Designed peptides with homochiral and heterochiral diproline templates as conformational constraints

Article abstract of DOI:10.1002/chem.200702029

Diproline segments have been advanced as templates for nucleation of folded structure in designed peptides. The conformational space available to homochiral and heterochiral diproline segments has been probed by crystallographic and NMR studies on model peptides containing L-Pro-L-Pro and D-Pro-L-Pro units. Four distinct classes of model peptides have been investigated: a) isolated D-Pro-L-Pro segments which form type II′ βturn; b) D-Pro-L-Pro-L-Xxx sequences which form type II′-I (β_II-I, consecutive β-turns) turns; c) D-Pro-L-Pro-D-Xxx sequences; d) L-Pro-L-Pro-L-Xxx sequences. A total of 17 peptide crystal structures containing diproline segments are reported. Peptides of the type Piv-D-Pro-L-Pro-L-Xxx-NHMe are conformationally homogeneous, adopting consecutive β-turn conformations. Peptides in the series Piv-D-Pro-L-ProD-Xxx-NHMe and Piv-L-Pro-L-Pro-L-Xxx-NHMe, display a heterogeneity of structures in crystals. A type Via β-turn conformation is characterized in Piv-L-Pro-L-Pro-L-Phe-OMe (18), while an example of a 5→1 hydrogen bonded α-turn is observed in crystals of Piv-D-Pro-L-Pro-D-Ala- NHMe (11). An analysis of pyrrolidine conformations suggests a preferred proline puckering geometry is favored only in the case of heterochiral diproline segments. Solution NMR studies, reveal a strong conformational influence of the C-terminal Xxx residues on the structures of diproline segments. In L-Pro-L-Pro-L-Xxx sequences, the Xxx residues strongly determine the population of Pro-Pro cis conformers, with an over-whelming population of the trans form in L-Xxx = L-Ala (19).

Full text of DOI:10.1002/chem.200702029

DOI: 10.1002/chem.200702029
Designed Peptides with Homochiral and Heterochiral Diproline Templates as
Conformational Constraints
Bhaswati Chatterjee,[a] Indranil Saha,[b] Srinivasarao Raghothama,[c]
Subrayashastry Aravinda,[b] Rajkishor Rai,[a] Narayanaswamy Shamala,*[b] and
Padmanabhan Balaram*[a]
Abstract: Diproline segments have
been advanced as templates for nuclea-
tion of folded structure in designed
peptides. The conformational space
available to homochiral and heterochi-
ral diproline segments has been probed
by crystallographic and NMR studies
on model peptides containing l-Pro-l-
Pro and d-Pro-l-Pro units. Four dis-
tinct classes of model peptides have
been investigated: a) isolated d-Pro-l-
Pro segments which form type II’ b-
turn; b) d-Pro-l-Pro-l-Xxx sequences
which form type II’-I (bII’-I, consecutive
b-turns) turns; c) d-Pro-l-Pro-d-Xxx
sequences; d) l-Pro-l-Pro-l-Xxx se-
quences. A total of 17 peptide crystal
structures containing diproline seg-
ments are reported. Peptides of the
type Piv-d-Pro-l-Pro-l-Xxx-NHMe are
conformationally homogeneous, adopt-
ing consecutive b-turn conformations.
Peptides in the series Piv-d-Pro-l-Pro-
d-Xxx-NHMe and Piv-l-Pro-l-Pro-l-
Xxx-NHMe, display a heterogeneity of
structures in crystals. A type VIa b-
turn conformation is characterized in
Piv-l-Pro-l-Pro-l-Phe-OMe (18), while
bonded a-turn is observed in crystals
of Piv-d-Pro-l-Pro-d-Ala-NHMe (11).
An analysis of pyrrolidine conforma-
tions suggests a preferred proline puck-
ering geometry is favored only in the
case of heterochiral diproline segments.
Solution NMR studies, reveal a strong
conformational influence of the C-ter-
minal Xxx residues on the structures of
diproline segments. In l-Pro-l-Pro-l-
Xxx sequences, the Xxx residues
strongly determine the population of
Pro-Pro cis conformers, with an over-
whelming population of the trans form
in l-Xxx = l-Ala (19).
an example of
a
5!1 hydrogen
Keywords: beta turns · conforma-
tion analysis · peptide folding ·
peptides
[a] B. Chatterjee, R. Rai, Prof. P. Balaram
Introduction
Molecular Biophysics Unit, Indian Institute of Science
Bangalore 560012 (India)
Diproline segments have emerged as important templates in
the design of synthetic peptides with defined structures.
While homochiral l-Pro-l-Pro segments have been ad-
vanced as potential nucleator of helical conformations,[1] the
heterochiral d-Pro-l-Pro segments are effective templates
for b-hairpin formation.[2] Proline is unique among the 20
genetically coded amino acids in possessing a covalent link-
age between the side chain and the backbone nitrogen
atom. The direct consequence of the formation of the pyrro-
Fax : (+91)80-2360-0683
E-mail: pb@mbu.iisc.ernet.in
[b] I. Saha, S. Aravinda, Prof. N. Shamala
Department of Physics, Indian Institute of Science
Bangalore 560012 (India)
Fax : (+91)80-23602602
E-mail: shamala@physics.iisc.ernet.in
[c] S. Raghothama
NMR Research Centre, Indian Institute of Science
Bangalore 560012 (India)
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author: Crystal and di-
fraction data and refinement details (Table S2); intramolecular hydro-
gen-bond parameters (Table S3), the puckering analysis of proline
rings (Tables S4–S6) and the packing diagrams for peptides 1–8, 10–
15, 18, 20 and 23 (Figures S7–S11); general procedures for peptide
synthesis of representative compounds 1, 9, 11; m/z values in ESI-MS
spectra (positive ion mode) for the peptides 1–23 (Table S1), ESI-MS
lidine ring, is the restriction imposed on the torsional free-
a
ꢀ
dom about the N C bond, limiting the dihedral angle f to
values of ꢀ60ꢁ308 for l-Pro. The restraint imposed on con-
formational freedom makes proline the most frequently ob-
served residue in turn segments of proteins[3] and also the
most widely used residue in the design of well structured
peptides.[4] Three distinct conformations have been charac-
terized, corresponding to i) polyproline (PII, f ~ꢀ708, y ~
+1208), ii) g-turn (C7, f ~ꢀ708, y ~+708) and iii) helical (aR,
spectra (positive ion mode) for all 1–23 peptides (Figures S1–S6), 13
chemical shifts for the peptides Piv-l-Pro-l-Pro-l-Xxx-NHMe in
CDCl3 (Table S7).
C
6192
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 6192 – 6204

FULL PAPER
f ~ꢀ608, y ~ꢀ308) structures. Proline has been considered
as a secondary structure breaker in proteins and is generally
disfavored in both regular a-helices and b-sheets.[5] This is
largely a consequence of the absence of the NH group, im-
peding intramolecular hydrogen bonding, an interaction
characteristic of polypeptide secondary structures. Proline is
however observed at the N-terminus (N-cap) region of heli-
ces.[6] This is readily rationalized by the fact that the back-
bone torsion angles required for a-helix formation are readi-
ly adopted by proline and that the N-terminus NH group
does not participate in intramolecular hydrogen bonding.
Proline is indeed a more decisive breaker of b-sheets, where
both the hydrogen bonding and torsion angle requirements
are not accommodated. The use of proline containing pep-
tides, which inhibit amyloid fibre formation is a notable ex-
ample in which the poor b-sheet forming propensity of this
residue has been cleverly exploited.[7] In proteins, proline
occurs widely in loops and in the turn regions which facili-
tate polypeptide chain reversal, thus permitting the forma-
tion of globular structures. Two residue turns (b-turns) are
the best characterized short range structural features in pro-
teins. Since their original description by Venkatachalam,
nearly four decades ago, based on an analysis of the sterical-
ly allowed intramolecular hydrogen bonded conformation of
the three linked trans planar peptide units,[8] a wide variety
of b-turn families have been characterized in protein crystal
structures.[3b] Several insights have emerged from this large
body of published work. Notably, l-Pro has the greatest pro-
pensity among the twenty genetically coded amino acids to
occur at the i+1 position of type I/III and type II b-turns.
When proline occurs at the i+2 position, the preceding X-
Pro peptide bond is invariably cis and the resultant intramo-
lecular 4!1 hydrogen bonded structure has been termed as
a type VI b-turn.[3b] These conformational properties of pro-
line residues have stimulated a large number of studies di-
rected towards the use of proline containing sequences in
the “first principles design” of folded peptide structures.
Diproline segments constitute a chain fragment with con-
siderably reduced conformational choices. Figure 1, illus-
trates the allowed regions of f, y space for both l- and d-
proline residues. The possible conformations of both homo-
chiral and heterochiral diproline segments are indicated, al-
lowing a choice of only PII/PII’ and aL/aR conformations. In a
study published as early as 1979 from this laboratory, a two
hydrogen bonded conformation was proposed for the model
sequence Piv-l-Pro-l-Pro-l-Ala-NHMe. The conformation
suggested on the basis of the NMR data available at that
time, consisted of two consecutive b-turns resulting in the
formation of an incipient 310-helical structure.[1a] In this
structure, all the three residues must occupy the aR region
of conformational space. No cis conformer was detected;
leading to the conclusion that the Pro-Pro bond exclusively
adopted a trans geometry. This finding was followed up by
Kemp and co-workers[1c–e,9] who designed a synthetic tem-
plate based on a diproline segment in which two pyrrolidine
rings were covalently linked by a thiomethylene bridge. The
success of Kempꢁs template in nucleating helix formation in
model peptide in aqueous solution is well documented.[1c–e]
Hanessian successfully explored the development of synthet-
ic diproline based organic templates in helix nucleation.[10]
A recent study from our laboratory revisited the use of un-
constrained diproline segments placed at the N-terminus of
the model hexapeptide (Piv-l-Pro-l-Pro-Aib-Leu-Aib-Phe-
OMe) in nucleating folded structures. While helix formation
could be demonstrated over the segments 2 to 6, Pro(1)
adopts a polyproline (PII) conformation in both crystals and
in solution.[1b]
Figure 1. Ramachandran maps for d-Pro and l-Pro residues showing the
available conformational space for proline. Arrows are used to roughly
indicate the locations of the i+1 and i+2 residues in f,y space. The let-
ters shown against the arrows indicate the diproline conformations. The
possible conformation for diproline segments are indicated A) PII–PII, B)
PII–aR, C) aR–aR, D) PII’–aR, E) PII–aL. Note that the aR–PII combination
is not observed in any diproline segment in the PDB.[1b]
The ability of d-Pro-Gly sequences to form type I’/II’ b-
turns, provided the first successful approach to the design of
stable, well characterized b-hairpin structures in short oligo-
peptides.[11] A large body of subsequent work has estab-
lished the utility of centrally positioned d-Pro-Xxx sequen-
ces to facilitate sharp polypeptide chain reversal, thus facili-
tating the design of isolated b-hairpins[2b,4c,d,11e–h,12] and
multi-stranded b-sheets.[2b,11h,13] The heterochiral d-Pro-l-
Pro unit is an extremely efficient nucleator of b-hairpin
structures since the segment has a marked propensity for
type II’ b-turn conformations.[14] Figure 2, illustrates the
crystallographic characterization of a d-Pro-l-Pro segment
as a nucleating unit in a synthetic hairpin and also provides
an example of a l-Pro-l-Pro segment at the N-terminus of a
helical segment in a protein.[15] Marshall and co-workers
have suggested that l-Pro-l-Pro and d-Pro-d-Pro segments
Chem. Eur. J. 2008, 14, 6192 – 6204
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6193

P. Balaram, N. Shamala et al.
Results and Discussion
Attempts were made to obtain diffraction quality single
crystals for all the peptides studied. Heterochiral diproline
peptides appeared to crystallize more readily than their ho-
mochiral counterparts. In some cases, enantiomeric peptides
were synthesized and recrystallization of peptide racemates
was attempted. A brief report on the use of racemic mix-
tures in facilitating recrystallization and exploring multiple
conformational states in proline peptides is presented else-
where.[17] The backbone torsion angles in all the crystalline
peptides and the relevant intramolecular hydrogen bonds
are listed in Table 1. A complete list of hydrogen bond pa-
rameters for the crystal structures is provided as Supporting
Information (Table S3).
Crystal structures of d-Pro-l-Pro segments
Type II’ b-turns: Figure 3 illustrates the molecular confor-
mations observed in crystals of peptides 1–5. Peptide 1, Piv-
d-Pro-l-Pro-NHMe is the prototypic, heterochiral diproline
segment. The structure determination of 1 reveals two mole-
cules in the asymmetric unit both of which adopt a type II’ b
turn conformation, with a single intramolecular 4!1 hydro-
gen bond between Piv C=O and NHMe (Table 1 and Sup-
porting Information, Table S3). Similar conformations are
observed in the tripeptide esters Piv-d-Pro-l-Pro-Xxx-OMe
(Xxx=l-Val 2, l-Phe 3, d-Ala 4). Interestingly, the crystal
structures of peptides 2 and 4 reveal multiple molecules in
the asymmetric unit, three in the case of 2 and two in the
case of 4. A structure determination of peptide 5, Piv-l-Pro-
d-Pro-l-Ala-OMe yielded as anticipated, a type II b-turn
structure (Figure 3e). The well-defined conformation of the
heterochiral diproline segment is apparent when the eight
independent molecules in the structure of peptides 1 to 4
are superposed yielding a RMSD value of 0.211 Š for the d-
Pro-l-Pro segment.
Figure 2. a) b-hairpin conformation of Piv-Leu-Phe-Val-d-Pro-l-Pro-Leu-
Phe-Val-OMe in crystals,[2c] b) Pro–Pro segment in an a-helix. PDB ID:
1QQF A (1250–1251),[15] c) Crystal state conformation of Piv-Pro-Pro-
Aib-Leu-Aib-Phe-OMe,[1b] d) Consecutive b-turn conformation (bII’-I) in
the crystal structure of Piv-d-Pro-l-Pro-l-Ala-NHMe (DPPAN).[14]
are poor turn constraining units, due to enhanced cis–trans
isomerism and small energy differences between turn-like
and extended structures.[16] A recent study from this labora-
tory has demonstrated the ability of the d-Pro-l-Pro-d-Ala
segment to form a three residue connecting loop in a well
registered b-hairpin.[2c] Despite the widespread interest in
diproline segments, systematic experimental investigations
of model systems are lacking.
We present in this report conformational studies on 23
peptides containing diproline segments.[17] The crystal struc-
tures of 17 peptides (1–8, 10–15, 18, 20 and 23) designed to
probe the allowed conformational space for both homochi-
ral and heterochiral diproline units are presented. We also
describe NMR studies, of some chosen sequences, to evalu-
ate the occurrence of cis Pro-Pro conformations and the for-
mation of type VI b-turns. The results demonstrate the im-
portance of the C-terminus flanking residues as conforma-
tional determinants. The peptides examined are: Piv-d-Pro-
l-Pro-NHMe (1); Piv-d-Pro-l-Pro-l-Xxx-OMe [l-Xxx = l-
Val (2), l-Phe (3)]; Piv-d-Pro-l-Pro-d-Ala-OMe (4); Piv-l-
Pro-d-Pro-l-Ala-OMe (5); Piv-d-Pro-l-Pro-l-Xxx-NHMe
[l-Xxx = l-Val (6), l-Leu (7), l-Phe (8), Gly (9), Aib (10)];
Piv-d-Pro-l-Pro-d-Xxx-NHMe [d-Xxx = d-Ala (11), d-Val
(12), d-Leu (13), d-Phe (14)]; Piv-l-Pro-d-Pro-l-Val-OMe
(15); Piv-l-Pro-l-Pro-l-Xxx-OMe [l-Xxx = l-Ala (16), l-
Val (17), l-Phe (18)]; Piv-l-Pro-l-Pro-l-Xxx-NHMe [l-Xxx =
l-Ala (19), l-Val (20), l-Leu (21), l-Phe (22), Aib (23)].
The use of the Piv group at the N-terminus restricts the
Piv-Pro tertiary amide bond to the trans conformation, a
feature essential for the formation of b-turn structures with
Pro(1) occupying the i+1 position.[1a,18]
Consecutive b-turns: Protected tripeptide N-methylamides
of the type Piv-d-Pro-l-Pro-l-Xxx-NHMe fold into a con-
secutive b-turn (bII’-I) conformation because of the presence
of a donor NH group at the C-terminus which permits the
formation of an additional 4!1 hydrogen bond. This struc-
tural feature was first characterized in Piv-d-Pro-l-Pro-l-
Ala-NHMe in crystals[14] (Figure 2d). The conformations de-
termined in crystals for peptides Piv-d-Pro-l-Pro-l-Val-
NHMe (6), Piv-d-Pro-l-Pro-l-Leu-NHMe (7), Piv-d-Pro-l-
Pro-l-Phe-NHMe (8) and Piv-d-Pro-l-Pro-Aib-NHMe (10)
are illustrated in Figure 4. In all the cases, the molecules
adopt the consecutive b-turn conformation (bII’-I), with two
intramolecular 4!1 hydrogen bonds (Table 1 and Table S3
in the Supporting Information). The superposition of the
five independent molecules in Figure 4e with the parent Piv-
d-Pro-l-Pro-l-Ala-NHMe establishes their close conforma-
tional similarity, with an RMSD of 0.387 Š.
6194
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 6192 – 6204

Conformational Constraints
FULL PAPER
Table 1. Torsion angles [8] and intramolecular hydrogen bonds.
Sequence
Residue 1
d-Pro/l-Pro
y
Residue 2
d-Pro/l-Pro
y
Residue 3
Intramolecular 4!1
Hydrogen bonds
Donor Acceptor N···O
Distance [Š]
f
w
f
w
f
y
w
type II’/II b-turns
1
molecule A
molecule B
2
56.3 ꢀ139.2
57.7 ꢀ134.7
178.4 ꢀ83.8
179.9 ꢀ81.5
13.9
5.3
173.1
175.3
–
–
–
–
–
–
N(3)
N(6)
O(0)
O(3)
2.903
2.926
molecule A
molecule B
molecule C
3
61.5 ꢀ133.2
180.0 ꢀ73.2
ꢀ7.3 ꢀ175.8 ꢀ115.5
177.6
178.3 N(3)
O(0)
O(4)
3.013
3.122
3.033
3.052
60.5 ꢀ138.1 ꢀ177.6 ꢀ87.4
59.5 ꢀ139.4 ꢀ179.8 ꢀ80.4
59.8 ꢀ137.6 ꢀ178.1 ꢀ81.4
14.1 ꢀ180.0
ꢀ4.4 ꢀ175.3
0.7 ꢀ179.2
ꢀ95.6
ꢀ72.5
ꢀ83.5
ꢀ46.6 ꢀ175.5 N(7)
135.6
177.2 ꢀ173.4 N(3)
176.8 N(11) O(8)
O(0)
4
molecule A
molecule B
5
consecutive b-turns
6
57.6 ꢀ144.7 ꢀ179.7 ꢀ82.9
57.0 ꢀ143.3 ꢀ179.0 ꢀ82.8
ꢀ0.6 ꢀ178.5
0.7 ꢀ175.2
55.9 ꢀ139.4
178.0 N(3)
O(0)
O(00)
O(0)
3.095
3.114
3.147
63.4 ꢀ149.9 ꢀ175.1 N(6)
ꢀ64.8
137.3
177.8
85.3
ꢀ6.1
ꢀ4.6
176.0
ꢀ68.9
160.4
176.2 N(3)
57.0 ꢀ135.3 ꢀ170.9 ꢀ72.7
175.8 ꢀ116.8
17.5
172.5 N(3)
N(4)
O(0)
O(1)
3.168
3.148
7
molecule A
65.0 ꢀ135.7 ꢀ178.8 ꢀ69.3
56.9 ꢀ141.7 ꢀ178.9 ꢀ68.3
62.1 ꢀ141.0 ꢀ175.2 ꢀ68.3
ꢀ6.7
ꢀ9.5
172.2
174.9
176.7
179.4
ꢀ92.7
ꢀ85.1
ꢀ75.5
ꢀ60.2
4.3
177.3 N(3)
N(4)
177.3 N(7)
N(8)
O(0)
O(1)
O(4)
O(5)
O(0)
O(1)
O(0)
O(1)
3.024
2.974
3.051
3.019
3.144
3.109
3.262
3.445
molecule B
ꢀ6.6
8
ꢀ12.3
ꢀ16.0
ꢀ17.9 ꢀ179.9 N(3)
N(4)
ꢀ35.7 ꢀ177.5 N(3)
N(4)
10
59.5 ꢀ153.6
177.7 ꢀ75.2
effect of residue 3[b]
11[a]
12
60.7 ꢀ153.5 ꢀ180.0 –85.7
66.4 ꢀ141.2 8.4 ꢀ78.1
30.5
158.5
168.8
170.8
129.4
34.1
175.3 N(3)
O(0)
–
3.291
–
77.7 ꢀ159.2 ꢀ176.8
–
94.1 ꢀ167.5 ꢀ179.9
13[17]
68.5 ꢀ165.5 ꢀ178.4 ꢀ64.8
137.7 ꢀ178.4
139.5 168.9
78.1 ꢀ153.0 ꢀ169.9
89.6
ꢀ2.7 ꢀ178.4 N(4)
O(1)
–
–
3.184
–
–
14[17]
63.9 ꢀ141.0
12.6 ꢀ76.9
ꢀ9.4
118.7 ꢀ151.8 ꢀ175.7
ꢀ87.1 169.4 176.8
–
–
15
ꢀ62.8
143.0
homochiral l-Pro-l-Pro sequences
18
molecule A
molecule B
20
23
ꢀ54.5
ꢀ52.4
ꢀ75.3
ꢀ95.9
141.4
141.4
137.4
173.3
11.0 ꢀ88.5
7.5 ꢀ90.8
2.9
8.1
177.6
168.3
ꢀ74.4 ꢀ156.3
ꢀ74.8 ꢀ3.1 ꢀ177.2 N(7)
116.0 ꢀ179.2
34.5 172.8 N(4)
177.9 N(3)
O(0)
O(4)
–
2.865
2.857
–
170.5 ꢀ63.5
168.8 ꢀ57.1
ꢀ30.5 ꢀ175.7 ꢀ135.5
–
136.4 174.1 57.1
O(1)
3.273
[a] In addition to the 4!1interaction there is also 5!1 intramolecular hydrogen bond between N(4) and O(0) with a N···O distance of 2.849 Š.
[b] Effect on the configuration of d-Pro-l-Pro segment.
Effect of residue 3 configuration on d-Pro-l-Pro conforma-
tions: Figure 5a shows the molecular conformation of Piv-d-
Pro-l-Pro-d-Ala-NHMe (11). The d-Pro(1)-l-Pro(2) seg-
ment in this structure forms a type II’ b-turn conformation
as anticipated, with a 4!1 (C10) hydrogen bond between
the Piv C=O and d-Ala(3) NH groups. The d-Ala residue
adopts an unusual conformation (f=129.4, y=34.18),
which lies in the sterically allowed region of f,y space. This
results in the formation of a strong 5!1 intramolecular hy-
drogen bond between the Piv C=O and methylamide NH
groups (N···O=2.849 Š). Consideration of the structure of
the peptide suggests that although the Piv C=O group par-
ticipates in bifurcated interactions with two hydrogen bond
donors, the dominant interaction is of the 5!1 (C13) type.
Such structures have been observed in proteins and are clas-
sified as a-turns.[19] Isolated a-turn structures have thus far
not been observed in crystal structures of short peptides and
peptide 11 provides the first example of such a conforma-
tional feature. An enantiomeric peptide Piv-l-Pro-d-Pro-l-
Ala-NHMe was synthesized and yielded single crystals with
identical cell dimensions suggesting an identity of molecular
conformation. The structure of peptide 11 may be compared
with the methyl ester analog Piv-d-Pro-l-Pro-d-Ala-OMe
(5) in which a single d-Pro(1)-l-Pro(2) type II’ b-turn is ob-
served. Clearly, the additional hydrogen bond donor in pep-
tide 11 is an important conformational determinant.
The peptide Piv-d-Pro-l-Pro-d-Val-NHMe[17] (12) pro-
vides an example of an unusual conformation in which the
d-Pro-l-Pro peptide bond adopts a cis geometry, with the
absence of any intramolecular hydrogen bonding interaction
(Figure 5b). Surprisingly, the enantiomeric peptide Piv-l-
Pro-d-Pro-l-Val-NHMe[17] reveals the expected type II b-
turn conformation at the l-Pro-d-Pro segment (Figure 5c).
Undoubtedly, the hydration of central peptide unit observed
Chem. Eur. J. 2008, 14, 6192 – 6204
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6195

P. Balaram, N. Shamala et al.
turn (Figure 5e). Instead, the l-Pro-d-Leu segment adopts a
type II b-turn conformation with d-Pro(1) taking up a d-
ProII conformation (f=68.88, y=ꢀ167.58). Notably, in the
d-Pro-l-Pro-d-Xxx-NHMe series, completely distinct con-
formers have been characterized for d-Xxx = d-Ala, d-Val
and d-Leu. In contrast, in the case of d-Pro-l-Pro-l-Xxx-
NHMe series, all the peptides studied (l-Xxx = l-Ala, l-
Val, l-Phe, l-Leu and Aib) yielded the same conformation
in crystals. In attempting to obtain suitable single crystals
for X-ray diffraction studies, we have investigated the utility
of racemic mixtures formed by using equimolar amounts of
pure peptide enantiomers. The crystal structures of three
racemic peptides in which the residue configurations alter-
nate (d-l-d/l-d-l) that have been discussed elsewhere may
provide a means of exploring polymorphic forms in which
multiple conformational states may be definitively charac-
terized.[17] Notably, for the racemic mixture involving l/d-
Phe[17] (14), the asymmetric unit contains a single peptide
molecule, which adopts an open structure lacking any intra-
molecular hydrogen bond (Figure 5f).
Homochiral l-Pro-l-Pro sequences: The structures of three
independent peptides Piv-l-Pro-l-Pro-l-Phe-OMe (18), Piv-
l-Pro-l-Pro-l-Val-NHMe (20) and Piv-l-Pro-l-Pro-Aib-
NHMe (23) are shown in Figure 6. In peptide 18, both mole-
cules in the crystallographic asymmetric unit adopt a type
VIa b-turn conformation[3c,4b,20] (fi+1 =ꢀ608, yi+1 =1208;
fi+2 =ꢀ908, yi+2 =08), in which the l-Pro-l-Pro peptide
bond is cis (w=11.08 in the case of molecule A and 7.58 in
the case of molecule B). A 4!1 hydrogen bond between
the Piv C=O and the Phe(3) NH is observed (Table 1 and
Supporting Information, Table S3). A similar conformation
has been characterized in the peptide ferrocenyl-Pro-Pro-
Phe-OBzl.[21] Interestingly, earlier studies on small linear
peptides and protein structures have revealed a tendency for
the Xaa-Pro peptide bond to adopt a cis geometry when the
Figure 3. Molecular conformation of peptides in crystals containing heter-
ochiral diproline segments a) Piv-d-Pro-l-Pro-NHMe (1), b) Piv-d-Pro-l-
Pro-l-Val-OMe (2), c) Piv-d-Pro-l-Pro-l-Phe-OMe (3), d) Piv-d-Pro-l-
Pro-d-Ala-OMe (4), e) Piv-l-Pro-d-Pro-l-Ala-OMe (5), f) superposition
of the structures of peptides 1–5.
proline residue is preceded by an aromatic residue.[22]
A
structure devoid of any intramolecular hydrogen bonds is
observed in peptide 20. Pro(1) adopts a PII conformation
with Pro(2) lying in the helical aR region and Val(3) in the
extended b-sheet region. In peptide 23, a Pro(2)-Aib(3) type
II b-turn is observed with a 4!1 intramolecular hydrogen
bond between Pro(1) C=O and the methylamide NH group.
Pro(1) adopts a conformation which is considerably distort-
ed from that observed for proline residues (f=ꢀ95.98, y=
173.38). In this case the Pro(1) ring exhibits an unusual
puckering (c1 =37.88, c2 =ꢀ33.98, c3 =17.08, c4 =7.58, q=
ꢀ28.58 for the first proline ring and c1 =ꢀ22.48, c2 =35.68,
c3 =ꢀ34.58, c4 =21.28, q=0.68 for the second proline ring).
Notably, in all the three cases of homochiral tripeptides 18,
20 and 23 a water of hydration is observed in the crystals.
This lone water molecule is hydrogen bonded to the Pro(1)
C=O group in all the three peptides. In the case of peptide
18, the asymmetric unit contains two molecules of the pep-
tide and only one molecule of water. In peptide 20, the
water molecule interacts with both the free NH groups of
Val(3) and the methylamide group. In peptide 23, the intra-
in the d-l-d peptide 12, result in an unfolded backbone con-
formation. In order to evaluate the role of terminal methyla-
mide NHMe group, the crystal structure of the correspond-
ing tripeptide ester Piv-l-Pro-d-Pro-l-Val-OMe (15) was
solved. The molecular conformation shown in Figure 5d re-
veals a cis l-Pro-d-Pro bond with the absence of any intra-
molecular hydrogen bond. This conformation is remarkably
similar to that observed in the tripeptide (Piv-d-Pro-l-Pro-
d-Val-NHMe). The crystal structure determination of a crys-
talline racemate containing both the l-d-l and the d-l-d
peptides revealed only the type II/II’ b-turn conformation
with the enantiomeric molecules related by inversion sym-
metry.[17] These observations suggest that the multiple con-
formational states that exist in solution may crystallize pref-
erentially under specific conditions. Interestingly, the crystal
structure of the peptide Piv-d-Pro-l-Pro-d-Leu-NHMe[17]
(13) does not show the anticipated d-Pro-l-Pro type II’ b-
6196
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 6192 – 6204

Conformational Constraints
FULL PAPER
molecular b-turn hydrogen
bond appears weak with
a
N···O distance of 3.273 Š, pre-
sumably as a consequence of a
strong bifurcated interaction of
Pro(1) C=O with the hydrating
water molecule.
Conformations of the proline
ring: The puckering of the
five-membered
pyrrolidine
ring in proline residues has
been the subject of investiga-
tion for almost four deca-
des.[23g]
Important
studies
which define proline ring con-
formations merit mention. In
1970 Ramachandran and co-
workers[23a] established two
major conformational states
which describe the relative po-
sitions of the Cg and the C’
atoms with respect to the
mean plane formed by the
Figure 4. Molecular conformation of peptides in crystals containing consecutive b-turns a) Piv-d-Pro-l-Pro-l-
Val-NHMe (6), b) Piv-d-Pro-l-Pro-l-Leu-NHMe (7), c) Piv-d-Pro-l-Pro-l-Phe-NHMe (8), d) Piv-d-Pro-l-Pro-
Aib-NHMe (10), e) superposition of the structures of peptides 6–10 and DPPAN.
other four atoms of the five-membered ring as A (Cg/C’
atoms on opposite side, Cg-exo) and B (Cg/C’ atoms on same
side, Cg-endo). Subsequently, an incisive study by Ashida
and Kakudo[23b] described Cs-Cg-exo, Cs-Cg-endo, Cs-Cb-exo,
Cs-Cb-endo, C2-Cg-exo-C2-Cb-endo and C2-Cg-endo-C2-Cb-exo.
In an important contribution Scheraga and co-workers[23c,d]
defined these states using the nomenclature up and down. A
generalized description of five-membered ring conforma-
tions introduced by Cremer and Pople[23e] which uses two
parameters (the amplitude q and the phase angle f) has
sometimes been used for the description of the proline ring.
Figure 5. Molecular conformation of peptides in crystals containing a d-
residue succeeding the heterochiral diproline segments a) Piv-d-Pro-l-
Pro-d-Ala-NHMe (11), b) Piv-d-Pro-l-Pro-d-Val-NHMe (12),[17] c) Piv-l-
Pro-d-Pro-l-Val-NHMe,[17] d) Piv-l-Pro-d-Pro-l-Val-OMe (15), e) Piv-d-
Pro-l-Pro-d-Leu-NHMe,[17] (13), f) racemic mixture of Piv-d-Pro-l-Pro-
d-Phe-NHMe + Piv-l-Pro-d-Pro-l-Phe-NHMe[17] (14). Only the d-l-d
enantiomer has been shown.
Figure 6. Molecular conformation of peptides in crystals containing ho-
mochiral diproline segments a) Piv-l-Pro-l-Pro-l-Phe-OMe (18), b) Piv-
l-Pro-l-Pro-l-Val-NHMe (20), c) Piv-l-Pro-l-Pro-Aib-NHMe (23).
Chem. Eur. J. 2008, 14, 6192 – 6204
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6197

P. Balaram, N. Shamala et al.
A study published in 1977 by De Tar and Luthra[23f] effec-
tively identified two major states of the proline ring. More
recent theoretical[23h] and database analysis[23i,j] studies have
focused on an analysis of the coupling between the proline
ring geometry and the polypeptide backbone. While many
descriptions of ring geometry have been advanced, it is con-
venient to adopt a classification scheme which is based on
endocyclic torsion angles since these readily permit visuali-
zation of the distortions of the five-membered pyrrolidine
ring from planarity. The availability of a large number of
structurally characterized diproline segments in the present
study prompted us to revisit the analysis of the proline ring
geometry. In particular, most observed forms of the proline
ring are simply visualized by considering the Ca-N-Cd plane
as a reference and describing the positions of the Cb and the
Cg atoms with respect to this plane. For example, values of
c4 close to zero immediately indicate that the Cg atom lies in
the plane and that the puckering must be localized at Cb.
Similarly, a value of q close to zero is an indicator of Cg
puckering with the Cb atom lying in the plane. Cases where
both c4 and q are greater than 108 are immediately diagnos-
tic of twisted proline geometry, in which both Cb and Cg
atoms lie out of plane. A few cases are also observed where
c4 ~q ~08, which corresponds to an almost flattened, nearly
planar ring conformation.
Peptide crystal structures, determined at atomic resolution
provide a wealth of data on proline rings. Table 2 summariz-
es the observations on proline rings in diproline segments
characterized crystallographically. A total of 44 diproline
segments have been examined in acyclic compounds out of
which 21 have been extracted from the Cambridge Crystal-
lographic Database.[24] The classification of ring conforma-
tions in individual residues and in diproline segments are
given as Supporting Information (Tables S4–S6). Investiga-
tion of the ring conformations reveals six distinct conforma-
tional states. Two states which are maximally populated cor-
respond to displacement of only one of the ring atoms from
the mean plane, that is, Cs-Cg-exo(19) and Cs-Cb-exo(21).
Significant populations of four more states are also noted.
These are Cs-Cg-endo, two twisted states in which both the
Cg and the Cb atoms are move out of the plane and a near
planar geometry of the five-membered ring. Significantly, an
almost planar five-membered ring is observed in as many as
13 out of 88 proline rings. Indeed a planar proline ring has
also been characterized in a high resolution structure of the
protein, triosephosphate isomerase.[25] Interestingly, there is
no example of a state which can be characterized as Cs-Cb-
endo. Examination of the conformations of proline rings in
diproline segments does not reveal any dramatic preference
for a specific state, in the case of homochiral segments.
However, in the case of heterochiral diproline segments
there is a preponderance of the Cg-exo/Cb-exo combination.
Figure 7 illustrates the six major classes of observed ring
conformational states. It is necessary to draw attention to
the structure of Boc-Pro-Pro-OH (CSD ID-BOCPRO01) in
which four independent dipeptide molecules constitute the
crystallographic asymmetric unit. While molecules 3 and 4
have been taken into the dataset for the preceding analysis,
molecules 1 and 2 are considered separately. In this case,
the following ring torsion angles are observed [molecule 1:
Pro(1): c1 =25.88,c2 =ꢀ34.48, c3 =29.78, c4 =ꢀ13.38, q=
ꢀ7.58, Pro(2): c1 =24.68,c2 =ꢀ15.28, c3 =0.08, c4 =16.68, q=
ꢀ25.88; molecule 2: Pro(1): c1 =26.38,c2 =ꢀ33.88, c3 =27.18,
c4 =ꢀ10.38, q=ꢀ9.78, Pro(2): c1 =24.18,c2 =ꢀ13.98, c3 =0.28,
c4 =14.48, q=ꢀ21.98]. In both the molecules Pro(1) is classi-
fied as twisted Cg-endo. However, Pro(2) in both the cases
has a c3 ~08 which is directly indicative of the fact that the
atoms N-Cd-Cg and Cb lie in a plane with the ring being dis-
torted by the movement of Ca atom out of the plane
(Figure 7). This is a rare example of a geometry which arises
from puckering at Ca. The pyrrolidine ring conformations
within diproline segments do not show a strong co-relation
to the backbone structural feature in which they are found.
It is likely that five membered rings have a considerable
plasticity of structure and are readily deformed in order to
accommodate a variety of energetically preferred backbone
conformations.
Solution conformations of peptides with diproline segments:
Crystallographic studies, described in the preceding section,
establish distinct conformational preferences in model pep-
tides containing d-Pro-l-Pro and l-Pro-l-Pro sequences. An
important feature to emerge from these studies is that con-
formational diversity is observed in the solid state, especially
in the case of d-Pro-l-Pro-d-Xxx and l-Pro-l-Pro-l-Xxx se-
quences, suggesting that in solution, multiple conformational
states are almost certainly likely to be populated. NMR
studies have therefore, been undertaken in order to probe
the nature of the conformations populated in solution, in or-
ganic solvents. The existence of hydrogen bonded conforma-
tions have been probed using solvent dependence of amide
NH chemical shifts, while specific nuclear Overhauser ef-
fects (NOEs) are used as a diagnostic for determining local
residue conformations.
Heterochiral d-Pro-l-Pro sequences: Table 3 lists the ob-
served chemical shift for the two amide protons and the
values for the change in chemical shift on going from pure
CDCl3 to a mixture containing an appreciable concentration
of high [D6]DMSO (21.7% v/v). The addition of varying
concentrations of the strongly hydrogen-bonded solvent
DMSO to the peptides in the poorly interacting solvent,
CDCl3 is expected to cause a large downfield shift of the
solvent exposed protons as a consequence of the interaction
with the added solvent. Inspection of Dd values listed in
Table 3 clearly reveals that in the d-Pro-l-Pro-l-Xxx series,
where l-Xxx = l-Val (6), l-Leu (7), l-Phe (8), Gly (9), Aib
(10), the Dd values are exceedingly small (0.02 to 0.11 ppm).
This strongly suggests that both the NH groups are solvent
shielded, supporting their involvement in strong intramolec-
ular hydrogen bonding. These results suggest that the con-
secutive type II’–I b-turn structures observed in crystals of 6,
7, 8 and 10 are indeed maintained in solution. NOE studies
of the peptide Piv-d-Pro-l-Pro-l-Phe-NHMe (8), reveals in-
6198
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 6192 – 6204

Conformational Constraints
FULL PAPER
Table 2. Puckering states of the proline ring.
Ring conformation
Number of
proline rings
Average ring torsional parameters [8]
Combination
No. of examples
No. of examples in
heterochiral diproline
segments[a]
[a]
in homochiral diproline
segments[a]
A)
Cs-Cg-exo
or
8+11=19
7+1=8
c1 =ꢀ20.4, c2 =32.9, c3 =ꢀ32.0,
c4 =19.5, q=0.4
AC
BB
AG
2
–
–
4+1
2
1 +1
Cg-exo
B)
Cs-Cg-endo
or
GC
BC
BF
GF
EC
1
–
–
3
1
1
2
2
1
c1 =22.6, c2 =ꢀ33.1, c3 =30.2,
c4 =ꢀ16.0, q=ꢀ4.0
Cg-endo
2+1
C)
Cs-Cb-exo
11+10=21
c1 =34.1, c2 =ꢀ36.7, c3 =24.3,
c4 =ꢀ2.2, q=ꢀ20.1
GE
AF
FC
1
–
–
–
2
2+1
D)
Cs-Cb-endo
–
–
AB
CC
1
1
–
–
E)
twisted Cg-exo-Cb-endo
or
FA
AE
FF
1
3
1
–
–
–
–
1
5+5=10
c1 =ꢀ29.7, c2=38.7, c3 =ꢀ32.2,
c4 =13.9, q=9.8
C2-Cg-exo-C2-Cb-endo
AA
F)
twisted Cg-endo-Cb-exo
or
FC
GG
EF
1
1
–
–
–
–
1
1
9+8=17
6+7=13
c1 =30.5, c2 =ꢀ38.1, c3 =30.4,
c4 =ꢀ11.6, q=ꢀ11.7
C2-Cg-endo-C2-Cb-exo
EG
G)
planar
c1 =6.2, c2 =ꢀ5.1, c3 =1.7,
c4 =2.6, q=ꢀ5.5
CA
1
–
[a] Numbers highlighted in bold font are examples from CSD.
the Pro(2)CaHQPhe(3)NH and Phe(3)NHQNHMe NOEs
are consistent with b-turn conformations.
Interestingly, in the case of the d-Pro-l-Pro-d-Xxx-NHMe
series, the Dd values observed for d-Xxx NH are extremely
small (ꢀ0.06 to 0.10 ppm), while the values for methylamide
NH are appreciably larger (0.26 to 0.36 ppm) indicative of a
greater degree of solvent exposure of the C-terminal amide.
Table 3. NMR parameters for the peptides Piv-d-Pro-l-Pro-l-Xxx-
NHMe and Piv-d-Pro-l-Pro-d-Xxx-NHMe.
[a]
Chemical shift [ppm]
[b]
Residues
Xxx
NH
Dd [ppm]
Xxx
Xxx
NHMe
NHMe
Piv-d-Pro-l-Pro-l-Xxx-NHMe
Gly (9)
Aib (10)
l-Val (6)
l-Leu (7)
l-Phe(8)
7.84
6.85
6.99
7.25
7.38
7.01
6.96
6.76
6.91
6.85
0.06
0.02
0.04
0.06
0.04
0.11
0.05
0.08
0.07
0.08
Piv-d-Pro-l-Pro-d-Xxx-NHMe
Figure 7. Puckering states of proline rings.
d-Ala (11)
d-Val (12)
d-Leu (13)
d-Phe(14)
7.52
7.62
7.48
7.46
7.01
7.02
7.11
6.99
0.03
ꢀ0.04
0.04
0.27
0.28
0.26
0.28
0.08
tense NOEs between Pro(1) CaHQPro(2)CdH confirming
the trans geometry about the Pro(1)-Pro(2) bond. Further,
[a] CDCl3. [b]Dd is chemical shift difference for NH protons in CDCl3
and 21.7% [D6]DMSO/CDCl3 (v/v) solutions.
Chem. Eur. J. 2008, 14, 6192 – 6204
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6199

P. Balaram, N. Shamala et al.
Significantly, the crystal structure determination of the three
members of this group d-Xxx = d-Ala (11), d-Val (12), d-
Leu (13) revealed three distinct structures. In the case of the
peptide d-Xxx = d-Ala (11), both the amide NH groups
are involved in intramolecular hydrogen bonding. But in the
case of d-Xxx = d-Val (12), neither of the amide NH(s)
were hydrogen bonded, while in d-Xxx = d-Leu (13), a
single intramolecular hydrogen bond involving a methyl
amide NH group was observed. The NMR results suggest
that in solution, the d-Pro-l-Pro b-turn is stabilized by a 4!
1 hydrogen bond between Piv C=O and d-Xxx NH group
and is maintained in all the cases with a degree of conforma-
tional variability involving the C-terminal residue. Notably,
in the three crystal structures which have been determined
in this series the d-Xxx, d-Ala residue, adopts a conforma-
tion of (f=1298 and y=348) in the relatively unpopulated,
but nevertheless allowed, regions of the Ramachandran
map. In contrast, the d-Leu residue in peptide 13, adopts an
aL conformation, while the d-Val residue in peptide 12
adopts an extended conformation, (f ~90–1108 and y ~140–
1608), in two distinct structures observed in pure enantio-
mers and racemic mixtures. Thus, the heterogeneity of the
conformations in the d-Pro-l-Pro-d-Xxx-NHMe series in-
volving residue 3 observed in the crystalline state undoubt-
edly manifests itself also in solution, as evidenced by moder-
ately high Dd values, observed in the solvent titration ex-
periments. NOE studies in the d-Pro-l-Pro-d-Xxx-NHMe
Homochiral l-Pro-l-Pro sequences: The delineation of sol-
vent shielded NH groups was carried out by monitoring the
downfield shift of NH resonances upon addition of strongly
hydrogen bonded solvent [D6]DMSO to the peptide solution
in CDCl3. Representative solvent titration curves are shown
ꢀ
in the Figure 8. The population of cis conformer about Pro
Pro bond was evident in all the sequences by the appearan-
ces of additional resonances. The assignment of the resonan-
ces of the cis form is based on the observation of a NOE be-
tween Pro(1) and Pro(2) CaH protons. Further, HSQC
(1H–13C) spectra of peptides 19 and 22, permitted the assign-
ment of Cb and Cg carbon resonances. In the case of the cis
conformer of peptide 22, the Cb resonances of Pro(2) ap-
pears at 31.6 ppm while Cg resonances appear at 21.2 ppm.
In contrast, trans form of peptide 19, the Cb chemical shift
of Pro(2) is at 29.3 ppm while Cg appears at 24.2 ppm (see
b
g
ꢀ
Supporting Information, Table S7). A large Dd (C C ) of
~10 ppm is characteristic of a cis X-Pro bond.[26] Table 4
summarizes the 1H NMR parameters determined for the
series of peptides Piv-l-Pro-l-Pro-l-Xxx-NHMe/OMe. In a
study of l-Xxx= l-Ala peptide (19) published over 25 years
ago, two strongly hydrogen bonded NH groups were identi-
fied and the consecutive type III–III b-turn conformation
was assigned. Reinvestigation of this peptide, synthesized
afresh, revealed that both NH groups are involved in intra-
molecular hydrogen bonding. Furthermore, the NMR spec-
trum revealed an overwhelming presence of trans confor-
mer, now conclusively established by the Pro(1)CaH
QPro(2)CdH NOEs. Careful examination of the spectrum
reveals that the cis conformer is populated to an extent of
5%. Interestingly, in all other peptides in these series, a sig-
nificantly higher proportion of cis conformer was observed.
series of the peptides revealed strong Pro(1) CaH–Pro(2)
d
ꢀ
C H NOEs, confirming that the Pro(1) Pro(2) bond is pre-
dominantly trans. Indeed, there is little evidence for the
presence of a minor cis conformation in all the peptides
studied. Curiously, the structure of the pure enantiomer,
Piv-d-Pro-l-Pro-d-Val-NHMe
(12) determined in single crys-
tals obtained from the mixture
of a ethyl acetate and petrole-
ꢀ
um ether reveals a cis Pro(1)
Pro(2) peptide bond in the
crystalline state. In peptides 11
and 12, relatively intense
NOEs were observed between
Pro(2) CdHQXxx NH, Pro(2)
CaHQ Xxx NH and Xxx
NHQNHMe. The simultaneous
observation of these NOEs is a
clear indication of the popula-
tion of multiple conformational
states in solution. While the d-
Pro-l-Pro sequence appears to
have a strong propensity for
forming type II’ b-turns, the
Pro(2)-d-Xxx(3) segment is
less constrained, with the
nature of d-Xxx residue being
a determining feature.
Figure 8. Representative experiments showing the solvent dependence of NH chemical shifts at varying con-
centrations of [D6]DMSO in CDCl3 to probe solvent exposed verses hydrogen bonded amides. * Corresponds
to cis conformation. a) 19, b) 20, c) 21, d) 23.
6200
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 6192 – 6204

Conformational Constraints
FULL PAPER
Table 4. NMR parameters for the peptides Piv-l-Pro-l-Pro-l-Xxx-NHMe/OMe in CDCl3 and [D6]DMSO[a]
solutions.
(22). In the case of Xxx=Aib
(23), it is the Aib NH which
shows a larger solvent depend-
ence (Dd=1.41 ppm), while
the methylamide NH is hydro-
gen bonded (Dd=0.30 ppm).
This is undoubtedly a conse-
quence of the formation of a
Pro(2)-Aib(3) b-turn confor-
mation, which has been widely
characterized in crystal struc-
ture of short peptides.[1b] In all
cases, the extent of solvent de-
pendence in the case of cis
conformers is much less for
both the sets of NH protons. A
type VI b-turn conforma-
tion,[22f,h,27] which might be an-
ticipated with l-Pro-l-Pro-l-
Xxx sequences with Pro-Pro
cis geometry, is expected to
result in solvent shielding of
the l-Xxx NH protons.
Chemical shift [ppm]
Residues
(l-Xxx)
NH
Dd
[ppm][b]
ACHTREUNG
N
Conformers
Xxx
7.33
NHMe
6.97
Xxx
NHMe
cis [%]
DG[c]
cis/trans
l-Ala-NHMe (19)
l-Val-NHMe (20)
l-Leu-NHMe (21)
l-Phe-NHMe (22)
Aib-NHMe (23)
l-Ala-OMe (16)
l-Val-OMe (17)
l-Phe-OMe (18)
trans
0.30
–
–
–
0.11
–
0.20
–
0.22
–
0.34
–
0.64
–
0.49
–
1.41
–
0.26
–
0.59
–
0.05
–
–
–
1.12
–
0.19
–
0.69
–
0.21
–
0.74
–
0.20
–
0.30
–
0.10
–
–
–
–
–
–
5
(–)[d]
–
1.76
N
A
(–)[d]
cis
trans
cis
6.65
(–)[d]
6.29
–
–
(–)[d]
7.20
(7.61)
–
54
(11)
–
ꢀ0.09
G
A
ACHTRE(UNG 1.25)
7.50
7.21
7.61
6.55
7.57
6.45
7.23
6.61
7.01
6.58
7.20
–
–
(8.15)
A
–
trans
cis
53
(19)
–
ꢀ0.07
(7.72)
(7.67)
ACHTRE(UNG 0.87)
–
–
(8.41)
(7.53)
–
trans
cis
62
(31)
–
ꢀ0.29
(7.76)
A
ACHTRE(UNG 0.48)
–
–
0.45
ACHTRE(UNG 1.46)
–
–
(8.53)
A
–
trans
cis
7.25
7.15
32
(8)
–
(8.21)
A
7.20
7.12
8.98
7.34
8.92
7.13
(7.90)
A
–
trans
cis
–
(–)
–
(–)
–
(–)
–
(–)
–
56
(21)
–
ꢀ0.14
(8.20)
ACHTRE(UNG 0.79)
ꢀ0.03
–
–
0.79
ACHTRE(UNG 1.25)
–
–
A
notable feature of the
(8.82)
–
–
data in Table 4 for the tripep-
tide N-methylamide in CDCl3
is that the percentage of the cis
form exceeds that of the trans
conformer in case of l-Xxx=
l-Val (20), l-Leu (21) and l-
Phe (22). In contrast, for
Xxx=Aib (23), there is consid-
erable reduction in the popula-
tion of cis form and most dra-
matically for l-Xxx = l-Ala
(19), the cis population drops
to ~5%. Thus, the nature of l-
trans
cis
0.30
–
21
(11)
–
(8.01)
–
–
–
–
–
–
–
ꢀ0.04
(8.69)
–
–
trans
cis
0.37
–
67
(33)
–
ꢀ0.42
R
(–)
–
(–)
ACHTRE(UNG 0.42)
ꢀ0.01
–
–
C
–
–
[a] The values in parentheses correspond to those in [D6]DMSO. [b]Dd is chemical shift difference for NH
protons in CDCl3 and 21.7% [D6]DMSO/CDCl3. [c]DG [kcalmolꢀ1] at 300 K. [d]Below the observable limit
of the present measurement.
Indeed, in the case of l-Xxx = l-Val (20), l-Leu (21) and l-
Phe (22), the cis conformer is populated to a greater extent
than the trans form. Figure 9 provides a comparison of par-
tial NOE spectra, CaHQCaH region which highlights the
difference between l-Ala (19) and l-Phe (22) peptides. The
crystal structure of Piv-l-Pro-l-Pro-l-Phe-OMe (18) re-
vealed that both molecules in the asymmetric unit adopt
Xxx residue has a profound effect on free energy differences
between the Pro-Pro cis and trans form. Interestingly, in
[D6]DMSO, there is a substantial reduction in the popula-
tion of cis conformer in all the cases. Indeed, in case of l-
Xxx= l-Ala, the cis form is undetectable. In earlier studies
of model peptides the population of X-Pro cis conformer
has been shown to increase in polar solvents.[28] Clearly, in
the series Piv-l-Pro-l-Pro-l-Xxx-NHMe, the stabilization of
the cis form in an apolar solvent like CDCl3 is undoubtedly
driven by the favourable energy of formation of an intramo-
lecular 4!1 hydrogen bond in a type VI b-turn conforma-
tion. The effect of the alanine side chain is particularly note-
worthy, in that an all trans, incipent 310 helix is stabilized. In-
terestingly, in the corresponding ester Piv-l-Pro-l-Pro-l-
Ala-OMe, the cis conformer is populated to the extend of
56% in CDCl3, clearly suggesting that the formation of
second hydrogen bond is critical in promoting an all trans
structure in the corresponding N-methylamide (19).
ꢀ
type VIa conformations, with Pro(1) Pro(2) having a cis ge-
ometry.
The solvent dependence data in Figure 8 (see also Dd
values in Table 4), revealed that in the trans form, two sol-
vent shielded NH groups appear to be present only in the
case of l-Xxx = l-Ala (19). In the case of l-Xxx= l-Val
(20), l-Leu (21) and l-Phe (22), the methylamide NH of the
trans form shows considerable solvent dependence, indica-
tive of a significant fraction of non-hydrogen bonded confor-
mations. There is a marked difference in the degree of expo-
sure of l-Xxx NH and methylamide NH in case of l-Val
(20) and l-Leu (21) peptides, whereas the two NH groups
appear to be more solvent exposed in the case of l-Phe
Chem. Eur. J. 2008, 14, 6192 – 6204
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6201

P. Balaram, N. Shamala et al.
case of d-Pro-l-Pro sequences, a variety of l-residues placed
at the C-terminus result in the formation of consecutive b-
turn structures (bII’-I) stabilized by the formation of two in-
tramolecular 4!1 hydrogen bonds. In contrast, when the C-
terminus residue has the d-configuration, the structures ob-
tained are diverse. Notably, in the case of Piv-d-Pro-l-Pro-
d-Ala-NHMe, a 5!1 hydrogen bonded a-turn (C13) struc-
ture is obtained. Proline peptides are also well disposed to-
wards adopting cis conformations about the Xaa-Pro pep-
tide bond. In the present study, cis pro-pro bonds have been
crystallographically characterized in two cases, one homo-
chiral and the other heterochiral. In the case of Piv-l-Pro-l-
Pro-l-Phe-OMe a type VIa b-turn conformation is observed
whereas in Piv-d-Pro-l-Pro-d-Val-NHMe a structure devoid
of any intramolecular hydrogen bond is observed with the
d-Pro-l-Pro peptide bond being cis. The ability to probe the
conformational space available to diproline segments is en-
hanced by the structural analysis of racemates in which
packing effects are different from those anticipated in the
case of crystals of pure enantiomers.
NMR studies in solution provide a means of characteriz-
ing multiple conformational states. In the present study, d-
Pro-l-Pro-l-Xxx sequences have been shown to be largely
conformationally homogeneous with the consecutive b-turn
also being maintained in solution. In sharp contrast, in the
l-Pro-l-Pro-l-Xxx series, the nature of the l-Xxx residue
has a significant effect on the cis–trans equilibrium about
the Pro-Pro peptide bond. l-Xxx= l-Ala is the only case
where the population of the cis conformers is extremely
small (5% in CDCl3). In all the other cases examined, cis
forms predominate in an apolar medium CDCl3, presumably
because of the formation of type VIa b-turn conformations
which are stabilized by a single 4!1 intramolecular hydro-
gen bond. In polar solvents like DMSO that compete for hy-
drogen bonding backbone sites, the population of cis con-
formers shows a dramatic decrease. The results of the pres-
ent study provide a detailed view of the possible conforma-
tional states of heterochiral and homochiral diproline se-
quences. The results also suggest that diproline segments
can be effectively used in the design of short well-structured
peptide sequences, especially when the role of the C-termi-
nus flanking residue is appreciated. The successful design of
a b-hairpin peptide with a three residue connecting loop is
an illustrative example of this approach.[2c] The body of evi-
dence presented in this paper on the conformations of
model peptides containing diproline sequences suggests that
unanticipated structures may yet be revealed by X-ray dif-
fraction analysis of short peptides. The results also empha-
size the subtle role of sequence effects in modulating the
conformations of short, constrained peptide segments.
Figure 9. Partial 500 MHz NOESY spectra of the peptides Piv-l-Pro-l-
Pro-l-Phe-NHMe (22) (top) and Piv-l-Pro-l-Pro-l-Ala-NHMe (19)
(bottom) illustrating Pro(1) CaHQ Pro(2) CaH and Pro(1) CaHQ Pro(2)
CdH in CDCl3. * Corresponds to cis conformation
Conclusion
Polypeptide chain folding can be directed by the imposition
of local backbone constraints.[4] Restricting short segments
to a limited range of conformational excursions permits the
nucleation of ordered structures. For example, type I/III b-
turns can serve as a nucleus for helical folding, with the for-
mation of an incipient 310 helix when two contiguous turns
occur over a three residue segment, with the central residue
being shared between the two turns. Prime b-turns like type
I’ and II’ serve as nuclei for the generation of registered b-
hairpin structures. Diproline segments can adopt only a lim-
ited range of conformations and are accommodated into
specific b-turn structures. In segments where the residues al-
ternate in chirality (heterochiral diproline segments), type II
b-turns are strongly favored: bII’ in the case of d-l segments
and bII in the case of l–d units. In homochiral l-Pro-l-Pro
sequences, the tendency to form type I/III turns competes
with the formation of semi-extended polyproline structures.
In both the homochiral and heterochiral diproline sequen-
ces, the nature of the flanking residue has a profound influ-
ence on the conformation of the diproline segment. In the
Experimental Section
Peptide synthesis: All the peptides reported were synthesized by conven-
tional solution phase methods using a fragment condensation strategy.
The pivaloyl (Piv) group was used for the N-terminus, while the C-termi-
6202
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 6192 – 6204

Conformational Constraints
FULL PAPER
nus was protected as an N-methylamide. Couplings were mediated by di-
cyclohexylcarbodiimide/1-hydroxybenzotriazole (DCC/HOBT).[29] All the
intermediates were characterized by 1H NMR (80 MHz) and TLC (silica
gel, chloroform/methanol 9:1) and were used without further purification.
The final peptides were purified by silica-gel column chromatography fol-
lowed by HPLC (C18, 5–10m), employing methanol/water gradients. The
homogeneity of the purified peptides was ascertained by analytical
HPLC. The purified peptides were characterized by electrospray ioniza-
tion mass spectrometry.[29]
J. P. Schneider, J. Am. Chem. Soc. 2005, 127, 17025–17029; e) R.
Fasan, R. L. A. Dias, K. Moehle, O. Zerbe, D. Obrecht, P. R. E.
Mittl, M. G. Grutter, J. A. Robinson, ChemBioChem 2006, 7, 515–
526; f) J. Spath, F. Stuart, L. Jiang, J. A. Robinson, Helv. Chim. Acta
1998, 81, 1726–1738; g) L. Jiang, K. Moehle, B. Dhanapal, D. Ob-
recht, J. A. Robinson, Helv. Chim. Acta 2000, 83, 3097–3112; h) M.
Favre, K. Moehle, L. Jiang, B. Pfeiffer, J. A. Robinson, J. Am.
Chem. Soc. 1999, 121, 2679–2685; i) A. Descours, K. Moehle, A.
Renard, J. A. Robinson, ChemBioChem 2002, 3, 318–323; j) S. C.
Shankaramma, K. Moehle, S. James, J. W. Vrijbloed, D. Obrecht,
J. A. Robinson, Chem. Commun. 2003, 1842–1843; k) Z. Athanas-
siou, R. L. A. Dias, K. Moehle, N. Dobson, G. Varani, J. A. Robin-
son, J. Am. Chem. Soc. 2004, 126, 6906–6913; l) J. K. Kretsinger,
L. A. Haines, B. Ozbas, D. J. Pochan, J. P. Schneider, Biomaterials
2005, 26, 5177–5186; m) B. Ozbas, K. Rajagopal, J. P. Schneider,
D. J. Pochan, Phys. Rev. Lett. 2004, 93, 268106; n) L. A. Haines, K.
Rajagopal, B. Ozbas, D. A. Salick, D. J. Pochan, J. P. Schneider, J.
Am. Chem. Soc. 2005, 127, 17025–17029; o) S. Hanessian, M. An-
giolini, Chem. Eur. J. 2002, 8, 111–117.
X-ray diffraction: Single crystals suitable for X-ray diffraction were ob-
tained by slow evaporation from petroleum ether/ethyl acetate and meth-
anol/water mixtures. X-ray diffraction data were collected at room tem-
perature on a Bruker AXS SMART APEX CCD diffractometer using
MoKa radiation (l=0.71073 Š). All the structures were solved by direct
methods using SHELXS-97[30a] and refined against F2, with full-matrix
least-squares methods using SHELXL-97.[30b] The crystal and diffraction
parameters and refinement statistics for 14 peptide crystal structures are
provided as Supporting Information (Table S2, Supporting Information).
CCDC 645999 (1), 646000 (2), 646001 (3), 646002 (4), 646003 (5), 646004
(6), 646005 (7), 646006 (8), 646007 (10), 646008 (11), 646009 (15), 646010
(18), 646011 (20) and 646012 (23) contain the supplementary crystallo-
graphic data for this paper. These data can be obtained free of charge
from The Cambridge Crystallographic Data Centre via www.ccdc.cam.
ac.uk/data_request/cif.
[3] a) P. Y. Chou, G. D. Fasman, J. Mol. Biol. 1977, 115, 135–175;
b) C. M. Wilmot, J. M. Thornton, J. Mol. Biol. 1988, 203, 221–232;
c) G. D. Rose, L. M. Gierasch, J. A. Smith, Adv. Protein Chem. 1985,
37, 1–109; d) J. S. Richardson, D. C. Richardson, In Prediction of
Protein Structure and the Principles of Protein Conformation (Ed.:
G. D. Fasman), Plenum Press, New York, 1989, pp. 1–98.
NMR spectroscopy: NMR experiments were carried out on a Bruker
DRX500 spectrometer. 1D and 2D spectra were recorded at a peptide
concentration of ~3 mm in CDCl3, at 300 K. Delineation of exposed NH
groups was achieved by titrating a CDCl3 solution with low concentra-
tions of DMSO-d6. Resonance assignments were carried out with the
help of 1D and 2D spectra. Residue specific assignments were obtained
from TOCSY experiments, while NOESY/ROESY spectra permitted se-
quence specific assignments. All 2D experiments were recorded in phase
sensitive mode using the TPPI (time proportional phase incrementation)
method. A data set of 1024”450 was used for acquiring the data. The
same data set was zero filled to yield a data matrix of size 2048”1024
before Fourier transformation. A spectral width of 6000 Hz was used in
both dimensions. Mixing times of 100 and 200 ms were used for TOCSY
and ROESY, respectively. Shifted square sine bell windows were used
while processing. All processing was done using BRUKER XWINNMR
software.
[4] a) W. F. DeGrado, Adv. Protein Chem. 1988, 39, 51–124; b) J. S. Ri-
chardson, Adv. Protein Chem. 1981, 34, 167–339; c) J. Venkatraman,
S. C. Shankaramma, P. Balaram, Chem. Rev. 2001, 101, 3131–3152;
d) S. Aravinda, N. Shamala, R. S. Roy, P. Balaram, Proc. Indian.
Acad. Sci. (Chem. Sci.) 2003, 115, 373–400.
[5] a) P. R. Schimmel, P. J. Flory, J. Mol. Biol. 1968, 34, 105–120; b) B.
Robson, E. Suzuki, J. Mol. Biol. 1976, 107, 327–356; c) P. Y. Chou,
G. D. Fasman, Biochemistry 1974, 13, 211–222; d) P. Y. Chou, G. D.
Fasman, Adv. Enzymol. Relat. Areas Mol. Biol. 1978, 47, 45–148;
e) M. Levitt, Biochemistry 1978, 17, 4277–4285; f) C. K. Smith, J. M.
Withka, L. Regan, Biochemistry 1994, 33, 5510–5517.
[6] a) L. G. Presta, G. D. Rose, Science 1988, 240, 1632–1641; b) J. S. Ri-
chardson, D. C. Richardson, Science 1988, 240, 1648–1652; c) R.
Aurora, G. D. Rose, Protein Sci. 1998, 7, 21–38; d) K. Gunasekaran,
H. A. Nagarajaram, C. Ramakrishnan, P. Balaram, J. Mol. Biol.
1998, 275, 917–932; e) A. R. Viguera, L. Serrano, Protein Sci. 1999,
8, 1733–1742.
[7] a) C. Soto, E. M. Sigurdsson, L. Morelli, R. A. Kumar, E. M. Casta-
no, B. Frangione, Nat. Med. 1998, 4, 822–826; b) L. D. Estrada, C.
Soto, Curr. Top. Med. Chem. 2007, 7, 115–126.
[8] a) C. M. Venkatachalam, Biopolymers 1968, 6, 1425–1436; b) For a
subsequent extension to the stereochemical analysis of homo and
heterochiral two residue turns see: R. Chandrasekaran, A. V. Laksh-
minarayan, U. V. Pandya, G. N. Ramachandran, Biochim. Biophys.
Acta Protein Struct. 1973, 303, 14–27.
[9] a) G. E. Job, B. Heitmann, R. J. Kennedy, S. M. Walker, D. S. Kemp,
Angew. Chem. 2004, 116, 5767–5769; Angew. Chem. Int. Ed. 2004,
43, 5649–5651; b) B. Heitmann, G. E. Job, R. J. Kennedy, S. M.
Walker, D. S. Kemp, J. Am. Chem. Soc. 2005, 127, 1690–1704.
[10] a) S. Hanessian, G. Papeo, M. Angiolini, K. Fettis, M. Beretta, A.
Munro, J. Org. Chem. 2003, 68, 7204–7218; b) S. Hanessian, H.
Sailes, A. Munro, E. Therrien, J. Org. Chem. 2003, 68, 7219–7233;
c) S. Hanessian, G. Papeo, K. Fettis, E. Therrien, M. T. P. Viet, J.
Org. Chem. 2004, 69, 4891–4899.
[11] a) S. K. Awasthi, S. Raghothama, P. Balaram, Biochem. Biophys.
Res. Commun. 1995, 216, 375–381; b) I. L. Karle, S. K. Awasthi, P.
Balaram, Proc. Natl. Acad. Sci. USA 1996, 93, 8189–8193; c) S. R.
Raghothama, S. K. Awasthi, P. Balaram, J. Chem. Soc. Perkin Trans.
2 1998, 137–143; d) T. S. Haque, J. C. Little, S. H. Gellman, J. Am.
Chem. Soc. 1996, 118, 6975–6985; e) T. S. Haque, S. H. Gellman, J.
Am. Chem. Soc. 1997, 119, 2303–2304; f) J. F. Espinosa, S. H. Gell-
man, Angew. Chem. 2000, 112, 2420–2423; Angew. Chem. Int. Ed.
2000, 39, 2330–2333; g) H. E. Stanger, S. H. Gellman, J. Am. Chem.
Soc. 1998, 120, 4236–4237; h) S. H. Gellman, Curr. Opin. Chem.
Acknowledgement
Research in this area has been funded by a grant from the Council of Sci-
entific and Industrial Research, India and a program grant from the De-
partment of Biotechnology, India in the area of Molecular Diversity and
Design. X- ray diffraction data were collected on the CCD facility
funded under the IRHPA program of the Department of Science and
Technology, India. S.A and R.R thank the Council of Scientific and In-
dustrial Research, India, and the Department of Biotechnology, India for
Research Assosiateships. B. C thanks Prof. R. Balaji Rao and Prof. R. L.
Gupta (Banaras Hindu University, Varanasi, India) for their interest and
encouragement.
[1] a) Y. V. Venkatachalapathi, P. Balaram, Nature 1979, 281, 83–84;
b) R. Rai, S. Aravinda, K. Kanagarajadurai, S. Raghothama, N. Sha-
mala, P. Balaram, J. Am. Chem. Soc. 2006, 128, 7916–7928; c) D. S.
Kemp, J. G. Boyd, C. C. Muendel, Nature 1991, 352, 451–454; d) D.
S Kemp, T. P. Curran, J. G. Boyd, T. J Allen, J. Org. Chem. 1991, 56,
6683–6697; e) D. S. Kemp, T. P. Curran, W. M. Davis, J. G. Boyd, C.
Muendel, J. Org. Chem. 1991, 56, 6672–6682.
[2] a) J. A. Robinson, Synlett 2000, 429–441; b) R. M. Hughes, M. L.
Waters, Curr. Opin. Struct. Biol. 2006, 16, 514–524; c) R. Rai, S. Ra-
ghothama, P. Balaram, J. Am. Chem. Soc. 2006, 128, 2675–2681;
d) L.A. Haines, K. Rajagopal, B. Ozbas, D. A. Salick, D. J. Pochan,
Chem. Eur. J. 2008, 14, 6192 – 6204
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6203

P. Balaram, N. Shamala et al.
Biol. 1998, 2, 717–725; i) D. Ramadan, D. J. Cline, S. Bai, C. Thorpe,
J. P. Schneider, J. Am. Chem. Soc. 2007, 129, 2981–2988.
[12] a) I. L. Karle, H. N. Gopi, P. Balaram, Proc. Natl. Acad. Sci. USA
2002, 99, 5160–5164; b) S. Aravinda, V. V. Harini, N. Shamala, C.
Das, P. Balaram, Biochemistry 2004, 43, 1832–1846; c) V. V. Harini,
S. Aravinda, R. Rai, N. Shamala, P. Balaram, Chem. Eur. J. 2005, 11,
3609–3620.
[13] a) C. Das, S. Raghothama, P. Balaram, J. Am. Chem. Soc. 1998, 120,
5812–5813; b) C. Das, S. Raghothama, P. Balaram, Chem. Commun.
1999, 967–968; c) J. Venkatraman, G. A. Nagana Gowda, P. Balar-
am, J. Am. Chem. Soc. 2002, 124, 4987–4994; d) H. L. Schenck,
S. H. Gellman, J. Am. Chem. Soc. 1998, 120, 4869–4870; e) J. D.
Fisk, S. H. Gellman, J. Am. Chem. Soc. 2001, 123, 343–344; f) J. D.
Fisk, M. A. Schmitt, S. H. Gellman, J. Am. Chem. Soc. 2006, 128,
7148–7149; g) C. M. Santiveri, J. Santoro, M. Rico, M. A. Jimenez,
Protein Sci. 2004, 13, 1134–1147; h) C. Das, V. Nayak, S. Raghotha-
ma, P. Balaram, J. Pept. Res. 2000, 56, 307–317.
[14] C. M. Nair, M. Vijayan, Y. V. Venkatachalapathi, P. Balaram, J.
Chem. Soc. Chem. Commun. 1979, 1183–1184.
[15] G. Zanotti, A. Bassetto, R. Battistutta, C. Folli, P. Arcidiaco, M.
Stoppini, R. Berni, Biochim. Biophys. Acta Protein Struct. Mol. En-
zymol. 2000, 1478, 232–238.
[16] a) Y. Takeuchi, G. R. Marshall, J. Am. Chem. Soc. 1998, 120, 5363–
5372; b) D. K. Chalmers, G. R. Marshall, J. Am. Chem. Soc. 1995,
117, 5927–5937.
218, 397–412; e) D. Pal, P. Chakrabarti, J. Mol. Biol. 1999, 294, 271–
288; f) W.-J. Wu, D. P. Raleigh, Biopolymers 1998, 45, 381–394;
g) C. M. Taylor, R. Hardre, P. J. B. Edwards, J. H. Park, Org. Lett.
2003, 5, 4413–4416; h) H. Y. Meng, K. M. Thomas, A. E. Lee, N. J.
Zondlo, Biopolymers (Peptide Sci.) 2006, 84, 192–204.
[23] a) G. N. Ramachandran, A. V. Lakshminarayanan, R. Balasubrama-
nian, G. Tegoni, Biochim. Biophys. Acta Protein Struct. 1970, 221,
165–181; b) T. Ashida, M. Kakudo, Bull. Chem. Soc. Jpn. 1974, 47,
1129–1133; c) S. Tanaka, H. A. Scheraga, Macromolecules 1974, 7,
698–705; d) F. A. Momany, R. F. McGuire, A. W. Burgess, H. A.
Scheraga, J. Phys. Chem. 1975, 79, 2361–2381; e) D. Cremer, J.A.
Pople, J. Am. Chem. Soc. 1975, 97, 1354–1358; f) D. F. De Tar, N. P.
Luthra, J. Am. Chem. Soc. 1977, 99, 1232–1244; g) V. Madison, Bio-
polymers 1977, 16, 2671–2692; h) B. K. Ho, E. A. Coutsias, C. Seok,
K.A. Dill, Protein Sci. 2005, 14, 1011–1018; i) E. J. Milner-White,
L. H. Bell, P. H. Maccallum, J. Mol. Biol. 1992, 228, 725–734; j) L.
Vitagliano, R. Berisio, A. Mastrangelo, L. Mazzarella, A. Zagari,
Protein Sci. 2001, 10, 2627–2632.
[24] F. H. Allen, Acta Crystallogr. Sect. B 2002, 58, 380–388.
[25] S. Donnini, G. Groenhof, R. K. Wierenga, A. H. Juffer, Proteins
2006, 64, 700–710.
[26] a) I. Z. Siemion, T. Wieland, K.-H. Pook, Angew. Chem. 1975, 87,
712–714; Angew. Chem. Int. Ed. Engl. 1975, 14, 702–703; b) D. E.
Dorman, F. A. Bovey, J. Org. Chem. 1978, 38, 2379–2383; c) R. Ri-
charz, K. Wüthrich, Biopolymers 1978, 17, 2133–2141.
[17] A preliminary report describing the crystal structure determination
of peptide racemates (12, 13 and 14) has been presented [I. Saha, B.
Chatterjee, N. Shamala, P. Balaram, Biopolymers (Peptide Sci.) in
press].
[18] H. Nishihara, K. Nishihara, T. Uefuji, N. Sakota, Bull. Chem. Soc.
Jpn. 1975, 48, 553–555.
[27] a) J. Yao, R. Bruschweiler, H. J. Dyson, P. E. Wright, J. Am. Chem.
Soc. 1994, 116, 12051–12052; b) J. Yao, H. J. Dyson, P. E. Wright, J.
Mol. Biol. 1994, 243, 754–766; c) J. Yao, V. A. Feher, B. F. Espejo,
M. T. Reymond, P. E. Wright, H. J. Dyson, J. Mol. Biol. 1994, 243,
736–753; d) L. Halab, W. D. Lubell, J. Am. Chem. Soc. 2002, 124,
2474–2484.
[19] a) D. V. Nataraj, N. Srinivasan, R. Sowdhamini, C. Ramakrishnan,
Curr. Sci. 1995, 69, 434–447; b) V. Pavone, G. Gaeta, A. Lombardi,
F. Nastri, O. Maglio, C. Isernia, M. Saviano, Biopolymers 1996, 38,
705–721; c) B. Dasgupta, L. Pal, G. Basu, P. Chakrabarti, Proteins
2004, 55, 305–315.
[28] a) V. Madison, J. Schellman, Biopolymers 1970, 9, 511–567; b) T. Hi-
gashijima, M. Tasumi, T. Miyazawa, Biopolymers 1977, 16, 1259–
1270.
[29] Synthetic procedures and ESI-MS spectra are provided as Support-
ing Information.
[20] P. N. Lewis, F. A. Momany, H. A. Scheraga, Biochim. Biophys. Acta
Protein Struct. 1973, 303, 211–229.
[30] a) G. M. Sheldrick, SHELXS-97, A program for automatic solution
of crystal structures, University of Gçttingen, Gçttingen (Germany),
1997; b) G. M. Sheldrick, SHELXL-97, A program for the refine-
ment of crystal structures, University of Gottingen, Gottingen (Ger-
many), 1997.
[21] H.-B. Kraatz, D. M. Leek, A. Houmam, G. D. Enright, J. Lusztyk,
D. D. M. Wayner, J. Organomet. Chem. 1999, 589, 38–49.
[22] a) C. Grathwohl, K. Wüthrich, Biopolymers 1976, 15, 2025–2041;
b) C. Grathwohl, K. Wüthrich, Biopolymers 1981, 20, 2623–2633;
c) D. E. Stewart, A. Sarkar, J. E. Wampler, J. Mol. Biol. 1990, 214,
253–260; d) M. W. MacArthur, J. M. Thornton, J. Mol. Biol. 1991,
Received: December 22, 2007
Published online: May 19, 2008
6204
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 6192 – 6204