Conformations of Heterochiral and Homochiral -Pseudoproline ConformationsSegments in of : Heterochiral Context and Dependent HomochiralCis–Trans Proline-PseudoprolinePeptide SegmentsBond Isomerization in Peptides: Context Dependent Cis–Trans Bond Isomerization

Kantharaju,1 Srinivasarao Raghothama,2 Upadhyayula Surya Raghavender,3 Subrayashastry Aravinda,3 Narayanaswamy Shamala,3 Padmanabhan Balaram1 1 Molecular Biophysics Unit, Indian Institute of Science, Bangalore, Karnataka 560012, India 2 NMR Research Centre, Indian Institute of Science, Bangalore, Karnataka 560012, India 3 Department of Physics, Indian Institute of Science, Bangalore, Karnataka 560012, India

Received 9 March 2009; accepted 7 April 2009 Published online 16 April 2009 in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/bip.21207

ABSTRACT: stabilized by two intramolecular hydrogen bonds. Further truncation of the sequence gives an appreciable rise in the The pseudoproline residue (CPro, L-2,2-dimethyl-1,3- D thiazolidine-4-carboxylic acid) has been introduced into population of cis conformers in the tripeptide Piv- Pro- C heterochiral diproline segments that have been previously Pro-Leu-OMe (6). In the homochiral segment Piv-Pro- C shown to facilitate the formation of b-hairpins, containing Pro-Leu-OMe (7)onlythecis form is observed with the b central two and three residue turns. NMR studies of the NMR evidence strongly supporting a type VIa -turn ? octapeptide Boc-Leu-Phe-Val-DPro-CPro-Leu-Phe-Val- conformation, stabilized by a 4 1 hydrogen bond OMe (1), Boc-Leu-Val-Val-DPro-CPro-Leu-Val-Val-OMe between the Piv (CO) and Leu (3) NH groups. The crystal CH,CH3 (2), and the nonapeptide sequence Boc-Leu-Phe- structure of the analog peptide 7a (Piv-Pro- Pro- Val-DPro-CPro-DAla-Leu-Phe-Val-OMe (3)established Leu-NHMe) confirms the cis peptide bond geometry for CH,CH3 well-registered b-hairpin structures in chloroform the Pro- Pro peptide bond, resulting in a type VIa b # solution, with the almost exclusive population of the trans -turn conformation. 2009 Wiley Periodicals, Inc. conformation for the peptide bond preceding the CPro Biopolymers (Pept Sci) 92: 405–416, 2009. residue. The b-hairpin conformation of 1 is confirmed by Keywords: pseudoproline; peptides; cis-trans isomeriza- single crystal X-ray diffraction. Truncation of the strand tion; NMR; turn length in Boc-Val-DPro-CPro-Leu-OMe (4) results in an increase in the population of the cis conformer, with a cis/ This article was originally published online as an accepted L preprint. The ‘‘Published Online’’ date corresponds to the trans ratio of 3.65. Replacement of CPro in 4 by Pro in 5, preprint version. You can request a copy of the preprint by results in almost exclusive population of the trans form, emailing the Biopolymers editorial office at biopolymers@wiley. resulting in an incipient b-hairpin conformation, com

INTRODUCTION he peptide bonds in polypeptide and struc- Additional Supporting Information may be found in the online version of this article. Correspondence to: Narayanaswamy Shamala; e-mail: [email protected] tures occur predominantly as the energetically more or Padmanabhan Balaram; e-mail: [email protected] favored trans conformer.1,2 However, a small but sig- Contract grant sponsor: Council of Scientific and Industrial Research, India Contract grant sponsor: Department of Biotechnology, India nificant number of peptide bonds occur in the cis VC 2009 Wiley Periodicals, Inc. T form and are of structural and functional impor- PeptideScience Volume 92 / Number 5 405 406 Kantharaju et al. tance.3–13 An interesting structural feature that is found in Homochiral (L,L) diproline segments and conformationally is the type-VIa b-turn, which possess a central cis constrained, covalently bridged templates have been shown amide linkage and facilitates chain reversal.14–23 The design to be inducers of helical folding in short peptides.48–54 Hetero- and synthesis of conformationally well-defined peptides as chiral (D,L) Pro-Pro sequences a have strong tendency to folding models and as analogs of biologically active peptides form type II0 b-turn structures,55,56 facilitating the nucleation requires strategies for the rational introduction of cis peptide of peptide b-hairpins, in which antiparallel strands are held bonds into polypeptide sequences. Mutter and coworkers together by registered cross-strand hydrogen bonds.57–59 have pioneered the use of pseudoproline (CPro) residues as Peptide hairpins containing a central three residue, DPro- a means of stabilizing cis peptide bonds in linear sequen- Pro-DAla, loop have also been characterized.57 In this article, ces.24–26 In the pseudoprolines (see Figure 1), the presence of we examine the effect of replacement of the LPro residue by d L geminal substituents at the C carbon atom of the five-mem- CPro (CPro, L-2,2-dimethyl-1,3-thiazolidine-4-carboxylic bered ring results in nonbonded interactions, which tilt the acid), in both the two residue (DPro-LPro) and three residue conformational equilibrium towards the cis conformer. In an (DPro-LPro-DAla) loop containing hairpins. We establish that elegant series of studies Mutter’s group has established the the CPro containing octapeptides Boc-Leu-Phe-Val-DPro- easy synthetic accessibility of a variety of pseudoprolines, CPro-Leu-Phe-Val-OMe (1), Boc-Leu-Val-Val-DPro-CPro- containing substituents at the Cd position and a heteroatom Leu-Val-Val-OMe (2), and the nonapeptide Boc-Leu-Phe- [O,S] at the c-position.27–31 Pseudoprolines have been incor- Val-DPro-CPro-DAla-Leu-Phe-Val-OMe (3) adopt b-hairpin porated into several model peptides and biologically active conformations with the DPro-CPro bond existing exclusively sequences including cyclosporin C,32,33 analogues of morphi- as the trans conformer. Studies with truncated sequences 34 D D ceptin and endomorphin-2, the gp120 V3 loop of HIV- Boc-Val- Pro-CPro-Leu-OMe (4) and Boc-Val- Pro-Pro- 1,35,36 the peptide ,37 and a d-conotoxin Leu-OMe (5) establish that folding into b-hairpins and the EVIA peptide analog.38 The incorporation of pseudoprolines resultant cross-strand hydrogen bond formation provides the into peptide backbone is also useful in limiting peptide energetic driving force for tilting the DPro-CPro peptide aggregation and facilitating chain extension in solid-phase bond towards the trans conformation. The effect of the con- .39–42 The Dumy and Mutter groups have figuration of the preceding Pro residue is established by com- also presented an extensive series of studies which establish paring the diastereomeric sequences Piv-DPro-CPro-Leu- the effect of Cd substitution on the cis-trans equilibrium.43–47 OMe (6) and Piv-Pro-CPro-Leu-OMe (7). Diproline templates have proved valuable in synthetic pep- tide design for the construction of both helices and hairpins. MATERIALS AND METHODS

Peptide Synthesis Peptides were synthesized by conventional solution-phase chemis- try, using a racemization free, fragment condensation strategy. The Boc group was used for N-terminal protection, and the C-terminus was protected as a methyl ester. Deprotection was performed by using 98% formic acid and by saponification for the N- and C-ter- minus, respectively. Couplings were mediated by N,N-dicyclohexyl- carbodiimide (DCC) and 1-hydroxy-1H-benzotriazole (HOBT). The intermediate Boc-DPro-Cys(CCH3,CH3Pro)-Leu-Phe-Val-OMe was obtained by the fragment condensation of Boc-DPro- CH3,CH3 Cys(C Pro)-OH and H2N-Leu-Phe-Val-OMe. The Boc-DPro-Cys(CCH3,CH3Pro)-OH was synthesized by the coupling D of Boc- Pro-OH and the L-2,2-dimethyl-1,3-thiazolidine-4-carb- oxylic acid (readily prepared by refluxing a mixture of L- hydrochloride monohydrate (7 g, 0.04 mol), 100 mL of 2,2-di- methoxypropane, and 500 mL of dry acetone for about 2 h. After cooling to room temperature, the resulting white platelets were col- lected by simple filtration), by using O-benzotriazole-N,N,N0,N0-tetra- methyluronium-hexafluorophosphate (HBTU)/HOBT as coupling agents and two equivalents of diisopropylethylamine. The final FIGURE 1 Trans and cis conformers of proline (top) and pseudo- step in the synthesis of the peptides was achieved by the frag- proline (bottom). ment condensation of Boc-Leu-Phe-Val-OH and HN-DPro-

Biopolymers (Peptide Science) Hetero and Homochiral Proline-Pseudoproline Segments in Peptides 407

CH3,CH3 Cys(C Pro)-Leu-Phe-Val-OMe. The Boc group was removed 5657 observed reflections with F0 >4r(F0) and 1366 parameters by using 1% TFA in dry dichloromethane for about 1 hr (detailed where the data-to-parameter ratio is 4.1:1.0 and S ¼ 1.063. The larg- procedures are provided as Supporting Information). All of the est difference peak was 0.59 e/A˚ 3 and the largest difference hole was intermediates were characterized by 1H NMR spectroscopy 0.31 e/A˚ 3. For peptide 7a, x scan type was used, with 2h ¼ 53.808, (500 MHz) and thin layer chromatography (TLC) on silica gel and for a total of 2493 unique reflections. The crystal size was 0.45 mm were used after purification using reverse-phase medium pressure 3 0.02 mm 3 0.04 mm. The space group was P31 with a ¼ b ¼ ˚ ˚ ˚ 3 liquid chromatography (C18, 40–60 lm). The final peptides were 14.6323(22) A, c ¼ 10.4359(32) A, V ¼ 1935.0(7) A , Z ¼ 3 for 3 purified by HPLC on a reverse phase C18 column (5–10 lm, 7.8– chemical formula C22H38N4O4S1, qcalc ¼ 1.17 g cm , l ¼ 0.16 250 mm) by using methanol–water gradients. Peptides 1–7 were mm1, and F(000) ¼ 738.0. The structure was solved by direct characterized by electrospray ionization mass spectrometry (ESI-MS) methods using SHELXD.60 This gave a fragment containing 28 on a Bruker Daltonics Esquire-3000 instrument and by complete atoms. The remaining atoms were located from difference Fourier assignment of the 700 MHz or 500 MHz 1H NMR spectra. Mass maps. Full-matrix least-squares refinement was carried out using + 61 spectral data (m/z): Peptide 1, 1091.6 [M+H] (Mcal ¼ 1090 Da); SHELXL-97. The hydrogen atoms were fixed geometrically in the + + + 1113.6 [M+Na] ; 1129.6 [M+K] ; 2, 995.6 [M+H] (Mcal ¼ idealized positions and refined in the final cycle of refinement as ri- 994.2 Da); 1017.6 [M+Na]+; 1033.5 [M+K]+; 3, 1162.6 [M+H]+ ding over the atoms to which they are bonded. The final R value + + (Mcal ¼ 1161.2 Da); 1184.5 [M+Na] ; 1200.6 [M+K] ; 4, 585.2 was 0.0587 (wR2 ¼ 0.159) for 1353 observed reflections with F0 > + + + [M+H] (Mcal ¼ 584.2 Da); 607.2 [M+Na] ; 623.2 [M+K] ; 5, 4r(F0) and 280 parameters where the data-to-parameter ratio is + + + ˚ 3 539.5 [M+H] (Mcal ¼ 538.8 Da); 561.4 [M+Na] ; 577.5 [M+K] ; 4.8:1.0 and S ¼ 0.899. The largest difference peak was 0.17 e/A and + + ˚ 3 6 and 7, 470.2 [M+H] (Mcal ¼ 469.1 Da); 492.2 [M+Na] . the largest difference hole was 0.16 e/A . The crystallographic coordinates for the structures are deposited at the Cambridge Crys- tallographic Data Centre with deposition numbers CCDC 714997 NMR Spectroscopy (1) and 714998 (7a). These data can be obtained free of charge via Experiments were carried out on a Bruker AV700MHz and DRX500 www.ccdc.cam.ac.uk/conts/retrieving.html (or from the Cambridge spectrometers. 1D and 2D spectra were recorded at a peptide con- Crystallographic Data Centre, 12 Union Road, Cambridge CB21EZ, centration of *5mM in CDCl3 at 300K. Delineation of exposed UK; fax: (+44) 1223-336-033; or e-mail: [email protected]). NH groups was achieved by titrating CDCl3 solutions with low con- centrations of DMSO-d6. Resonance assignments were carried out with the help of 1D and 2D spectra. Residue specific assignments RESULTS AND DISCUSSION were obtained from TOCSY experiments, whereas ROESY spectra permitted sequence specific assignments. All 2D experiments were recorded in phase sensitive mode using the TPPI (time proportional Conformation of Peptides 1–3 D phase incrementation) method. A data set of 1024 3 450 was used The octapeptides Boc-Leu-Phe-Val- Pro-CPro-Leu-Phe-Val- for acquiring the data. The same data set was zero filled to yield a OMe 1 and Boc-Leu-Val-Val-DPro-CPro-Leu-Val-Val-OMe 2 data matrix of size 2048 3 1024 before Fourier transformation. A and nonapeptide Boc-Leu-Phe-Val-DPro-CPro-DAla-Leu- spectral width of 6000 and 8700 Hz was used in both dimensions at Phe-Val-OMe 3 exhibited well dispersed 1H NMR spectra in 500 and 700 MHz, respectively. Mixing times of 100 and 200 ms CDCl . In all the three cases there was no evidence for the were used for TOCSY and ROESY, respectively. Shifted square sine 3 bell windows were used while processing. All processing was done slowly exchanging minor conformer, suggesting that the D using BRUKER TOPSPIN software. Pro-CPro peptide bond adopted a single conformation. Sequence specific resonance assignments could be readily established using a combination of TOCSY and ROESY spec- X-Ray Diffraction tra. Chemical shifts and relevant NMR parameters are sum- The crystals of peptides 1 and 7a were grown from a methanol/ water solution by slow evaporation. X-ray diffraction data were col- marized in Table I. Figure 2 shows the partial ROESY spec- lected on a Bruker AXS SMART APEX CCD diffractometer using trum, highlighting key NOEs between backbone protons of MoKa radiation (k ¼ 0.71073 A˚ ) for both the peptides. For peptide peptide 1. Inspection of the amide NH resonances establishes 1, x scan type was used, with 2h ¼ 46.518, for a total of 9859 unique the wide dispersion of NH chemical shifts, with the Val(3), reflections. The crystal size was 0.45 mm 3 0.15 mm 3 0.01 mm. Leu(6), and Val(8) NH resonances appearing at the lowest The space group was C2 with a ¼ 34.646(20) A˚ , b ¼ 15.337(22) A˚ , field. This is a characteristic of octapeptide hairpins which c ¼ 25.553(53) A˚ , b ¼ 103.3878, V ¼ 13207.77 A˚ 3, Z ¼ 8 for chemi- 3 contain the Leu-Phe-Val segment as the N- and C-terminal cal formula C57H86N8O11S1 1.5H2O, qcalc ¼ 1.1 g cm , l ¼ 57,62,63 0.11 mm1, and F(000) ¼ 4800.0. The structure was solved by direct b-strand units. In registered b-hairpins the Leu (1), methods using SHELXD.60 Two fragments were obtained containing Val (3), Leu (6), and Val (8) NH groups are involved in inter- 66 and 58 atoms. The remaining atoms and solvent molecules were strand hydrogen bonding as shown schematically in Figure 3. located from difference Fourier maps. Full-matrix least-squares Delineation of internally hydrogen bonded NH groups was refinement was carried out using SHELXL-97.61 The hydrogen atoms were fixed geometrically in the idealized positions and refined carried out using solvent titration experiments in CDCl3- in the final cycle of refinement as riding over the atoms to which DMSO-d6 mixtures (see Figure 4). Addition of the hydrogen they are bonded. The final R value was 0.115 (wR2 ¼ 0.328) for bonding solvent DMSO results in large downfield shifts of

Biopolymers (Peptide Science) 408 Kantharaju et al.

Table I NMR Parameters for Peptides 1 to 3 in CDCl3

Peptide 1 Chemical Shift [ppm] Peptide 2 Chemical Shift [ppm] Peptide 3 Chemical Shift [ppm]

a 3 a a a 3 a b a 3 a b Residue NH C H JNHC H Dd NH C H JNHC H Dd NH C H JNHC H Dd

Leu 1 5.12 4.21 8.60 0.68 5.61 4.09 8.44 0.59 5.12 4.10 8.34 0.56 Phe 2/Val 2 6.69 5.42 6.58 0.88 6.46 4.81 9.31 0.92 6.47 5.44 9.33 1.25 Val 3 8.74 4.52 9.72 0.06 8.66 4.47 9.57 0.04 8.73 5.06 9.36 0.06 DPro 4 — 4.72 — — — 4.67 — — — 4.62 — — CPro 5 — 5.43 — — — 5.40 — — — 5.41 — — Leu 6/DAla 6 7.86 4.46 8.25 0.08 7.84 4.40 8.72 0.05 7.32 4.76 9.32 0.12 Phe 7/Val 7/Leu 7 6.53 4.68 8.24 1.19 6.37 4.72 9.31 0.90 7.55 4.23 8.00 0.11 Val 8/Phe 8 7.67 4.25 9.03 0.10 8.41 4.53 8.74 0.04 6.57 4.98 8.39 1.51 Val 9 7.60 4.26 9.09 0.14

a Dd ¼ d(26.83% DMSO-d6/CDCl3) d(CDCl3). b Dd ¼ d(20.64% DMSO-d6/CDCl3) d(CDCl3). solvent exposed NH groups, while those involved in intramo- chains face one another in the nonhydrogen bonding posi- lecular interactions are perturbed to a much smaller extent. tion of the hairpin. In such an orientation, ring current The solvent shift values (Dd, ppm) are extremely small for effects due to proximal aromatic rings perturb the chemical the Val (3), Leu (6), and Val (8) NH protons (Table I). The shift position. This has been established for Phe/Phe N-terminus Leu (1) NH exhibits a somewhat larger Dd value pairs62,63 and Tyr/Trp pairs.64,65 The b-hairpin conformation as a consequence of fraying of the ends of the b-hairpin. The of peptide 1 is supported by the observation of key cross- wide dispersion of the aromatic proton resonances and the upfield position of the resonances at 6.7d is a characteristic feature of b-hairpins, in which Phe (2) and Phe (7) side

FIGURE 3 Schematic representation of the proposed b-hairpin FIGURE 2 Partial 700 MHz ROESY spectrum of peptide 1 in structure of peptide 1. The expected hydrogen bonds are shown by a CDCl3 highlighting C H$NH and NH$NH region NOEs. Key broken lines. Observed NOEs are highlighted by double edged cross peaks are boxed. arrows.

Biopolymers (Peptide Science) Hetero and Homochiral Proline-Pseudoproline Segments in Peptides 409

Peptide 2 in which Phe (2) and Phe (7) are replaced by the aliphatic Val residue was designed to eliminate cross- strand aromatic interactions. The observed NMR chemical shifts of the backbone CaH$NH protons and the Dd values (Table I) strongly suggests that peptide 2 adopts a conforma- tion very similar to peptide 1. The occurrence of the b-hair- pin conformation in peptide 2 is further confirmed by observing all the cross-strand NOEs as noted above for pep- tide 1 (Supporting Information Figures S1 and S2). The chemical shift pattern (Table I) and NOEs observed for the nonapeptide 3 are very similar to that observed previ- ously for the corresponding Pro analog,57 supporting a b- hairpin conformation with a central three residue DPro- LPro-DAla loop. Solvent titration experiments establish that the Val (3), DAla (6), Leu (7), and Val (9) NH protons are solvent shielded (Supporting Information Figure S3), as evi- denced by extremely low Dd values. Three of these NH groups, Val (3), Leu (7), and Val (9), are involved in cross- strand hydrogen bonds, whereas DAla (6) participates in the hydrogen bond stabilized three residue turn. The Leu (1) NH resonance exhibits an intermediate Dd value consistent with FIGURE 4 Solvent dependence of NH chemical shifts in peptide a moderate degree of fraying of the b-hairpin at the termini. D C 1 Boc-Leu-Phe-Val- Pro- Pro-Leu-Phe-Val-OMe in CDCl3-DMSO Figure 7 illustrates the NOE evidence which establishes both mixtures. the trans geometry of the DPro-CPro bond and registry of strand NOEs: dNN, Leu (1)$Val (8), Val (3)$Leu (6); daN, the antiparallel strands. The former is established by the ob- Phe (2)$Val (8) and daa, Phe (2)$Phe (7) (Figures 2 and servation of the strong NOEs between the methyl groups of 5). The trans geometry of the DPro-CPro peptide bond is confirmed by the strong NOEs observed between DPro CaH and the geminal methyl groups on the thiazolidine ring (see Figure 5). The b-hairpin conformation of peptide 1 with the trans geometry for the DPro-CPro amide bond has also been confirmed in the solid state by single crystal X-ray diffraction (Figure 6A). The two molecules observed in the crystallo- graphic asymmetric unit adopt very similar hairpin confor- mations for the peptide backbone. The backbone torsion angles are listed in Table II. Figure 6B shows a view of the orientation of the aromatic rings that occur in the facing positions across the antiparallel strands. The short distance between the centroids of the aro- matic rings in both the independent molecules (5.11–5.39 A˚ ) is suggestive of a significant aromatic–aromatic interaction. When peptide 1 was designed there was the expectation that introduction of the CPro residue might favor the cis ge- ometry for the DPro-CPro bond, resulting in the perturba- tion of the parent b-hairpin conformation. The observation of the trans DPro-CPro geometry, with retention of the b- hairpin conformation, suggests that the forces stabilizing folding, like cross-strand hydrogen bonding and aromatic FIGURE 5 Partial 700 MHz ROESY spectrum of peptide 1 illus- interactions, may have overwhelmed the intrinsic tendency a aCH3 a a trating C H$C and C H$C H NOEs observed in CDCl3.Key of CPro to favor a cis geometry. cross peaks are boxed.

Biopolymers (Peptide Science) 410 Kantharaju et al.

FIGURE 6 (A) Molecular conformation of peptide 1 (Boc-Leu-Phe-Val-DPro-CPro-Leu-Phe- Val-OMe) in crystals. Two molecules in the asymmetric unit are indicated as Molecule-A and Mole- cule-B. (B) The schematic view of Phe-Phe interaction in peptide 1. the thiazolidine ring of CPro and DPro CaH protons (4a/ Conformations of Peptides 4 and 5 0 5d,d ). The latter is supported by the strong NOE between the Figure 8 compares the peptide NH resonances observed in Phe (2) CaH and Phe (8) CaH protons (2a/8a) (Supporting the CPro peptide 4 (Boc-Val-DPro-CPro-Leu-OMe) and its Information Figure S4). Thus, insertion of CPro at the posi- Pro analog 5 (Boc-Val-DPro-LPro-Leu-OMe). Although pre- tion of LPro in both two and three residue loops results in dominantly only one set of NH resonances is observed in retention of the parent b-hairpin conformation, with a trans peptide 5, two sets of resonances are seen in the CPro con- peptide geometry for the DPro-CPro segment. To assess the taining sequence 4. Figure 9 shows the partial ROESY spec- energetic contributions of cross-strand hydrogen bonds in trum of peptide 4. Inspection of the NOEs in 4 reveals the determining the geometry of the DPro-CPro bond, we strong DPro (2) CaH$CPro (3) CaH NOE for the major turned to the truncated peptide analogs 4 and 5. species establishing its identity as the cis conformer. The

Biopolymers (Peptide Science) Hetero and Homochiral Proline-Pseudoproline Segments in Peptides 411

D a L d Table II Torsion Angles (deg) for Peptide 1 evidenced by the strong Pro (2) C Hto Pro (3) C H2 NOEs (Figure 10). Solvent titration experiments establish Residue Angle Molecule A Molecule B very low Dd values for Val (1) NH (0.07 ppm) and Leu (4) Backbone NH protons (0.01 ppm) in the minor, trans, conformer, sug- Leu (1) / 133.7 114.2 gestive of their involvement in intramolecular hydrogen w 131.1 141.4 bonds. The corresponding Dd values for peptide 5 are Val (1) x 175.3 178.7 NH 0.09 ppm and Leu (4) NH 0.04 ppm (Supporting In- Phe (2) / 126.1 126.7 formation Table SI). Thus, the trans peptide geometry for the w 146.5 138.8 central DPro-LPro/CPro segment facilitates the formation of x 166.1 167.0 Val (3) / 111.6 112.0 an incipient b-hairpin structure containing two cross-strand w 92.5 103.4 hydrogen bonds. Further evidence for this conformer is x 166.2 175.2 obtained by the observation of the NOEs between the Val (1) D Pro (4) / 53.0 62.8 NH and Leu (4) NH protons in peptide 5 (Supporting Infor- w 135.8 134.2 mation Figure S5) and in the minor, trans, conformation of x 177.1 177.7 CPro (5) / 89.3 84.1 peptide 4 (Supporting Information Figure S6). H-D w 22.1 1.9 exchange experiments were carried out at 288 K in CDCl3 x 173.5 173.2 solution by adding approximately 20 % v/v CD3OD. Under Leu (6) / 94.8 93.2 these conditions, no loss of NH resonance intensity was w 120.9 126.1 observed for peptide 5 even after several hours (Supporting x 178.5 171.3 Information Figure S7). In the case of peptide 4, the Val (1) Phe (7) / 86.6 82.6 w 124.1 125.4 NH of the major cis conformer showed significant exchange. x 173.7 167.5 The other three NH resonances, trans Val (1) NH and Leu Val (8) / 126.2 136.5 (4) and cis Leu (4) NH, did not show evidence for significant w 1.0 21.1 exchange (Supporting Information Figure S8). These results x 179.0 141.6 are consistent with the two hydrogen bonded, incipient Side chains Leu (1) v1 171.9 60.5 hairpin in the trans conformer of peptide 4.Incontrast, v2 67.7, 166.7 67.8, 160.7 both resonances in the cis conformer of 4 exhibited signifi- Phe (2) v1 73.6 74.9 cant downfield shift with increase in concentration, indic- 2 v 75.9, 102.8 76.2, 101.9 ative of the absence of a significant population of intra- 1 Val (3) v 55.2, 178.0 67.2, 173.6 molecularly hydrogen bonded structures (Supporting DPro (4) v1 20.5 20.4 v2 28.8 34.7 Information Figure S9). In peptide 5 both Val (1) and Leu v3 24.2 34.5 (4) NH resonances were largely unaffected by addition of 4 v 10.6 15.2 DMSO-d6 to CDCl3 solutions (Supporting Information h 5.6 1.8 Figure S10), consistent with the two hydrogen bonded 1 CPro (5) v 39.4 42.5 hairpin conformation. v2 38.1 35.4 A comparison of the results obtained for the CPro con- v3 25.3 20.2 v4 5.6 3.0 taining octapeptides 1 and 2 and tetrapeptide 4 suggests that h 22.1 29.8 truncation of the strand segments results in a dramatic shift Leu (6) v1 172.0 176.2 in the equilibrium population of the cis and trans conformer 2 v 68.9, 167.5 118.5, 118.1 about the DPro-CPro bond. When the two additional cross- 1 Phe (7) v 177.4 178.9 strand hydrogen bonds are present the trans conformer is v2 46.2, 136.4 59.5, 120.1 Val (8) v1 58.3, 69.8 61.3, 151.3 exclusively favored. In the truncated peptide 4, the equilibrium favors the cis form over the trans form by a factor of 3.65, corresponding to a DG * 5.56 kJ mol1 (1 kcal mol1 ¼ trans peptide bond geometry in the minor species is con- 4.12 kJ mol1). firmed by the observation of the strong NOE between DPro In the sequences described above, we have examined the (2) CaH and the gem dimethyl group on the thiazolidine DPro-LCPro segment. To evaluate the role of the configura- ring. Integration of the resonances establishes the popula- tion of the preceding residue on CPro, we turned to the tions as 78.5 % cis and 21.5 % trans in peptide 4. In the pro- model sequences Piv-DPro-CPro-Leu-OMe 6 (D,L) and Piv- line analog 5, the trans form is overwhelmingly populated as Pro-CPro-Leu-OMe 7 (L, L).

Biopolymers (Peptide Science) 412 Kantharaju et al.

FIGURE 7 Partial 700 MHz ROESY spectrum of peptide 3 Boc-Leu-Phe-Val-DPro-CPro-DAla- a dCH3 a a Leu-Phe-Val-OMe in CDCl3 highlighting C H$C and C H$C H, NOEs.

FIGURE 8 NH proton resonances in peptides 4 (A) and 5 (B) in CDCl3 (500 MHz).

Biopolymers (Peptide Science) Hetero and Homochiral Proline-Pseudoproline Segments in Peptides 413

FIGURE 9 Partial 700 MHz ROESY spectrum of peptide 4 Boc- D a dCH3 a a FIGURE 10 Partial 500 MHz ROESY spectrum of peptide 5 Boc- Val- Pro-CPro-Leu-OMe illustrating C H$C and C H$C H Val-DPro-LPro-Leu-OMe, illustrating CaH$CdH2 NOEs observed NOEs observed in CDCl3 at 288 K. Key NOEs are boxed. in CDCl3 at 300 K. Key NOEs are boxed.

FIGURE 11 Partial 500 MHz NMR spectra (CDCl3 at 300 K) of peptides 6 (bottom) and 7 (top) showing NH resonances. Partial ROESY spectrum of peptide 7 (inset) highlighting key NOEs observed in the cis conformer {Pro1 (CaH)$CPro2 (CaH) and Pro1 (CaH)$Leu3 (NH)} is also shown.

Biopolymers (Peptide Science) 414 Kantharaju et al.

NMR Analysis of Peptides 6 and 7 177.0. The x value for the amide bond between the N-ter- Peptide 6 which contains the heterochiral DPro-CPro unit minus pivaloyl blocking group and Pro (1) is 1808. These val- reveals two sets of resonances for the backbone CaH and NH ues deviate significantly from those anticipated in the ideal protons suggesting that both cis and trans conformers are type VIa b-turn (60.0, 120.0, 90.0, 0.0).5 The observed pa- populated (Figure 11, bottom). The major conformation cor- rameters for the 4?1 hydrogen bond between Piv responds to the cis form as evidenced by a strong NOE C¼¼O HN Leu (3) are: N O ¼ 2.839 A˚ ,HO ¼ 1.997 between the DPro CaH and CPro CaH protons. The assign- A˚ , ffNH O ¼ 166.258, ffC¼¼O N ¼157.448 and ment of the minor species to the trans form is confirmed by ffC¼¼O H ¼ 154.68. The Pro (1) CaH-Leu (3) NH distance the observation of a NOE between the DPro CaH and Cd in the crystal state conformation is 2.842 A˚ . The observed methyl substituents on the CPro ring (Supporting Informa- conformation in crystals of peptide 7a is thus completely con- tion Figure S11). The population ratio (cis/trans) of 3.81 cor- sistent with the intramolecularly hydrogen bonded conforma- responds to a free energy difference (DG)of* 5.97 kJ tion anticipated from the NMR data for peptide 7. mol1. In contrast, peptide 7 which contains the homochiral In previously published work by Mutter and coworkers, Pro-CPro segment shows only one set of NMR resonances, the propensity of the pseudoproline residue to facilitate corresponding to the cis conformer, as evidenced by a strong introduction of cis peptide bonds into polypeptide back- NOE between the LPro CaH and CPro CaH protons (Figure bones was established. The energetic differences between the 11, top). A notable feature of peptide 7 is the extremely low cis and trans forms of the Xxx-CPro peptide linkage are large field position of the Leu (3) NH proton (9.2 ppm), suggestive enough to favor the cis conformation in the absence of any of its involvement in a strong hydrogen bond. The observa- compensating interactions. This energy difference can be tion of the strong NOE between the Pro (1) CaH and Leu (3) compensated in large peptides by medium and long range NH protons confirms their proximity, which is anticipated in interactions which provide additional stabilizing factors for atypeVIab-turn conformation. The crystal structure deter- specific folded conformations. In the present study, we have mination (Figure 12) of an analog peptide 7a, which differs provided evidence for a trans peptide bond in the hetero- from 7 in possessing only one methyl group at the Cd position chiral DPro-CPro segment, in the context of well defined of the thiazolidine ring (Piv-Pro-CH,CH3Pro-Leu-NHMe), octapeptide b-hairpins, which are stabilized by three cross-strand permitted characterization of the type VIa b-turn conforma- hydrogen bonds. Truncation of the strand residues and tion for this homochiral Pro-CPro segment. In peptide 7a reduction of the number of stabilizing hydrogen bonds the configuration of the Ca carbon atom in both the Pro and results in an increase in the population of the cis conformers CPro residues is L (R). The asymmetric centre at the Cd posi- in tetrapeptide sequences. In the tripeptide Piv-DPro-CPro- tion of CPro has the R configuration. The observed backbone Leu-OMe there is a single hydrogen bond formed in a type torsion angles, /, w, x (8) are: Pro (1), 67.2, 141.7, 1.5; II0 b-turn conformation, with a central trans peptide bond. CH, CH3Pro (2), 109.3, 50.5, 176.9; Leu (3), 68.6, 17.7, The population of the trans form is only 20.8% and the cis form is predominant, 79.2%. In the homochiral sequence Piv-Pro-CPro-Leu-OMe only the cis form is observed and the type VIa b-turn structure, stabilized by an intramolecular 4?1 hydrogen bond, is strongly favored.

CONCLUSIONS Pseudoproline residue (CPro, L-2,2-dimethyl-1,3-thiazol- idine-4-carboxylic acid) has been introduced as a residue that stabilizes the cis conformation of the preceding amide bond. This study with Pro-CProsegmentsasananalogfordi- proline units leads to the following conclusions. In the homo- chiral sequence, Piv-Pro-CPro-Leu-OMe, only the cis form is observed and the type VIa b-turn structure, stabilized by an intramolecular 4?1 hydrogen bond, is strongly favored. In homochiral sequences, both internal hydrogen bond forma- FIGURE 12 Molecular conformation of peptide 7a, Piv-Pro-CH,CH3 tion and the intrinsic tendency of the CPro residue to favor Pro-Leu-NHMe, in crystals. cis conformations operate in tandem, facilitating type VIa

Biopolymers (Peptide Science) Hetero and Homochiral Proline-Pseudoproline Segments in Peptides 415 b-turn structures. In heterochiral sequences internal hydro- 24. Keller, M.; Sager, C.; Dumy, P.; Schutkowski, M.; Fischer, G. S.; gen bond formation is possible only within the context of a Mutter, M. J Am Chem Soc 1998, 120, 2714–2720. type II0 b-turn structure, which constrains the Xxx-CPro 25. Keller, M.; Lehmann, C.; Mutter, M. Tetrahedron 1999, 55, 413– 422. bond to adopt a trans conformation. Pseudoproline residues 26. Wo¨hr, T.; Wahl, F.; Nefzi, A.; Rohwedder, B.; Saxo, T.; Sun, X.; are valuable tools for design strategies, which require the Mutter, M. J Am Chem Soc 1996, 118, 9218–9227; and referen- rational introduction of cis peptide bonds into polypeptide ces cited therein. backbones. Contextual effects must be taken into account in 27. Wo¨hr, T.; Mutter, M. Tetrahedron Lett 1995, 36, 3847–3848. assessing the relative preferences for cis and trans geometries 28. Savrda, J. In 14th Proceedings of European Peptide Symposium, Wepion, Belgium, edited by Albert Loffet. April 11–17, 1976, at amide bonds preceding pseudoproline residues. p 653. 29. Nefzi, A.; Schenk, K.; Mutter, M. Protein and Peptide Lett 1994, Kantharaju is supported by the award of a DBT Postdoctoral Fel- 1, 66–69. lowship from the Department of Biotechnology (DBT), India. U.S. 30. Lewis, N. J.; Inloes, R. L.; Hes, J. J Med Chem 1978, 21, 1070– Raghavender and S. Aravinda thank the University Grants Commis- 1073. sion and Council of Scientific and Industrial Research, India, for 31. Guichou, J-F.; Patiny, L.; Mutter, M. Tetrahedron Lett 2002, 43, Senior Research Fellowship and Research Associateship, respectively. 4389–4390. X-ray diffraction data were collected at the CCD facility funded 32. Patiny, L.; Guichou, J-F.; Keller, M.; Turpin, O.; Ruckle, T.; under the IRHPA program of the Department of Science and Tech- Lhote, P.; Buetler, T. M.; Ruegg, U. T.; Wengera, R.; Mutter, M. nology, India. Tetrahedron 2003, 59, 5241–5249. 33. Keller, M.; Wo¨hr, T.; Dumy, P.; Patiny, L.; Mutter, M. Chem Eur J 2000, 6, 4358–4363. REFERENCES 34. Keller, M.; Boissard, C.; Patiny, L.; Chung, N. N.; Lemieux, C.; 1. Ramachandran, G. N.; Sasisekharan, V. Adv Protein Chem Mutter, M.; Schiller, P. W. J Med Chem 2001, 44, 3896–3903. 1968, 23, 283–438. 35. Wittelsberger, A.; Keller, M.; Scarpellino, L.; Patiny, L.; Acha- 2. Dugave, C.; Demange, L. Chem Rev 2003, 103, 2475–2532. Orbea, H.; Mutter, M. Angew Chem Int Ed 2000, 39, 1111–1115. 3. Ramachandran, G. N.; Mitra, A. K. J Mol Biol 1976, 107, 36. Liu, C.-F.; Rao, C.; Tam, J. P. J Am Chem Soc 1996, 118, 307–312. 85–92. 37. Wittelsberger, A.; Patiny, L.; Slaninova, J.; Barberis, C.; Mutter, 4. Wilmot, C. M.; Thornton, J. M. J Mol Biol 1988, 203, 221–232. M. J Med Chem 2005, 48, 6553–6562. 5. Rose, G. D.; Gierasch, L. M.; Smith, J. A. Adv Protein Chem 38. Chierici, S.; Jourdan, M.; Figuet, M.; Dumy, P. Org Biomol 1985, 37, 1–109. Chem 2004, 2, 2437–2441. 6. Brandts, J. F.; Halvorson, H. R.; Brennan, M. Biochemistry 39. Mutter, M.; Nefzi, A.; Sato, T.; Sun, X.; Wahl, F.; Wo¨hr, T. Pep- 1975, 14, 4953–4963. tide Res 1995, 8, 145–153. 7. MacArthur, M. W.; Thornton, J. M. J Mol Biol 1991, 218, 397–412. 40. Haack, T.; Mutter, M. Tetrahedron Lett 1992, 33, 1589–1592. 8. Mu¨ller, G.; Gurrath, M.; Kurz, M.; Kessler, H. Proteins: Struct 41. White, P.; Keyte, J.W.; Bailey, K.; Bloomberg, G. J Peptide Sci Funct Genet 1993, 15, 235–251. 2004, 10, 18–26. 9. Stewart, D. E.; Sarkar, A.; Wampler, J. E. J Mol Biol 1990, 214, 42. Sampson, W. R.; Heather, P.; Nicholas, E. J. J Pept Sci 1999, 5, 253–260. 403–409. 10. Milner-White, E. J.; Bell, L. H.; Maccallum, P. H. J Mol Biol 43. Thomas, R.; Pierre, De. L.; Keller, M.; Dumy, P.; Mutter, M. Tet- 1992, 228, 725–734. rahedron 1999, 55, 11281–11288. 11. Pal, D.; Chakrabarti, P. J Mol Biol 1999, 294, 271–288. 44. Mutter, M.; Wo¨hr, T.; Gioria, S.; Keller, M. Biopolymers 1999, 12. Wathen, B.; Jia, Z. J Proteome Res 2008, 7, 145–153. 51, 121–128. 13. Dugave, C. Ed. Cis-Trans Isomerization in Biochemistry; Wein- 45. Chierici, S.; Figuet, M.; Dettori, P.; Dumy, P. Chimia 2005, 8, heim: Wiley-VCH, 2006. 875–880. 14. Fischer, G. Angew Chem Int Ed Engl 1994, 33, 1415–1436. 46. Dumy, P.; Keller, M.; Ryan, D. E.; Rohwedder, B.; Wo¨hr, T.; Mut- 15. Liu, J.; Chen, C.-M.; Walsh, C. T. Biochemistry 1991, 30, 2306– ter, M. J Am Chem Soc 1997, 119, 918–925. 2310. 47. Jamet, H.; Jourdan, M.; Dumy, P. J Phys Chem B 2008, 112, 16. Fischer, S.; Michnick, S.; Karplus, M. Biochemistry 1993, 32, 9975–9981. 13830–13837. 48. Venkatachalapti, Y. V.; Balaram, P. Nature 1979, 281, 83–84. 17. Kallen, J.; Walkinshaw, M. D. FEBS Lett 1992, 300, 286–290. 49. Rai, R.; Aravinda, S.; Kanagarajadurai, K.; Raghothama, S.; Sha- 18. Johnson, M. E.; Lin, Z.; Padmanabhan, K.; Tulinsky, A.; Kahn, mala, N.; Balaram, P. J Am Chem Soc 2006, 128, 7916–7928. M. FEBS Lett 1994, 337, 4–8. 50. Kemp, D. S.; Boyd, J. G.; Muendel, C. C. Nature 1991, 352, 451– 19. Smith, J. A.; Pease, L. G. CRC Crit Rev Biochem 1980, 8, 315–399. 454. 20. Richardson, J. S. Adv Protein Chem 1981, 34, 167–339. 51. Kemp, D. S.; Curran, T. P.; Boyd, J. G.; Allen, T. J. J Org Chem 21. Meng, H. Y.; Thomas, K. M.; Lee, A. E.; Zondlo, N. J. Biopoly- 1991, 56, 6683–6697. mers (Peptide Sci) 2006, 84, 192–204. 52. Kemp, D. S.; Curran, T. P.; Davis, W. M.; Boyd, J. G.; Muendel, 22. Yao, J.; Dyson, H. J.; Wright, P. E. J Mol Biol 1994, 243, 754–766. C. J Org Chem 1991, 56, 6672–6682. 23. Yao, J.; Feher, V. A.; Espejo, B. F.; Reymond, M. T.; Wright, P. E.; 53. Job, G. E.; Heitmann, B.; Kennedy, R. J.; Walker, S. M.; Kemp, Dyson, H. J. J Mol Biol 1994, 243, 736–753. D. S. Angew Chem Int Ed 2004, 43, 5649–5651.

Biopolymers (Peptide Science) 416 Kantharaju et al.

54. Heitmann, B.; Job, G. E.; Kennedy, R. J.; Walker, S. M.; Kemp, 60. Schneider, T. R.; Sheldrick, G. M. Acta Crystallogr D 2002, 58, D. S. J Am Chem Soc 2005, 127, 1690–1704. 1772–1779. 55. Nair, C. M.; Vijayan, M.; Venkatachalapathi, Y. V.; Balaram, P. J 61. Sheldrick, G. M. SHELXL-97, Program for the Refinement of Chem Soc Chem Commun 1979, 1183–1184. Crystal Structures. Germany: Universitat Gottingen, 1997. 56. Chatterjee, B.; Saha, I.; Raghothama, S.; Aravinda, S.; Rai, R.; 62. Das, C.; Naganagowda, G. A.; Karle, I. L.; Balaram, P. Biopoly- Shamala, N.; Balaram, P. Chem Eur J 2008, 14, 6192–6204. mers 2001, 58, 335–346. 57. Rai, R.; Raghothama, S.; Balaram, P. J Am Chem Soc. 2006, 128, 63. Aravinda, S.; Shamala, N.; Rajkishore, R.; Gopi, H. N.; Balaram, 2675–2681. P. Angew Chem Int Ed 2002, 41, 3863–3865. 58. Robinson, J. A. Acc Chem Res 2008, 41, 1278–1288. 64. Mahalakshmi, R.; Raghothama, S.; Balaram, P. J Am Chem Soc 59. Shankaramma, S. C.; Athanassiou, Z.; Zerbe, O.; Moehle, K.; 2006, 128, 1125–1138. Mouton, C.; Bernardini, F.; Vrijbloed, J. W.; Obrecht, D.; Robin- 65. Zhao, C.; Polavarapu, P. L.; Das, C.; Balaram, P. J Am Chem Soc son, J. A. Chembiochem 2002, 3, 1126–1133. 2000, 122, 8228–8231.

Biopolymers (Peptide Science)