Titanium catalysis in the Ugi reaction of α-amino acids with aromatic aldehydes

Article abstract of DOI:10.1021/ol048850x

(Chemical Equation Presented) Ugi reaction between an (S)-α-amino acid, an aromatic aldehyde, and an isonitrile proceeds best under catalysis by TiCl₄ in MeOH. The sense of diastereoinduction is (S,S).

Full text of DOI:10.1021/ol048850x

ORGANIC
LETTERS
2004
Vol. 6, No. 19
3281-3284
Titanium Catalysis in the Ugi Reaction
of r-Amino Acids with Aromatic
Aldehydes
Thomas Godet,† Yannick Bonvin,† Guillaume Vincent,† Daphne´ Merle,‡
,†
Alain Thozet,‡ and Marco A. Ciufolini*
Laboratoire de Synthe`se et Me´thodologie Organiques, CNRS UMR 5181, and Centre
Commun de Diffractome´trie, UniVersite´ Claude Bernard Lyon 1 and Ecole Supe´rieure
de Chimie, Physique et Electronique de Lyon, 43, Bd du 11 NoVembre 1918,
69622 Villeurbanne, France
ciufi@cpe.fr
Received June 17, 2004
ABSTRACT
Ugi reaction between an (S)-r-amino acid, an aromatic aldehyde, and an isonitrile proceeds best under catalysis by TiCl4 in MeOH. The sense
of diastereoinduction is (S,S).
An ongoing synthetic project required a method for the rapid
assembly of hybrid arylglycine-pyroglutamic acid intermedi-
ates 3. These substances may be prepared by amidoalkylation
chemistry1 and related reactions.2 An attractive alternative
involves an Ugi condensation of (S)-glutamic acid 5-methyl
ester, 1, an aromatic aldehyde, and an isonitrile (Scheme 1).
Ugi reactions3 of R-amino acids are documented in pioneer-
ing reports by Ugi himself,4 but overall they have received
only sporadic attention,5 and their sense of diastereoinduction
has been the subject of conflicting reports.6 Moreover, the
literature records only instances of such reactions with
aliphatic aldehydes. This prompted us to investigate the
desired transformations in some detail. Herein, we describe
initial results in this area.
Reaction of 1 with 4 and 5 under Ugi conditions4 was
slow. The reaction was accelerated by heating to 40 °C,
whereupon even after 4 days some starting material (5-10%)
remained. Compound 6 was formed as the major product of
a 6:1 mixture of diastereomers but in a modest 30% yield
(Scheme 2). The stereochemical assignment rests on an X-ray
† Laboratoire de Synthe`se et Me´thodologie Organiques.
‡ Centre Commun de Diffractome´trie.
(1) Cf.: Roth, E.; Altman, J.; Kapon, M.; Ben-Ishai, D. Tetrahedron
1995, 51, 801.
Scheme 1. Possible Avenue to Desired Intermediate 3
(2) E.g., Gosselin, F.; Roy, A.; O’Shea, P. D.; Chen, C.; Volante, R. P.
Org. Lett. 2004, 6, 641.
(3) Recent review: Do¨mling, A.; Ugi, I. Angew. Chem., Int. Ed. 2000,
39, 3168.
(4) For instance: (a) Demharter, A.; Ho¨rl, W.; Herdtweck, E.; Ugi, I.
K. Angew. Chem., Int. Ed. Engl. 1996, 35, 173. (b) Ugi, I.; Ho¨rl, W.;
Hanusch-Kompa, C.; Schmid, T.; Herdtweck, E. Heterocycles 1998, 47,
955.
(5) For instance: (a) Janvier, P.: Sun, X.; Bienayme´, H.; Zhu, Z. J. Am.
Chem. Soc. 2001, 124, 2560. (b) Dyker, G.; Breitenstein, K.; Henkel, G.
Tetrahedron: Asymmetry 2002, 13, 1929. Brief review: (e) Dyker, G.
Angew. Chem., Int. Ed. Engl. 1997, 36, 1700.
(6) Cf. Park, S. J.; Keum, G.; Kang, S. B.; Koh, H. Y.; Kim, Y.
Tetrahedron Lett. 1998, 39, 7109.
10.1021/ol048850x CCC: $27.50 © 2004 American Chemical Society
Published on Web 08/18/2004

cal to those of the uncatalyzed process, suggesting that no
proton source is really necessary beyond the substrate. By
contrast, a number of Lewis acids had a beneficial effect on
yields and rates. Best results were observed with TiCl4 as
the catalyst.10 Thus, the transformation proceeded to comple-
tion in less than 24 h and the yield of desired product doubled
relative to the uncatalyzed reaction. On the other hand, only
small changes in diastereomeric ratios were detected as a
function of the promoter, and transformations catalyzed by
TiCl4 afforded essentially the same diastereomeric ratios (6:1
in favor of the (S,S)-diastereomer) as the uncatalyzed process.
This seems to imply that the catalyst is involved primarily
in the formation of an iminium ion from 1 and 4 but that
probably it plays a marginal role in the nucleophilic addition
of the isonitrile thereto.
Scheme 2. Uncatalyzed Ugi Rection of Glutamate 1
Complex mixtures of products were again obtained when
diffractometric study of crystalline 8, a congener of 3
obtained by refluxing a toluene solution of 6. In accord with
Ugi’s work, the sense of diastereoinduction is (S,S), but the
kinetics, yield, and diastereoselectivity of the reaction were
significantly poorer than expected on the basis of those
reports. However, Ugi’s published results with aliphatic
aldehydes were readily duplicated, signaling that the in-
efficiency of the transformation leading to 6 is attributable
to the aromatic nature of aldehyde 4.
Table 1. TiCl4-Catalyzed Ugi Reactions of 5-Methyl
Glutamate
Reasoning that the observed difficulties were likely to
ensue from the reduced electrophilicity of aromatic aldehydes
relative to aliphatic ones and, consequently, from a slower
rate of formation of an iminium ion intermediate from 1 and
4, we explored the effect of solvents and of Lewis acid
promoters on the process.
Alternative reaction solvents were no cure for the problem.
In particular, isonitrile-based multicomponent condensations
may proceed more efficiently in trifluoroethanol (TFE)6 or
in water.7 In the present case, the desired material was
obtained as a component of a complex mixture of products
when the reaction was conducted in TFE, whereas in water,
formation of the desired product was suppressed altogether,
perhaps because of the poor solubility of 4 in an aqueous
medium. The use of surfactants as a means to circumvent
such solubility issues8 would introduce needless complica-
tions and was not explored.
Turning then to the effect of acidic catalysts, we screened
2 protonic acids (MsOH and TFA) and 13 Lewis acids9 as
promoters of the reaction of 1 with 4 and 5. These
experiments were carried out at a constant 5 mol % amount
of catalysts, over reaction times ranging from 12 to 144 h.
Bronsted acids proved to be ineffective promoters, and indeed
they may be harmful. For instance, no desired product was
obtained from a reaction run in the presence of 5 mol %
MsOH, perhaps due to polymerization of the isonitrile,
whereas the weaker TFA produced results essentially identi-
(7) Pirrung, M. C.; Das Sarma, K. J. Am. Chem. Soc. 2004, 126, 444.
(8) Cf., e.g.: Saito, K.; Tago, T.; Masuyama, T.; Nishide, H. Angew.
Chem., Int. Ed. 2004, 43, 730
(9) Lewis acids have been used in certain multicomponent condensations
involving isonitriles; cf.: Keung, W.; Bakir, F.; Patron, A. P.; Rogers, D.;
Priest, C. D.; Darmohusodo V. Tetrahedron Lett. 2004, 45, 733.
a All reactions carried out in a 0.15 M solution at room temperature for
12 h. b Chromatographed yield of the mixture of 9 and 10. c In parentheses:
chromatographed yield of the mixture of 9/10 for the uncatalyzed reaction
run under the conditions of refs 4.
3282
Org. Lett., Vol. 6, No. 19, 2004

the reaction was carried out in TFE instead of MeOH. On
the other hand, the use of t-Bu-NC, in lieu of 5, resulted in
an appreciable improvement in yields (>75%) and kinetics
(completion in 6-12 h). A summary of several reactions of
1 run under optimized conditions appears in Table 1.11 Yields
are significantly higher and reaction times are considerably
shorter relative to the uncatalyzed reactions run under Ugi’s
original conditions.4
The optimal procedure devised for 1 proved to be quite
suitable for analogous reactions of other amino acids with
aromatic aldehydes. Representative examples with valine,
serine, phenylalanine, and tryptophane appear in Table 2.
the in situ generation of strong protonic acids such as HCl.
We tend to ascribe the catalytic effect of the adjuvant
primarily to its Lewis acidity. The notion that the efficacy
of the promoter may be attributable to in situ release of HCl
generated through methanolysis of TiCl4 would be incon-
sistent with the observation that strong Bronsted acids are
poor promoters for the reaction. Indeed, a reaction between
1, 4, and t-Bu-NC run in the presence of 5% HCl in MeOH13
showed little improvement in yield (42%) or selectivity (6:
1) relative to the uncatalyzed process, just as observed with
TFA. Furthermore, Ti(OiPr)4, which cannot generate HCl,
displayed catalytic efficacy comparable to that of TiCl4
(Scheme 3). However reactions catalyzed by Ti(OiPr)4
Table 2. TiCl4-Catalyzed Ugi Reactions of Other Amino Acids
Scheme 3. Catalytic Effect of Ti(OiPr)4
proceeded somewhat slower and afforded slightly lower
stereoselectivities, rendering TiCl4 the promoter of choice.
Amides 9 and 10 undergo a useful dealkylative cyclization
under acidic conditions. For instance, brief refluxing of 9d
in trifluoroacetic acid induced conversion to 14 (Scheme 4),
Scheme 4. Dealkylative Cyclization of Amides 9
a All reactions carried out in a 0.15 M solution at room temperature for
12 h. b Chromatographed yield of the mixture of 12 and 13. c In parentheses:
chromatographed yield of the mixture of 12/13 for the uncatalyzed reaction
run under Ugi conditions (refs 4).
a crystalline substance whose structure was confirmed by
single-crystal X-ray diffractometry.14 Notice again the struc-
tural analogy between 14 and 3. The use of the weaker AcOH
in this reaction resulted only in formation of 15, a substance
more readily obtained by plain thermal activation of the Ugi
product (cf. Scheme 2).
Again, the yields of the much faster catalyzed reaction (12
h or less vs 4 days) were double or triple those of the
uncatalyzed process conducted as detailed in refs 4.12
The unusual nature of our reaction medium begs the
question of whether the catalytic efficacy of TiCl4 is primarily
attributable to its Lewis acidity or to its being an agent for
(11) Typical Procedure. A 0.15 M solution of 1 and an aldehyde in
MeOH was stirred at rt for 20 min, and then 1.2 equiv of t-Bu-NC was
added, followed by 5 mol % TiCl4. The mixture was stirred at rt for 12 h,
and then it was diluted with EtOAc and washed with water. The organic
phase was dried (Na2SO4), filtered, and concentrated. The residue was
purified by silica gel chromatography.
(12) The TiCl4/MeOH procedure is applicable to reactions involving
aliphatic aldehydes. While excellent yields of products are obtained, the
substantial increase of the rate of such reactions results in significant erosion
of diastereoselectivity relative to the original Ugi procedure, consistent with
the reactivity-selectivity principle.
(13) Prepared by addition of acetyl chloride to dry MeOH.
(14) The same product was obtained upon refluxing a solution of 9e in
HCOOH but in a diminished 30% yield.
(10) Other Lewis acids examined in the course of this work: AlMe3 (72
h, 39% yield, diastereomeric ratio S,S/S,R ) 10), BF3OEt2 (complex
mixture); Cu(OAc)2 (144 h, 37%, dr ) 10), Cu(OTf)2 (144 h, 37%, dr )
3), Eu(hfc)3 (72 h, 39%, dr ) 11), Sc(OTf)3 (24-96 h, 40-45%, dr ) 10),
Cp2TiCl2 (72 h, 33%), YbCl3 (48-72 h, 45%, dr ) 5-9), Yb(OTf)3 (48 h,
40%, dr ) 5), Yb(hfc)3 (24-96 h, 46-48%, dr ) 9), Yb(fod)3 (48 h, 40%,
dr ) 10), ZnCl2OEt2 (72 h, 50%, dr ) 9).
Org. Lett., Vol. 6, No. 19, 2004
3283

On a final note, we entertained the possibility of creating
the desired 3 by an Ugi-like condensation of 4 and 5 with
pyroglutamic acid, 16. Ugi-type reactions in which the
nitrogenous component is an amide or a lactam appear to
be undocumented, hence the interest of exploring such an
opportunity. Whereas direct formation of 3 in this guise
proved ultimately not to be possible, an interesting trans-
formation took place when 16, 4, and 5 were admixed in
MeOH at 40 °C for 48 h: a 6:1 mixture of two diastereomers
of imide 1715 emerged in moderate yield (Scheme 5). The
Scheme 6. Mechanistic Hypothesis
Scheme 5. Attempted Ugi Reaction with Pyroglutamic Acid
anion of 16 to 21 produces 22. At lower temperatures, 22
may be sufficiently long-lived to react with methanol,
resulting in formation of methyl pyroglutamate and 19. On
the other hand, thermal activation promotes cyclization of
22 to 23, which then suffers reversible loss of xylidine to
give 24. Being an activated ester, 24 reacts with the liberated
amine to yield 17. This step is presumed to involve
diastereoselective protonation of the putative intermediate
24 as the event that determines the configuration of 17.
In summary, TiCl4 catalysis extends the scope of the Ugi
reaction of R-amino acids to the aromatic series of aldehydes.
The new process should find widespread application in the
stereoselective synthesis of nitrogenous substances. Inves-
tigations in that sense are ongoing in our laboratory and will
be the subject of future reports.
sense of diastereoinduction is now (S,R). Conducting the
reaction in EtOH under otherwise identical conditions led
to 18, whereas at room temperature the Passerini product
19 was obtained.16 Contrary to the previous case, Lewis acids,
had no effect on the yield and the selectivity of this unusual
reaction,17 a mechanism for which is proposed in Scheme
6.
We presume that oxonium ion 20 results reversibly upon
acid-promoted combination of 4 with, e.g., MeOH. Sequen-
tial addition of the isonitrile to 20 and of the carboxylate
Acknowledgment. We thank the MRT, the CNRS, and
the Re´gion Rhoˆne-Alpes for support of our research. We
are grateful to Ms. Laurence Rousset and Dr. Denis Bouchu
for the mass spectral data, to Dr. Bernard Fenet for the NMR
data, and to Mr. Sylvain Canesi for helpful discussion.
M.A.C. is the recipient of a Merck & Co. Academic
Development Award.
Supporting Information Available: Experimental pro-
cedures and spectral data of key compounds. This material
is available free of charge via the Internet at http://pubs.acs.org.
(15) The major isomer, a crystalline substance, was characterized by
X-ray diffractometry.
(16) The structure of this racemic product was confirmed by X-ray
diffractometry.
(17) Bronsted acids were also ineffective promoters.
OL048850X
3284
Org. Lett., Vol. 6, No. 19, 2004