Carbohydrate amino acids: The intrinsic conformational preference for a β-turn-type structure in a carbopeptoid building block

Article abstract of DOI:10.1021/ja0607133

Infrared ion-dip spectroscopy coupled with DFT and ab initio calculations are used to establish the intrinsic conformational preference of the basic structural unit of a peptide mimic, a cis-tetrahydrofuran-based "carbopeptoid" (amide-sugar-amide), isolat

Full text of DOI:10.1021/ja0607133

Published on Web 12/09/2006
Carbohydrate Amino Acids: The Intrinsic Conformational
Preference for a â-Turn-Type Structure in a Carbopeptoid
Building Block
Rebecca A. Jockusch,*,†,§ Francis O. Talbot,†,| Paul S. Rogers,‡ Michela I. Simone,‡
George W. J. Fleet,‡ and John P. Simons†
Contribution from the Chemistry Department, Physical and Theoretical Chemistry Laboratory,
South Parks Road, Oxford OX1 3QZ, UK, and Chemistry Department, Chemistry Research
Laboratory, Mansfield Road, Oxford OX1 3TA, UK
Received January 30, 2006; E-mail: rebecca.jockusch@utoronto.ca
Abstract: Infrared ion-dip spectroscopy coupled with DFT and ab initio calculations are used to establish
the intrinsic conformational preference of the basic structural unit of a peptide mimic, a cis-tetrahydrofuran-
based “carbopeptoid” (amide-sugar-amide), isolated at low temperature in the gas phase. The carbopeptoid
units form a â-turn-type structure, stabilized by an intramolecular NH f OdC hydrogen bond across the
sugar ring, thus forming a 10-membered, C10 turn. Despite the clear preference for C10 â-turn structures in
the basic unit, however, the presence of multiple hydrogen-bond donating and accepting groups also
generates a rich conformational landscape, and alternative structures may be populated in related molecules.
Calculations on an extended, carbopeptoid dimer unit, which includes an alternating amide-sugar-amide-
sugar-amide chain, identify conformers exhibiting alternative hydrogen-bonding arrangements that are
somewhat more stable than the lowest-energy double â-turn forming conformer.
1. Introduction
Solution NMR and IR studies of the cis-THF-based monomer
units 1 and 2 provide strong evidence for the existence of such
a 10-membered turn in the acetyl protected monomer 2 and
provide some indication that the unprotected monomer 1 may
also exist as a â-turn. Evidence for repeating â-turn structure
in the fully acetylated tetramer 4 has been provided by an NOE-
NMR investigation in CDCl3 and dimethyl sulfoxide (DMSO)
solutions, coupled with the results of a molecular mechanics
calculation based upon the CHARMM force field.4 Molecular
dynamics simulations with the GROMOS96 force field suggest,
however, that a repeating â-turn motif in the acetylated tetramer
4 is transient, appearing ∼15% of the time in CHCl3 solution
and less frequently in DMSO.7 The near IR spectrum of another
carbopeptoid tetramer based on an acetonide bridged trans-THF
building block 5, in a chloroform solution, gave no indication
of NH f OdC bonding.3 On the other hand, NMR and IR
investigations of the (longer) octameric carbopeptoid based on
acetonide bridged unit 5 indicated that it resembles a π helix,
exhibiting repeating 16-membered (C16) turns stabilized by NH
f OdC hydrogen bonds, which span one additional THF-amide
unit as compared to the 10-membered â-turn.3 To date, however,
there has been no direct experimental evidence for NH f Od
C hydrogen-bonded stabilization in the isolated carbopeptoids
free of any solvent interactions, which may perhaps offer a better
model of the peptide mimic structure in a biological environ-
ment,8 for example, when it is bound to a protein.
A key factor in the development of peptide mimics that might
be used to provide successful peptidic drugs is the control of
their conformation. One approach to this exploits the synthesis
of so-called “carbopeptoid” mimics,1 polymers constructed from
carbohydrate amino acids to adopt a specific target conformation
presenting secondary folding characteristics that resemble those
in a protein.2
Solution studies of oligomers based upon sterically controlled
tetrahydrofuran (THF) sugar-amide scaffolds (Scheme 1)
provide support for the existence of â-turn-type structures,
stabilized by intramolecular NH f OdC hydrogen bonding
(Scheme 2) across the THF ring.3-6
† Physical and Theoretical Chemistry Laboratory.
‡ Chemistry Research Laboratory.
§ Present address: Chemistry Department, University of Toronto, 80 St.
George Street, Toronto, Ontario M5S 3H6, Canada.
| Present address: Laboratoire Francis Perrin, CNRS, CEA-Saclay,
Baˆtiment 522, 91191 Gif sur Yvette, Cedex, France.
(1) Nicolaou, K. C.; Florke, H.; Egan, M. G.; Barth, T. A.; Estevez, V. A.
Tetrahedron Lett. 1995, 36, 1775-1778.
(2) Gellman, S. H. Acc. Chem. Res. 1998, 31, 173-180.
(3) Claridge, T. D. W.; Long, D. D.; Baker, C. M.; Odell, B.; Grant, G. H.;
Edwards, A. A.; Tranter, G. E.; Fleet, G. W. J.; Smith, M. D. J. Org. Chem.
2005, 70, 2082-2090.
(4) Claridge, T. D. W.; Long, D. D.; Hungerford, N. L.; Aplin, R. T.; Smith,
M. D.; Marquess, D. G.; Fleet, G. W. J. Tetrahedron Lett. 1999, 40, 2199-
2202.
(5) Hungerford, N. L.; Claridge, T. D. W.; Watterson, M. P.; Aplin, R. T.;
Moreno, A.; Fleet, G. W. J. J. Chem. Soc., Perkin Trans. 1 2000, 3666-
3679.
(6) Smith, M. D.; Claridge, T. D. W.; Sansom, M. S. P.; Fleet, G. W. J. Org.
Biomol. Chem. 2003, 1, 3647-3655. Smith, M. D.; Claridge, T. D. W.;
Tranter, G. E.; Sansom, M. S. P.; Fleet, G. W. J. Chem. Commun. 1998,
2041-2042.
Ion dip vibrational spectroscopy coupled with density func-
tional and ab initio theoretical calculations provides a powerful
(7) Baron, R.; Bakowies, D.; van Gunsteren, W. F. Angew. Chem., Int. Ed.
2004, 43, 4055-4059.
9
10.1021/ja0607133 CCC: $33.50 © 2006 American Chemical Society
J. AM. CHEM. SOC. 2006, 128, 16771-16777
16771

A R T I C L E S
Scheme 1
Jockusch et al.
Scheme 2
means of investigating the gas-phase structures of peptides9-11
and carbohydrates (including oligo-saccharides),8,12 and its
success encourages extension of this strategy to investigate
carbohydrate amino acid peptide mimics. The incidence of any
localized hydrogen-bonded interactions involving N-H or O-H
modes, to which fully resolved near-infrared spectra are
extremely sensitive, should readily be detected. In this way,
infrared spectral analysis supported by reference to ab initio
and density functional theoretical computation can establish
overall molecular structure and, in the present case, determine
whether â-turn forming NH f OdC hydrogen bonds are present
in the intrinsically favored conformation(s) of carbopeptoid
structural units.
the more extended carbopeptoid dimer unit, 3 (Scheme 1), a
penta-peptide mimic.
2. Methods
2.1. Calculations. Starting structures of molecules 1-3 for higher-
level calculations were produced using the MacroModel molecular
modeling program (v. 8.5, Schrodinger Inc.).13 Conformational searches
were performed using the Monte Carlo Multiple Minimum (MCMM)
conformation searching routine14 followed by energy minimization using
the MMFFs force field15 included with the MacroModel package. The
pucker of the sugar ring was not constrained in the searches, nor were
constraints placed on exocyclic bonds with the exception of the amides,
which were constrained to a trans geometry in molecules 2 and 3 (but
allowed to rotate in molecule 1). For molecule 1, an initial conforma-
tional search of 1000 MCMM steps was performed followed by MMFFs
energy minimization. All conformers from this search with energies
lying within 50 kJ mol-1 of the global minimum (a total of 26 different
structures) were saved and used as starting structures for higher-level
calculations. A second conformational search on molecule 1, consisting
of 10 000 MCMM steps followed by MMFFs energy minimization,
was subsequently performed. From the second search, all structures
within 20 kJ mol-1 of the global minimum were examined, and those
exhibiting types of interactions not found in the previous search were
selected as additional starting structures for higher level calculations,
resulting in a further 10 starting conformations of 1. The correlation
between relative energies computed with the MMFFs forcefield and
The results of applying this strategy to a structural investiga-
tion of the “free” and acetylated (“protected”) carbopeptoid
building blocks, 1 and 2, are presented in this Article, together
with a computational study of the conformational landscape of
(8) Jockusch, R. A.; Kroemer, R. T.; Talbot, F. O.; Snoek, L. C.; C¸ arc¸abal, P.;
Simons, J. P.; Havenith, M.; Bakker, J. M.; Compagnon, I.; Meijer, G.;
von Helden, G. J. Am. Chem. Soc. 2004, 126, 5709-5714.
(9) Chin, W.; Compagnon, I.; Dognon, J. P.; Canuel, C.; Piuzzi, F.; Dimicoli,
I.; von Helden, G.; Meijer, G.; Mons, M. J. Am. Chem. Soc. 2005, 127,
1388-1389. Chin, W.; Dognon, J. P.; Canuel, C.; Piuzzi, F.; Dimicoli, I.;
Mons, M.; Compagnon, I.; von Helden, G.; Meijer, G. J. Chem. Phys. 2005,
122, 054317. Chin, W.; Mons, M.; Dognon, J. P.; Piuzzi, F.; Tardivel, B.;
Dimicoli, I. Phys. Chem. Chem. Phys. 2004, 6, 2700-2709.
(10) Chin, W.; Dognon, J. P.; Piuzzi, F.; Tardivel, B.; Dimicoli, I.; Mons, M.
J. Am. Chem. Soc. 2005, 127, 707-712. Chin, W.; Piuzzi, F.; Dimicoli, I.;
Mons, M. Phys. Chem. Chem. Phys. 2006, 8, 1033-1048.
(13) 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, 440-467.
(14) Chang, G.; Guida, W. C.; Still, W. C. J. Am. Chem. Soc. 1989, 111, 4379-
4386. Saunders, M.; Houk, K. N.; Wu, Y. D.; Still, W. C.; Lipton, M.;
Chang, G.; Guida, W. C. J. Am. Chem. Soc. 1990, 112, 1419-1427.
(15) Halgren, T. A. J. Comput. Chem. 1996, 17, 616-641. Halgren, T. A. J.
Comput. Chem. 1996, 17, 553-586. Halgren, T. A. J. Comput. Chem. 1996,
17, 520-552. Halgren, T. A. J. Comput. Chem. 1996, 17, 490-519.
Halgren, T. A.; Nachbar, R. B. J. Comput. Chem. 1996, 17, 587-615.
(11) Chin, W.; Piuzzi, F.; Dognon, J. P.; Dimicoli, I.; Mons, M. J. Chem. Phys.
2005, 123, 084301. Chin, W.; Piuzzi, F.; Dognon, J. P.; Dimicoli, L.;
Tardivel, B.; Mons, M. J. Am. Chem. Soc. 2005, 127, 11900-11901. Hu¨nig,
I.; Kleinermanns, K. Phys. Chem. Chem. Phys. 2004, 6, 2650-2658.
Unterberg, C.; Gerlach, A.; Schrader, T.; Gerhards, M. J. Chem. Phys. 2003,
118, 8296-8300.
(12) Simons, J. P.; Jockusch, R. A.; C¸ arc¸abal, P.; Hu¨nig, I.; Kroemer, R. T.;
Macleod, N. A.; Snoek, L. C. Int. ReV. Phys. Chem. 2005, 24, 489-531.
9
16772 J. AM. CHEM. SOC. VOL. 128, NO. 51, 2006

Carbohydrate Amino Acids
A R T I C L E S
Scheme 3 a
a (i) PhCH2NH2, MeOH (57%); (ii) H2, 10% Pd/C, dioxane:H2O (2:1); then Ac2O, pyridine (72%); (iii) K2CO3, MeOH (84%).
those from subsequent ab initio calculations (see Supporting Information
Figure S3) suggests that MMFFs force field with 16 kJ mol-1 cutoff is
adequate to identify low-energy conformations for this type of molecule;
none of the higher energy structures selected (beyond the 16 kJ mol-1
cutoff) had energies within 10 kJ mol-1 of the global minimum at the
MP2/6-311+G(d,p)//B3LYP/6-31+G(d) level, giving us reasonable
confidence that the chosen energy cutoff levels did not result in missed
low-energy conformations.
A similar 10 000 step MCMM/MMFFs search was used to identify
candidate conformations for molecule 2. All conformations with
energies within 50 kJ mol-1 of the global minimum were saved and
examined. Despite this initial filtering, there were too many different
conformers for it to be feasible to perform ab initio calculations on all
of them, so a subset of conformers was selected, based upon the relative
energies determined from the molecular mechanics calculation, with
the goal of examining a variety of structures through further, higher
level calculations. The conformers selected for the carbopeptoid building
block 2 included all those within 16 kJ mol-1 of the global minimum
(32 conformers) as well as a few higher energy structures (4 conformers)
which displayed different types of hydrogen-bonded interaction. None
of the higher energy structures selected (beyond the 16 kJ mol-1 cutoff)
had energies within 10 kJ mol-1 of the global minimum at the MP2/
6-311+G(d,p)//B3LYP/6-31+G(d) level, again giving us reasonable
confidence that the chosen energy cutoff levels, although low, were
adequate to identify the most stable conformations.
For the double building block (or penta-peptide mimic) 3, the
X-cluster program (Schrodinger, Inc.)16 was used to cluster all structures
with relative energies lying within 19 kJ mol-1 of the global minimum
(50 structures) into families based on RMS distances of all atoms,
excluding the terminal methyl and phenyl groups. The lowest-energy
member of each of these families (22 in toto at the clustering level
chosen) was selected for further calculations; some additional structures
from low-energy families were also selected to confirm correct
identification of the lowest-energy structures in each family. The
potential energy surface for dimer 3 is extremely complicated, and,
while the global minimum conformation may not have been identified
using this computational procedure, effort was made to identify the
most stable double â-turn conformer; all double â-turn conformers
among the catalogued structures (those within 50 kJ mol-1 cutoff) were
selected for high-level calculations.
Each of the selected structures was subsequently optimized using
hybrid method density functional theory at the B3LYP/6-31+G(d) level,
and their single point energies were evaluated using perturbation theory
at the MP2/6-311+G(d,p) level. This resulted in 31, 26, and 29 unique
conformers for molecules 1, 2, and 3, respectively. Harmonic vibrational
frequencies were calculated at the B3LYP/6-31+G(d) level for 23 of
the 31 remaining conformer of 1. Because of the computational expense
for the (much larger) protected monomer molecule 2 and the dimer,
molecule 3, the frequency calculations were restricted to 10 of each of
their conformers. These included all those with energies lying within
12 kJ mol-1 of the minima determined from the MP2 perturbation theory
calculations. All B3LYP and MP2 calculations were done using the
Gaussian 03 suite of programs.17 The relative energies of conformers
presented in this report were all obtained from single point MP2/6-
311+G(d,p) calculations and include a correction for zero-point
vibrational energy from the harmonic B3LYP/6-31+G(d) frequency
calculations. Calculated infrared frequencies are scaled by 0.9734 for
comparison with experimental frequencies. This scaling factor has been
found to provide good agreement for weakly bonded OH stretches in
mono- and disaccharides.12 However, the use of a constant scaling
factor, and indeed the harmonic approximation, is less good for more
strongly perturbed stretches. Nevertheless, a consistent scaling factor
was used for simplicity, and matches between calculated and measured
infrared spectra were determined on the basis of patterns of IR bands,
rather than absolute position, especially for the more perturbed modes.
2.2. Synthesis. The tetrahydrofuran-based protected 2 and unpro-
tected 1 carbopeptoids were prepared from the azido ester 6.6 Reaction
of 6 with benzylamine gave the azido amide 7 (Scheme 3). Catalytic
hydrogenation of azide 7 to the corresponding amine, followed by
acetylation, afforded the protected derivative 2, which on treatment
with potassium carbonate in methanol underwent ester exchange to give
the unprotected carbopeptoid 1. Supplementary synthetic information
is available in the Supporting Information.
2.3. Spectroscopy. The carbopeptoid units 1 and 2 were vaporized
by heating ca. 40 mg samples of each monomer in a 200 °C oven
attached to the expansion side of a pulsed nozzle (General Valve, pulsed
at 10 Hz). The vaporized molecules were entrained and cooled in the
expanding pulse of gas (Ar and/or Xe), backing pressure of 3.5 (Xe)
to 6 (Ar) bar. The pulsed expansion was skimmed, after which its path
was crossed with one or two laser beam(s) in the extraction region of
a linear time-of-flight mass spectrometer (Jordan). Resonant two photon
ionization (R2PI) spectra were recorded using the frequency doubled
output of an Nd:YAG-pumped dye laser (Spectra Physics GCR 170 at
355 nm/LAS 205, pulsed at 10 Hz). For the IR ion-dip (IRID)
experiments, this laser was tuned onto successive transitions of interest
and used as the probe laser, while 150 ns prior to the probe laser pulse,
an IR burn pulse was scanned. The IR radiation (∼3000-4000 cm-1
,
0.4 cm-1 bandwidth, 4-5 mJ pulse-1) was generated by difference
frequency mixing the fundamental of a Nd:YAG laser (Powerlite
Precision 8000) and that of a pumped dye laser (ND6000, Continuum)
in a LiNbO3 crystal. When both lasers are in resonance with the same
ground state (i.e., the same conformation), this results in a dip in the
ion signal measured by the probe laser because population has been
transferred out of the monitored ground state by the burn laser. The
signal monitored by the probe then corresponds to the IR spectrum of
each selected conformer.
3. Results and Discussion
3.1. Calculations. Figure 1 shows a selection of the lowest-
energy conformers examined for the unprotected monomer unit
1 including the two most stable structures identified, Ia and Ib.
They both exhibit a 10-membered (C10) â-turn H-bond, as do
several other low-energy structures, differing from Ia or Ib
primarily by the torsion of the phenyl ring and the pucker of
the sugar ring (see the Supporting Information).
The lowest-energy structure of 1 that does not form a
hydrogen-bonded â-turn, conformer Ic, lies significantly higher
in energy (11 kJ mol-1 above the global minimum, conformer
(16) Shenkin, P. S.; Mcdonald, D. Q. J. Comput. Chem. 1994, 15, 899-916.
(17) Frisch, M. J.; et al. Gaussian 03, revision B.03; Gaussian, Inc.: Pittsburgh,
PA, 2003.
9
J. AM. CHEM. SOC. VOL. 128, NO. 51, 2006 16773

A R T I C L E S
Jockusch et al.
Figure 1. A selection of low-energy structures of the carbopeptoid monomer units 1 and 2 with their relative energies in kJ mol-1 calculated at the MP2/
6-311+G(d,p)//B3LYP/6-31+G(d) level of theory. Relative energies of the methylated analogues of 1 are given in parentheses. Conformers Ia and Ib (IIa
and IIb), the two most stable conformers identified for 1 (2), each exhibit a C10 â-turn. Conformers Ic and IIc are the lowest-energy structures that do not
form â-turns.
Ia) and is unlikely to be populated in the molecular beam.
Several other conformers lying at energies slightly above Ic
exhibit different hydrogen-bonded interactions, for example, an
alternative cross-ring bond that creates an eight-membered, C8
turn (see Scheme 2); conformer Id is the most stable structure
of this type. In view of their high energy, significant population
of the C8 conformers is improbable for the unprotected monomer
unit 1 (although the corresponding conformers of the extended
dimer 3 are quite stable, see later discussion).
The presence of the benzyl moiety (used as a chromophore
in these experiments) rather than a second methyl terminus does
not change the global minimum conformation, but it might
influence the overall conformational landscape. Replacement
of the benzyl group in 1 by a methyl group does indeed alter
the relative conformational energies among conformers with
differing benzyl group orientations; conformer Id, for example,
becomes the second most stable conformer. However, the
relative energies of conformers with similar benzyl group
orientations are little changed, and the general ordering of
Figure 2. Comparisons between â-turn structures of (a) and (c), the
carbopeptoid monomer units, 1 and 2, and (b) and (d), the computed â III
and â II′ structures of the alanine tripeptide.
relative energies among stable conformers, shown in parentheses
in Figure 1, is largely retained.
The most stable structure calculated for the extended,
unprotected dimeric structure 3, conformer IIIa, is shown in
Figure 3. It does not display a repeated, double â-turn
interaction. Instead, an NHB f NA hydrogen bond (linking the
central and terminal units) is followed by a “head-to-tail” NHA
f OdC hydrogen bond, creating a C14 turn, see Figure 3. In
contrast to the monomer units, 1 and 2, the lowest-energy
conformer exhibiting a repeating C10 â-turn structure, IIIc, is
calculated to lie 10 kJ mol-1 above the most stable conformation
found. Conformer IIIc features two successive âIII-type turns,
which overlay well with conformer Ia of the monomer.
A similar distribution of conformers is found for the acetyl
protected monomeric unit 2 (Figure 1). Again, the two most
stable conformers, IIa and IIb, each exhibit C10 â-turn structures
and are of similar energy. The most stable conformation that
does not have a C10 â-turn, conformer IIc, lies at an energy 11
kJ mol-1 above the global minimum, close to the energy gap
for the unprotected monomer. The second most stable non-â-
turn conformer is conformer IId; this conformation is stabilized
by a C8 turn, similar to conformer Id. Once again, non-â-turn
conformers are unlikely to be populated in the molecular beam.
Despite the preference for non-â-turn-type structure in the
dimer 3, in even longer carbopeptoid chains, it is likely that
regular extensible hydrogen-bonding patterns such as repeated
â-turns will be favored by co-operative dipole alignment. NMR
investigations of the acetylated tetramer 4 and of other extended
carbopeptoid chains indicated the existence of such repeating
turn structures.4,5
The monomer units are capable of forming different types
of â-turn analogue structure, as measured by the overall bend
in the chain of the backbone. Conformer Ia, for example, is a
âIII-type turn, while the structure of conformer IIb lies closest
to a âII′-turn, see Figure 2. This effects the “angular propaga-
tion” of the continuing backbone chain; in conformers Ia and
IIa, for example, the terminal methyl group is oriented into the
page, but it points out of the page in conformers Ib and IIb, see
Figure 1.
3.2. Spectroscopy. R2PI spectra of 1 and 2 and the associated
IRID spectra, recorded in the N-H and O-H stretch region,
are shown in Figures 4 and 5. Both the R2PI and the IRID
9
16774 J. AM. CHEM. SOC. VOL. 128, NO. 51, 2006

Carbohydrate Amino Acids
A R T I C L E S
Figure 3. Three low-energy conformations of the carbopeptoid dimer 3 with relative energies (kJ mol-1) calculated at the MP2/6-311+G(d,p)//B3LYP/
6-31+G(d) level of theory. The top row shows 3-D molecular representations, while the cartoons below highlight the stabilizing hydrogen-bonding interactions,
indicated by the red arrows. The lowest-energy conformer found that displays a repeating C10 â-turn motif is conformer IIIc.
Figure 4. R2PI spectra of the carbopeptoid monomer units, 1 and 2. Vertical
arrows indicate wavenumbers at which IRID spectra were measured.
spectra of the unprotected monomer unit 1 are well resolved,
and the IRID spectrum clearly shows the expected four bands
(corresponding to two O-H stretches and two N-H stretches).18
Virtually identical IRID spectra were recorded at several UV
probe wavelengths (indicated by the vertical arrows in Figure
4), suggesting that UV bands correspond to the same conforma-
tion, that is, suggesting the population of one predominant
conformer in the molecular beam expansion. The R2PI spectrum
of the acetylated carbopeptoid monomer unit 2 was not well
resolved, and its IRID spectrum was measured with the UV
probe set to the R2PI maximum. The IRID spectrum, displaying
the two bands expected from N-H stretches, indicates the
population of a single predominant conformer at the probe
wavelength, although there may be a hint of a minor contribution
from a second conformer (see later discussion). The overall
shape of the R2PI spectrum of molecule 2 closely resembles
Figure 5. Experimental and computed IRID spectra of the carbopeptoid
that of the better resolved R2PI of molecule 1, which suggests
that it too has only a single conformer.
monomer units, 1 and 2, with calculated relative energies of low-lying
conformers in kJ mol-1
.
The two low-energy, C10 â-turn conformers of the carbopep-
toid unit 1, Ia and Ib (and others illustrated in the Supporting
(18) An overtone of the amide I band with low intensity is also distinguishable
at ∼3420 cm-1
.
9
J. AM. CHEM. SOC. VOL. 128, NO. 51, 2006 16775

A R T I C L E S
Jockusch et al.
Information), each have similar stretching motions: one weakly
H-bonded (weakly perturbed) O-H stretch (O4-H), one
relatively free N-H stretch (NA-H), an O-H stretch that is
strongly perturbed (through an O3H f OdC hydrogen bond),
and an N-H stretch that is strongly perturbed by an NBH f
OdC hydrogen bond, linking the terminal amide groups across
the sugar ring. The latter interaction completes the 10-membered
â-turn. These common interactions generate a characteristic
vibrational signature, as can be seen in the close similarity of
the near IR spectra calculated for each of the two â-turn forming
conformers, see Figure 5, frames Ia and Ib; their vibrational
patterns are both in good agreement with the experimental IRID
signature. In contrast, the O-H and N-H stretches in the low
energy non-â-turn forming structures, conformers Ic and Id (as
well as others illustrated in the Supporting Information), are
involved in different types of interactions and exhibit quite
different spectral patterns that do not look at all like that of the
measured spectrum. The good match between band patterns of
measured and calculated infrared spectra for conformers Ia and
Ib and lack of match for higher-energy conformers gives us
added confidence that important low-energy conformations were
not missed in the original conformational search procedure used
to identify possible conformations. Thus, the experimental IRID
spectrum can be assigned to a â-turn conformational structure
associated with conformer Ia, or possibly Ib, although Ia is the
more favored in view of its lower relative energy.
alter measured IRID intensities. In conformer IIb of the protected
monomer 2, the NAH group is weakly bonded to the carbonyl
oxygen of the neighboring acetyl protecting group, while the
other group, NBH, is involved in the (more strongly perturbing)
NBH f OdC â-turn forming a hydrogen bond. It is possible
that conformer IIa is also present in the molecular beam
contributing to the broad R2PI spectrum displayed in Figure
4;19 note the possible blending of the more strongly shifted (and
broadened) NB-H band at 3350 cm-1. The weak IRID feature
at 3515 cm-1 might be associated with an N-H mode, but it
lies at too high wavenumber to correspond to a “free” NH. Much
more probable is its association with an overtone of the amide
I band, also visible in the IR spectrum of the unprotected
monomer 1 at 3420 cm-1
.
In natural peptide structures, â-turn-type I occurs the most
frequently, but types II and III are also common.20 The favored
conformations of the monomeric carbopeptoid building blocks,
Ia (1) and IIb (2), correspond best to â-turns of type III and
type II′, respectively (see Figure 2 where the structures were
compared to those calculated for the alanine tripeptide), although
for both monomer units, carbopeptoid building block conforma-
tions corresponding to â-turn types I′, II′, and III were also
calculated to lie at relatively low energies. Mons and co-
workers,9,10 who have examined the conformations of several
capped di- and tripeptides, isolated in the gas phase, have found
evidence for multiple varieties of â-turn structure, populated in
the gas phase, including types I, II, and II′ and VIa.
In the protected monomer unit 2, the O-H groups of
molecule 1 are replaced by acetyl protecting groups whose
stretching motions lie at much lower wavenumbers; only the
two N-H stretch modes remain in the IR region scanned, see
Figure 5. The band at 3455 cm-1 indicates a slightly perturbed
N-H vibration, but the band displaced to lower wave number
at 3350 cm-1 indicates a much stronger interaction. The pattern
of the measured IR ion dip spectrum matches well that
calculated for all conformers containing one NH group, which
acts as a hydrogen-bond donor, including conformers IIa, IIb,
and IId. Conformer IIc does not have an NH acting as a
hydrogen-bond donor; consequently, its calculated spectrum
does not resemble the measured spectrum, and significant
population of this conformer can be ruled out. Distinguishing
among other possible conformers is more difficult. The spectral
data alone do not permit an unequivocal conformational
assignment due to the similarity of the calculated spectral
patterns. When the large energy difference between the â-turn
and the non-â-turn structures is taken into account, however,
assignment to one (or more) of the â-turn conformer(s) is, again,
strongly favored. The low-lying conformer IIb provides the
closest match with experiment (see Figure 5), because the higher
wavenumber N-H band appears at 3455 cm-1, ∼35 cm-1 below
that of the corresponding N-H band measured for monomer
1. The shift to lower wavenumber as well as the increased
intensity of the band both suggest an N-H group that is involved
in a more perturbing interaction than the corresponding “free”
N-H group in the unprotected monomer 1. The integrated
intensities of the experimental bands of molecule 2 provide a
reasonable match with the calculated relative intensities of
conformer IIb, although band patterns and relative positions
(here, analyzed in comparison to molecule 1) are a more reliable
indicator of conformation than band intensities because experi-
mental factors such as saturation of the infrared transition can
In â-turn conformers of small peptides, bands corresponding
to the N-H stretches participating in the C10 turns have been
observed between 3460 and 3370 cm-1 10
significantly higher
,
in wavenumber than the corresponding bands observed here for
carbopeptoids 1 and 2, which appear at ∼3300 and 3350 cm-1
.
This suggests a stronger perturbation of the C10 N-H stretches
in the carbopeptoids as compared to the peptides. Calculations
indicate that this is not due just to shorter or more linear H-bonds
in the carbopeptoids; the H‚‚‚O bond lengths calculated for the
â-turn carbopeptoid conformers range between 1.99 Å (Ia) and
2.14 Å (IIb), similar to those calculated for peptides by Mons
and co-workers. The N-H‚‚‚O bond angles, ranging between
150° and 160°, are also similar to or of lower linearity than
those calculated for peptides. An alternative explanation for the
relatively large shifts to lower wavenumber in the carbopeptoids
is the proximity of the oxygen in the furanose ring, which is
2.0-2.2 Å away from the N-H hydrogen, close enough to
confer some additional hydrogen-bonding character. Whatever
the cause, the strong perturbation of this mode in the carbopep-
toids relative to that in peptides is correctly predicted by the
calculations, although the scaling factor (which corrects for
anharmonicity and deficiencies in the computational method)
that is required to reconcile the absolute magnitude of the
calculated and experimental shift between the relatively “free”
N-H and O-H modes and the strongly perturbed modes
increases with the degree of displacement in all of these systems.
In other words, the computational methods used, by our group
as well as others, systematically underestimate the band shift
(19) The broad R2PI spectrum of the protected monomer likely results from
difficulties associated with efficiently cooling the molecule, which are
significantly larger than the unprotected monomer.
(20) Creighton, T. E. Proteins: Structures and Molecular Properties; W. H.
Freeman and Co.: New York, 1984.
9
16776 J. AM. CHEM. SOC. VOL. 128, NO. 51, 2006

Carbohydrate Amino Acids
A R T I C L E S
to lower wavenumbers associated with more strongly bonded
interactions.21
balance among intramolecular interactions, which can determine
conformational choice; unlike the smaller monomer units, the
favored conformation of the dimer does not exhibit â-turns. In
many ways, this reflects the natural situation: the functional
folds of proteins are often only marginally stable, and short
amino acid sequences adopt a variety of conformations in
peptides and proteins, although within this there are known
conformational preferences and structural motifs.
4. Conclusions
The cis-tetrahydrofuran-based carbopeptoid units, 1 and 2,
provide relatively robust peptidomimetic building blocks,
exhibiting â-turn structures when isolated at low temperature
in the gas phase. Calculations also indicate a rich conformational
landscape, however, presenting structures that are quite flexible
and likely to be influenced by interactions with a surrounding
environment as well as by intramolecular interactions with
neighboring groups. Replacement of the terminal phenyl ring
in the unprotected monomer unit 1, for example, by a methyl
group, significantly reduces the gap between the (global
minimum energy) â-turn forming conformer and the lowest-
energy non-â-turn conformer, although the â-turn conformer
remains the most stable structure.
Acknowledgment. We appreciate helpful remarks made by
Dr. Michel Mons and the support provided by the Leverhulme
Trust (Grant No. F/08788D), EPSRC, the Royal Society, for a
USA Research Fellowship (R.A.J.), Linacre College (R.A.J.),
the European Community’s Human Potential Programme under
contract HPRN-CT-2002-00173 (to M.I.S.), the CLRC Laser
Loan Pool, and the Physical and Theoretical Chemistry Labora-
tory at Oxford.
Calculations on the carbopeptoid dimer 3 further exemplify
the flexibility of these structures and reveal a fortiori the delicate
Supporting Information Available: Carbopeptoid confor-
mational and vibrational calculations. Comparison of MMFFs
energy and MP2 energies for molecule 1. Details of synthesis.
Complete ref 17. This material is available free of charge via
the Internet at http://pubs.acs.org.
(21) The tripeptide calculations of Mons and coworkers, which show cosmeti-
cally better agreement, were scaled by a factor of 0.96 (ref 10). All of the
carbopeptoid calculations employed a “weaker” scaling factor of 0.9734,
a value which has provided good agreement with experiment in very weakly
hydrogen-bonded systems (ref 12), but which results in poorer agreement
among bands at lower wavenumber.
JA0607133
9
J. AM. CHEM. SOC. VOL. 128, NO. 51, 2006 16777