Solid-state and solution-phase conformations of pseudoproline-containing dipeptides

Article abstract of DOI:10.1071/CH09151

The conformations of 14 threonine-derived pseudoproline-containing dipeptides (including four d-allo-Thr derivatives) have been investigated by NMR. In solution, the major conformer observed for all dipeptides is that in which the amide bond between the p

Full text of DOI:10.1071/CH09151

RESEARCH FRONT
Full Paper
CSIRO PUBLISHING
www.publish.csiro.au/journals/ajc
Aust. J. Chem. 2009, 62, 711–719
Solid-State and Solution-Phase Conformations
of Pseudoproline-Containing Dipeptides
Jack K. Clegg,A James R. Cochrane,A Nima Sayyadi,A Danielle Skropeta,A
Peter Turner,A and Katrina A. JolliffeA,B
ASchool of Chemistry, The University of Sydney, NSW 2006, Australia.
BCorresponding author. Email: jolliffe@chem.usyd.edu.au
The conformations of 14 threonine-derived pseudoproline-containing dipeptides (including four d-allo-Thr derivatives)
have been investigated by NMR. In solution, the major conformer observed for all dipeptides is that in which the amide
bond between the pseudoproline and the preceding amino acid is cis. For dipeptides in which the N-terminus is protected,
the ratio of cis- to trans-conformers does not depend significantly on the side chain of the N-terminal amino acid, or the
stereochemistry of the Thr residue. However, for dipeptides bearing a free N-terminus, there are significant differences in
the ratios of cis- to trans-conformers depending on the side chain present.Three dipeptides were crystallized and their X-ray
structures determined. In two cases, (benzyloxycarbonyl (Cbz)-Val-Thr(ꢀMe,Mepro)-OMe and Cbz-Val-Thr(ꢀMe,Mepro)-
OH), the dipeptides adopt a trans-conformation in the solid state, in contrast to the structures observed in solution. In
the third case, (9-fluorenylmethoxycarbonyl (Fmoc)-Val-d-allo-Thr(ꢀMe,Mepro)-OH), a cis-amide geometry is observed.
These structural differences are attributed to crystal-packing interactions.
Manuscript received: 14 March 2009.
Final version: 5 June 2009.
Introduction
synthesis because they prevent aggregation of growing peptide
chains[3,4] and have been found to significantly increase the
yields of difficult peptide sequences[5–9] by improving solvation
and peptide coupling kinetics. This is attributed to a conforma-
ꢀ
Pseudoprolines (ꢀR,R pro) are oxazolidine or thiazolidine
derivatives of serine, threonine, and cysteine residues that form
on cyclocondensation of the amino acid with an aldehyde or
ketone (Fig. 1). Their five-membered ring structure is reminis-
cent of a proline residue. These modified amino acid residues
were introduced by Mutter and coworkers as temporary pro-
tecting groups for peptide synthesis and were found to exert a
pronounced effect on the peptide backbone conformation.[1,2]
The oxazolidine derivatives are now extensively used as tools to
impart better solubility to a growing peptide chain in solid-phase
tional preference for cis-amide bond formation. The ability of
ꢀ
ꢀ
R,R pro residues to induce cis-amide bonds is well established
by experimental[4,10,11] and theoretical[12] studies. In particu-
lar, disubstitution at the 2-C position (e.g. ꢀMe,Mepro) strongly
favours the cis-conformer and can be used to tailor peptide back-
bone conformation for specific applications. We have recently
exploited this conformational ability to improve the head-to-tail
cyclization yields of short peptide sequences by incorporation
of ꢀMe,Mepro residues in the linear precursors.[13,14] We have
found that, in some cases, cyclization yields vary substantially
depending on the side chain of the amino acids preceding the
R
OH
H
N
RꢂHN
ꢀ
Me,Mepro residues, suggesting that these may influence the
O
cis–trans isomerization of the amide bond.[15] This prompted us
to investigate the ratios of cis- and trans-conformers of several
model dipeptides with and without N- and C-termini protect-
ing groups present, to examine the influence of both side chain
and protecting groups on the pseudoproline conformation. Pre-
vious studies on the conformations of similar dipeptides have
O
ORꢁ
O
, Hꢀ
focussed on examining the effects of changing the nature of the
ꢀ
ꢀ
R,R pro derivative (whether it is derived from Ser, Thr, and Cys
ORꢁ
and changing the 2-C substituents R and Rꢀ),[10] but the influence
of protecting groups and side chain bulk has not been inves-
tigated and is of significance in applying the results obtained
from model studies to the conformations of longer peptides. We
report here the results of an investigation of the solution-phase
cis:trans ratios of several Xaa-Thr(ꢀMe,Mepro) derivatives, vary-
ing both the Xaa substituents and the protecting groups (or lack
thereof) at both N- and C-termini and including several novel
O
O
R
R
N
RꢂHN
N
O
RꢂHN
O
ORꢁ
O
O
trans-amide
cis-amide
Fig. 1. Synthesis of an Xaa-Thr(ꢀMe,Mepro) dipeptide and the two con-
formers accessible by rotation about the Xaa-Thr amide bond.
© CSIRO 2009
10.1071/CH09151
0004-9425/09/070711

RESEARCH FRONT
712
J. K. Clegg et al.
d-allo-Thr(ꢀMe,Mepro) derivatives, and their comparison with
solid-state structures for three of the compounds as determined
by X-ray crystallography.
uncapped N- and C-termini. We positioned Val, Phe, and Gly
in the Xaa position to examine the effect the Xaa side chain
had on the cis:trans ratio of the Xaa-Thr(ꢀMe,Mepro) amide
bond. We included four d-allo-threonine derivatives (5, 6, 9,
and 12) because, while the influence of inverting the relative
stereochemistry at the α-position of the residue preceding a
Thr(ꢀMe,Mepro) has been reported,[4] the effect of changing
the relative stereochemistry between the α- and β-positions
on Thr(ꢀMe,Mepro) formation and conformation has not been
previously investigated.
Results and Discussion
In order to investigate the cis–trans isomerization of the peptide
bond in differently protected ꢀMe,Mepro-containing dipeptides,
we prepared several derivatives 1–14 (Fig. 2) with either full
protection (both N- and C-termini), N-terminus protection or
Our standard conditions for the synthesis of Xaa-
Thr(ꢀMe,Mepro) derivatives involve treatment of the Thr-
containing dipeptides with 2-methoxypropene at 0◦C in the
presence of an acid catalyst.[13,14] Compounds 1–4 were readily
synthesized from the corresponding fully protected dipeptides
according to this general method. In the case of the d-allo-Thr
derivatives 5 and 6, this method gave only low yields of the corre-
sponding pseudoprolines, but these were substantially improved
using more forcing conditions (2,2-dimethoxypropane, pyri-
dinium p-toluenesulfonate (PPTS), toluene, 80◦C). Hydrolysis
of the methyl esters of 1, 2, and 5, and hydrogenolysis of the ben-
zyl ester of 6 were performed under standard conditions[13,14] to
yield the carboxylic acids 7–9 and 12, respectively. Hydrogenol-
ysis of the benzyloxycarbonyl (Cbz)-groups of 7 and 8 was
performed under standard conditions[13] to give 13 and 14,
respectively, Compounds 10 and 11 were prepared according
to the method of Mutter et al.[3]
O
O
RꢂHN
RHN
O
O
N
N
R
O
O
ORꢂ
ORꢁ
1
2
R ꢃ CH(CH3)2, Rꢂ ꢃ Cbz, Rꢁ ꢃ Me
R ꢃ CH2Ph, Rꢂ ꢃ Cbz, Rꢁ ꢃ Me
5
6
9
R ꢃ Cbz, Rꢂ ꢃ Me
R ꢃ Fmoc, Rꢂ ꢃ Bn
R ꢃ Cbz, Rꢂ ꢃ H
3
4
7
8
R ꢃ H, Rꢂ ꢃ Cbz, Rꢁ ꢃ Me
R ꢃ CH2Ph, Rꢂ ꢃ Fmoc, Rꢁ ꢃ CH2CCI3
R ꢃ CH(CH3)2, Rꢂ ꢃ Cbz, Rꢁ ꢃ H
R ꢃ CH2Ph, Rꢂ ꢃ Cbz, Rꢁ ꢃ H
12 R ꢃ Fmoc, Rꢂ ꢃ H
10 R ꢃ CH(CH3)2, Rꢂ ꢃ Fmoc, Rꢁ ꢃ H
11 R ꢃ CH2Ph, Rꢂ ꢃ Fmoc, Rꢁ ꢃ H
13 R ꢃ CH(CH3)2, Rꢂ ꢃ Rꢁ ꢃ H
14 R ꢃ CH2Ph, Rꢂ ꢃ Rꢁ ꢃ H
Fig. 2. Structures of the Xaa-Thr(ꢀMe,Mepro) dipeptides.
We determined the cis:trans ratios of the amide bonds in
dipeptides 1–14 using NMR spectroscopic techniques. In the
fully protected dipeptides 1–4, a major and minor set of reso-
nances were clearly observed in both the 1H and 13C NMR spec-
tra, indicating the presence of two conformers in slow exchange
(Fig. 3). In all cases, the major conformer was determined to be
that with a cis-amide bond by the presence of typical nuclear
Overhauser effect (NOE) cross peaks observed by 2D NMR
rotating frame Overhauser effect spectroscopy (ROESY) and
nuclear Overhauser effect correlation spectroscopy (NOESY)
experiments (i.e. αHi-1–αHi and αHi-1–βHi crosspeaks) that
reflect the spatial proximity of the αHi-1 and αHi protons in the
cis-form.[16] The cis:trans ratios are given inTable 1. NMR spec-
tra of a representative dipeptide 1 were obtained in a variety of
trans
cis
6.0
5.8
5.6
5.4
ppm
Fig. 3. Portion of the 300 MHz 1H NMR spectrum of 4 illustrating the
peaks observed for the carbamate NH of the cis- and trans-conformers
(cis:trans = 85:15).
Table 1. Ratios of cis:trans conformers as determined by 1H NMR
Dipeptide
Compound no.
Solvent
cis:trans ratio
CbzNH-Val-ꢀMe,MeThr-OMe
CbzNH-Val-ꢀMe,MeThr-OMe
CbzNH-Val-ꢀMe,MeThr-OMe
CbzNH-Val-ꢀMe,MeThr-OMe
CbzNH-Val-ꢀMe,MeThr-OMe
CbzNH-Phe-ꢀMe,MeThr-OMe
CbzNH-Gly-ꢀMe,MeThr-OMe
FmocNH-Phe-ꢀMe,MeThr-OTce
CbzNH-Val-d-allo-ꢀMe,MeThr-OMe
FmocNH-Val-d-allo-ꢀMe,MeThr-OBn
CbzNH-Val-ꢀMe,MeThr-OH
CbzNH-Phe-ꢀMe,MeThr-OH
CbzNH-Val-d-allo-ꢀMe,MeThr-OH
FmocNH-Val-ꢀMe,MeThr-OH
FmocNH-Phe-ꢀMe,MeThr-OH
FmocNH-Val-d-allo-ꢀMe,MeThr-OH
H2N-Val-ꢀMe,MeThr-OH
1
1
1
1
1
2
3
4
5
6
7
8
9
10
11
12
13
14
CD3CN
[D6]DMSO
CDCl3
[D8]toluene
CD3OD
CD3CN
CDCl3
CD3CN
CDCl3
CDCl3
CD3CN
CD3CN
CDCl3
CD3CN
CD3CN
CD3OD
CD3CN
CD3CN
85:15
85:15
85:15
85:15
85:15
85:15
75:25
85:15
95:5
90:10
80:20
80:20
80:20
>95:<5
>95:<5
>95:<5
65:35
90:10
H2N-Phe-ꢀMe,MeThr-OH

RESEARCH FRONT
Pseudoproline-Containing Dipeptides
713
solvents (CD3CN, CDCl3, [D8]toluene, [D6]DMSO, CD3OD)
and at varying temperatures (300–370 K in [D8]toluene) but the
ratio of cis:trans conformers did not change significantly under
these conditions. Substitution of the N-terminal Val with Phe (2)
did not affect the ratio of cis:trans conformers, but a slightly
lower proportion of cis-amide conformer (75:25) was observed
for the Gly-containing dipeptide 3. This is consistent with the
proposal that steric interactions between the methyl groups of the
the dipeptides bearing Fmoc protecting groups, ester hydrolysis
resulted in significant increases in the amount of cis-conformer
present, with only a single (cis) conformer observed by 1H and
13C NMR for 10, 11 and the d-allo-Thr derivative 12.
In two cases, the effect of deprotection of the N-terminus on
dipeptide conformation was also investigated. Fully deprotected
dipeptides 13 and 14 had significantly different ratios of cis:trans
conformers in comparison with the fully protected analogues
1 and 2. In contrast to the results observed for the fully pro-
tected analogues and N-protected peptides where the side chain
of the Xaa amino acid had little effect on the ratio of amide bond
conformers, for the fully deprotected dipeptides, the cis:trans
ratio depends significantly on the side chain of the N-terminal
amino acid, with the Val derivative 13 having a significantly
lower proportion of the cis-conformer than the Phe derivative
14. This difference cannot easily be explained by steric interac-
tions, although the increased proportion of cis-conformer in 14,
which has an aromatic side chain, reflects a similar propensity
for Xaa-Pro dipeptides that has been attributed to stabilization
of the cis-confomer as a result of CH–π interactions between
the aromatic side chain and the protons at the α-position of the
proline ring.[17] A similar interaction between the protons of the
Thr(ꢀMe,Mepro) methyl groups and the aromatic side chain of
the Phe residue would provide a similar stabilization and may
explain the observed differences between 13 and 14, although
we did not observe any evidence for such an interaction.
In all 14 Xaa-Thr(ꢀMe,Mepro) dipeptides that we examined
by NMR spectroscopy, the cis-conformer was predominant, with
cis:trans ratios varying from 65:35 to >95:<5. For dipeptides
bearing N-protecting groups, this ratio did not significantly
depend on the side chain of the Xaa residue or the stereochem-
istry of theThr residue. However, it was affected by the presence
of protecting groups at the N- and C-termini, suggesting that
studies of longer peptides are required to provide a better under-
standing of the influence of the side chain on the conformation
of Thr(ꢀMe,Mepro) peptides.
ꢀ
Me,MePro and the side chain (or peptide backbone in the case of
Gly) of the preceding amino acid are predominantly responsible
for the favoured cis-conformation of these peptides.[3] Dipep-
tide 4, with alternative N- and C-terminal protecting groups
(9-fluorenylmethoxycarbonyl (Fmoc)- and trichloroethyl ester,
respectively) had an identical cis:trans ratio to the Cbz-, methyl
ester protected analogue 2.
For the d-allo-Thr derivatives 5 and 6, a major set of
resonances was observed in the 1H and 13C NMR spectra, with
a second set of resonances from a second conformer present
but difficult to distinguish owing to their low intensity and sig-
nal overlap. The observed major conformers were identified as
those having a cis-amide bond by the presence of αHi-1–αHi
NOE crosspeaks observed by 2D NMR ROESY and NOESY
experiments (Fig. 4). Despite the differences in stereochemistry
with the previously discussed systems, an analysis of CPK mod-
els indicates that in 5 and 6, the αHi-1 and αHi protons in the
cis-conformer are in close proximity, whereas those in the trans-
conformer are much further away from each other, so an NOE
interaction is likely to be observed only for the conformer with
a cis-amide bond. This proximity in the cis-conformer is evident
in the X-ray structure of the carboxylic acid derivative 12 (see
below).
Removalofthe C-terminalprotectinggroupsresultedinsome
changes in the cis:trans ratios. For the Cbz-protected dipep-
tides 7–9, a slightly lower amount of the cis-conformer was
observed than for the analogous methyl esters. More notably, for
Solid-State Studies
Although several solution-phase studies of (ꢀR,R pro)-
ꢀ
ppm
containing peptides have been performed, little is known about
the conformations of these peptides in the solid state. To the best
of our knowledge, only one X-ray structure of a (ꢀMe,Mepro)-
containing dipeptide has been reported previously, indicating
that Fmoc-Ala-Cys(ꢀMe,Mepro)-OH retains the cis-amide con-
formation in the solid state, although the amide bond is slightly
twisted (ωi = −9.7◦).[18] To obtain further information about
the conformations of our Xaa-Thr(ꢀMe,Mepro) dipeptides, we
investigated the solid-state structures of three compounds.
Colourless prismatic crystals of 1 suitable for X-ray crys-
tallography were grown by slow diffusion of diethyl ether
into a methanolic solution (see Accessory Publication for
ORTEP plot). The X-ray analysis indicates that in the solid
state, the amide bond preceding the Thr(ꢀMe,Mepro) adopts a
trans-conformation (Fig. 5). This is in contrast to the NMR
experiments, which indicate that in the solution-phase, the
cis-conformer predominates (cis:trans 85:15) independently of
solvent or temperature. In the solid state, the amide bond (N(1)–
C(9)) is 1.3517(13) Å in length, which is similar to the length of
the bond between the Val nitrogen atom and the Cbz protecting
group (N(2)–C(14) = 1.3474(14) Å). The amide bond is slightly
twisted (ωi = 171.36(11)◦), indicating the steric hindrance that
4.0
4.5
5.0
5.5
5.5
5.0
4.5
4.0
ppm
Fig. 4. 400 MHz nuclear Overhauser effect (NOE) correlation spectrum
of 6 indicating the NOE interaction between the Val α-H (3.83 ppm)
and the d-allo-Thr(ꢀMe,Mepro) α-H (4.95 ppm). Small baseline peaks are
attributable to the trans-conformer.

RESEARCH FRONT
714
J. K. Clegg et al.
would result between the Thr(ꢀMe,Mepro) methyl substituents
and the Val side chain if it were planar.
Colourless plate-like crystals of 7, the free-acid analogue of
1, were isolated by slow diffusion of diethyl ether into a methanol
solution and subjected to a crystallographic study (see Acces-
sory Publication for ORTEP plot). The X-ray analysis revealed
that 7 adopts a conformation (ωi = 174.27(17)◦) in the solid state
that is almost identical to that of 1 (Fig. 6), again in contrast to
the solution-phase conformation of this molecule. In fact, both 1
and 7 crystallize in the same chiral space group (orthorhombic
P212121) with similar unit cells. In both structures, the a and
b axis lengths are comparable, while the c axis lengths differ
by only 2.1 Å. Given their chemical similarity, their structural
similarity is not surprising, but suggests the possibility of use
of species such as these in crystal engineering studies. Such a
possibility is further demonstrated by the analysis of the crystal
packing.
An analysis of the crystal packing shows a strong car-
bamate (N(2)) to Thr(ꢀMe,Mepro) oxygen (O(1)) hydrogen
bond (see Table 2) as the most significant motif present
within the lattice. This interaction results in the formation
of infinite one-dimensional polymeric chains of molecules
that propagate down the crystallographic b-axis. A sec-
tion of one of these chains is shown in Fig. 5. Given
the energy difference between the cis- and trans-conformers
of ꢀMe,Mepro dipeptides is ∼62–75 kJ mol−1 [10]
,
the sta-
bilization introduced in the solid state by this hydrogen-
bonding interaction may explain the differences observed
between the solution-phase and solid-state conformations of this
dipeptide.
c-axis
c-axis
Fig. 6. A schematic representation of the one-dimensional polymeric chain
formed in 7 through carbamate–oxygen hydrogen bonds. Dashed lines
indicate hydrogen-bonding interactions.
Fig. 5. A schematic representation of the one-dimensional polymeric chain
formed in 1. Dashed lines indicate hydrogen-bonding interactions.
Table 2. Hydrogen bond geometry from X-ray structures of 1, 7 and 12 where D indicates the H-bond
donor and A the acceptor
D
H
A
D–H [Å]
0.850(14)
H–A [Å]
D–A [Å]
DHA [◦]
1
N(2)
H(2N)
O(1)A
2.221(15)
3.0534(13)
166.1(13)
7
H(2O)
N(2)
H(2O)
H(2N)
O(3)B
O(1)A
0.83(3)
0.82(3)
1.87(3)
2.34(3)
2.684(2)
3.146(2)
164(3)
168(2)
12
N(2)
N(4)
H(1N)
H(2N)
O(8)C
O(4)C
0.89(5)
0.81(4)
1.88(5)
2.32(4)
2.760(3)
3.060(3)
171(4)
152(4)
Symmetry operators: A−x, y + 1/2, −z + 1/2; B−x + 1/2, −y, z + 1/2; Cx − y, −y, −z + 2/3.

RESEARCH FRONT
Pseudoproline-Containing Dipeptides
715
Much like 1, molecules of 7 pack along the crystallographic
b-axistoformaone-dimensionalpolymericchainthroughcarba-
mate (N(2)) to pseudoproline oxygen (O(1)) hydrogen-bonding
interactions (see Table 2). A portion of this chain is shown in
Fig. 6.
The presence of the free carboxylic acid group in 7 (com-
pared with 1) adds an additional site for hydrogen bonding.
Although this site does bind to the corresponding carboxylic site
in adjacent molecules, unexpectedly, it does not bind in the clas-
sic carboxylic acid dimer and instead forms a polymeric chain
(see Fig. 7), which propagates along the crystallographic a-axis
at ∼90◦ to the amide–oxygen hydrogen-bonding chain. Overall,
these two sets of hydrogen-bonding interactions combine to form
infinite two-dimensional sheet-like arrays that stack parallel to
the crystallographic ab-plane.
The d-allo-Thr derivative 12 also crystallizes in the
orthorhombic P212121 space group. Colourless plate-like crys-
tals were grown by the slow evaporation of a chloroform solution
and there are two chloroform solvent molecules present in the
lattice for each peptide molecule. In contrast to the previous
two structures, in this case the amide bond has the expected cis
geometry (ωi = −1.7(5)◦). The structure is given in Fig. 8 and
illustrates the short distance between the two α-protons (H(2)
and H(9)) observed in the NOESY spectrum of this molecule.
Once again, the dominant crystal-packing effects are
hydrogen-bonding interactions. However, in contrast to 1 and 7,
these do not involve the Thr(ꢀMe,Mepro) oxygen. Interestingly,
again despite the presence of a carboxylic acid, the dimeric car-
boxylic acid motif is not present. Instead the carboxylic acid
binds to the carbamate (N(2)) nitrogen and amide (O(4)) oxygen
of theVal residue (Table 2), forming an infinite one-dimensional
polymer (Fig. 9) along the a-axis. There is also a weak methyl–
π interaction within the lattice. There are several non-classical
hydrogen-bonding interactions present in the lattice, with the
chloroform solvent molecules acting as H-bond donors and oxy-
gen atoms as acceptors (Fig. 10). In particular, the presence of
an H-bond between chloroform and the Thr(ꢀMe,Mepro) oxygen
(O(3)) is notable, as this prevents the formation of the carbamate
to Thr(ꢀMe,Mepro) oxygen H-bonds that are the predominant
crystal packing interactions in 1 and 7. This may explain the
conformational differences observed for 12 (cis-amide) in com-
parison with 1 and 7 (trans-amide) in the solid state. It is notable
that no solvent is present in the crystal lattices of 1 and 7, despite
their crystallization from methanol, a solvent that is more likely
to form H-bonds than chloroform.
Conclusions
In solution, the predominant conformer for Xaa-Thr(ꢀMe,Mepro)
dipeptides is that in which the amide bond between the Xaa
and Thr(ꢀMe,Mepro) residues adopts a cis geometry. For dipep-
tides bearing both N- and C-terminal protecting groups, the
ratio of cis:trans conformers does not depend strongly on sol-
vent, temperature or the size of the Xaa side chain. It is notable
that the relative stereochemistry of the Thr(ꢀMe,Mepro) residue
does not have a significant impact on the cis:trans amide bond
ratio, with d-allo-Thr-containing peptides having similar (or
enhanced) conformational preferences for the cis-isomer to the
l-Thr derivatives in solution. It has previously been shown that
the relative stereochemistry between the Xaa and (ꢀMe,Mepro)
α-carbons does not have a significant influence on the amide
bond conformation.[4] Minor differences in the cis:trans ratios
are observed for N-protected dipeptides with a free carboxylic
acid, depending on whether the N-protecting group is Cbz or
Fmoc, with Fmoc-protected dipeptides having a higher propor-
tion (>95%) of cis-conformer. However, for dipeptides with
free N- and C-termini, the cis:trans ratio is strongly influenced
by the Xaa side chain, with 14, which has an aromatic side
c-axis
Fig. 7. A schematic representation of the one-dimensional polymeric chain formed via carboxylic acid–carboxylic acid hydrogen bonds in 7. Dashed lines
indicate hydrogen-bonding interactions.

RESEARCH FRONT
716
J. K. Clegg et al.
C(19)
C(18)
C(20)
C(21)
C(4)
C(17)
C(23)
C(16)
C(15)
O(2)
C(3)
C(14)
O(6)
C(24)
O(5)
C(2)
C(22)
O(3)
O(1)
C(1)
C(7)
C(25)
C(13)
C(27)
N(1)
C(8)
C(26)
C(5)
C(9)
N(2)
O(4)
C(6)
C(12)
C(10)
C(11)
Fig. 8. An ORTEP representation of 12 shown with 50% probability ellipsoids. Chloroform solvent molecules omitted for clarity.
Fig. 10. A schematic representation of 12 illustrating the H-bonding
interactions observed with CHCl3 solvent molecules.
systems to avoid any influence of the protecting groups on
conformation.
In the solid state, 1 and 7 adopt a trans-amide bond con-
formation, although in both cases the amide bond is slightly
twisted. This directly contrasts with their solution structures and
can be attributed to crystal-packing effects, with X-ray struc-
tures of both 1 and 7 showing a predominant carbamate (N)
to Thr(ꢀMe,Mepro) oxygen hydrogen-bonding interaction in the
crystal lattice.The stabilization provided in the solid state by this
ꢄb-axis
hydrogen-bonding interaction and other crystal lattice effects
is clearly greater than the 62–75 kJ mol−1 stabilization of the
Fig. 9. A schematic representation of the one-dimensional polymeric chain
cis-conformer relative to the trans-conformer in solution.[10]
formed in 12. Dashed lines indicate hydrogen-bonding interactions. Val side
In contrast, 12 adopts a cis-amide bond in both solution and
chains and Thr(ꢀMe,Mepro) methyl groups have been removed for clarity.
the solid state. In this case, the co-crystallized chloroform
chain, having a significantly higher proportion of cis-conformer
than 13, which bears an isopropyl side chain. This suggests
that an analysis of the conformation of longer (ꢀMe,Mepro)-
containing peptides must be performed on the fully deprotected
solvent molecules act as hydrogen bond donors to the d-allo-
Thr(ꢀMe,Mepro) oxygen hydrogen-bond acceptor, preventing
the formation of the stabilizing carbamate–pseudoproline oxy-
gen hydrogen bonds observed for 1 and 7. An investigation

RESEARCH FRONT
Pseudoproline-Containing Dipeptides
717
of the solution and solid-state structures of longer ꢀMe,Mepro-
containing peptides and the relationship between peptide con-
formation and head-to-tail cyclization yields is ongoing in our
laboratories.
Cbz-L-Val-D-allo-Thr(ΨMe,Mepro)-OH 9
NaOH (400 mg, 0.100 mol) was dissolved in distilled water
(2 mL) and added to a solution of 5 (650 mg, 1.60 mmol) in
THF (3 mL) and methanol (3 mL). The reaction mixture was
stirred for 24 h before the solution was partitioned between 1 M
HCl (50 mL) and dichloromethane (50 mL). The aqueous phase
was extracted with dichloromethane (2 × 20 mL) and the com-
bined organic phases were washed with brine (2 × 100 mL), then
dried (Na2SO4). Removal of the solvent under reduced pressure
gave the title compound 9 (0.390 g, quantitative) as a yellow
foam. δH (200 MHz, CDCl3) 7.33 (5H, s, ArH), 5.47 (1H, d,
J 9.3, Val(NH)), 5.10 (2H, s, Cbz(CH2)), 5.02 (1H, d, J 6.1,
Thrꢀ(α-CH)), 4.47–4.41 (1H, Thrꢀ(β-CH)), 3.94–3.65 (1H, m,
Val(α-CH)), 2.0–1.96 (1H, m, Val(β-CH)), 1.84 (3H, s, Me),
1.59 (3H, s, Me), 1.37 (3H, d, J 6.1, Thr(γ-CH3)), 0.93 (3H, d, J
6.7, Val(γ-CH3)), 0.87 (3H, d, J 6.7, Val(γ-CH3)). δC (75 MHz,
CDCl3) 173.1, 170.4, 156.9, 136.3, 128.6, 128.2, 127.8, 96.7,
72.6, 67.1, 63.6, 59.6, 29.5, 24.9, 23.6, 19.6, 18.2, 15.3. m/z (ESI)
415.1 [M + Na]+ (100%), 393 (10), 357 (5), 353 (10), 341 (5),
335 (15); m/z (HRMS ESI, MNa+) Calc. for C20H28N2O6·Na
415.1857; found 415.1857.
Experimental
Synthesis
Preparative column chromatography was carried out using Ajax
Finechem silica gel (SiO2, 0.040–0.063 mm) with the indicated
solvents. Melting points were determined using a Gallenkamp
melting point apparatus and are reported in degrees Celsius
(uncorrected). NMR spectra were recorded on a 200, 300, or
400 MHz spectrometer. The solvent 1H and 13C signals, δH 7.26
for residual CHCl3 and δC 77.0 for CDCl3; δH 3.31 and δC 49.0
for [D4]MeOH; δH 2.50 and δC 39.5 for [D6]DMSO; δH 1.94
and δC 1.3 for CD3CN were used as internal references. Sig-
nal assignments are based on a combination of 1D (including
1H selective homonuclear decoupling and 13C Distortionless
Enhancement by Polarisation Transfer (DEPT)) and 2D spec-
tral data (including H,H-correlation spectroscopy, heteronuclear
multiple bond correlation, heteronuclear single-quantum cor-
relation, and gradient-NOESY (gr-NOESY)). Optical rotations
were measured on a dual-wavelength polarimeter in a 0.25-dm
cell at 22◦C using the indicated spectroscopic grade solvents.
Elemental analyses were performed by Campbell Microanalyt-
ical Laboratories. Compounds 1–4, and 7–8 were synthesized
from the appropriately protected dipeptides according to pre-
viously reported methods.[13–15] Compounds 10 and 11 were
prepared according to the method of Mutter and coworkers.[3]
Compounds 13 and 14 were prepared on removal of the Cbz
protecting groups from 7 and 8, respectively, by hydrogenoly-
sis under previously reported conditions[13] and used without
purification for 1H NMR studies.
Fmoc-Val-D-allo-Thr(ΨMe,Mepro)-OBn 6
Dimethoxypropane (472 µL, 3.85 mmol) and pyridinium-p-
toluene sulfonate (60.0 mg, 0.231 mmol) were added to a solu-
tion of Fmoc-Val-d-allo-Thr-OBn (408 mg, 0.769 mmol) dis-
solved in toluene (10 mL) and the mixture was heated at reflux
for 16 h. The solution was cooled to ambient temperature before
diluting with EtOAc (100 mL). The solution was then washed
with saturated aqueous NaHCO3 (100 mL), then the aqueous
layer was re-extracted with EtOAc (2 × 50 mL). The combined
organic layers were then washed with brine (2 × 100 mL), dried
(Na2SO4), and the solvent removed under reduced pressure. The
crude product was purified by flash chromatography (2:1 v/v
hexane/EtOAc) to give the title compound 6 as a yellow oil
(329 mg, 75%). [α]D +29◦ (c 0.92 in CHCl3). δH (400 MHz,
CDCl3) 7.76 (2H, d, J 7.6, ArH), 7.57, (2H, d, J 7.6, ArH), 7.42–
7.28 (9H, m, ArH), 5.25 (1H, d, J 9.4, Val(NH)), 5.26 (1H, d, J
12.3, Cbz(CH2)), 5.16 (1H, d, J 12.3, Cbz(CH2)), 4.95 (1H, d, J
6.0, Thrꢀ(α-CH)), 4.37 (3H, m), 4.22 (1H, t, J 6.8, Fmoc(CH)),
3.83 (m, 1H, Val(α-CH)), 1.86 (1H, m, Val(β-CH)), 1.74 (3H, s,
Me), 1.59 (3H, s, Me), 1.22 (3H, d, J 6.7,Thr(γ-CH3)), 0.83 (3H,
d, J6.7,Val(γ-CH3)), 0.71(3H, J6.7,Val(γ-CH3)). δC (100 MHz,
CDCl3) 169.4, 169.3, 156.4, 143.7, 143.1, 134.9, 128.9, 128.7,
127.7, 125.1, 125.0, 96.3, 72.7, 67.3, 67.0, 63.4, 59.0, 47.1, 30.5,
24.9, 23.6, 19.4, 17.7, 15.2. m/z (ESI) 570 [M + Na]+ (100%),
571 (35); m/z (HRMS ESI, MNa+) Calc. for C34H38N2O6·Na
593.2622; found 593.2614.
Cbz-L-Val-D-allo-Thr(ΨMe,Mepro)-OMe 5
Dimethoxypropane (1.00 mL, 8.14 mmol) and pyridinium-p-
toluene sulfonate (124 mg, 0.540 mmol) were added to a solu-
tion of Cbz-l-Val-d-allo-Thr-OMe (0.533 mg, 1.46 mmol) dis-
solved in toluene (25 mL) and the mixture stirred at 80◦C
for 16 h. The solution was cooled to ambient temperature and
dichloromethane (100 mL) was added. The solution was then
washed with saturated aqueous NaHCO3 (100 mL) and the
aqueous layer re-extracted with dichloromethane (2 × 30 mL).
The combined organic layers were then washed with brine
(2 × 100 mL), dried (Na2SO4), and the solvent removed under
reduced pressure. The crude product was purified by flash chro-
matography (4:1 v/v dichloromethane/EtOAc). Concentration
of the appropriate fractions (RF 0.74) yielded title peptide 5
(474 mg, 80%) as a yellow oil. [α]D +8.40◦ (c 1.0 in CHCl3).
δH (200 MHz, CDCl3) 7.33 (5H, m, ArH), 5.09 (1H, d, J 9.1,
Val(NH)), 5.01 (2H, s, Cbz(CH2)), 4.98 (1H, d, J 6.1, Thrꢀ(α-
CH)), 4.40 (1H, m,Thrꢀ(β-CH)), 3.85 (1H, m,Val(α-CH)), 3.78
(3H, s, Me), 1.91–1.74 (1H, m, Val(β-CH)), 1.68 (3H, s, Me),
1.57 (3H, s, Me), 1.28 (3H, d, J 6.3, Thr(γ-CH3)), 0.93 (3H, d, J
6.7, Val(γ-CH3)), 0.85 (3H, d, J 6.7, Val(γ-CH3)). δC (75 MHz,
CDCl3) 170.2, 169.9, 156.6, 136.4, 128.5, 128.2, 127.9, 96.4,
72.5, 70.0, 63.6, 57.0, 52.0, 27.9, 24.9, 23.6, 19.7, 18.0, 15.2.
m/z (electrospray ionization (ESI)) 429 [M + Na]+ (100%),
356 (40), 349 (30); m/z (high resolution mass spectroscopy
(HRMS) ESI, MNa+) Calc. for C21H30N2O6·Na 429.1996;
found 429.1990.
Fmoc-Val-D-allo-Thr(ΨMe,Mepro)-OH 12
Compound 6 (200 mg, 0.350 mmol) was dissolved in 10 mL
of dry THF and 10% Pd/C catalyst was added. The reaction
flask was purged with H2 and evacuated three times and the
solution was left stirring under an atmosphere of H2 for 48 h.
The solution was then filtered through a pad of Celite and the
solvent removed under reduced pressure. The residue was puri-
fied by flash chromatography (100:5:1 CHCl3/MeOH/AcOH) to
give the desired peptide as a colourless foam. Recrystallization
(CHCl3) afforded the title compound 12 as colourless needles
(138 mg, 82%). Mp 95–98◦C. [α]D +19◦ (c 0.99 in CHCl3).
(Found C 67.2, H 6.6, N 5.6. C27H32N2O6 requires C 67.5, H
6.7, N 5.8%.) δH (400 MHz, MeOD) 7.75 (2H, d, J 7.9, ArH),

RESEARCH FRONT
718
J. K. Clegg et al.
7.58 (2H, d, J 7.9, ArH), 7.4–7.3 (4H, m, ArH), 5.05 (1H, d, J
6.1, Thrꢀ(α-CH)), 4.45 (1H, m, Thrꢀ(β-CH)), 4.40 (2H, d, J
6.8, Fmoc(CH2)), 4.24 (1H, t, J 6.8, Fmoc(CH)), 3.80 (1H, m,
Val(α-CH)), 1.87 (1H, m, Val(β-CH)), 1.77 (3H, s, Me), 1.61
(3H, s, Me), 1.38 (3H, d, J 6.0, Thr(γ-CH3)), 0.94 (3H, d, 6.6,
Val(γ-CH3)), 0.88 (3H, d, J 6.6, Val(γ-CH3)). δC (100 MHz,
CDCl3) 173.3, 170.0, 156.6, 143.6, 141.3, 127.7, 127.1, 125.1,
120.0, 96.5, 72.4, 67.2, 63.4, 59.4, 47.0, 31.6, 24.8, 23.5, 19.4,
18.1, 15.2. m/z (ESI) 503 [M + Na]+ (100%), 504 (30); m/z
(HRMS EI, MNa+) Calc. for C27H32N2O6·Na 503.2158; found
503.2161.
9, −19 19, −26 26, N 19587, Nind 4776(Rmerge 0.0399), Nobs
4232(I > 2σ(I)), Nvar 266, residuals* R1(F) 0.0356, wR2(F2)
0.0952, GoF(all) 1.132, ρmin,max −0.232, 0.287 e− Å−3
.
*R1 = ꢁ||Fo| – |Fc||/ꢁ|Fo| for Fo > 2σ(Fo); wR2 = [ꢁw(F2o −
F2c)2/ꢁ(wFc2)2]1/2 all reflections w = 1/[σ2(F2o) + (0.05P)2 +
0.3P] where P = (F2o + 2Fc2)/3.
Crystal Data for 12
Formula C29H34Cl6N2O6, M 719.28, orthorhombic, space
group P212121(#19), a 13.335(4), b 14.547(4), c 17.684(5) Å,
V 3430.2(16) Å3, Dc 1.393 g cm−3, Z 4, crystal size 0.400
by 0.350 by 0.100 mm, colourless, habit plate, tempera-
Crystallography
ture 150(2) K, λ(MoKα) 0.71073 Å, µ(MoKα) 0.543 mm−1
,
Data for 1, 7, and 12 were collected with ω scans to ∼56◦
2θ using a Bruker SMART 1000 diffractometer employing
graphite-monochromated Mo-Kα radiation generated from a
sealed tube (0.71073 Å) at 150(2) K. Data integration and reduc-
tion were undertaken with SAINT and XPREP.[19] Subsequent
computations were carried out using the teXsan, WinGX-32, and
XTAL graphical user interfaces.[20–22]
T(SADABS)min,max 0.860, 0.947, 2θmax 56.72, hkl range −17
17, −19 19, −22 22, N 34065, Nind 8220(Rmerge 0.0614), Nobs
5400(I > 2σ(I)), Nvar 399, residuals* R1(F) 0.0639, wR2(F2)
0.1454, GoF(all) 1.010, ρmin,max −0.515, 0.664 e− Å−3
.
*R1 = ꢁ||Fo| – |Fc||/ꢁ|Fo| for Fo > 2σ(Fo); wR2 = [ꢁw(F2o −
F2c)2/ꢁ(wFc) ]
2 2 1/2 all reflections w = 1/[σ2(F2o) + (0.0590P)2 +
2.5336P] where P = (F2o + 2Fc2)/3.
Structures were solved by direct methods using SIR97.[23]
Multiscan empirical absorption corrections were applied to the
dataset using the program SADABS.[24] Gaussian[25] adsorption
corrections were applied with XPREP.[19] Data were refined and
extended with SHELXL-97.[26] In general, non-hydrogen atoms
with occupancies greater than 0.5 were refined anisotropically.
Carbon-bound hydrogen atoms were included in idealized posi-
tions and refined using a riding model. Oxygen- and nitrogen-
boundhydrogenatomswerefirstlocatedinthedifferenceFourier
map before refinement. Where these hydrogen atoms could not
be located, they were not modelled. Disorder was modelled
using standard crystallographic methods including constraints
and restraints where necessary. Crystal and structure refinement
data, including any specific refinement details are summarized
below. Crystallographic information files reported in the present
manuscript have been deposited with the Cambridge Crystallo-
graphic Data Centre as supplementary publication numbers
CCDC 723749–723751. Copies of the data are available free of
charge from the CCDC, 12 Union Road, Cambridge, CB2 1EZ,
UK (fax: (+44) 1223 336 033; email: deposit@ccdc.cam.ac.uk).
Accessory Publication
NOESY NMR spectrum of 12 and ORTEP plots of 1 and 7 are
available from the Journal’s website.
Acknowledgements
We thank the Australian Research Council for financial support and for the
award of a Queen Elizabeth II research fellowship to K.A.J.
References
[1] T. Haack, M. Mutter, Tetrahedron Lett. 1992, 33, 1589.
doi:10.1016/S0040-4039(00)91681-2
[2] T. Wöhr, M. Mutter, Tetrahedron Lett. 1995, 36, 3847. doi:10.1016/
0040-4039(95)00667-2
[3] T. Wöhr, F. Wahl, A. Hefzi, B. Rohwedder, T. Sato, X. Sun, M. Mutter,
J. Am. Chem. Soc. 1996, 118, 9218. doi:10.1021/JA961509Q
[4] P. Dumy, M. Keller, D. E. Ryan, B. Rohwedder, T. Wöhr, M. Mutter,
J. Am. Chem. Soc. 1997, 119, 918. doi:10.1021/JA962780A
[5] I. Coin, M. Beyermann, M. Bienert, Nat. Protocols 2007, 2, 3247.
doi:10.1038/NPROT.2007.454
[6] F. García-Martín, P. White, R. Steinauer, S. Cote, J. Tulla-Puche,
F. Albericio, Biopolymers 2006, 84, 566. doi:10.1002/BIP.20564
[7] G.-A. Cremer, H. Tanae, A. F. Delmas, J. Pept. Sci. 2006, 12, 437.
doi:10.1002/PSC.746
Crystal Data for 1
Formula C21H30N2O6, M 406.47, orthorhombic, space group
P212121(#19), a 6.894(2), b 14.248(2), c 21.863(2) Å, V
2147.5(7) Å3, Dc 1.257 g cm3,
Z
4, crystal size 0.493
by 0.438 by 0.219 mm, colourless, habit prism, tempera-
ture 150(2) K, λ(MoKα) 0.71073 Å, µ(MoKα) 0.092 mm−1
[8] A. Abedini, D. P. Raleigh, Org. Lett. 2005, 7, 693. doi:10.1021/
OL047480+
[9] M. Keller, A. D. Miller, Bioorg. Med. Chem. Lett. 2001, 11, 857.
doi:10.1016/S0960-894X(01)00085-3
,
T(Gaussian)min,max 0.961, 0.983, 2θmax 56.60, hkl range −9
9, −18 19, −29 29, N 21262, Nind 5151(Rmerge 0.0295), Nobs
4802(I > 2σ(I)), Nvar 272, residuals* R1(F) 0.0301, wR2(F2)
[10] M. Keller, C. Sager, P. Dumy, M. Schutkowski, G. S. Fischer,
M. Mutter, J. Am. Chem. Soc. 1998, 120, 2714. doi:10.1021/
JA973966S
[11] M. Mutter, T. Wöhr, S. Gioria, M. Keller, Biopolymers 1999,
51, 121. doi:10.1002/(SICI)1097-0282(1999)51:2<121::AID-BIP2>
3.0.CO;2-O
0.0708, GoF(all) 1.088, ρmin,max −0.212, 0.200 e− Å−3
.
*R1 = ꢁ||Fo| – |Fc||/ꢁ|Fo| for Fo > 2σ(Fo); wR2 = [ꢁw(F2o −
F2c)2/ꢁ(wFc2)2]1/2 all reflections w = 1/[σ2(F2o) + (0.03P)2 +
0.3P] where P = (F2o + 2Fc2)/3.
[12] Y. Che, G. R. Marshall, Biopolymers 2006, 81, 392. doi:10.1002/
BIP.20431
[13] D. Skropeta, K. A. Jolliffe, P. Turner, J. Org. Chem. 2004, 69, 8804.
doi:10.1021/JO0484732
[14] N. Sayyadi, D. Skropeta, K. A. Jolliffe, Org. Lett. 2005, 7, 5497.
doi:10.1021/OL0522891
[15] N. Sayyadi, Ph.D. Thesis: Pseudoprolines: A Versatile New Tool for
the Synthesis of Cyclic Peptides, University of Sydney, 2008.
[16] K. Wüthrich, NMR of Proteins and Nucleic Acids 1986, p. 117 (John
Wiley & Sons, Inc.: New York, NY).
Crystal Data for 7
Formula C20H28N2O6, M 392.44, Orthorhombic, space group
P212121(#19), a 6.8768(17), b 14.555(4), c 19.733(5) Å,
V 1975.2(9) Å3, Dc 1.320 g cm−3, Z 4, crystal size 0.285
by 0.275 by 0.101 mm, colourless, habit plate, tempera-
ture 150(2) K, λ(MoKα) 0.71073 Å, µ(MoKα) 0.098 mm−1
,
T(Gaussian)min,max 0.974, 0.993, 2θmax 56.88, hkl range −9

RESEARCH FRONT
Pseudoproline-Containing Dipeptides
719
[17] D. Pal, P. Chakrabarti, J. Mol. Biol. 1999, 294, 271. doi:10.1006/
JMBI.1999.3217
[18] A. Nefzi, K. Schenk, M. Mutter, Protein Pept. Lett. 1994, 1, 66.
[19] SMART SAINT and XPREP, Bruker 1995 (Bruker Analytical X-ray
Instruments Inc.: Madison, WI).
[20] teXsan for Windows: Single Crystal Structure Analysis Software
1997–1998 (Molecular Structure Corporation: The Woodlands, TX).
[21] L. J. Farrugia, J. Appl. Cryst. 1999, 32, 837. doi:10.1107/
S0021889899006020
[22] Xtal3.6 System (Eds S. R. Hall, D. J. du Boulay, R. Olthof-Hazekamp)
1999 (University of Western Australia).
[23] A. Altomare, M. C. Burla, M. Camalli, G. L. Cascarano,
C. Giocavazzo, A. Guagliardi, A. G. C. Moliterni, G. Polidori,
R. Spagna, J.Appl. Cryst. 1999, 32, 115. doi:10.1107/S002188989800
7717
[24] G. M. Sheldrick, SADABS: Empirical Absorption and Correction
Software 1999–2003 (University of Göttingen: Germany).
[25] P. Coppens, L. Leiserowitz, D. Rabinovich, Acta Crystallogr. 1965,
18, 1035. doi:10.1107/S0365110X65002487
[26] G. M. Sheldrick, SHELXL-97: Programs for Crystal Structure Anal-
ysis 1997 (University of Göttingen: Germany).