Synthesis of amides from unprotected amino acids by a simultaneous protection-activation strategy using dichlorodialkyl silanes

Article abstract of DOI:10.1016/S0040-4039(02)02275-X

Reaction of L-phenylalanine (1) with dichlorodimethylsilane or other dichlorosilane derivatives 6b-d and primary amines leads to the formation of amides probably via a cyclic silyl intermediate. It is also possible to use β-amino acids and N-alkylated amino acids (peptoid building blocks) as well as the amino dicarboxylic acid L-aspartic acid. The latter leads to almost exclusive formation of the α-amide.

Full text of DOI:10.1016/S0040-4039(02)02275-X

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 43 (2002) 9203–9207
Synthesis of amides from unprotected amino acids by a
simultaneous protection–activation strategy using
dichlorodialkyl silanes
S. H. van Leeuwen,a P. J. L. M. Quaedflieg,b Q. B. Broxtermanc and R. M. J. Liskampa,*
aDepartment of Medicinal Chemistry, Utrecht Institute for Pharmaceutical Sciences, Utrecht University, PO Box 80082,
3508 TB Utrecht, The Netherlands
bDSM Fine Chemicals Netherlands, R&D Centre Venlo, PO Box 81, 5900 AB Venlo, The Netherlands
cDSM Fine Chemicals—Advanced Synthesis and Catalysis (DFC-ASC), PO Box 18, 6160 MD Geleen, The Netherlands
Received 3 July 2002; accepted 4 October 2002
Abstract—Reaction of
amines leads to the formation of amides probably via a cyclic silyl intermediate. It is also possible to use b-amino acids and
N-alkylated amino acids (peptoid building blocks) as well as the amino dicarboxylic acid -aspartic acid. The latter leads to
almost exclusive formation of the a-amide. © 2002 Elsevier Science Ltd. All rights reserved.
L-phenylalanine (1) with dichlorodimethylsilane or other dichlorosilane derivatives 6b–d and primary
L
The coupling of amino acids without the requirement
of separate protection/deprotection and coupling steps
is one of the great challenges in peptide synthesis. For
instance, the number of steps involved in the synthesis
of a dipeptide would then be greatly reduced and
therefore the cost-price for industrial production would
be drastically reduced.
In this communication the first step is described
towards using cost-effective silyl reagents for the prepa-
ration of peptide amides in a simultaneous protection–
activation strategy.4
When
L
-phenylalanine (1) was treated in dichloro-
methane with commercially available dichlorodimethyl-
silane followed by addition of benzylamine under the
conditions reported by Barlos et al. for the protection
of histidine,5 neither (silyl) intermediate nor product
could be detected. However, it was speculated that the
In order to meet this challenge, a ‘simultaneous N-pro-
tection and carboxyl-activation’ strategy is called for.
The very interesting approaches already reported in the
literature represent significant steps towards realization
of this strategy. In the amino acid N-carboxy anhydride
(NCA) approach, phosgene is used to protect simulta-
neously the amino function and to activate the car-
boxylic acid moiety.1 As well as problems concerning
the use of phosgene on a large scale, the carboxylic acid
moiety is activated to such an extent that it is difficult
to limit the coupling to, e.g. the dipeptide stage, and
consequently it is difficult to avoid polymerization to
oligo- and polypeptides.2 The elegant approach by
Burger and Rudolph3 has as a disadvantage the use of
the toxic and expensive hexafluoroacetone, which pre-
cludes this approach e.g. for the preparation of the
dipeptide-methyl ester aspartame on an industrial scale.
very poor solubility of L-phenylalanine as well as the
liberation of HCl might impede the formation of and/
or cause premature decomposition of the silyl interme-
diate 2 and consequently prevent the formation of
L
-phenylalanine-benzylamide 3a (Scheme 1). Therefore,
the solvent was changed to pyridine, in which -phenyl-
L
alanine almost completely dissolves and which can act
as an acid scavenger. Now, the reaction with dichloro-
dimethylsilane and benzylamine proceeded almost
quantitatively at room temperature to afford amide 3a.6
Carrying out the reaction at higher temperatures led to
lower yields and no product could be detected at reflux
temperature.7 It turned out that pyridine was the best
solvent/base for this reaction.8 It was also found that
this reaction is very sensitive to the bulkiness of the
incoming amine nucleophile. As a result, primary
amines with their amino group attached to a primary
carbon atom gave excellent yields of amides 3a–d
(Scheme 1). However, the dipeptide 3e, for which
* Corresponding author. Tel.: +31 30 253 7396; fax: +31 30 253 6655;
e-mail: r.m.j.liskamp@pharm.uu.nl
0040-4039/02/$ - see front matter © 2002 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(02)02275-X

9204
S. H. 6an Leeuwen et al. / Tetrahedron Letters 43 (2002) 9203–9207
Scheme 1.
H–Phe–OMe-containing a primary amine attached
to a secondary carbon atom was used as the amine
nucleophile, was obtained in a much lower yield (17%).
The use of isopropylamine or 2-aminobutane resulted
in even lower yields of 3f and 3g (55%, Scheme 1)
and starting material 1 was recovered.
after approximately 0.6 equiv. dichlorodimethylsilane,
the final activated compound became visible. This
observation corroborates that the carboxylic acid func-
tionality of
L-phenylalanine first reacts with dichloro-
dimethylsilane to give the linear C(O)ꢀOꢀSi(CH3)2ꢀCl
structure, which then cyclizes to give structure 2.
Since all attempts to isolate the reactive intermediate-
Realizing that the silyl intermediate 2 was too unstable
postulated to be the cyclic silane 2 failed, indirect
to be isolated,
L
-phenylalanine was now treated with
sterically
mechanistic evidence was collected. Firstly,
a
di-tert-butyldichlorosilane to obtain
a
crowded silyl intermediate (4: R3=R4=tert-Bu,
Scheme 2) which it was hoped would be less susceptible
to nucleophilic attack at room temperature and there-
fore could be isolated. However, di-tert-butyldichlorosi-
lane is apparently sterically crowded to such an extent
that instead of 4, the disilylated cyclic derivative 5 was
isolated in 40% yield (Scheme 2).10 This derivative
however appeared to be resistant towards nucleophilic
opening and treatment with benzylamine or methyl-
amine at various temperatures did not lead to any
dichlorodimethylsilane-mediated coupling reaction of
acetic and benzoic acid with benzylamine did not fur-
nish the amide product, indicating that the activated
system in the case of
dicarboxydimethylsilyl (C(O)ꢀOꢀSi(CH3)2ꢀOꢀC(O)) or
C(O)ꢀOꢀSi(CH3)2ꢀCl structure. Secondly,
dichlorodimethylsilane-mediated coupling reaction of
N-benzyloxycarbonyl- -phenylalanine, N,N-dibenzyl-
phenylalanine and N-phthaloyl- -phenylalanine with
L-phenylalanine has not simply a
a
a
L
L-
L
benzylamine did not give the amide product either,
indicating that a free a-amine functionality is required
indirectly proving that 2 is indeed the reactive interme-
diate. Thirdly, NMR experiments9 were performed with
detectable
L-phenylalanine-amide product. Neverthe-
less, it appeared to be possible to use other commer-
cially available dichlorosilane derivatives (6b–d, Scheme
2).11 Derivatives 6c–d were less bulky than di-tert-
butyldichlorosilane but more so than dichlorodimethyl-
silane and consequently resulted in lower yields of the
adduct. As expected only the somewhat less bulky
chloromethylsilane 6b gave a comparable yield of 3a.
For industrial use, however, dichlorodimethylsilane is
the reagent of choice since being the raw material for
silicones it is very cost-effective, available in large quan-
tities and easy to handle.
L
-phenylalanine in pyridine-d5 to which varying
amounts of dichlorodimethylsilane were added, giving
independent support for the existence of a cyclic active
intermediate. When 1.2 equiv. of dichlorodimethyl-
silane was added, one single activated compound was
1
formed9 according to the H NMR spectrum bearing
one dimethylsilyl functionality per amino acid residue
and which was neither the
L-phenylalanine–HCl salt
nor its acid chloride congener. However, when
dichlorodimethylsilane was titrated in the NMR tube it
was observed that initially
transformed to another intermediate (also bearing one
dimethylsilyl functionality per amino acid residue) and
L-phenylalanine was first
Interestingly, it was also possible to apply this dichloro-
dialkylsilane-mediated amidation strategy to other

S. H. 6an Leeuwen et al. / Tetrahedron Letters 43 (2002) 9203–9207
9205
Scheme 2.
amino acid types.11 As examples the reactions of N-
dipeptide sweetener aspartame (
L
-aspartyl-
L
-phenylala-
methyl glycine (7: sarcosine) and b-
L
-alanine (8) are
nine methyl ester: APM).15,16 For the preparation of
this dipeptide it is crucial that selective amidation
occurs at the a-carboxylic acid moiety.
shown here (Scheme 3).12 N-Methyl glycine is an N-
alkylated amino acid derivative, which is used in the
synthesis of a very important class of peptoid pepti-
domimetics.13 b-Amino acids like b-
L
-alanine are the
Indeed, it was found that it was possible to direct the
regiochemistry of the reaction of L-aspartic acid using
building blocks of the increasingly important b-pep-
tides.14 The lower yield of b-alaninebenzylamide 10 in
dichlorodimethylsilane. Use of pyridine as the base and
solvent as described above led to an about equal (52:48)
mixture of a- and b-products (12 and 13, respectively)
in an overall yield of 44%. Pleasingly, when using
triethylamine (TEA) as the base and solvent instead, a
very good selectivity in favor of the a-product 12 was
observed (a:b ratio of 99.4:0.6 with an overall amide
yield of 51.3%). The highly selective formation of the
a-product is consistent with the assumption that the
stronger base TEA leads to the preferred formation of
the kinetically favored five-membered cyclic intermedi-
ate (Scheme 4).17 In the case of the weaker base pyri-
the reaction of b-L-alanine as compared to the yield of
peptoid amide 9 can be explained assuming the forma-
tion of a six-membered cyclic silyl intermediate, which
may be less easily formed or less reactive than a five-
membered cyclic silyl intermediate (Scheme 3).
The next challenge was to investigate the regiochem-
istry if two carboxylic acid moieties are present in the
starting amino acids as is the case in
glutamic acid (see Scheme 4). We focused on
L
-aspartic acid and
-aspartic
L
acid 11 since this amino acid is a constituent of the
commercially important high-intensity low caloric
dine, the formation of a significant amount of
Scheme 3.

9206
S. H. 6an Leeuwen et al. / Tetrahedron Letters 43 (2002) 9203–9207
Scheme 4.
phenylalanine (330 mg, 2.0 mmole) was suspended in
pyridine (10 mL, distilled from CaH2) under a nitrogen
atmosphere at ambient temperature (approx. 23°C) and
dichlorodimethylsilane (267 mL, 2.2 mmole) was added in
one portion. During 2 min the mixture turned clear and
the temperature rose from 23 to 28°C. Subsequently, 6
mmole of the selected amine was added (in most cases
giving a suspension) and the reaction mixture was stirred
for another 16 h and monitored with TLC (using chloro-
form/methanol/ammonia 60/45/20 v/v/v, sec-butanol/for-
mic acid/water 75/15/10 v/v/v and n-butanol/acetic
acid/ethyl acetate/water 1/1/1/1 v/v/v/v). The reaction
mixture was concentrated in vacuo and the resulting
residue chromatographed on silica gel (eluent:
dichloromethane/methanol 100/0 to 95/5 v/v) to furnish
b-product is consistent with the assumption that more
of the thermodynamically favored six-membered ring
intermediate is formed.
In conclusion, we have shown that it is possible to
prepare amino acid amides via a simultaneous protec-
tion and activation strategy using commercially
available silylating agents, in particular dichloro-
dimethylsilane. Moreover, it is possible to prepare the
a-amide derived from L-aspartic acid with high regiose-
lectivity. Obviously the scope of this method has to be
further investigated and widened to the preparation of
dipeptide derivatives.
the pure L-phenylalanine-amide.
7. The silyl intermediate probably easily decomposes at
higher temperatures.
References
1. See for example and references cited therein: (a) Pfaen-
der, P.; Kuhnle, E.; Krahl, B.; Backmansson, A.;
Gnauck, G.; Blecher, H. Hoppe-Seyler’s Z. Physiol.
Chem. 1973, 354, 267–285; (b) Katakai, R.; Oya, M.;
Uno, K.; Iwakura, Y. J. Org. Chem. 1972, 37, 327–329;
(c) Hirschmann, R.; Schwam, H.; Strachan, G.; Schoe-
newaldt, E. F.; Barkemeyer, H.; Miller, S. M.; Conn, J.
B.; Garsky, V.; Veber, D. F.; Denkewalter, R. G. J. Am.
Chem. Soc. 1971, 93, 2746–2754.
8. Besides pyridine the following solvents were tried in the
coupling reaction of
using dichlorodimethylsilane: THF, dichloromethane,
N,N-dimethyl-acetamide (DMA), N,N-dimethylfor-
L-phenylalanine with benzylamine
mamide, N-methylpyrrolidone (NMP), 1,4-dioxane, 1,2-
dichloroethane, acetone, DMSO, chloroform, acetonitrile
and sulpholane; the maximum
amide yield obtained was 25% (using DMA and NMP).
When triethylamine (2 equiv. based on -phenylalanine)
L-phenylalanine-benzyl-
L
2. Deming, T. J. Nature 1997, 390, 386–389.
was added before the addition of dichlorodimethylsilane,
the yields in these solvents significantly increased (for
instance to 80% in DMA and NMP, to 50% in acetoni-
trile and 48% when triethylamine was used as the solvent
itself) but they were still lower than in pyridine (98%).
3. Burger, K.; Rudolph, M. Chem. Ztg. 1990, 114, 249–251.
4. Quaedflieg, P. J. L. M.; Broxterman, Q. B.; Van
Leeuwen, S. H.; Liskamp, R. M. J. PCT Int. Appl. 2000,
21 pp. WO 0037484 A1 20000629 CAN 133:59104 AN
2000:441809.
5. Dichlorodimethylsilane was used under these conditions
by Barlos et al. in the simultaneous protection of the
amino and carboxylic acid moiety of histidine in order to
protect selectively its imidazole function with a trityl
protecting group: (a) Barlos, K.; Papaioannou, D.;
Theodoropoulos, D. J. Org. Chem. 1982, 47, 1324–1326;
(b) Eleftheriou, S.; Gatos, D.; Panagopoulos, A.; Statho-
poulos, S.; Barlos, K. Tetrahedron Lett. 1999, 40, 2825–
2828.
9. A solution of
L-phenylalanine in pyridine-d5 was trans-
ferred into a dry NMR tube under a nitrogen atmosphere
and dichlorodimethylsilane was added via a micro-
syringe. When 1.2 equiv. of dichlorodimethylsilane was
added in one portion, the single activated compound
displayed the following chemical shifts (in pyridine-d5,
with the most downfield pyridine-H at 8.50 ppm): 7.2–6.9
ppm (5 H, m), 4.42 ppm (1 H, dd), 3.18 ppm (1 H, dd),
2.84 ppm (1 H, dd) and 0.40 (6 H, bs). To prove that this
compound was not simply the
L-phenylalanine acid chlo-
6. General experimental procedure for the preparation of
ride·HCl salt, the latter was separately synthesized from
L
-phenylalanine-amides 3a–g as shown in Scheme 1: L-

S. H. 6an Leeuwen et al. / Tetrahedron Letters 43 (2002) 9203–9207
-phenylalanine and PCl5 in dichloromethane; in pyridine
9207
L
Kania, R. S.; Zuckermann, R. N.; Huebner, V. D.;
Jewell, D. A.; Banville, S.; Ng, S.; Wang, L.; Rosenberg,
S.; Marlowe, C. K. D.; Spellmeyer, C.; Tan, R.; Frankel,
A. D.; Santi, D. V.; Cohen, F. E.; Barlett, P. A. Proc.
Natl. Acad. Sci. USA 1992, 89, 9367–9371; (d) Kruijtzer,
J. A. W.; Hofmeyer, L. J. F.; Heerma, W.; Versluis, C.;
Liskamp, R. M. J. Chem. Eur. J. 1998, 4, 1570–1580.
14. See for example: (a) Appella, H.; Christianson, L. A.;
Klein, D. A.; Richards, M. R.; Powell, D. R.; Gellman,
S. H. J. Am. Chem. Soc. 1996, 118, 13071; (b) Seebach,
D.; Ciceri, P.; Overhand, M.; Juan, B.; Rigo, D.; Oberer,
L.; Hommel, U.; Amstutz, R.; Widmer, H. Helv. Chim.
Acta 1996, 79, 2043; (c) Seebach, D.; Matthews, J. L.
Chem. Commun. 1997, 2015; (d) Appella, H.; Christian-
son, L. A.; Klein, D. A.; Richards, M. R.; Powell, D. R.;
Gellman, S. H. J. Am. Chem. Soc. 1999, 121, 7574; (e)
Wang, X.; Espinosa, J. F.; Gellman, S. H. J. Am. Chem.
Soc. 2000, 122, 4821.
this compound rapidly polymerized.
10. Compound 5: 1H NMR (pyridine-d5): l 1.20 (bs, 18H,
CH3, tBu), 1.25 (bs, 18H, CH3, tBu), 2.93 (dd, 1H, CHA,
J=8.8 Hz, J=13.2 Hz), 3.36 (dd, 1H, CHB, J=4.8 Hz,
J=13.6 Hz), 3.96 (dd, 1H, aCH, J=4.8 Hz, J=8.8 Hz),
7.24–7.56 (m, 5H, Ph). 13C NMR (CDCl3): l 19.8 (Cq,
tBu), 21.2 (Cq, tBu), 27.3, 27.6, 27.7 (CH3, tBu), 40.9
(CH2), 57.0 (aCH2), 126.4–129.3 (CH, Ph), 137.4 (Cq,
Ph), 173.2 (C=O). Mass C25H45O3NSi2 mw 463; [M+H+]
464.4; Mass C17H27O2NSi mw 305; [M+H+] 306.
11. General experimental procedure for the preparation of
L
-phenylalanine-benzylamide using the dichlorodialkylsi-
lane reagents as shown in Scheme 2. -Phenylalanine (330
L
mg, 2.0 mmole) was suspended in pyridine (10 mL,
distilled from CaH2) under a nitrogen atmosphere at
ambient temperature (approx. 23°C) and 2.2 mmole of
the dichlorodialkylsilane was added in one portion. After
1 min 3–5 equiv. of benzylamine was added and the
reaction mixture was stirred for another 16 h, monitored
with TLC and worked up (as above).
15. (a) Mazur, R. H.; Schlatter, J. M.; Goldkamp, A. H. J.
Am. Chem. Soc. 1969, 91, 2684–2694; (b) Ager, D. J.;
Pantaleone, D. P.; Henderson, S. A.; Katritzky, A. R.;
Prakash, I.; Walters, D. E. Angew. Chem., Int. Ed. 1998,
37, 1802–1817.
12. The same experimental procedure was applied as given in
Ref. 6.
16. For an NCA-route to aspartame: Tou, J. S.; Vineyard, B.
D. J. Org. Chem. 1985, 50, 4982–4984.
13. See for example and references cited therein: (a) Zucker-
mann, R. N.; Martin, E. J.; Spellmeyer, D. C.; Stauber,
G. B.; Shoemaker, K. R.; Kerr, J. M.; Figliozzi, G. M.
D.; Goff, A.; Siani, M. A.; Simon, R. J.; Banville, S. V.;
Brown, E. G.; Wang, L.; Richter, L. S.; Moos, W. H. J.
Med. Chem. 1994, 37, 2678–2685; (b) Nguyen, J. T.;
Turck, C. W.; Cohen, F. E.; Zuckermann, R. N.; Lim,
W. A. Science 1998, 282, 2088–2092; (c) Simon, R. J.;
17. For the coupling experiments with
similar experimental procedure as given in Ref. 6 was
used. The a-product (a- -aspartic acid-benzylamide) and
the b-product (b- -aspartic acid-benzylamide) were deter-
L-aspartic acid, a
L
L
mined on HPLC by comparison of the peak areas with
those of reference solutions of known concentration.