Competition between Amide Stacking and Intramolecular H Bonds in Î³

ARTICLE pubs.acs.org/JPCA

Competition between Amide Stacking and Intramolecular H Bonds in γ-Peptide Derivatives: Controlling Nearest-Neighbor Preferences William H. James, III,†,‡ Evan G. Buchanan,† Li Guo,§,|| Samuel H. Gellman,§ and Timothy S. Zwier*,† †

Department of Chemistry, Purdue University, West Lafayette, Indiana 47907-2084, United States Department of Chemistry, University of Wisconsin, Madison, Wisconsin 53706, United States

§

bS Supporting Information ABSTRACT: Resonant two-photon ionization (R2PI), IR-UV holeburning (IR-UV), and resonant ion-dip infrared spectroscopy (RIDIRS) have been used to record mass-selected, singleconformation ultraviolet and infrared spectra of three simple diamide derivatives of γ-amino acids as isolated molecules cooled in a supersonic expansion. This work builds on an earlier study of Ac-γ2-hPhe-NHMe (James, W. H., III, et al. J. Am. Chem. Soc. 2009, 131, 14243), which showed that this methylcapped γ-peptide forms amide-stacked conformations that are similar in stability to H-bonded conformations containing a C9 ring and more stable than C7 H-bonded ring structures. Among the γ-peptides discussed here, Ac-γ2-hPhe-N(Me)2 contains an additional methyl group relative to the previously studied Ac-γ2-hPhe-NHMe and therefore lacks the amide NH group responsible for C9 ring formation. Three conformations of Ac-γ2-hPhe-N(Me)2 are observed, all of which are amide-stacked structures. In a second new molecule, Ac-γ2-hPhe-NH(iPr), the C-terminal NHMe group of Ac-γ2-hPhe-NHMe is replaced with an NH(iPr) group. Three conformations of Ac-γ2-hPhe-NH(iPr) are observed, all of which are C9 H-bonded structures. The dramatic difference between C-terminal NHMe and NH(iPr) reveals the delicate balance of noncovalent forces within these γ-peptides. The third molecule we examined is a gabapentin-derived diamide (designated 1), which contains a phenylacyl group at the N-terminus and an N(Me)2 group at the C-terminus; the latter precludes C9 H bonding. Comparison of 1 with Ac-γ2-hPhe-N(Me)2 allows us to examine the impact of the backbone substitution pattern (monosubstitution at carbon-2 vs disubstitution at carbon-3) on the competition between the C7 H-bonded and the amide-stacked conformation. In this case, only C7 rings are observed. The different gas-phase behaviors observed among the molecules analyzed here offer insight on the intrinsic conformational propensities of the γpeptide backbone, information that provides a foundation for future foldamer design efforts.

I. INTRODUCTION Synthetic foldamers are oligomers that display discrete conformational propensities. Molecules of this type can mimic the shapes and functions of natural folded oligomers (proteins and nucleic acids) or display new shapes and functions. An almost infinite range of synthetic oligomeric backbones can be imagined, but conformational behavior must be elucidated in each case. Among the most widely studied foldamer systems to date are oligomers that contain β- and/or γ-amino acid residues, exclusively or in combination with the α-amino acid residues that serve as fundamental subunits in proteins. Residues derived from β- or γ-amino acids contain one or two additional backbone carbon atoms, respectively, relative to α-residues. These extra backbone atoms increase the distance between backbone H-bonding groups (the secondary amides that link residues) and provide sites for additional side chains, which can be used to modulate folding propensity.1 6 γ-Peptides have received increasing attention recently,1,3,5 10 in part due to improvements in the ease with which γ-amino acids with varying substitution patterns can be synthesized.11,12 r 2011 American Chemical Society

A γ-amino acid residue can display enhanced flexibility relative to an α-residue because there are two backbone dihedral angles (designated here as θ1 and θ2) in addition to the ϕ and ψ angles associated of α-residues (Figure 1). Understanding the intrinsic conformational propensities of isolated γ-residues is critical to the long-term goal of designing γ-containing foldamers that adopt stable and predictable secondary and tertiary structures. Recently, we studied the single-conformation spectroscopy of a series of small γ-peptides as isolated molecules in the gas phase in order to directly probe their inherent conformational preferences in the absence of solvent.13 15 The smallest of these compounds, Ac-γ2-hPhe-NHMe (Figure 1), displayed three conformational isomers under jet-cooled conditions. On the basis of previous experience with α- and β-peptides, we anticipated formation of structures stabilized by amide amide H bonds involving either a seven-membered ring (C7) linking Received: August 23, 2011 Revised: September 19, 2011 Published: September 20, 2011 11960

dx.doi.org/10.1021/jp2081319 | J. Phys. Chem. A 2011, 115, 11960–11970

The Journal of Physical Chemistry A

ARTICLE

Figure 1. Chemical structures of the parent compound Ac-γ2-hPheNHMe and three γ-peptide derivatives studied in this work.

interior NH and CdO groups or a nine-membered ring (C9) involving the exterior groups. However, no C7 structures were observed. Instead, two of the conformers contained C9 H bonds with different positions of the side chain phenyl ring (C9(a) or C9(g-)) (Figure 2a), while the third conformer, constituting ∼20% of the population, did not contain any H bond but instead showed clear spectroscopic signatures that the two amide groups were stacked on top of one another in a fashion analogous to the stacking of aromatic rings. In this conformation the two amide groups were arranged in a stacked, antiparallel manner that allows near-perfect alignment of the complementary partial charge sites within each amide group, as shown in Figure 2b. This amide-stacked conformation is stabilized by a combination of electrostatic and dispersive interactions between the two amide groups, interactions that are facilitated by the minimal torsional strain within the γ-peptide backbone when the amide groups are properly juxtaposed. The discovery of amide stacking in Ac-γ2-hPhe-NHMe raises several important questions. First, what role, if any, does amide stacking play in polypeptides composed of α-amino acid residues? Are there circumstances in which the close approach of amide groups allows dipole dipole attractions even in the absence of H bonding and thereby contributes to the stability of the structures so produced? Could amide stacking play a role in tertiary interactions or in the aggregation of proteins? These questions could be addressed by bioinformatic searches of the protein database. Second, does amide stacking survive in the presence of H2O molecules, or does the successful competition with H bonding, as observed in Ac-γ2-hPhe-NHMe, reflect a unique consequence of an isolated molecule’s inherent preferences? One approach to this question involves the study of γ-peptide-(H2O)n clusters, an avenue we are currently pursuing. Third, does amide stacking play a role in the conformational behavior of large γ-peptides, i.e., those containing more than one residue? We have recently begun to address this question via single-conformation study of triamides that contain two γ-amino acid residues: Ac-γ2-hPhe-γ2-hAla-NHMe and Ac-γ2-hAla-γ2hPhe-NHMe.13 The observed conformers were an interesting mix of anticipated and unexpected structures. Of the seven conformers detected, four contained two sequential C9 rings,

Figure 2. DFT M05-2X/6-31+G(d)-optimized structures for the structures of the three observed conformers of Ac-γ2-hPhe-NHMe. The lower panel contains an additional view of the amide-stacked structure with the phenyl ring replaced with a hydrogen atom for clarity.

one contained a C9/C14 bifurcated double ring, and the remaining two contained C7/C7/C14 H-bond networks. This last H-bonding pattern is unique to the γ-peptides; the relatively high flexibility of the γ-peptide backbone is necessary for formation of three amide amide H bonds within a di-γ-peptide segment. The N-terminal and C-terminal amide groups both fold in toward the central amide group in near-stacking configurations, but formation of a “tristacked” amide arrangement is precluded by formation of a C14 H bond between the N- and the C-terminal amide groups. Finally, how is the balance among C7, C9, and amide-stacked conformations influenced by variations in substitution pattern with a single γ-amino acid-derived unit? This is the question that motivates the present work. In order to understand more clearly the balance of intramolecular forces at play within γ-residues, we carried out singleconformation studies on three diamides, each derived from a single γ-amino acid. The three new compounds (Figure 1) can be compared to the previously studied Ac-γ2-hPhe-NHMe. Ac-γ2hPhe-NMe2 contains an additional N-methyl group at the C-terminus, which precludes H-bond donation from this site, thereby preventing C9 ring formation. Ac-γ2-hPhe-NH(iPr) retains the two amide NH groups necessary for formation of C9 and C7 H bonds, but the C-terminal NH group is adjacent to a bulky isopropyl substituent, in contrast to the smaller methyl substituent at the analogous position in Ac-γ2-hPhe-NHMe. The third new compound is a derivative of gabapentin, which incorporates a spiro-cyclohexyl unit at carbon-3 and can thus be considered a γ3,3-amino acid. The gabapentin derivative we studied bears an N,N-dimethylamide at the C-terminus, once again precluding formation of a C9 H bond. In addition, our gabapentin-derived diamide contains a phenylacyl group at the N-terminus, which contains the aromatic chromophore necessary for IR/UV double resonance. As shown below, each of these three γ-peptides has a strong and characteristic preference for a particular type of conformation, amide-stacked (Ac-γ2-hPheNMe2), C9 (Ac-γ2-hPhe-NH(iPr)), or C7 (1). Collectively, these 11961

dx.doi.org/10.1021/jp2081319 |J. Phys. Chem. A 2011, 115, 11960–11970

The Journal of Physical Chemistry A findings demonstrate the strong influence of particular substitution patterns on local γ-peptide folding propensity.

II. EXPERIMENTAL AND COMPUTATIONAL METHODS Ac-γ2-hPhe-NH(iPr) was prepared from the corresponding β-substituted δ-nitrobutanol derivative.11 Jones oxidation of this alcohol provided the corresponding γ-nitro butyric acid derivative, O2N-γ2-hPhe-COOH. This γ-nitro butyric acid derivative (1.5 mmol) was added directly to a solution containing isopropyl amine (1.8 mmol), (N,N-dimethylamino)propyl-3-ethylcarbodiimide hydrochloride (1.2 equiv), and N,N-diisopropylethylamine (DIEA; 1.0 mmol) in CH2Cl2. The resulting solution was stirred for 2 days at room temperature. The reaction mixture was diluted with EtOAc, and the resulting solution was washed with 10% aqueous citric acid, saturated aqueous NaHCO3, and then brine. The organic layer was dried over MgSO4, filtered, and concentrated to give the desired amide, which was purified via SiO2 column chromatography. The nitro group was subsequently reduced via hydrogenation with 10% Pd/C in methanol at an H2 pressure of 50 psi to yield the corresponding amine, which was not purified. The amine (0.8 mmol) was dissolved in CH2Cl2, and N,N-diisopropylethylamine (0.96 mmol) and acetic anhydride (0.96 mmol) were added at 0 °C. The mixture was stirred at room temperature for 3 h and concentrated to give a crude product in 80% yield, which was purified via column chromatography eluting with CH2Cl2/MeOH. The product was further purified by recrystallization from a methanol/EtOAc/ hexane mixture. 1H NMR (300 MHz, CDCl3): δ 7.30 7.12 (m, 5H), 5.90 (m, 2H), 3.97 (o, J = 6.8 Hz, 1H), 3.45 (m, 1H), 3.09 (m, 1H), 2.92, 2.69 (AB of ABX, JAB = 13.5 Hz, JAX = 8.7 Hz, JBX = 6.0 Hz, 2H), 2.30 (m, 1H), 1.96 (s, 3H), 1.91 (m, 1H), 1.60 (m, 1H), 1.09 (d, J = 6.6 Hz, 3H), 0.92 (d, J = 6.6 Hz, 3H). 13C NMR (75.4 MHz, CDCl3): δ 173.72, 170.92, 139.78,129.18, 128.61, 126.54, 47.43, 41.43, 39.39, 38.13, 32.82, 23.52, 22.79, 22.68. HRMS m/z (ESI): calcd for C16H24N2O2, [M]+, 276.1833, found 276.1833. Ac-γ2-hPhe-NMe2 was synthesized using a procedure analogous to that described above for Ac-γ2-hPhe-NH(iPr). The synthetic route to gabapentin derivative 1 is outlined below. Gabapentin was purchased from TCI. Solid gabapentin NH2Gabapentin-COOH (2.0 mmol) was suspended in anhydrous CH2Cl2 (20 mL) and stirred vigorously. TMSCl (4.0 mmol) was added in one portion, and the reaction mixture was stirred at room temperature for 1 h. The mixture was cooled to 0 °C, and DIEA (1.2 equiv) and Boc2O (2.4 mmol) were added sequentially. The reaction mixture was allowed to warm to room temperature and stirred for 6 h. The resulting mixture was concentrated to provide a yellow oil, which was dissolved in EtOAc (50 mL). To this solution, water was added (20 mL), and the mixture was acidified with 1 N HCl. The separated organic layer was dried over MgSO4, filtered, and concentrated to give a white solid. The Boc-protected gabapentin was purified via column chromatography eluting with EtOAc/hexane (1:10 to 1:1; v/v) to give pure product NHBoc-Gabapentin-COOH as a white solid in 89% yield. This Boc-protected gabapentin (1.5 mmol) was added directly to a solution containing dimethyl amine (1.8 mmol), (N,N-dimethylamino)propyl-3-ethylcarbodiimide hydrochloride (1.8 mmol), and N,N-diisopropylethylamine (1.8 mmol) in 15 mL of CH2Cl2. The resulting solution was stirred for 2 days at room temperature. The reaction mixture was diluted with 20 mL of EtOAc, and the resulting solution was

ARTICLE

washed with 30 mL of 10% aqueous citric acid, 30 mL of saturated aqueous NaHCO3, and then 30 mL of brine. The organic layer was dried over MgSO4, filtered, and concentrated to give the desired amide NHBoc-Gabapentin-N(Me)2, which was purified via SiO2 column chromatography to give the desired product in 92% yield. NHBoc-Gabapentin-N(Me)2 (0.8 mmol equiv) was then treated with 4.0 M HCl in dioxane (∼10 equiv). The mixture was stirred for 30 min and then concentrated under a nitrogen gas stream to give HCl salt of gabapentin-dimethylamide as a white solid. This material (0.8 mmol) was dissolved in CH2Cl2, and DIEA (0.96 mmol) and phenylacetyl chloride (1.2 mmol) were added at 0 °C. The mixture was stirred at room temperature for 3 h and concentrated to give a residue that purified via column chromatography eluting with EtOAc/hexane to provide gabapentine derivative 1 in 82% yield. 1H NMR (300 MHz, CDCl3): δ 7.39 7.19 (m, 5H), 3.54 (s, 2H), 3.29 (d, J = 6.2 Hz, 2H), 3.02 (s, 3H), 2.91 (s, 3H), 2.16 (s, 2H), 1.70 1.14 (m, 11H). 13C NMR: (75.4 MHz, CDCl3): δ 172.42, 171.27, 136.05, 129.46, 128.84, 127.01, 45.18, 44.46, 41.46, 38.75, 38.54, 35.87, 35.17, 26.15, 21.89. HRMS m/z (ESI): calcd for C19H28N2O2Na, [M + Na]+, 339.2043, found 339.2034. The vacuum chamber, lasers, and experimental methods used to record single-conformation ultraviolet and infrared spectra have been described previously.16 21 Details specific to the current study are presented here. The solid peptide samples were wrapped in glass wool and placed in a glass insert in an effort to reduce thermal decomposition and then inserted into a stainless steel sample holder and heated to approximately 175 °C to increase the sample vapor pressure. The sample holder was located directly behind a pulsed valve (Parker General Valve, Series 9) fitted with a 400 μm nozzle orifice. The sample was entrained in a carrier gas composed of a 70/30% neon/helium mixture at a backing pressure of 1.5 bar and expanded into vacuum. The sample molecules were cooled through collisions with the carrier gas and skimmed to form a molecular beam prior to entering the ionization region of a timeof-flight mass spectrometer. Short gas pulses (300 500 μs in duration) were utilized to minimize interference of the gas pulse with the conical skimmer. One-color resonant two-photon ionization (R2PI) spectra were recorded by tuning the UV laser while monitoring ion signal in the parent mass channel of interest in the time-of-flight mass spectrum. Electronic spectra were recorded using resonant two-photon ionization (R2PI) spectroscopy, while conformation-specific UV spectra were recorded using the double-resonance method of IR/ UV holeburning spectroscopy.21 In IR-UV holeburning, a unique IR absorption is used as hole-burn transition to selectively remove population from the zero-point level of a given conformational isomer. The hole-burn spectrum is recorded by tuning the UV laser, operating at 20 Hz, through the region of interest while monitoring the difference in ion signal with or without the IR laser present using active baseline subtraction. The method requires the presence of an IR absorption unique to each conformer to serve as IR hole-burning wavelength, an assumption that must be checked against the single-conformation IR spectra. Single-conformation IR spectra were obtained in the amide I (1600 1800 cm 1) and amide NH stretch (3200 3500 cm 1) spectral regions using resonant ion-dip infrared spectroscopy (RIDIRS).22 RIDIR scans differ from IR-UV holeburning by switching which laser is tuned (IR) and which is fixed in wavelength (UV laser). 11962


The Journal of Physical Chemistry A For molecules of the size considered here (∼30 50 atoms), it is possible to carry out a systematic search of conformational space to find all possible structural motifs and assess their relative energies. Such a search was carried out using the MMFFs force field23 within the MACROMODEL24 suite of programs. The conformational minima so found within a 50 kJ/mol energy window were then used as starting structures for higher level calculations to compute optimized geometries, relative energies, vibrational frequencies, and infrared intensities. Initial calculations employed density functional theory with the B3LYP functional25 and a 6-31+G(d) basis set using the GAUSSIAN 03 suite of programs.26 The vibrational frequencies were then compared to the experimentally recorded, conformation-specific infrared spectra, and tentative initial structural assignments were made. In order to better account for dispersive interactions, the DFT B3LYP calculations were followed up with calculations at the DFT M05-2X level with the same basis set. The M05-2X functional has been developed by Truhlar and co-workers explicitly to better account for dispersive interactions.27,28 Optimizations used the scf = tight and ultrafine grid options.26 This level of theory has been shown in several recent studies to provide a more accurate set of relative energies for the minima.29,30 The resultant DFT M05-2X vibrational frequencies were then compared to the experimental infrared spectra, leading to assignments of the observed conformers. Scale factors (0.9600 in the NH stretch region and 0.9400 in the amide I region) were chosen so as to match the free NH and free amide I fundamentals of conformer A of Ac-γ2-hPhe-NHMe, respectively. In all cases, the structural assignments were unchanged from those tentatively made based on the B3LYP functional. As a further test of the extent to which dispersive interactions play a role in the assigned conformers of γ-peptide derivatives, the following dispersioncorrected density functional methods were used and compared with one another: B3LYP-D/SVP,31,32 M06-2X/6-31+G(d),33 and ωB97X-D/6-31+G(d).34,35 Relative energies for a subset of the structures for which the full range of calculations were carried out are compared in the Supporting Information. Assignments made in what follows are robust to the level of theory chosen for comparison with experiment. In the rest of the paper, the M052X calculations are used as a primary point of comparison due to the fact that a larger number of structures were optimized at this level of theory. Vertical excitation energies for the conformers were determined from single-point TDDFT calculations at the optimized ground state geometry using the M05-2X/6-31+G(d) level of theory.

III. RESULTS A. R2PI and IR-UV Holeburning Spectra. Figure 3b, 3c, and 3d displays the R2PI spectra of Ac-γ2-hPhe-NMe2, Ac-γ2-hPheNH(iPr), and the gabapentin derivative 1 in the S0 S1 origin region of the phenylalanine side chain, respectively. The corresponding spectrum of the parent compound Ac-γ2-hPhe-NHMe (Figure 3a) is included for comparison. Initial inspection of the R2PI spectra shows that the modifications made to Ac-γ2-hPheNHMe in the design of each of the derivatives range from subtle (Ac-γ2-hPhe-NH(iPr)) to dramatic (gabapentin derivative 1). The R2PI spectrum of Ac-γ2-hPhe-NH(iPr) (Figure 3b) is very similar in appearance to that of Ac-γ2-hPhe-NHMe above it. Its two most intense transitions (A and B) occur at 37 564 and

ARTICLE

Figure 3. R2PI spectra of (a) Ac-γ2-hPhe-NHMe, (b) Ac-γ2-hPheNMe2, (c) Ac-γ2- hPhe-NH(iPr), and (d) gabapentin derivative 1 in the S0 S1 origin region. The band marked with an asterisk in b has been tentatively assigned as a hot band of A.

37 481 cm 1, respectively, positions very close to the S0 S1 origins of the two C9 conformers of Ac-γ2-hPhe-NHMe, with shifts of 20 and 3 cm 1. No transition corresponding to the amide-stacked C(S(a)) of Ac-γ2-hPhe-NHMe is observed. The R2PI spectrum of Ac-γ2-hPhe-NMe2 (Figure 3c) is dominated by transition A at 37 502 cm 1, with weak transitions to both lower and higher frequency. Its position is similar to those of B and C in Ac-γ2-hPhe-NHMe but cannot be due to a C9 conformer, which is blocked from formation by double methylation of the C-terminus. Finally, the S0 S1 origin region of the spectrum of the gabapentin derivative 1 (Figure 3d) is shifted to lower frequency by approximately 250 cm 1 from the activity in the parent molecule Ac-γ2-hPhe-NHMe, reflecting a substantially different local environment for the aromatic ring due to its unique location relative to the γ-peptide backbone (Figure 1). The spectrum also shows a significant increase in vibronic activity compared to the three others, including a long progression in a 9 cm 1 mode built off the most intense transition (labeled A) at 37 227 cm 1. Figure 4a c presents IR-UV holeburning spectra of the major conformers of each molecule. These spectra were recorded using IR frequencies unique to each conformer for IR holeburning taken from the RIDIR spectra (Figure 4). Comparison of the R2PI spectrum of Ac-γ2-hPhe-NH(iPr) with the two holeburning spectra below it (Figure 4a) shows that the majority of transitions are accounted for by the two conformers A and B. However, a third conformer C, with S0 S1 origin at 37 590 cm 1, does not burn with either A or B. Due to its small intensity, no holeburning scans were attempted on peak C. A small vibronic transition 24 cm 1 above the C origin (connected by a tie line) also is tentatively assigned to conformer C. The holeburn spectra of both A and B are dominated by their respective S0 S1 origin transitions with low-frequency vibronic bands at 25 and 44 cm 1 (for A) and 22 and 30 cm 1 (for B) ascribable to low-frequency motions of the peptide backbone relative to the phenyl ring. 11963



ARTICLE

Figure 4. R2PI spectra (top) and IR-UV holeburning spectra of the major conformers of (a) Ac-γ2-hPhe-NH(iPr), (b) Ac-γ2-hPhe-NMe2, and (c) gabapentin derivative 1 in the S0 S1 origin region.

A single IR-UV holeburning scan of transition A in Ac-γ2hPhe-NMe2 (Figure 4b) accounts for all but two transitions, marked B and *. As we shall see, transition B has a unique IR spectrum that is clearly ascribable to a unique conformer. The band marked by an asterisk has an NH stretch infrared absorption only slightly shifted from that in A; hence, given its location immediately to the red of the A 0°0 it is tentatively assigned as a hot band of A. Finally, the holeburning spectrum of conformer A of 1 (Figure 4c), recorded using an IR hole-burn frequency of 3365 cm 1, suggests that much of the observed vibronic structure in the R2PI spectrum of 1 can be ascribed to a single conformer with S0 S1 origin at 37 227 cm 1. However, the dense vibronic structure present in A makes it possible that more than one conformer with IR absorption at 3365 cm 1 could be present. We will return to this possibility after considering the predictions of calculations. The weak transition labeled B does not burn with the others, identifying it as a second conformer. It is likely that the small transition 7 cm 1 to the red of the S0 S1 origin of A which burns with A in the IR-UV holeburn spectrum is a hot band of A, indicating that the 9 cm 1 vibration in S1 is even lower in frequency in S0 (7 cm 1). Table 1 summarizes the observed S0 S1 origin positions and relative intensities for all the observed conformers of the three γ-peptide derivatives and compares them with Ac-γ2-hPhe-NHMe. B. RIDIR Spectra. The transitions suspected as possible S0 S1 origins in each R2PI spectrum were used as monitor transitions when recording conformation-specific infrared spectra in the NH stretch region (Figure 5) and amide I region (Figure 6) using RIDIR spectroscopy. Amide I spectra were only recorded for the major conformers of each molecule due to the small intensity in the R2PI spectrum. The wavenumber positions of the infrared transitions are included in Table 1. The left panel of Figure 5 shows the RIDIR spectra of Ac-γ2hPhe-NH(iPr), while the corresponding spectra in the amide I region are shown in Figure 6a. In both regions, the three spectra closely resemble one another. They also are nearly identical with the corresponding spectra of the two C9 conformers (A and B) of Ac-γ2-hPhe-NHMe. One NH stretch fundamental appears near 3350 cm 1 indicative of an NH 3 3 3 OdC H bond and one near 3475 cm 1 due to a free amide NH. In the amide I region (Figure 6a), the free CdO stretch absorbs in the 1704 1708 cm 1 region and the C9-bound CdO stretch at

1687 1689 cm 1. The two bands have similar intensity, indicating that the intensity enhancement characteristic of the NH stretch when it is in a hydrogen bond does not carry over to the amide I region. The stick spectra below each experimental spectrum show scaled harmonic vibrational frequencies and infrared intensities for conformers best matching experiment. As anticipated from the close resemblance of the R2PI spectra with the C9 conformers of Ac-γ2-hPhe-NHMe, the best-fit structures for all three conformers of Ac-γ2-hPhe-NH(iPr) are C9 structures, all nearly isoenergetic with one another. It would be difficult on the basis of the amide NH stretch region alone to distinguish between C9 and C7 structures, since both support H-bonded NH stretch fundamentals near 3350 cm 1. However, in Ac-γ2-hPhe-NH(iPr), the first conformer to have any C7 H bonding is one 4.5 kJ/ mol above the global minimum. However, this structure might be better labeled as amide stacked, somewhat perturbed by the presence of the isopropyl group. Its amide NH stretch spectrum has far too small a shift for the H-bonded NH stretch (∼3450 cm 1). In what follows, we label this structure as C7/ S to denote this mixed C7/stacked character. The first C7 structure with a typical C7 ring is 9.7 kJ/mol higher in energy, but here an NH 3 3 3 π H bond shifts the free amide NH stretch down by about 30 cm 1 relative to experiment. Thus, the preference for C9 over C7 structures in Ac-γ2-hPhe-NH(iPr) is clear. The C9 NH stretch fundamentals of conformers A, B, and C appear at 3363, 3333, and 3351 cm 1, indicating small but systematic differences in the strengths of the C9 H bonds formed. The frequencies of the C9 NH stretch fundamentals can be used to distinguish between possible C9 conformers. The assigned structures (Figure 7a) follow the experimental shifts, leading to assignments of A to C9(a), B to C9(g-), and C to a second C9(a) conformer in which the isopropyl group is in a different local conformation. Note that these assignments for conformers A and B maintain the same peptide backbone dihedral angles and chromophore positions as conformers A and B of Ac-γ2-hPhe-NHMe, consistent with the near-identical positions of their S0 S1 origin transitions in the UV spectrum. The assignment of conformer C to a second C9(a) conformer, differing only in the orientation of the C-terminal isopropyl cap, is also consistent with the small shift of its S0 S1 origin relative to that of A, since the different 11964



ARTICLE

Table 1. Key Spectroscopic Data for the Observed Conformations of Ac-γ2-hPhe-NHMe, Ac-γ2-hPhe-NMe2, Ac-γ2-hPheNH(iPr), and Gabapentin Derivative 1 family

assignment

exp. S0 S1

TDDFT vert.

(cm 1)a

(cm 1)b

R2PI intensity

exp. CdO

exp. N H

(cm 1)

(cm 1)

Ac-γ2-hPhe-NHMe A B

C9 C9

C9(a) C9(g-)

37 584 (0) 37 484 ( 100)

C

S

S(a)

37 471 ( 113)

0 111

strong strong

1692, 1709 1693, 1710

3372, 3481 3357, 3476

123

weak

1716

3469, 3480

1660, 1715

3457

Ac-γ2-hPhe-NMe2 A

S

S(a)

37 502 (0)

0

strong

B

S

S(a)

37 576 (74)

17

weak

3446

2

Ac-γ -hPhe-NH(iPr) A

C9

C9(a)

37 564 (0)

B

C9

C9(g-)

37 481 ( 83)

C

C9

C9(a)

37 576 (12)

A

C7

C7

37 227 (0)

B

C7

C7

37 309 (82)

0

strong

1688, 1708

3363, 3481

strong

1689, 1704

3333, 3475

weak

1687, 1706

3351, 3481

0

strong

1652, 1707

3374

152

weak

108 5 gabapentin derivative 1

3341

Relative shift from conformer A for a given γ-peptide derivative in parentheses. b TDDFT vertical excitations caclulated using M05-2X/6-31+G(d). Relative shifts are calculated from conformer A for a given γ-peptide derivative using the following scale factors: 0.8340 for Ac-γ2-hPhe-NHMe, 0.8305 for Ac-γ2-hPhe-NHMe2, 0.8342 for Ac-γ2-hPhe-Nh(iPr), and 0.8409 for the gabapentin derivative. a

Figure 5. Single-conformation IR spectra of conformers A C of Ac-γ2-hPhe-NH(iPr) (left), conformers A and B of Ac-γ2-hPhe-NMe2 (middle), and conformers A and B of gabapentin derivative 1 (right) in the amide NH stretch region. Stick diagrams calculated at the DFT M05-2X/6-31+G(d) level of theory; 0.9600 scale factor.

isopropyl configuration occurs far from the aromatic ring. According to the calculations, conformers A and B are the two lowest energy conformers of Ac-γ2-hPhe-N(iPr), with B the global minimum (Figure 7b). The somewhat smaller intensity of conformer C likely reflects its slightly higher energy. It is also possible that cooling in the expansion moves population from C into A, especially if the barrier to reorientation of the isopropyl group is small. The middle panel of Figure 5 shows the single-conformation IR spectra of transitions A and B of Ac-γ2-hPhe-NMe2. Due to the NMe2 cap at the C-terminus, a single NH group is present in this derivative, producing a single NH stretch fundamental. Interestingly, the lone NH stretch fundamental in both conformers is located above 3440 cm 1 (Table 1) and is therefore not H bonded, leading to the conclusion that both conformers

A and B of Ac-γ2-hPhe-NMe2 are amide stacked (Figure 7b). As anticipated, both conformers are amide-stacked structures differing in the position of the aromatic ring. Only the spectrum of conformer A was recorded in the amide I region (Figure 6b). Its two CdO stretch fundamentals appear at 1713 and 1660 cm 1. This surprisingly large splitting (53 cm 1) is a consequence of a shift in the amide I frequency inherent with NMe2 substitution and is properly accounted for by the calculations. The corresponding splitting in Ac-γ2-hPhe-NHMe is predicted to be only 16 cm 1. However, in that case, only a single CdO stretch band was observed at 1716 cm 1, with the lower frequency band carrying little intensity, presumably due to coupling between the antiparallel CdO groups in the two amide planes. In Ac-γ2-hPhe-NMe2, the dimethyl cap has the dual effect of shifting the adjacent CdO stretch fundamental to lower 11965



ARTICLE

Figure 6. Single-conformation IR spectra of (a) conformers A C of Ac-γ2-hPhe-NH(iPr), (b) conformer A of Ac-γ2-hPhe-NMe2, and (c) conformer A of gabapentin derivative 1 in the amide I region. Stick diagrams calculated at the DFT M05-2X/6-31+G(d) level of theory; 0.9400 scale factor.

frequency and partially localizing the amide I vibrations, thereby turning on intensity in both. While the peak at 1660 cm 1 due to the methyl-capped amide is still the weaker of the two transitions, it is nonetheless clearly visible (Figure 6b). Finally, the single-conformation IR spectra of the two conformers of gabapentin derivative 1 are shown in the right panel of Figure 5. At first glance, the spectrum of conformer B looks qualitatively similar to the C9 conformers of the parent compound; however, the N(Me)2 cap in 1 prevents C9 formation. Instead, the two conformers of the gabapentin derivative are engaged in amide amide H bonding via the one remaining NH group, with a broadened NH stretch fundamental below 3400 cm 1 consistent only with C7 ring formation. Stick diagrams associated with two low-energy C7 structures which best match experiment are shown in Figure 5. As we shall see shortly, there are several C7 conformers among the low-energy structures of 1, and these have calculated spectra very similar to one another. The spectrum of conformer B is quite unique in having a H-bonded NH stretch shifted down to 3340 cm 1, and a C7 conformer with relative energy 13.6 kJ/mol is the only good match. Given its high energy, we postulate that the IR frequency used for IR-UV holeburning of A is one shared by two or more of the low-energy C7 conformers. This is consistent with the dense vibronic structure and unusual intensities appearing in the IR-UV holeburning spectrum recorded at 3365 cm 1 (Figure 4c). In this case, it was not possible to record UVHB spectra36 to further dissect this spectrum. In the amide I region (Figure 6c), only the spectrum of conformer A was recorded. Once again the NMe2 cap present in 1 shifts the lower frequency CdO stretch fundamental to lower frequency, resulting in two nearly local modes with similar integrated intensity at 1707 and 1653 cm 1. The lower frequency of this latter band is consistent with C7 ring formation involving this group. The calculations correctly track the shifts of both free and C7 H-bonded CdO groups (Figure 6c) relative to the stacked structure above it (Figure 6b). Having made these assignments, the experimental S0 S1 origins are compared with vertical S0 S1 excitation energies computed at the TDDFT M05-2X/6-31+G(d) level of theory.

Figure 7. Assigned structures of (a) Ac-γ2-hPhe-NH(iPr), (b) Ac-γ2hPhe-NMe2, and (c) gabapentin derivative 1 (right) with relative energies (kJ/mol) calculated at the M05-2X/6-31+G(d) level of theory.

This comparison is included in Table 1. Note that the relative wavenumber positions of the conformers in the R2PI track the experimental trends, providing further confirming evidence for the conformational assignments.

IV. DISCUSSION As discussed in section III, the combination of IR-UV holeburning and RIDIR spectroscopy in the amide NH and amide I regions provided firm assignments for the seven observed conformers of the three γ-peptides examined in this work. The structures of these assigned conformers are shown in Figure 7. Interestingly, each of the three γ-peptides possesses a unique conformational preference, with Ac-γ2-hPhe-N(iPr) selectively forming C9 structures, Ac-γ2-hPhe-NMe2 forming amidestacked structures, and gabapentin derivative 1 forming C7 structures. We seek to understand, based on the calculations, how the covalent differences among these molecules lead to the observed conformational preferences. We were motivated to study these three compounds primarily by the unanticipated presence of an amide-stacked conformation in the previously studied molecule Ac-γ2-hPhe-NHMe.13,14 In that case the data suggested that amide-stacked conformations are energetically comparable with the most stable conformations 11966



Figure 8. Energy level diagram comparing the relative energies of lowenergy structures within the first 15 kJ/mol of the global minimum for Ac-γ2-hPhe-NHMe, Ac-γ2-hPhe-NH(iPr), Ac-γ2-hPhe-NMe2, and gabapentin derivative 1 calculated at the M05-2X/6-31+G(d) level of theory. Energy levels labeled ‘Ex’ are extended structures without intramolecular H bonds.

containing amide amide H bonds. As a way of visualizing this fact, the left-hand column of Figure 8 presents an energy level diagram of the low-energy conformations of Ac-γ2-hPhe-NHMe. Amide-stacked conformation S(a) is within 1 kJ/mol of the lowest lying C9 conformations, and all of these conformations are more stable than any C7 conformation by ∼11 kJ/mol. Several factors contribute to the stability of the amide-stacked conformation in Ac-γ2-hPhe-NHMe: (i) attractive dipolar interactions arising from the antiparallel stacked arrangement of the amide groups, (ii) attractive dispersive interactions between the π clouds of the two amide groups, much as would occur in aromatic ring stacking, and (iii) minimal torsional strain in the γ-peptide backbone when configured so as to bring the two amide groups into antiparallel arrangement.14,18 Our goal in studying the three γ-peptides introduced here was to determine the conformational impact of changing the covalent structure of Ac-γ2-hPhe-NHMe in modest and well-defined ways; we anticipated that the results would help us understand the balance of intramolecular forces that underlies the competition between amide-stacked and H-bonded conformations. Figure 8 offers a global comparison by presenting the calculated energies of the low-lying conformational minima of the three γ-peptides studied here along with the minima of the previously studied Ac-γ2-hPhe-NHMe. In each diagram, the energy level associated with the most stable member of each conformational family (amide-stacked, C7, or C9) is identified. A specific goal of this work was to identify backbone modifications that would lead to a γ-peptide strongly predisposed toward amide-stacked structures. This goal was achieved with

ARTICLE

Ac-γ2-hPhe-NMe2, which contains an additional N-methyl group at the C-terminus, relative to Ac-γ2-hPhe-NHMe, which precludes formation of a C9 H-bonded ring. This modification leaves only C7 H-bonded conformers to compete with amide-stacked conformers. Since the additional methyl group is larger than the hydrogen it replaced, this alteration could potentially perturb the amide-stacking interaction. However, only amide-stacked conformations are observed experimentally, indicating a clear energetic preference for amide-stacked over C7 structures. This behavior is reflected in the calculations for Ac-γ2-hPhe-NMe2 (Figure 8), which predict that the two lowest energy structures are both amide stacked, with the lowest energy C7 structure ∼13 kJ/mol higher in energy at the DFT M05-2X/6-31+G(d) level of theory. It can be seen that removal of the C9 structures leads to a sparse energy level diagram with two amide-stacked conformations as the only structures within the first 10 kJ/mol. Ac-γ2-hPhe-NH(iPr) differs from Ac-γ2-hPhe-NHMe in that the former displays a clear preference for C9 H-bonded conformations, with amide-stacked conformations not able to compete for measurable population. Both of these γ-peptides contain the two amide NH groups necessary for formation of C9 and C7 H bonds, but in Ac-γ2-hPhe-NH(iPr) the C-terminal NH group is adjacent to a bulky substituent (isopropyl). In this case, three C9 conformers were observed; no amide-stacked structures were detected. This behavior is consistent with the energy level diagram in Figure 8, in which the four lowest energy conformations for Ac-γ2-hPhe-NH(iPr) are all C9 structures. One can surmise on this basis that the isopropyl group destabilizes the amide-stacked conformation, presumably by sterically inhibiting close approach of the two amide planes. This destabilization is not prohibitive, however, since the first stacked structure is predicted to be only 2.0 kJ/mol less stable than the C9 global minimum. Thus, the calculations suggest that it might be possible to detect population of an amide-stacked conformation under different expansion conditions. It is noteworthy that the disfavoring of amide-stacked structures in Ac-γ2-hPhe-NH(iPr) does not lead to formation of C7 structures, since only C9 conformers were observed. Our calculations predict that the energy gap between lowest energy C9 and C7 structures (9.75 kJ/mol) is nearly the same as in Ac-γ2-hPheNHMe (11.05 kJ/mol). Gabapentin derivative 1 introduces backbone constraints distinct from those in the other two γ-peptides studied here by incorporating a spiro-cyclohexyl ring at carbon-3. The geminal substitution at this position forces the backbone bonds between carbon-2 and -3 and between carbon-3 and -4 into gauche conformations, with θ1 and θ2 ≈ (60° (Figure 1), thereby facilitating formation of a localized turn, an effect analogous to that often observed upon incorporation of α-aminoisobutyric acid (Aib) residues into α-peptides.37 Comparable local folding behavior has been observed in crystal structures of many short gabapentin-containing oligomers.5,6 The three types of intramolecular amide amide interactions available to the γ-peptides examined here, C7 H bonding, C9 H bonding, and amide stacking, all involve θ1 and θ2 backbone angles that appear to be favorable for the gabapentin backbone (Table 2). In order to prevent C9 H bonding in 1, we placed a dimethylamide group at the C-terminus, as in Ac-γ2-hPhe-NMe2. Thus, we anticipated that C7 and amide-stacked conformations would compete with one another in gabapentin derivative 1. The single-conformation data for 1 demonstrate a strong preference for C7 conformations, with no measurable population 11967



ARTICLE

Table 2. Calculated γ-Peptide Backbone Dihedral angles (u, θ1, θ2, ψ), N 3 3 3 C Interplane Stacking Distances, and Amide Amide H-bond Structural Parameters for the Lowest-Energy Structures of each Conformational Family Type in Ac-γ2-hPhe-NHMe and the Three γ-Peptide Derivativesa amide stacked dihedral angles

interior N 3 3 3 C (Å)

exterior N 3 3 3 C (Å)

Ac-γ2-hPhe-NHMe

( 102, +55,

76, +136)

2.90

3.05

Ac-γ2-hPhe-NMe2

( 106, +58,

72, +138)

2.87

3.11

Ac-γ2-hPhe-NH(iPr)

( 117, +66,

68, +129)

2.90

3.18

gabapentin derivative 1

( 121, +69,

62, +123)

2.96

3.18

C7 dihedral angles (deg) Ac-γ2-hPhe-NHMe

( 176,

Ac-γ2-hPhe-NMe2

( 72,

Ac-γ2-hPhe-NH(iPr)c

( 170, +70,


(+94, +47, +49, +98)

C7 RH 3 3 3 O (Å)

63, +92, +136) 70, +57, +86) 67, +138)

CO 3 3 3 H angle (deg)

NH 3 3 3 O angle (deg)

1.90

114

153

2.16

95

126

2.18

90

136

2.04

109

132

CO 3 3 3 H angle (deg)

NH 3 3 3 O angle (deg)

134

156

133

153

C9 dihedral angles (deg) Ac-γ2-hPhe-NHMe

(+99,

70,

73, +105)


2.00 not observedb

Ac-γ2-hPhe-NMe2 Ac-γ2-hPhe-NH(iPr)

C9 RH 3 3 3 O (Å)

(+99,

71,

73, +107)

2.03 not observedb

a Calculations performed at the M05-2X/6-31+G(d) level of theory. b C9 structures are not possible due to the dimethylated C-terminus. c A mixed C7/stacked structure (C7/S in Figure 8).

of amide-stacked structures. As Figure 8 shows, the three lowest energy structures predicted for 1 all contain a C7 H bond, differing from one another by changes in the position of the phenyl group and in backbone dihedral angles. These conformations all have very similar predicted amide NH stretch and amide I spectra and may contribute to the IR-UV holeburning spectrum of A in Figure 4c, as mentioned previously. The calculations predict a lowest energy stacked structure that is 7.4 kJ/mol less stable than the global minimum C7 conformation (8.2 kJ/mol with ZPE correction). Table 2 compares key structural parameters of the lowest energy C9, C7, and amide-stacked structures of Ac-γ2-hPheNHMe and the three γ-peptides studied in this work. We characterize the amide amide H bonds using three measures: the H 3 3 3 O distance, the NH 3 3 3 O angle, and the CdO 3 3 3 H angle. Surveys of amide amide H bonds in proteins and of NH 3 3 3 O H bonds between small molecules have provided a general understanding of the favored geometries for such H-bond formation.38,39 The H-bond distance is most critical, with values near 2.0 Å typical in proteins. The N H 3 3 3 O H bond prefers a linear approach (N H 3 3 3 O angle near 180°), with the NH bond pointing directly at the acceptor oxygen atom of the carbonyl group. Typical values in proteins are near 160°. The CdO group accepts H bonds over a much broader range of CdO 3 3 3 H angles. In the absence of steric constraints, an inplane approach along the sp2 lone pairs is preferred, but H bonds of substantial strength are possible for CdO 3 3 3 H angles from 90° to 180°, with out-of-plane angles up to 40° not uncommon.38 Since two of the molecules introduced here cannot form C9 H bonds, because of a C-terminal dimethylamide, C9 structures are possible only in Ac-γ2-hPhe-NHMe and Ac-γ2-hPhe-NH(iPr)

among the γ-peptides we examined. The lowest energy C9 conformations in these two γ-peptides have nearly identical H-bond lengths (O 3 3 3 H ≈ 2.00 Å) and backbone dihedral angles (Table 2); in each case this C9 conformation represents the global energy minimum according to the calculations. The C9 ring so formed brings the C-terminal NH group into a favorable position for forming a strong H bond with the N-terminal carbonyl. In the C9 rings, the N H 3 3 3 O angle is near ∼160°, much as it is in the helices, β-sheets, and long-range H bonds present in proteins.39 Furthermore, the C9 rings have the NH group approaching nearly in plane, with an H OdC angle (133°) approximately along the CdO lone pairs. This preference for C9 nearest-neighbor conformations mirrors previous observations with model γ-peptides in nonpolar solvents.40 The γ-peptides studied here display striking differences in their propensity to support amide stacking. Both Ac-γ2-hPheNHMe and Ac-γ2-hPhe-NMe2 can engage in optimal amide stacking involving nearly parallel planes and antiparallel dipole orientation, with the two aligned interplane N 3 3 3 C distances of 2.90/2.87 and 3.05/3.11 Å, respectively. The dihedral angles of the γ-peptide backbone that support optimal stacking are included in Table 2 and are nearly identical for these two molecules. By contrast, in Ac-γ2-hPhe-NH(iPr) and gabapentin derivative 1, the two amide planes cannot achieve antiparallel orientations in the best calculated nearly stacked conformations, bending out from interior N 3 3 3 C distances from 2.90/2.96 to 3.18/3.18 Å at the exterior position. In 1, this behavior may result from dispersive interactions between the cyclohexyl ring and the phenyl group, which may compete with amide stacking for stabilization. 11968


The Journal of Physical Chemistry A The structural differences among the lowest energy C7 conformations of the four γ-peptides are dramatic, and these differences appear to be largely responsible for the substantial variations in conformational preference among these molecules. C7 ring formation occurs at the cost of some steric replusion in the γ-peptides, much as it does in α-peptides, where C7 rings are not observed in solution when other options exist.40 In the γ-peptides under study in this work, there appear to be two aspects to C7 ring stabilization, achieving optimal H-bond distance/ orientation, and achieving optimal Phe side chain position (axial versus equatorial) relative to the C7 ring; these factors may compete with one another. In Ac-γ2-hPhe-NHMe, the lowest energy C7 ring contains an unusually short NH 3 3 3 O distance (1.90 Å), with N H 3 3 3 O and CdO 3 3 3 H angles not so different from those in the C9 rings. However, this H-bonding configuration can occur only with the Phe side chain in the axial position relative to the C7 ring. In the well-studied C7 conformations formed among α-peptides, axial side chain placement is known to be less favorable than equatorial placement.41,42 By contrast, in Ac-γ2-hPhe-NMe2 (Table 2), the lowest energy C7 conformation forms a weak H bond, with a long H 3 3 3 O distance (2.16 Å), highly bent N H 3 3 3 O angle (126°), and near-perpendicular CdO 3 3 3 H approach angle (90°). At the same time, the Phe side chain is in an axial position, resulting in a double destabilization that pushes the energy of the most stable C7 structure some 13.7 kJ/mol above that of the amide-stacked global minimum. We noted in section III that the lowest energy structure in Acγ2-hPhe-NH(iPr) with any C7 H-bonding interaction is also one with a weak C7 H bond and was better labeled as a mixed C7/ stacked structure (C7/S in Figure 8). Its highly bent H bond is apparent in the structural parameters listed in Table 2. The more typical C7 structure is almost 10 kJ/mol above the global minimum (Figure 8). In the C7/S structure, the backbone dihedral angles are close to those of the amide-stacked structures listed above it in Table 2. This combination of weak C7 H bond and amide stacking, when combined with a more favorable equatorial Phe side chain position, bring its energy within 5 kJ/ mol of the C9 global minimum. Despite this, neither C7 nor C7/S structures of Ac-γ2-hPhe-NH(iPr) are observed experimentally. In gabapentin derivative 1, C7 ring structures compete only with amide stacking, since C9 formation is not possible. The γ-peptide dihedrals in the lowest energy C7 conformer of gabapentin derivative 1 are quite different from those in the other derivatives due to the unique backbone substitution pattern and rigid cyclohexane ring. The lowest energy conformation of 1 has a C7 ring analogous to the C7eq ring in α-peptides, with a similar H-bond distance (2.04 Å) and N H 3 3 3 O angle (132°, Table 2). The combined effects of varying the covalent structure among the four γ-peptides on the C9, C7, and amide-stacked conformations account for the observed shifts in population in each molecule relative to the others. Ac-γ2-hPhe-NHMe can form C9 rings and allows favorable amide stacking, but the C7 conformations available to this molecule have poor properties; this combination of factors rationalizes the experimental observation of both C9 and amide-stacked conformers but not C7 conformers. In Ac-γ2-hPhe-NH(iPr), the isopropyl group interferes with stacking and also leads to poor H-bonding properties in the C7 conformation, which explains why only C9 conformations are observed. In Ac-γ2-hPhe-NMe2 the C9 option has been eliminated; unfavorable features of the available C7

ARTICLE

conformations in combination with the near optimal amidestacking geometry explained why only amide-stacked conformations are observed for this molecule. Gabapentin derivative 1 cannot form a C9 H bond, and in this case, the γ-peptide backbone appears not to allow favorable amide-stacking geometries. Since the gabapentin backbone forms a favorable C7 H-bonded ring, this intramolecular interaction determines the observed conformations.

V. CONCLUSIONS The molecules studied in this work provide examples in which each of the three nearest-neighbor interactions that are possible within a γ-peptide backbone plays a dominant conformationdetermining role: C9 H bonding in Ac-γ2-hPhe-N(iPr), C7 H bonding in gabapentin derivative 1, and amide stacking in Ac-γ2hPhe-NMe2. In the molecule that inspired this work, Ac-γ2-hPheNHMe, C9 conformers comprise the majority of the gas-phase population. Nearest-neighbor C9 H bonding has commonly been observed for gabapentin residues in peptidic oligomers for which crystal structures have been determined.5,6,43 In contrast, other types of γ-amino acid residues seem to prefer longer range H-bonding patterns, especially CdO(i) H N(i+3), when embedded in peptidic oligomers.43 46 These preferences for longer range H bonding support formation of helical secondary structures. In Ac-γ2-hPhe-NMe2 and gabapentin derivative 1, placement of a dimethylamide group at the C-terminus makes C9 H bonding impossible, which enables other nearest-neighbor interactions (C7 or amide stacking) to influence conformational preferences. Our goal of creating circumstances in which amide stacking is a dominant intramolecular factor was realized in Ac-γ2hPhe-NMe2. In future design efforts, it would be interesting to try to create peptidic oligomers in which amide stacking rather than amide H bonding specifies a particular secondary structure. Our finding that amide stacking can compete with H bonding within a small γ-peptide leads one to wonder whether amide stacking could play a role in stabilizing protein or foldamer secondary or tertiary structures. Re-examining recent literature on γ-peptides in light of the above results suggests that amide stacking may indeed contribute to folding in these systems. For example, backbone torsion angles predicted for mixed H14/12 helical γ-peptides47 are similar to those necessary for amide amide-stacking interactions. Thus, in addition to the H bonds, amide-stacking interactions may contribute to helix stability in this case. Amide-stacking interactions may also play a role whenever a close side-on approach of two amide groups is feasible (e.g., glycine-rich regions of proteins), especially if H-bonding sites are already occupied.48 Indeed, the recent studies of DeGrado and co-workers49 point to just such interactions playing an important role in the association of transmembrane helices. When glycine residues are prevalent at the interface between helices, dimerization is facilitated in an antiparallel fashion. The authors identify van der Waals forces enabled by the close approach of the two helices as key to the observed preference for antiparallel dimer formation. Whether amide stacking is at play in the close approach is an open question worthy of further investigation. More generally, our results suggest that efforts to use amide stacking as a foldamer design element may be particularly fruitful when the backbone contains γ-residues. 11969



’ ACKNOWLEDGMENT W.H.J., E.G.B., C.W.M., J.C.D., and T.S.Z. acknowledge support for this research from the National Science Foundation (NSFCHE0909619). C.W.M. would also like to thank the “Deutsche Akademie der Naturforscher Leopoldina” for a postdoctoral scholarship (grant number BMBF-LPD 9901/8-159 of the “Bundesministerium f€ur Bildung und Forschung”). S.H.G. and L.G. were supported by NSF grant CHE-0848847. D.K. and L.V.S. acknowledge support from NSF grant CHE-CAREER-0955419. ’ ASSOCIATED CONTENT

bS

Supporting Information. Relative energies of the assigned structures of Ac-γ2-hPhe-NHMe and the three γ-peptide derivatives calculated at the B3LYP, M05-2X, M06-2X, ωB97X-D and B3LYP-D levels of theory. This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected]. Present Addresses ‡

)

SCHOTT North America, Incorporation, 400 York Avenue, Duryea, Pennsylvania 18642, United States. The Dow Chemical Company, Formulation Science, 1712 Building 23-1, Midland, Michigan 48667, United States.

’ REFERENCES (1) Horne, W. S.; Gellman, S. H. Acc. Chem. Res. 2008, 41, 1399. (2) Gellman, S. H. Acc. Chem. Res. 1998, 31, 173. (3) Seebach, D.; Beck, A. K.; Bierbaum, D. J. Chem. Biodiversity 2004, 1, 1111. (4) Seebach, D.; Cardiner, J. Acc. Chem. Res. 2008, 41, 1366. (5) Vasudev, P. G.; Chatterjee, S.; Shamala, N.; Balaram, P. Acc. Chem. Res. 2009, 42, 1628. (6) Vasudev, P. G.; Chatterjee, S.; Shamala, N.; Balaram, P. Chem. Rev. 2011, 111, 657. (7) Seebach, D.; Brenner, M.; Rueping, M.; Jaun, B. Chem.—Eur. J. 2002, 8, 573. (8) Seebach, D.; Hook, D. F.; Glattli, A. Biopolymers 2006, 84, 23. (9) Vasudev, P. G.; Ananda, K.; Chatterjee, S.; Aravinda, S.; Shamala, N.; Balaram, P. J. Am. Chem. Soc. 2007, 129, 4039. (10) Vasudev, P. G.; Shamala, N.; Ananda, K.; Balaram, P. Angew. Chem., Int. Ed. 2005, 44, 4972. (11) Chi, Y.; Guo, L.; Kopf, N. A.; Gellman, S. H. J. Am. Chem. Soc. 2008, 130, 5608. (12) Guo, L.; Chi, Y. G.; Almeida, A. M.; Guzei, I. A.; Parker, B. K.; Gellman, S. H. J. Am. Chem. Soc. 2009, 131, 16018. (13) Buchanan, E. G.; James III, W. H.; Gutberlet, A.; Dean, J. C.; Guo, L.; Gellman, S. H.; Zwier, T. S. Faraday Discuss. 2011, 150. (14) James, W. H., III; M€uller, C. W.; Buchanan, E. G.; Nix, M. G. D.; Guo, L.; Roskop, L.; Gordon, M. S.; Slipchenko, L. V.; Gellman, S. H.; Zwier, T. S. J. Am. Chem. Soc. 2009, 131, 14243. (15) James III, W. H.; Buchanan, E. G.; Mueller, C. W.; Dean, J. C.; Kosenkov, D.; Slipchenko, L. V.; Guo, L.; Gellman, S. H.; Zwier, T. S. J. Phys. Chem. A 2011, accepted. (16) Baquero, E. E.; James, W. H.; Choi, S. H.; Gellman, S. H.; Zwier, T. S. J. Am. Chem. Soc. 2008, 130, 4784. (17) Baquero, E. E.; James, W. H.; Choi, S. H.; Gellman, S. H.; Zwier, T. S. J. Am. Chem. Soc. 2008, 130, 4795. (18) Shubert, V. A.; Baquero, E. E.; Clarkson, J. R.; James, W. H.; Turk, J. A.; Hare, A. A.; Worrel, K.; Lipton, M. A.; Schofield, D. P.; Jordan, K. P.; Zwier, T. S. J. Chem. Phys. 2007, 127, 234315.

ARTICLE

(19) James, W. H., III; Baquero, E. E.; Choi, S. H.; Gellman, S. H.; Zwier, T. S. J. Phys. Chem. A 2010, 114, 1581. (20) James, W. H., III; Baquero, E. E.; Shubert, V. A.; Choi, S. H.; Gellman, S. H.; Zwier, T. S. J. Am. Chem. Soc. 2009, 131, 6574. (21) Zwier, T. S. J. Phys. Chem. A 2006, 110, 4133. (22) Zwier, T. S. Annu. Rev. Phys. Chem. 1996, 47, 205. (23) Halgren, T. A. J. Comput. Chem. 1999, 20, 730. (24) Mohamadi, F.; Richards, N. G. J.; Guida, W. C.; Liskamp, R.; Lipton, M.; Caufield, C.; Chang, G.; Hendrickson, T.; Still, W. C. J. Comput. Chem. 1990, 11, 440. (25) Becke, A. D. J. Chem. Phys. 1993, 98, 5648. (26) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheesemann, J. R.; Montgomery, J. A.; Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li, X.; KNox, J. E.; Hratchian, H. P.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A.; Gaussian 03, Revision E.01; Gaussian, Inc.: Wallingford, CT, 2004. (27) Zhao, Y.; Truhlar, D. G. J. Chem. Theory Comput. 2006, 2, 1009. (28) Zhao, Y.; Truhlar, D. G. J. Chem. Theory Comput. 2007, 3, 289. (29) Baquero, E. E.; James, W. H.; Choi, S. H.; Jordan, K. D.; Zwier, T. S. J. Phys. Chem. A 2008, 112, 11115. (30) James, W. H.; Baquero, E. E.; Shubert, V. A.; Choi, S. H.; Gellman, S. H.; Zwier, T. S. J. Am. Chem. Soc. 2009, 131, 6574. (31) Grimme, S. J. Comput. Chem. 2006, 27, 1787. (32) Ahlrichs, R.; Bar, M.; Haser, M.; Horn, H.; Kolmel, C. Chem. Phys. Lett. 1989, 162, 165. (33) Zhao, Y.; Truhlar, D. G. Theor. Chem. Acc. 2008, 120, 215. (34) Chai, J. D.; Head-Gordon, M. Phys. Chem. Chem. Phys. 2008, 10, 6615. (35) Chai, J. D.; Head-Gordon, M. J. Chem. Phys. 2008, 128. (36) Zwier, T. S. J. Phys. Chem. A 2001, 105, 8827. (37) Venkatraman, J.; Shankaramma, S. C.; Balaram, P. Chem. Rev. 2001, 101, 3131. (38) Baker, E. N.; Hubbard, R. E. Prog. Biophys. Mol. Biol. 1984, 44, 97. (39) Grishaev, A.; Bax, A. J. Am. Chem. Soc. 2004, 126, 7281. (40) Dado, G. P.; Gellman, S. H. J. Am. Chem. Soc. 1994, 116, 1054. (41) Chin, W.; Mons, M.; Dognon, J. P.; Mirasol, R.; Chass, G.; Dimicoli, I.; Piuzzi, F.; Butz, P.; Tardivel, B.; Compagnon, I.; von Helden, G.; Meijer, G. J. Phys. Chem. A 2005, 109, 5281. (42) Chin, W.; Piuzzi, F.; Dimicoli, I.; Mons, M. Phys. Chem. Chem. Phys. 2006, 8, 1033. (43) Guo, L.; Zhang, W.; Reidenbach, A. G.; Giuliano, M. W.; Guzei, I. A.; Spenser, L. C.; Gellman, S. H. Angew. Chem., Int. Ed. 2011in press. (44) Hanessian, S.; Luo, X. H.; Schaum, R.; Michnick, S. J. Am. Chem. Soc. 1998, 120, 8569. (45) Seebach, D.; Abele, S.; Gademann, K.; Guichard, G.; Hintermann, T.; Jaun, B.; Matthews, J. L.; Schreiber, J. V. Helv. Chim. Acta 1998, 81, 932. (46) Guo, L.; Almeida, A. M.; Zhang, W.; Reidenbach, A. G.; Choi, S. H.; Guzei, I. A.; Gellman, S. H. J. Am. Chem. Soc. 2010, 132, 7868. (47) Baldauf, C.; Gunther, R.; Hofmann, H. J. Helv. Chim. Acta 2003, 86, 2573. (48) Matsushima, N.; Yoshida, H.; Kumaki, Y.; Kamiya, M.; Tanaka, T.; Izumi, Y.; Kretsinger, R. H. Curr. Protein Peptide Sci. 2008, 9, 591. (49) Zhang, Y.; Kulp, D. W.; Lear, J. D.; DeGrado, W. F. J. Am. Chem. Soc. 2009, 131, 11341. 11970


Competition between Amide Stacking and Intramolecular H Bonds in Î³

Recommend Documents