Ile6 bonded to the carbonyl oxygen of Phe3, is consistent with the data of Table I1 and with those of Table 111, which shows the most relevant interresidue ROEs. Phe3 N H could form a 3 + 1y-turn, bonding the Pro" CO, or a somehow distorted 3 --* 9 @-turn with Val9 CO. J , ROE, and 6 values are not sufficient to identify a unique molecular model of complexed CLA but are certainly good enough to build reasonable starting models for energy minimization procedures. All models were based on the presence, in the backbone, of a type 1 6 --* 3 ,&turn, i.e., involving residues 36, and a 3 --* 1 y-turn. Inclusion of the strongest ROEs and of the closure constraints drastically limit the choice of starting structures. T h e number of acceptable conformations was further restricted by the necessary condition t h a t hydrophilic groups must be accessible t o the cation. All structures topologically consistent with these conditions were subjected

Table IV Backbone Angles of the Ba" Complexed Form of CLA Shown in Figure 6" cp

Pro' -53.2 (-60.0) Pro2 -79.8 (-90.3) Phe3 -90.8 (-97.3) Phe4 -59.5 (-86.0) Leu5 -79.7 (-61.5) Ile6 -166.4 (-55.2) 1ie7 -100.0 (-114.7) -69.6 (55.1) Leu' vai9 -84.5 (-125.0)

* -50.0 (160.2) 21.7 (-17.7) 85.8 (-48.7) -26.8 (56.7) 16.5 (-28.2) 154.0 (-32.7) -68.6 (21.7) -26.3 (48.4) 124.1 (74.8)

w

170.9 (10.0) -178.5 (-175.9) 175.7 (-166.9) 168.8 (-179.5) -166.2 (178.9) -169.0 (179.4) -163.3 (175.6) 169.0 (169.2) -169.4 (175.5)

The values in parentheses refer to the solid state conformation of CLA.''

to energy minimization by means of the package INSIGHT/DISCOVER, using the consistent valence force fieldz4and a variety of minimization algorithms (see Methods), including prominent ROEs as minimization constraints. The final structure was then checked for consistency with all observable

Table I11 Relevant Interresidue ROEs of the Ba2+/CLAComplex

P'

P2

F3

F4

L5

I6

I7

L'

V9

Figure 6. Schematic model of CLA as derived from the minimization procedure. Solid circles represent the five oxygen atoms occupying t.he most favorable position for interaction with Ba2+ion.

ION BINDING OF CYCLOLINOPEPTIDE A

ROES and with the internal rotation angles derived from J couplings. Table IV summarizes all relevant internal coordinates of the final model, together with the corresponding values found for the solid state conformation. T h e main conformational features of this model of the equimolar Ba2+/CLA complex, shown in Figure 6, are that the backbone contains only trans peptide bonds, the retention of the type I 6 3 @-turnsuggested by the Ile6 NH temperature coefficient, and the arrangement of residues 9-3 in a local conformation intermediate between a 3 -+1 y-turn and a 3 + 9 0-turn. The global shape is consistent with the accommodation of Ba2+in a hydrophilic cavity, as emphasized by the representation of' Figure 6, in which only the oxygens that can be involved in the complexation are highlighted the shape of the complexed peptide can in fact be described as a bowl, with the concave (polar) side hosting Ba2' and a n opposite convex side predominantly apolar. I t is interesting t o note that the structure of the free peptideI2contains a Pro'-Pro2 cis peptide bond and is characterized by a much more compact global shape, in which most of the oxygen atoms used to bind Ra" in the complex are involved in intramolecular hydrogen bonds. -+

REFERENCES 1. Kaufmann, H. P. & Tobshirbel, A. ( 1959) Chem. Ber. 92, 2805-2809. 2. Prox, A. & Weygand, F. (1962) in Peptides. Proceeding (if the

8th European Peptide Symposium, Beyerman, H. C., van de Linde, A. & van den Brink, W. M., Eds., Elsevier/ North Holland Biomedical Press, Amsterdam and New York, pp. 158-172. 3. Kessler, H., Kelin, M., Muller, A,, Wagner, K., Bats, .J. W., Ziegler, K. & Frimmer, M. (1986) Angew. Chem. lnt. Ed. Engl. 25,997-999. 4. Ziegler, K., Frimmer, M., Kessler, H., Damm, I., Eiermann, V., Koll, S. & Zarbock, J. (1985) Biochim. Biophys Acta 845, 86-93. 5. Tonelli, A. E. (1971) Proc. Natl. Acad. Sci. U S A 68, 1203-1207.

767

6. Naider, F., Benedetti, E. & Goodman, M. (1971) Proc. Natl. Acad. Sci. U S A 68, 1195-1198. 7. Brewster, A. I. & Bovey, F. A. (1971) Proc. Natl. Acad. Sci. U S A 68,1199-1202. 8. Siemion, I. Z., Klis, W. A., Sucharda-Sobczyk, A. & Obermeier, R. (1977) Rocniki Chem. 51, 1489-1498. 9. Balasubramanian, D., Chopra, P., Ardeshir, F. & Rajappa, S. (1976) F E B S Lett. 65,69-72. 10. Di Blasio, B., Benedetti, E., Pavone, V. & Pedone, C. ( 1987) Biopolymers 26, 2099-2101. 11. Tancredi, T., Zanotti, G., Rossi, F., Benedetti, E., Pedone, C. & Temussi, P. A. (1989) Biopolymers 28, 513-523. 12. Di Blasio, B., Rossi, F., Benedetti, E., Pavone, V., Pedone, C., Temussi, P. A., Zanotti, G. & Tancredi, T. ( 1989) J. A m . Chem. SOC.111,9089-9098. 13. Klis, W. A., Obermeier, R., Siemion, I. Z., SuchardaSobczyk, A. & Gatner, K. (1977) Rocniki Chem. 51, 1499-1509. 14. Chatterji, D., Sankaram, M. B. & Balasubramanian, D. (1987) J. Biosci. 11,473-484. 15. Marion, D. & Wuthrich, K. (1983) Biochim. Biophys. Res. Commun. 113, 967-974. 16. Grimaldi, M., Rossi, F., Saviano, M., Benedetti, E., Pavone, V. & Pedone, C. ( 1990) Biopolymers 29, in press. 17. Rance, M., Ssrensen, 0. W., Bodenhausen, G., Wagner, G., Ernst, R. R. & Wuthrich, K. ( 1983) Biochem. Biophys. Res. Commun. 117, 479-485. 18. Braunschweiler, L. & Ernst, R. R. (1983) J. Magn. Reson. 53,521-528. 19. Freeman, R. & Morris, G. A. (1978) J . Chem. SOC. Chem. Commun. 684-686. 20. Bothner-By, A. A., Stephens, R. L., Lee, J., Warren, C. D. & Jeanloz, R. W. (1984) J. Am. Chem. SOC. 106,811-813. 21. Dorman, E. D. & Bovey, F. A. (1973) J. Org. Chem. 38, 2379-2383. 22. Macura, S. & Ernst, R. R. (1979) Mol. Phys. 41,91101. 23. Bystrov, V. F. (1976) Prog. N M R Spectrosc. 10, 4162. 24. Dauber-Osguthorpe, P., Roberts, V. A., Osguthorpe, D. J., Wolff, J., Genest, M. & Hagler, A. T. (1988) Proteins: Struct. Funct. Genet. 4, 3 1-4 7.

Received August 1, 1990 Accepted October 17, 1990