J. Peptide Res. 49, 1997, IS-22 Printed in UK - all rights reserved

Copyright 0 Munksgaard 1997 JOURNAL OF

PEPTIDE RESEARCH ISSN 1397-002X

Conformational analysis of cyclo (2,9)-Ac-QCRSVEGSCG-OH from the C-terminal loop of human growth hormone

D.S. SEETHARAMA JOIS, MATTHEW W. CONRAD, SOMA CHAKRABARTI and TERUNA J. SIAHAAN

Department of Pharmaceutical Chemistry, University of Kansas, Lawrence, Kansas, USA

Received 5 March, revised 24 April, accepted for publication 23 June 1996

A 10 amino acid residue cyclic peptide, cyclo(2,9)-Ac-Glnl -Cys2-Arg3-Ser4-Val5-Glu6-Gly7- Ser8-Cys9-Glyl0, from the C-terminal region of human growth hormone (hGH) was synthesized and studied by 2D proton NMR and molecular dynamics (MD) simulations. The solubility of the peptide was low in water; hence, NMR studies were done in two solvent mixtures, water and deuterated dimethyl sulfoxide. NOE-constrained molecular dynamics and MD simulations resulted in major and minor conformers in solution. The major conformer has a type I p-turn at Glnl-Cys2-Arg3-Ser4 and a loop structure around Glu6-Gly7-Ser8. Comparison of the conformation of this peptide with the peptide fragment 181-190 in the intact hGH protein X-ray crystal structure indicated that the synthetic peptide retains some structural similarity to the intact protein. Since the C-terminal region is important in binding the hCH protein to its receptor, the conformation of the synthetic peptide could be useful in understanding the binding mechanism. 0 Munksgaard 1997.

Key words: human growth hormone; hGH; hGH-peptide conformation; C-terminal loop of hGH; NMR; molecular dynamics

Human growth hormone (hGH) is a protein of 191 amino acids that functions in humans as a regulator of growth. The X-ray structure of the hGH complex with the hGH receptor shows that hGH binds to a dimer form of the receptor (1). A small part of the C-terminal loop of hGH contributes to the binding between the hGH protein and its receptors. Antibodies have also been used as an analytical tool to determine the correct folded structure of the recom- binant product of hGH (2, 3 ) . There are several antibodies that recognize different regions of hGH, but the mechanism of antibody recognition is sensitive to amino acid sequence change in hGH (4). For example, the C-terminal region of hGH is recognized by monoclonal antibody Mab-12; a mutation of the Arg-183 residue with an Ala residue on hGH reduces the binding affinity of Mab-12 to hGH (5).

In order to get some insight into the structural feature of the antibody binding region, we have synthesized a peptide with a sequence similar to that of the fragment 181-190 amino acids in the C- terminal loop of the hGH protein. This region was chosen because of the conformational restrictions due to the disulfide bond and the importance of Arg183 noted above. Structural studies of different fragments of human growth hormones show that they have a

secondary structure similar to that of the intact proteins. Gooley et al. (6, 7) have studied the con- formation of fragments 96-133 of bovine growth hormone by NMR and molecular modeling and found that the formation of an omega loop in the peptide fragment was similar to the intact protein. Roongta et a/. (8) studied a peptide fragment, 1-28 of hGH, and found an a-helix conformation from residue 8 to 24 similar to the helix-1 of hGH. In the present work we report the synthesis and conforma- tional study of a cyclic peptide cyclo(2,9)-Ac- QCRSVEGSCG-OH (hGH peptide, Fig. 1) derived from the C-terminal region of the hGH protein. The aim of the study is to provide additional structural information on the C-terminal loop that is important in binding to its receptor and monoclonal antibody using NMR and molecular dynamics (MD) simulations.

EXPERIMENTAL PROCEDURES

Synthesis of the peptide. Peptide hGH was synthesized by the standard solid-phase method, using a phenyl- acetamidomethyl (PAM) resin and tert-butyloxy- carbamate (BOC) amino acid chemistry. The peptide was cleaved from the resin using a mixture of tri-

15

D.S.S. Jois et a1

OH 1 J-

O 7 C n l

4L

FIGURE 1 Chemical structure of cyclo(Ac-QCRSVEGSCG-OH )

fluoromethylsulfonic acid/ trifluoroacetic acid (TFMSA/TFA) in the presence of scavengers (phenol and ethanedithiol ) and then precipitated in cold ether to give the crude product. The crude product was dissolved in water at a concentration of 0.02 $1, and the pH was adjusted to 7.8 with ammonium hydrox- ide. Potassium ferricyanide solution (2.0 mg/mL) was added to this solution to form a disulfide bond between Cys2 and Cys9 and give the cyclic peptide. The water was removed by high vacuum evaporation followed by lyophilization. The crude cyclic peptide was purified by preparative high-performance liquid chromatography (HPLC) !sing a CIS column (21.4x250mm, 12 pm, 300 A) with a gradient of solvent A (95 : 5 = 0.1% TFA/H,O : acetonitrile) and solvent B (100% acetonitrile). The peptide was detected by UV at i.=220 nm. Fractions were col- lected in test tubes using an automatic fraction col- lector. The purity of each fraction was determined by analytical HPLC using a CI8 column (4.6x250mm, 5pm, 300A) with the same solvent system as in the preparative HPLC. The cyclic peptide was then analyzed by fast atom bombardment mass spectrometry (FAB-MS) to give M S 1 = 1066; this result was confirmed by two-dimensional NMR.

N M R spectroscopj-. NMR studies of the hGH peptide were carried out using two solvent mixtures. Samples for NMR spectra in water were prepared by dissol- ving 3 mg ( 5 . 5 mM) of peptide in 500 ILL of 90Y0 H20/10% D,O at pH 3.5. Two samples at different

16

concentrations were prepared in DMSO-d6: one 9 . 9 m ~ and one 20 mM. An additional sample with 70% H 2 0 and 30% DMSO-d, (8 mM) was also prepared for NMR studies. The one- and two- dimensional NMR experiments were carried out using 500 MHz Briiker AM-500 and Varian Unity 500 MHz NMR spectrometers with variable temperature probes. Homonuclear Hartmann-Hahn (HOHAHA) ( 9) and nuclear Overhauser enhancement and exchange spectroscopy (NOESY) ( 10) experiments were done by presaturation of water during relaxation delay. Data were collected in time-proportional phase incrementation (TPPI) (11) or states method with a sweep width of 5000Hz. Typically, 48 or 64 scans were used with 2K data points in the t2 dimension and 256 or 300 experiments in the t , dimension. Data sets consisting of 512 r , increments of 2K complex data points were collected for double quantum filtered correlated spectroscopy (DQF-COSY) ( 12) experi- ments. A primitive exclusive correlated spectroscopy (PE-COSY ) ( 13) experiment was done with 4K com- plex data points in the t , dimension and 1024 experi- ments along the t , dimension. The NMR spectra were processed using FELIX software, version 2.30 (Biosym Technologies, San Diego, CA) with the final matrix of 1K x 1K real data points. For PE-COSY data, 4K x 4K complex data points were used. A sine bell filter shifted by 45 or 60" was used in both dimensions prior to Fourier transformation. Baseline corrections were applied after phase correction using the convolution method available in the FELIX soft- ware. Spectra were represented in the absorption mode in the case of NOESY and the phase-sensitive mode in the case of DQF-COSY and PE-COSY experiments.

The temperature dependence of the amide proton chemical shift was measured by collecting data from 10 to 40 'C in steps of 5 "C using a variable temper- ature probe. Coupling constants (3JHN,) were meas- ured from the DQF-COSY spectra at 25 "C, and 3J,s were measured using a cross-section of PE-COSY cross-peak (14). NOESY spectra were acquired at 100. 200, 300 and 400 ms mixing times. The buildup of NOE up to 300 ms NOESY in this series of experiments ruled out spin diffusion artifacts ( 15). NOE cross-peak volumes were measured using NOESY spectra with 200 ms mixing time of an 8 mM sample in 70% H,O, 30% DMSO-d6. The NOE observed between the gemiqal protons of glycine (with a fixed distance of 1.78 A) was used to calibrate NOE cross-peak intensities using a two spin approxi- mation (16). NOESY buildup rates were obtained by least-squares fitting of the four measured points. The correlation coefficient for the linear least-square fit was greater than 0.9. In the case of dN?(i,i), dmN(i,i+ 1 ) and dN,(i,i+ 1) the upper and lower bounts of dis- tances were generated by adding k0.3 A to the calculated exact distances from NOE data to take

CYCIO (2,9)-Ac-QCRSVEGSCG-OH

Several conformations were selected for the final analysis using two criteria: (a) the conformation ha$ an interproton distance error of less than 0.5 A compared to the calculated distances from NOE data; and (b) the conformation had q5 angles within 30' of the calculated &values from 3JHNa (20). These were taken as viable NMR solution conformations and were minimized with solvent molecules using the conjugate gradient method without crcss-terms until the r.m.s. derivative was 0.3 kcal/mol A. The energy- minimized structures were analyzed in terms of dihed- ral angles, energy and observed NOE distances.

into account the error in the measurement of volume of NOESY cross-peaks. The upper bounds of dis- tances dNa(i,Q, dEN(i,i+ 1) and dNN(i,i+ 1) were <3.0, 3.6 and 4.5 A, respectively. The Cys9 side-chain pro- tons were stereospecifically assigned using PE-COSY and relative NOE cross-peak intensities and hence NOE restraints were applied. All other side chains of amino-acids in the peptide were not stereospecifically assigned due to either overlapping of HP NOE cross- peaks or average 3J.p coupling constants. Therefore, NOE restraints were calculated by considering pseudo atoms (17).

Computational methods Conformational space was searched for this cyclic peptide using the Discover program, version 2.35 (Biosym Technologies, San Diego, CA), to identify conformations consistent with the experimental NOE and coupling constant data. Calculations were performed on an Indigo, Silicon Graphics computer. Consistence valence force field (CVFF ) was used for these calculations. Initially, the peptide was cyclized using NOE and disulfide bond constraints by running MD simulations in vucuo for 20 ps at 900 K with a dielectric constant of unity. A penalty force of 100 kcal mol-' was used to bring the protons to observable NOE distances. From the 20 ps MD simulations, several structures with reasonable disulfide geometry were chosen, and the disulfide bond was formed to make cyclic structures. These structures were then subjected to 900K MD for 100 ps to explore the possibility of several conformations that the peptide can acquire. The trajectories were updated every 100 fs. During this high-temperature simulation, all of the peptide bonds were held fruns, since the peptide does not contain proline residues or N-methylated residues (18, 19). The trajectory from high-temperature dynamics was analyzed for similarities between the structures by comparing the root mean square (r.m.s.) deviations between each possible pair of structures. The r.m.s. deviations between the backbone atoms were plotted in a 100 x 100 matrix with the x- and y-axes repres- enting structure numbers. Elements near the diagonal represented structures with low r.m.s. values. The plot (data not shown) revealed a number of areas of clusters along the diagonal, indicating that several distinct structures were encountered during the MD simulations. These areas were divided into seven conformational families. The r.m.s. deviation of the backbone atoms within theo family compared to the average structure was 0.4 A. The average structure was taken from each family and subjected to 100 steps of energy minimization, using the steepest des- cents method to relax the molecule. Each structure w;t,s then soaked with water molecules (a layer of SA), followed by 100 ps MD simulations at 300K with all the NOE constraints. Trajectories were updated every 100 fs during the MD simulations.

RESULTS AND DISCUSSION

Proton NMR assignments Because of the low aqueous solubility of the peptide, NMR data were collected in the nonaqueous solvent DMSO-d6 as well as in aqueous-nonaqueous mix- tures. Figure2 shows the 500MHz 'H NMR HOHAHA spectrum of the hGH peptide in a H,0/DMSO-d6 (70%/30%) solvent mixture. Assignment of all the proton resonances was accomp- lished using standard NMR procedures (17). The chemical shift data for the peptide in different solvent systems are given in Table 1. Although NOES were observed in H,O/D,O, H,O/DMSO-d, proved to be quite advantageous in studying the peptide, as there was less overlap of signals and more intense NOE cross-peaks. NOESY spectra were used to obtain inter-residue connectivities and to distinguish between equivalent spin systems. C"H-NH (i,i+ 1 ) connectivi-

I R 3

I , 8 . 4 8 : 1 7 . 8

D I ( p p m )

FIGURE 2 500 MHz 'H NMR HOHAHA spectrum of the hGH peptide at 70 ins mixing time at 25 "C. Cross-peaks between amide protons and H", HB, HY and Hd protons are shown with assignments.

17

D.S.S. Jois et al.

TABLE 1 Proton chemical shfts of the IiGHpepmIe at 25 C 111 9 W H 2 0 1094

DZO (pH 3 S), 7#% HZO 3trX DMSO-d,j uI1cl lo@% DMSO-rl,

Residue Proton Chemical shift (ppm)

Gln 1

cys2

Arg3

Ser4

Val5

Glu6

Gly7

Ser8

cys9

GlylO

NH C'H CBH C*H C'H NH C"H CBH NH C'H CpH C,H C*H NH C"H CPH NH C'H CpH C'H NH C'H CBH C,H NH C"H NH C'H CPH NH C'H CPH NH C'H

8 33 4 43 199 2 16 245 245 6 81 7 29 8 56 4 52 3 15 3 15 8 62 4 92 175 1 9 5 170 1 65 3 16 8 05 4 50 381 391 8 21 4 16 2 14 0 9 6 0 9 5 8 33 4 25 195 2 36 8 20 3 35 3 12 8 09 4 54 3 8 7 3 8 7 8 33 4 77 3 3 0 3 10 8 58 3 20 3 20

8.06 4.02 1.91 1.72 2.15,2.15 6.71 7.25 8.09 4.46 2.90 3.06 8.48 4.23 1.68 1.68 l.54 1.43 3.02 7.80 4.29 3.53 3.53 8.12 4.02 2.04 0.77 8.00 4.16 1.80 1.60 2.05 7.86 3.88 3.45 7.74 4.40 3.57 3.57 8.29 4.61 3.12 2.87 7.94 3.66 3.45

8 17 4 25 1 90 I 90 2 19 2 19 6 88 7 28 8 17 4 54 3 0 3 18 8 53 4 34 1 74 1 66 150 1 50 3 15 7 80 4 48 3 68 3 68 8 17 4 12 2 06 0 84 0 82 8 06 4 12 180 2 06 7 94 3 52 3 98 7 88 4 36 3 60 3 60 8 36 4 70 3 20 2 98 8 02 3 60 3 43

ties were observed from Cys2 to Ser4 and Ser8 to GlylO (Fig. 3). GlylO resonances could be distingu- ished from Gly7 by its NH-to-C"H of Cys9 connectiv- ity as well as its Cys9-GlylO NH-NH connectivity (Fig. 4). Similarly, Cys2, and Ser4 resonances could be distinguished from Cys9 and Ser8 by their connect- ivities to Arg3 in the NOESY spectrum. One of the long-range NOES observed between the NH of Arg3 and the NH of GlylO indicated the folded peptide structure.

Conformation of the peptide One-dimensional NMR of the peptide showed good dispersion of chemical shift over a range of chemical shifts (0.7 ppm) in all three solvent systems studied, indicative of a preferable conformation of the peptide.

18

58-58 a

-54 a

CS-C9

8 : 4 8 : 2 81 0 71 8 DI (pprn)

FIGURE 3 Portion of 200 ms mixing time NOESY spectrum of the hGH peptide ( 8 mM, 70% H20/30% DMSO-d,, 25 "C) showing the NH-C'H region with sequential connectivities.

815 8.0 7: 5 7 : 0 DI (pprn)

FIGURE 4 Amide region of the 200 ms NOESY spectrum in 70% Hz0/30% DMSO-d, at 25 C showing NH-NH connectivities. Gua refers to guanidine group.

One of the glycine (Gly7) C"H enantiotopic protons had a Ad value around 0.4 ppm, indicating that these protons are not in the same environment. However, the coupling constant 3JHN3 values were in the range 6-8 Hz, indicating that the peptide might be under- going conformational exchange (21). GlylO has a Ad value of 0.2ppm, indicating flexibility in the C- terminal region of the peptide.

Most information about tertiary structures was obtained from NOE data in a 70% H20/30% DMSO- d6 solvent mixture. Comparison of chemical shifts of

Glnl ~ Cya? - Arg3 - Ser4 - Val5 - Glu6 - Gly7 - Ser8 - Cys9 - GlylO

I===

rn m - - mm m = - -

FIGURE 5 Summary of ‘H-’H NOE connectivities for the hGH peptide in 70% H,0/30% DMSO-d, at 25 “C. Intensity of NOE cross-peaks is indicated by the thickness of the lines, grouped into strong, medium. and weak.

proton resonances (Table 1) and NOEs indicated that the backbone conformation of the peptide is con- served in different solvent systems. The observed NOEs are schematically represented in Fig. 5. From the NOE-restrained MD simulations and energy min- imization, seven families of structures were generated. Two familes of the structures satisfied most of the observed NOEs, while the remaining five familes satisfied only 60% of the observed NOE distance constraints. The observation of both dNN(i,i+ 1) and darN(i,i + 1 ) NOEs indicate that the peptide backbone exists in a tight turn. All of the seven families of structures were examined, since there is a possibility that the cyclic peptide may exist as a mixture of conformers in fast exchange on the chemical-shift timescale. In all of the seven conformers, the dihedral angles around Cys2 and Arg3 residues remained the same, indicating the well defined structure of the peptide in this region. The major change in the backbone conformation was around Ser4 and Va15, giving rise to observed violation of NOE distance constraints. NOE data, temperature dependence of chemical shift and vicinal coupling constants indi- cated the possibility of a type I p-turn at Glnl-Cys2-Arg3-Ser4. To check the stability of this turn in MD-simulated structures, the distance between the carbonyl oxygen of Glnl and the NH of Ser4 was plotted (Fig. 6a) from 100 ps, 300 K MD trajectory. From Fig. 6a it is clear that, while most of the time the hydrogen bonding is intact, it is destabilized during the course of dynamics. This is due to the movement of the carbonyl group of Glnl and the NH of Ser4 away from each other because of the flexibilty of the structure around Ser4, thus destabilizing the turn. However, 4, $ angles around Cys2 and Arg3 varied within +40° from the standard values for a type I p-turn during the course of dynamics. Hence the major conformation of popula-

h

Y

a> n L)

v) .3

I

0.0 101.0 Time(ps)

0 7

0.6 I

0.0 ! I

1 3 5 7 4

Residue number

FIGURE 6 Plots of (a) distance between the C=O of Glnl and the NH of Ser4 during 100 ps MD simulation, showing the stability of the intramolecular hydrogen bond; (b) average pairwise r.m.s. devi- ation for each residue of 10 structures to the average structure from the hGHl family, (0) for backbone atoms, (+) all atoms.

tion has a p-turn structure. Two families of structures that satisfied more than 95% of NOE constraint were chosen to represent the final conformation of the peptide. Each family consisted of ten structures that satisfied all of the NMR constraints. These ten struc- tures were superimposed on the average structure in each family to verify convergence to the structure. The r.m.s. deviations of these structures from the average structure were calculated. In the hGHl family, the ave!age r.m.s. deviation of the backboGe atoms was 0.5 A and of the overall atoms was 0.8 A. In the hGH2 family, the average r.m.s. deviFtion was 0.5 A between the backbone atoms and 1.0 A between the overall atoms. Figure 6b shows the plot of average pairwise r.m.s. deviations of 10 structures from the average structure for each residue in the peptide for hGHl family. From the figure, it is clear that the peptide structure is stable in the 8-turn region and near the loop, and flexible around the fourth and fifth residues.

Conformer hGHl (Fig. 7a) is characterized by a p-turn at Glnl-Cys2-Arg3-Ser4. The p-turn is stabil- ized by the hydrogen bonding between the NH of

19

D.S.S. Jois et al.

FIGURE I Stereoview of the energy minimized average structure of the hGH peptide: (a) conformation hGH1: ( b ) conformation hGH2

Ser4 and the Glnl C=O. The dihedral angle at Cys2 and Arg3 indicated that this is a type I p-turn (22). The possibility of a j-turn at Gln 1 -Cys2-Arg3-Ser4 is supported by the observed NOE between the NH of Cys2 and the NH of Arg3 as well as between the NH of Arg3 and the NH of Ser4. The possibility of intramolecular hydrogen bonding from the NH of Ser4 to the C=O of Glnl is also indicated by the relatively low temperature coefficient of chemical shift of the NH of Ser4 (4.2 ppb/K) compared to other amide protons in the peptide (Table 2) . The proposed conformer also indicated the possibility of two more intramolecular hydrogens in the peptide, one between the NH of Cys2 and the carbonyl group of the acetyl group and the other between the Glu side chain and the C=O of Ser8. The dihedral angles at Glnl indicated that there is an inverse ;'-turn that is stabil- ized by intramolecular hydrogen bonding between the NH of Cys2 and the carbonyi group of the acetyl group. However, the temperature coefficient of the chemical shift value of the NH of Cys2 was quite high, indicating the solvent-exposed amide proton. The hydrogen bonding between the Glu6 side chain and the C=O of Ser8 forms a loop-like structure that stabilizes the C-terminal region of the peptide. The overall conformer hGH 1 is a turn-extended-loop type of structure with the turn and loop stabilized by intramolecular hydrogen bonding. The side chains of Glu6 and SerS are exposed to one side of the peptide,

20

TABLE 2 Trniperciture dependent e of t k uniide resonances, and coupling co~istuiit~ 3JH, , of the IiGHpeptide in 90% H,OjIff% D 2 0 ( p H 3 S),

7Q': H,O 30% DMSO-cl, und 100% DMSO-d,

Residue H,O D20 H,O/DMSO-d, DMSO-d, 9O'Yo 10% 100/;~/30'% ~

Afi /AT J A 6 , A T .P ASJAT J (ppb/K)

(ppb K ) (PPb/K)

Gln 1

Arg3 Ser4 Val5 Glu6 Gly7

Ser8 Cys9 GlylO

cys2 5.5 6.1 5.8 1.5 8.5 1.3 6.3 1.2 1.6 4.6 1.6 6.0 1.6 5.1 4.2 6.1 5.0 6.4 5.0 6.5 1.5 6.6 1.0 6.0 1.5 8.0 5.8 5.6 1.8 4.9 5.0 5.0 4.8 4.1 4.5

5.9 5.8 5.9 6.1 7.2 5.5 6.4 I. I 1.9 5.8 8.2 6.1 6.5 6.3 1.8 6.0 5.0 6.4 5.0 5.1 4.5

6.0 6.0 5.5

J refers to coupling constant 'JHNZ in Hz.

creating a negative potential; Arg3 is exposed to the other side of the peptide, creating a positive potential.

The second possible conformer, hGH2 (Fig. 7b), is characterized by a )'-turn at Ser4-Val5-Glu6. The ;*-turn is stabilized by intramolecular hydrogen bond-

CYCIO ( 2,9)-Ac-QCRSVEGSCG-OH

relatively due to overlap, and hence the Ser4 side chain could not be assigned to a single conformer. Cys2, Arg3 and Ser8 showed average coupling con- stants (5.5 and 6.5 Hz), indicating rotational mobility around xl.

To compare the conformation of the synthetic fragment of the hGH peptide with the peptide frag- ment (181-190) in the intact protein, we overlaid the conformation of the hGH peptide with the fragment (amino acids 181-190) from the X-ray crystal struc- ture of the hGH protein (Fig. 8) (1). The synthetic peptide fragment maintains a structural similarity to the analogous region of intact protein. The r.m.s. deviations of the backbone atoms between con- formers hGHl and hGH2 of the hGH peptide and thatofrom fragment 181-190 of intact protein were 2.8 A and 3.1 A, respectively. In the case of the hGH protein, there is a type I1 /l-turn at Va15-Glu6-Gly7-

ing between Glu6 NH and Ser4 C = 0. The dihedral angles at Val5 indicated that this is an inverse y-type of turn (22). However, the intramolecular hydrogen bonding is not supported by temperature dependence data from NMR.

The method outlined by Wagner et al. (23) was used to make stereospecific assignments of side chain protons from 3Jmp coupling constants using a PE-COSY spectrum and from intensities of cross- peaks in the NH-/lH region of NOESY spectrum. Ser4 and Cys9 resonances showed one small (5 Hz) and one large (1 1 Hz) coupling constant (3JaP) , indic- ating a gauche-trans or trans-gauche conformation (24) with x1 = 180" or -60". Cys9 showed a weak H"-HP2 NOE, a strong H"-HP3 NOE, a strong NH-HP2 NOE, and a weak NH-Hp3 NOE, resulting in x1 acquiring a t2g3 conformation (i l l = - 60"). Ser4 Hfl cross-peak intensities could not be measured

TABLE 3 Dihedral angles for the energy minimized average structure of the conformers hGH1, hGH2 and for residues 181-190 froin the crystal

structure of the hGH protein. Calculated dihedral angfes (in ") from coupling constant (3J,,v,) are also given for comparison

Residue Calculated 4 hGH 1 hGH2 "hGH ( 181-1 90)

4 i 4 i 4 t i /

Gln 1 70 - 85 62 110 - 70 - 38 cys2 70,50, -90, - 155 - 60 - 24 - 65 130 - 65 - 27 Arg3 80,40, -84, - 155 - 55 - 18 - 95 - 55 - 92 - 33 Ser4 80,30, -70, - 160 -110 - 136 - 83 135 - 80 - 42 Val5 75,40, -85, - 150 - 140 86 -71 80 -110 99 Glu6 78,41, -85, - 155 ~ 145 - 36 - 150 53 - 77 133 Gly7 138,50, -50,138 128 - 74 120 - 65 96 7 Ser8 77,42, -86, - 153 - 94 130 - 73 - 55 -71 -51 cys9 77,42, -85, - 153 - 104 125 - 90 165 - 98 - 26 Gly 10 141, 48, -48, 141 - 170 175 138

a Dihedral angles for residues 181-190 were obtained from crystal structure of hGH protein (1); 4 and $ angles In degrees.

FIGURE 8 Stereoview of an overlay of conformation hGHl (stick) of the hGH peptide with the residues 181-190 in the X-ray crystal structure of the hGH protein (ball and stick).

21

D.S.S. Jois et al.

Ser8, with intramolecular hydrogen bonding between the N H of Ser8 and the C=O of Va15. In conformer hGH1, a type I p-turn is observed at the residues Glnl-Cys2-Arg3-Ser4. The dihedral angles (Table 3 ) around Cysl82-Argl83 in the hGH protein indicated that there is a type I p-turn. There is no intramolecu- lar hydrogen bonding, although the distance betwe5n the NH of Ser184 and the C=O of Gln181 is 2.7 A. The carbonyl group of Gln 18 1 was involved in 5 -+ 1 intramolecular hydrogen bonding with the NH of Va1185, and hence moved away from the NH of Ser184. The conformation of a small peptide fragment in solution is similar to that of intact protein; this may be because of the restriction of the conformation from the disulfide bond and the p-turn. The orienta- tion of side chains and the b-turn structure may be important in binding the hGH protein to its receptor and to the monoclonal antibody.

In conclusion, the synthetic peptide hGH exhibits a major and a minor conformation in solution: one with a ?-turn, type I p-turn and one with a loop structure. The conformer with a /3-turn is similar to the conformer of the peptide fragment (181-190) in the intact protein. This structure may be important in binding the hGH protein to its receptor with proper side chain orientation.

ACKNOWLEDGEMENT

We thank the Pharmaceutical Manufactures Association for an Undergraduate Research Fellowship to M.W C.. NSF EPSCoR (EHR 92-55213) for financial support to T.J.S.. and Dr. Om Prakash at Kansas State University for assistance with NOESY and PE-COSY NMR spectra.

1.

2.

3.

4.

5 .

REFERENCES

De Vos. A.M.. Ultsch. M. & Kossiakoff. A.A. (1992) Science 255. 3 0 6 3 12 Surowy, T.K.. Bartholomew. R.M. & VanderLaan. W.P. (1984) Mol. h7nZUi70l. 21. 345-353 Geysen, H.M., Tainer, J.A.. Rodda. S.J . . Mason. T.J.. Alexander. H.. Getzoff, E.D. & Lerner. R.A. ( 1987) Science

Cunningham, B.C.. Jhurani. P.. Ng. P. & Wells. J.A. (1989) Sciencc~ 243. 1330-1336 Abdelmeguid, S.S.. Bentle. L.A.. Dayringer. H.E.. Shieh. H.-S.

235. I 184- I 190

6

7

8

9 LO

I I

12.

13. 14

15

16

17

18 19

20 11 12

23

Smith. W.W. & Violand, B.N. (1987) Proc. Natl. Acad. Sci.

Gooley. P.R., Carter, S.A., Fagerness, P.E. & MacKenzie, N.E. (1988) Proteins: Struct. Funct. Genet. 4, 48-55 Gooley, P.R., Plaisted, S.M., Brems, D.N. & MacKenzie, N.E. ( 1988) Biochen,isrrj, 27. 802-809 Roongta. V., Powers, R., Jones, C., Beakage, M.J., Shields, J.E. & Gorenstein, D.G. (1989) Biochemistry 28, 355-360 Bax. A. & Davis, D.G. (1985) J. Magn. Reson. 65, 355-360 Jeener, J., Meier, B.H., Bachman, P. & Ernst, R.R. (1979) J. Cheni. P h ~ s . 71, 4546-4553 Marion. D. & Wiithrich, K. (1983) Biochem. Biophys. Res. Co~i~r i iun. 113. 967-974 Rance. M. Sorensen, O.W., Bodenhausen,G., Wagner, G., Ernst. R.R. & Wiithrich, K. (1983) Biochem. Biophys. Res. Coriniiurr. 117, 479-485 Mueller, L. (1987) J. Magit. Reson. 72, 191-196 Smith. J.L., Sutcliffe, M.J., Redfield, C. & Dobson, M.C. (1991 ) Biochenri.stry 30, 986-996 Clore. G.M.. Martin. S.R. & Gronenborn, A.M. (1986) J. Mol. Biol. 191, 553-561 Thomas. P.D.. Basus, V.J. & James, T.L. (1991) Proc. Nutl.

Wiithrich. K. ( 1986) N M R qfProreins and Nucleic Acids, John Wiley & Sons. New York Tonelli. A.E. (1976) Biopolymers 15, 1615-1622 Ramachandran, G.N. & Mitra, A.K. (1976) J. Mol. Bid.

Bystrov. V.F. (1976) Progr. N M R Spectrosc. 10, 41-81 Kessler. H. (1982) Angen'. Chem., Int. Ed. Engl. 21, 512-523 Rose. G.D., Gierasch, L.M. & Smith, J.A. (1985) Adv. Profein C h n . 37. 1-109 Wagner. G., Braun, W., Havel, T.F., Schaumann, T., Go, N. & Wuthrich. K. (1987) J ,4401. Biol. 196, 611-639

USA 84. 64346437

A c d SC~. USA. 88. 1237- I241

107, 85-92

24. Nagayama. K. & Wuthrich, K. (1981) Eur. J. Biochem. 114, 365-314

Address

Dr. Terurtn J. Sialiaatr Assistant Professor Department of Pharmaceutical Chemistry Simons Laboratories The University of Kansas Lawrence. KS 66047 USA

Tel (913)463-7327 e-mail: siahaan@smissman.hbc.ukans.edu

22