Stereoselective synthesis of Ψ[CH2O] pseudodipeptides and conformational analysis of a PheΨ[CH2O]Ala containing analogue of the drug desmopressin

Article abstract of DOI:10.1016/S0960-894X(02)00016-1

A method for the synthesis of XaaΨ[CH₂O]Ala/Gly pseudodipeptides in good yields and excellent diastereoselectivity from azido alcohols and (R)-2-chloropropionic acid or tert-butyl bromoacetate has been developed. Insertion of one of the pseudodipeptide building blocks in the peptide drug desmopressin revealed that methylene either isosteres may have only a minor influence on the secondary structure of peptides.

Full text of DOI:10.1016/S0960-894X(02)00016-1

Bioorganic & Medicinal Chemistry Letters 12 (2002) 841–844
Stereoselective Synthesis of W[CH2O] Pseudodipeptides and
Conformational Analysis of a PheW[CH2O]Ala Containing
Analogue of the Drug Desmopressin
Mattias Hedenstrom, Lotta Holm, ZhongQing Yuan,
Hans Emtenas and Jan Kihlberg*
˚ ˚
Organic Chemistry, Department of Chemistry, Umea University, SE-901 87 Umea, Sweden
Received 18October 2001; revised 12 December 2001; accepted 13 December 2001
Abstract—A method for synthesis of XaaÉ[CH2O]Ala/Gly pseudodipeptides in good yields and excellent diastereoselectivity from
azido alcohols and (R)-2-chloropropionic acid or tert-butyl bromoacetate has been developed. Insertion of one of the pseudodi-
peptide building blocks in the peptide drug desmopressin revealed that methylene ether isosteres may have only a minor influence
on the secondary structure of peptides. # 2002 Elsevier Science Ltd. All rights reserved.
Replacement of backbone amide bonds by isosteric
units stabilizes biologically active peptides towards
enzymatic degradation and may also alter other phar-
macokinetic properties in a favourable manner.1 In
addition, incorporation of amide bond isosteres is a
convenient way to elucidate the role of selected amide
bonds in receptor binding, as well as their influence on
the secondary structure of peptides. The methylene
ether isostere constitutes an interesting amide bond
surrogate since the calculated Cia–Cai+1 distance of
É[CH2O] pseudodipeptides (3.7 A) is almost identical to
that in a dipeptide (3.8A ).2 É[CH2O] pseudodipeptides
are, however, substantially more flexible than normal
dipeptides. Several synthetic routes towards É[CH2O]
pseudodipeptides suitable for incorporation into peptides
have been reported.2À6 The most straightforward of
these methods is the Williamson’s ether synthesis that
involves intermolecular2 or intramolecular3 substitution of
an a-halo acid with an alkoxide obtained from an amino-
alcohol. The intermolecular route has predominantly been
employed for synthesis of XaaÉ[CH2O]Gly pseudodi-
peptides,2 but was found to give poor diastereoselectivity
when applied to a LeuÉ[CH2O]Ala pseudodipeptide.3
Unfortunately, in our hands both the intermolecular
and the intramolecular route led to formation of
diastereomeric mixtures as well as to difficulties in
purification of the target pseudodipeptides (cf., below).
We have therefore developed an improved procedure
for synthesis of XaaÉ[CH2O]Ala/Gly pseudodipeptide
building blocks and incorporated two of these in
biologically active peptides. Conformational studies of
one of the resulting modified peptides are also
reported.
Studies based on NMR spectroscopy and/or energy
calculations have revealed that peptide hormones, or
their analogues, such as the drug desmopressin7 (Mpa-
Tyr-Phe-Gln-Asn-Cys-Pro-d-Arg-Gly-NH2, Mpa and
Cys form a disulphide bridge), Leu-enkephalin8,9 (Tyr-
Gly-Gly-Phe-Leu) and LHRH10 (pGlu-His-Trp-Ser-
Tyr-Gly-Leu-Arg-Pro-Gly-NH2) may adopt g- or b-
turn structures in aqueous solution (residues involved in
turns have been underlined). We have therefore become
interested in introduction of turn mimetics in the above
peptides in attempts to establish if turns are also
involved in their interactions with G-protein coupled
receptors.11,12 Our group has also investigated the role
of a glycopeptide from type II collagen (Ac-Ile-Ala-
Gly-Phe-Hyl(b-d-Gal)-Gly-Glu-Gln-NH2: CII260–267)
in induction of rheumatoid arthritis in a mouse
model.13 In both of these projects it was essential to
determine the role of certain amide bonds in receptor
binding and initially we focused our attention on pre-
paration of PheÉ[CH2O]Ala, TyrÉ[CH2O]Gly, and
IleÉ[CH2O]Ala pseudodipeptides. Since it is known that
Gln4 in desmopressin can be replaced by Ala without
*Corresponding author. Fax: +46-90-138885; e-mail: jan.kihlberg@
chem.umu.se
0960-894X/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(02)00016-1

842
M. Hedenstrom et al. / Bioorg. Med. Chem. Lett. 12 (2002) 841–844
¨
having a major effect on the biological activity,14 the
PheÉ[CH2O]Ala pseudodipeptide was to be incorpo-
rated instead of residues Phe3-Gln4 of desmopressin.
The TyrÉ[CH2O]Gly pseudodipeptide can be incorpo-
rated both in Leu-enkephalin and LHRH, whereas
IleÉ[CH2O]Ala matches the first two residues of the
CII260–267 glycopeptide. Synthesis of a PheÉ[CH2O]Ala
pseudodipeptide was first attempted by substitution of
(R)-2-bromopropionic acid with l-phenylalaninol under
basic conditions, but the product was obtained in poor
yield and as a mixture of diastereomeres that could not
be separated.15 Intramolecular formation of methylene
ether isosteres has been reported to proceed without
formation of diastereomers3 and the next attempt
therefore involved an intramolecular route (Scheme 1).
Compound 1 was prepared and then treated with
potassium hydride in THF to give morpholinone 2 in
1
good yield (74%).15 Unfortunately, H NMR spectro-
scopy revealed that 2 was obtained as a distereomeric
mixture which was difficult to purify. Ring-opening of 2
in refluxing 6 M HCl gave 3 which was then protected
with an Fmoc group to give 4. Compounds 3 and 4 were
both very difficult to handle during workup and pur-
ification by flash column chromatography, and neither
of them could be obtained in diastereomerically pure
form.
Scheme 1. (a) KH, THF/DMF, 0 ꢁC; 74%; (b) 6 M aq HCl, 100 ꢁC;
(c) Fmoc-Cl, 10% aq Na2CO3, dioxane, 0 ꢁC.
The difficulties encountered in preparation of 2–4
prompted us to find an improved synthetic approach to
XaaÉ[CH2O]Ala/Gly pseudodipeptides (Schemes 2 and
3). To circumvent the workup problems encountered
with 3 and 4, an azido group was chosen as protective
group for the a-amino group of the pseudodipeptides.
The limited size of the azido group also serves to reduce
steric hindrance in subsequent reactions. In addition,
the azido group can easily be reduced to an amine after
the pseudodipeptide has been coupled to a peptide.16
Aminoalcohols 5 and 6 were converted to azides 7 and 8
in 68and 90% yields, respectively, via a copper-cata-
lyzed diazo transfer reaction.17 In a first attempt, alky-
lation of azidoalcohol 7 with (R)-2-bromopropionic
acid was carried out in dioxane with sodium hydride as
base. This gave pseudodipeptide 9 in 68% yield, but the
product was found to consist of a mixture of diaster-
Scheme 2. (a) TfN3, DMAP, CuSO4, CH2Cl2, rt; 68% for 7 and 90%
for 8; (b) (R)-2-chloropropionic acid, NaH, dioxane, rt; 48% for 9 and
63% for 10.
1
eomers (ꢀ4:1) according to H and 13C NMR spectro-
scopy. In an attempt to improve the diastereoselectivity
in this reaction, the bromopropionic acid was exchan-
ged for (R)-2-chloropropionic acid.18 Under conditions
that were otherwise identical to when 2-bromopropionic
acid was used as alkylating agent pseudodipeptides
919,20 and 1021 could be obtained in 48and 63% yields,
respectively, with only 1–3% of the diastereomeric
PheÉ[CH2O]-d-Ala and IleÉ[CH2O]-d-Ala being
1
formed as side-products according to H NMR spec-
Scheme 3. (a) iBuOCOCl, NMM, THF, À10 ꢁC; (b) NaBH4, MeOH,
0 ꢁC; 90% over two steps; (c) morpholine, rt; 81%; (d) TfN3, DMAP,
CuSO4, CH2Cl2, rt; 64%; (e) BrCH2CO2tBu, NaOH, QHSO4, ben-
zene, rt, 98%; (f) HCOOH, rt, 68%.
troscopy. TyrÉ[CH2O]Gly (15) was synthesized in a
similar manner as 9 and 10 (Scheme 3). Thus, conver-
sion of Fmoc-Tyr(tBu)-OH (11) into a mixed anhydride
by treatment with isobutylchloroformate in the presence
of N-methyl morpholine, followed by reduction using
sodium borohydride,22 and Fmoc-deprotection with
morpholine gave l-tyrosinol (12, 72%). Compound 12
was then converted into azide 13 by diazo transfer17
(64%) after which O-alkylation of 13 with tert-butyl
bromoacetate under phase-transfer catalysis gave pro-
tected methylene ether isostere 14 (98%). Deprotection
of the tert-butyl ester, as well as the tert-butyl ether, in
formic acid gave TyrÉ[CH2O]Gly pseudodipeptide 15
(68%).23 Finally, pseudodipeptides 9 and 10 were
incorporated in desmopressin analogue 1624 and the
type II collagen derived glycopeptide 1725 on solid phase
using conditions described previously (Scheme 4; 23 and
28% yields, respectively).11,13 After attachment of pseu-
dodipeptides 9 and 10 to the solid phase, the azido
Scheme 4. (a) Solid-phase peptide synthesis.

M. Hedenstrom et al. / Bioorg. Med. Chem. Lett. 12 (2002) 841–844
¨
843
the ROESY spectrum and used in simulated annealing
calculations using the program X-PLOR.28 As revealed
by comparison of the lowest energy structures of 16
and desmopressin the two peptides have a high degree
of similarity (Fig. 1). The rmsd when superimposing
all backbone atoms in the macrocyclic ring of the
two lowest energy conformations is only 0.66 A. The
methylene ether isostere in 16 does not allow formation
of the hydrogen bond between Phe3 CO and Asn5 NH,
which is found in the inverse g-turn of desmopressin.
Despite this the backbone conformation is almost iden-
tical for residues Tyr2-Cys6 in the two peptides. In
addition, the positions of the side chains of residues
Tyr2-Cys6 deviate only slightly between 16 and des-
mopressin. Altogether this suggests that insertion of
methylene ether isosteres into peptides may not have a
major influence on their conformations. Methylene
ether isosteres, as well as, for example, aminomethylene,
ketomethylene, and vinylic amide bond replacements,
should therefore be useful tools in attempts to establish
the role of amide bonds in interactions between peptides
and receptors.29À31 Future studies in our laboratory will
focus on incorporation of methylene ether isosters in
biologically active peptides such as hormones, and the
influence this has on structural features and receptor
binding.
Figure 1. Comparison of the lowest energy structures found in aqu-
eous solution for desmopressin and peptide 16. Desmopressin is
represented in green and 16 in red. The backbone atoms (N, Ca, CO)
of residues 1–6 were used for the superimposition and the rmsd
calculations.
In conclusion, an improved method for synthesis of
XaaÉ[CH2O]Ala/Gly pseudodipeptides in good yields
and excellent diastereoselectivity has been developed.
This method is based on reaction of azido alcohols,
which are readily available from amino acids, with (R)-
2-chloropropionic acid or tert-butyl bromoacetate. The
versatility of the pseudodipeptide building blocks pre-
pared in this manner was demonstrated by incorpora-
tion in the drug desmopressin and a T cell stimulatory
glycopeptide from type II collagen. Conformational
studies of the resulting desmopressin analogue revealed
a close similarity to desmopressin, suggesting that
replacement of amide bonds in peptides by methylene
ether isosteres can be done without having a major
effect on the structure of the substituted peptide.
groups were conveniently reduced16 by treatment with
tin(II) chloride in the presence of thiophenol and tri-
ethylamine. However, it was found to be essential to
wash the resin with 20% piperidine in DMF before
proceeding with the synthesis. If washing with piperi-
dine was omitted, incorporation of the following amino
acid did not reach completion. This was assumed to be
due to complexation of tin-salts to the liberated amino
group in the attached pseudodipeptide.
Little is known on how replacement of amide bonds by
methylene ether isosteres affects the secondary structure
of peptides. The structure of desmopressin (Mpa-Tyr-
Phe-Gln-Asn-Cys-Pro-d-Arg-Gly-NH2, Mpa and Cys
form a disulphide bridge) in aqueous solution has been
determined by NMR spectroscopy.7 Investigation of the
structure of analogue 16 would therefore allow an
assessment of the structural influence of insertion of a
methylene ether isostere into a peptide. In desmopressin
the backbone of residues Phe3-Cys6 was found to be
rigid and contained an inverse g-turn centered at Gln4.
Residues Mpa1 and Tyr2 were somewhat more mobile,
whereas the disulphide bridge and Pro7-Gly9 in the
acyclic tail were even more flexible. The structure of
peptide 16 was determined in aqueous solution under
the same conditions as for desmopressin using two-
dimensional NMR techniques. The 1H NMR reso-
nances of 16 were assigned according to standard tech-
niques using COSY, TOCSY and ROESY spectra.26,27
A total of 55 distance restraints were then derived from
Acknowledgements
This work was funded by grants from the Swedish
Research Council and the programme ‘Glycoconjugates
in Biological Systems’ sponsored by the Foundation for
Strategic Research. We are grateful to Kambiz Kateb-
zadeh for helping with the synthesis of 15.
References and Notes
1. Spatola, A. F. In Chemistry andBiochemistry of Amino
Acids, Peptides, and Proteins; Weinstein, B., Ed.; Marcel Dek-
ker: New York, Basel, 1983; Vol. 7, p 267.
2. Rubini, E.; Gilon, C.; Selinger, Z.; Chorev, M. Tetrahedron
1986, 42, 6039.
3. TenBrink, R. E. J. Org. Chem. 1987, 52, 418.
4. Ho, M.; Chung, J. K. K.; Tang, N. Tetrahedron Lett. 1993,
34, 6513.

844
M. Hedenstrom et al. / Bioorg. Med. Chem. Lett. 12 (2002) 841–844
¨
5. Norman, B. H.; Kroin, J. S. J. Org. Chem. 1996, 61, 4990.
6. Breton, P.; Monsigny, M.; Mayer, R. Int. J. Peptide Protein
Res. 1990, 35, 346.
CH3). FAB MS m/z calcd for C12H15N3O3 (M+H+) 250.119,
found 250.119.
21. Compound 10 had: H NMR (400 MHz, CDCl3): d 7.97
1
7. Walse, B.; Kihlberg, J.; Drakenberg, T. Eur. J. Biochem.
1998, 252, 428.
8. Smith, G. D.; Griffin, J. F. Science 1978, 199, 1214.
9. Amodeo, P.; Naider, F.; Picone, D.; Tancredi, T.; Temussi,
P. A. J. Peptide Sci. 1998, 4, 253.
10. Guarnieri, F.; Weinstein, H. J. Am. Chem. Soc. 1996, 118,
5580.
11. Brickmann, K.; Yuan, Z.; Sethson, I.; Somfai, P.; Kihl-
berg, J. Chem. Eur. J. 1999, 5, 2241.
(br s, 1H, COOH), 4.07 (q, J=6.9 Hz, 1H, CHOCH3), 3.88
(dd, J=9.0 and 2.2 Hz, 1H, CH2O), 3.47 (dd, J=9.0 and 8.4
Hz, 1H, CH2O), 3.50 (m, 1H, CHN3), 1.58(m, 1H,
CHCH2CH3), 1.51 (d, J=6.9 Hz, 3H, CHOCH3), 1.48(m,
1H, CHCH2CH3), 1.22 (m, 1H, CHCH2CH3), 0.92 (d, J=6.8
Hz, 3H, CHCH3), 0.90 (t, J=7.6 Hz, CH2CH3). FAB MS m/z
calcd for C9H17N3O3 (M+H+) 216.135, found 216.136.
22. Kokotos, G. Synthesis 1990, 1990, 299.
1
23. Compound 15 had: H NMR (400 MHz, CDCl3): d 7.09
12. Kreye, P.; Kihlberg, J. Tetrahedron Lett. 1999, 40, 6113.
13. Broddefalk, J.; Backlund, J.; Almqvist, F.; Johansson, M.;
Holmdahl, R.; Kihlberg, J. J. Am. Chem. Soc. 1998, 120, 7676.
14. Cort, J. H.; Fric, I.; Carlsson, L.; Gillessen, D.; Bystricky,
S.; Skopkova, J.; Gut, V.; Studer, R. O.; Mulder, J. L.; Blaha,
K. Mol. Pharmacol. 1976, 12, 313.
15. Emtenas, H. Masters Thesis, Umea University, 1998.
16. Bartra, M.; Romea, P.; Urpi, F.; Vilarrasa, J. Tetrahedron
1990, 46, 587.
(d, J=8.5 Hz, 2H, Ph), 6.80 (d, J=8.5 Hz, 2H, Ph), 4.21 (s,
2H, OCH2CO), 3.75 (m, 1H, CHN3), 3.67 (dd, J=3.7 and 9.8
Hz, 1H, CHN3CH2O), 3.55 (dd, J=6.7 and 9.8Hz, 1H,
CHN3CH2O), 2.83 (dd, J=6.5 and 14.0 Hz, CH2PhOH), 2.77
(dd, J=7.5 and 14.0 Hz, CH2PhOH). FAB MS m/z calcd for
C11H13N3O3 (M+Na+) 274.079, found 274.083.
24. Peptide 16 had: FAB MS calcd for C44H63N12O11S2 999
(M+H), found 999.
25. Glycopeptide 17 had: ES MS calcd for C46H74N10O18
1055 (M), found 1055.
26. Wuthrich, K. NMR of Proteins andNucleic Acids ; John
Wiley & Sons: New York, 1986.
27. The NMR sample was prepared by dissolving 16 (5 mg) in
aqueous phosphate buffer (40 mM, 550 mL) containing 10%
D2O so that a 9 mM solution at pH 6.6 was obtained. All
NMR experiments were conducted at 278K on a Bruker
600 MHz DRX spectrometer.
28. Brunger, A. T. X-PLOR, Version 3.1. A System for X-ray
Crystallography andNMR ; Yale Universtity Press: New
Haven, CT, 1992.
29. Hocart, S. J.; Nekola, M. V.; Coy, D. H. J. Med. Chem.
1988, 31, 1820.
30. Chandrakumar, N. S.; Yonan, P. K.; Stapelfeld, A.;
Savage, M.; Rorbacher, E.; Contreras, P. C.; Hammond, D. J.
Med. Chem. 1992, 35, 223.
17. Alper, P. B.; Hung, S.-C.; Wong, C.-H. Tetrahedron Lett.
1996, 37, 6029.
18. Andersson, L.; Kenne, L. Carbohydr. Res. 1998, 313, 157.
19. Conditions used for synthesis of pseudodipeptide 9: A
mixture of 7 (89 mg, 0.50 mmol), (R)-2-chloropropionic acid
(125 mg, 1.15 mmol) and NaH (50% in mineral oil, 610 mg,
12.7 mmol) in 1,4-dioxane (25 mL) was stirred for 16 h at rt.
Then H2O (9 mL) was added to destroy the excess of NaH.
The upper phase was was washed with heptane, acidified with
2 M HCl, and extracted with CH2Cl2. The resulting organic
phase was then dried with MgSO4. After filtration and con-
centration in vacuo, the crude product was purified by flash
chromatography (heptane/ethyl acetate 6:1+2% HOAc) on
silica gel to give 9 as a colorless oil (60 mg, 48%).
1
20. Compound 9 had: H NMR (400 MHz, CDCl3): d 10.11
(br s, 1H, COOH), 4.05 (q, J=6.9 Hz, 1H, CHCH3), 3.83 (m,
2H, CH2O and CHN3), 3.41 (dd, J=10.3 and 8.0 Hz, 1H,
CH2O), 2.77 (dd, J=13.5 and 8.0 Hz, 1H, PhCH2), 2.68(dd,
J=13.5 and 5.8Hz, 1H, PhC H2), 1.52 (d, J=7.0 Hz, 3H,
31. Angelastro, M. R.; Marquart, A. L.; Mehdi, S.; Koehl,
J. R.; Vaz, R. J.; Bey, P.; Peet, N. P. Bioorg. Med. Chem. Lett.
1999, 9, 139.