Conversion of glucuronic acid glycosides to novel bicyclic β-lactams

Article abstract of DOI:10.1021/ol017026v

(matrix presented) Methodology for the conversion of glucuronic acid glycosides to novel bicyclic β-lactams is reported. Using this strategy, we prepared two novel templates suitable for use in combinatorial chemistry strategies for the construction of a

Full text of DOI:10.1021/ol017026v

ORGANIC
LETTERS
2002
Vol. 4, No. 1
135-138
Conversion of Glucuronic Acid
Glycosides to Novel Bicyclic â-Lactams
Timothy B. Durham and Marvin J. Miller*
Department of Chemistry and Biochemistry, UniVersity of Notre Dame,
Notre Dame, Indiana 46556-5670
marVin.j.miller.2@nd.edu
Received November 9, 2001
ABSTRACT
Methodology for the conversion of glucuronic acid glycosides to novel bicyclic â-lactams is reported. Using this strategy, we prepared two
novel templates suitable for use in combinatorial chemistry strategies for the construction of a number of interesting â-lactam motifs. Key
features of this strategy include a diastereoselective Ferrier reaction of a glucuronic acid glucal, selective â-lactam ring formation using a
cyclic allylic alcohol, and a chemoselective benzylic oxidation.
The use of chiral pool molecules to prepare â-lactams has
become routine.1,2 Within this context, the use of carbohy-
drates as synthons is especially appealing given the density
of functionality in these substrates that can ultimately be
incorporated into the final targets.3 We recognized that the
carbohydrate template could be useful for the construction
of a number of types of functionalized 3-hydroxy, 3,4-
bicyclic, 1,4-bicyclic, and polycyclic â-lactams and oxam-
azins4 (Scheme 1). The successful incorporation of the
inherent oxygen groups and stereochemical information
contained in the carbohydrates into the â-lactam system
presents intriguing synthetic challenges, the most daunting
of which is facile, differential protection of the oxygen
functionalities. We felt that development of suitable chemical
technologies/strategies for overcoming these obstacles would
provide opportunities to expand the frontiers of â-lactam
chemistry as well as generate new 2-azetidinone templates
for biological investigations.
Our synthetic strategy using carbohydrate derivatives to
prepare â-lactams features the N1-C4 bond closure strategy
of â-hydroxy hydroxamates pioneered by our research
group.5 Faced with the daunting task of using such highly
functionalized starting materials, we immediately recognized
the necessity of masking the hydroxyl moieties that would
not be directly involved in the N1-C4 cyclization step to
form the â-lactam ring system. While a number of strategies
are available to allow selective protection of the various
oxygen functionalities of carbohydrate scaffolds, these strate-
gies may require a large number of synthetic steps, especially
if each functional group is to be differentiated. One goal of
this research was to devise as general a strategy as possible
to the targeted bicyclic intermediates, allowing access to the
largest number of possible derivatives from a common core
structure. In consideration of these requirements, we recog-
nized an interesting possibility would be to mask the C2-
C3 hydroxyls as an olefin moiety. This would allow
additional reaction manifolds in later steps to be more readily
(1) For two classic examples, see: (a) Woodward, R. B.; Heusler, K.;
Gosteli, J.; Naegeli, P.; Oppolzer, W.; Ramage, R.; Ranganathan, S.;
Vorbru¨ggen, H. J. Am. Chem. Soc. 1966, 88, 852-853. (b) Salzmann, T.
N.; Ratcliffe, R. W.; Christensen, B. G.; Bouffard, F. A. J. Am. Chem. Soc.
1980, 102, 6161-6163.
(2) Miller, M. J. Acc. Chem. Res. 1986, 19, 49-55.
(3) Storz, T.; Bernet, B.; Vasella, A. HelV. Chim. Acta 1999, 82, 2380-
2412.
(5) (a) Mattingly, P. G.; Kerwin, J. F. J.; Miller, M. J. J. Am. Chem.
Soc. 1979, 101, 3983-3985. (b) Miller, M. J.; Mattingly, P. G.; Morrison,
M. A.; Kerwin, J. K. J. J. Am. Chem. Soc. 1980, 102, 7026-7032.
(4) Boyd, D. B.; Eigenbrot, C.; Indelicato, J. M.; Miller, M. J.; Pasini,
C. E.; Woulfe, S. R. J. Med. Chem. 1987, 30, 528-536.
10.1021/ol017026v CCC: $22.00 © 2002 American Chemical Society
Published on Web 12/13/2001

Scheme 1. Retrosynthetic Analysis of Targets 1 and 2
accessed and the preparation of additional derivatives
and Thiem.9 Unfortunately, none of the reported aglycons
were suitable for our purposes since we hoped to be able to
incorporate an alcohol bearing a group amenable to selective
manipulation later in the synthesis. However, given the harsh
Lewis acid conditions needed to carry out Ferrier reactions7,9
and the need to differentiate other protected oxygen func-
tionalities later in our synthetic strategy, we were left with
few viable options. Two potentially useful candidate nucleo-
philes for the proposed Ferrier reaction were trichloroethanol
(TCE) and benzyl alcohol. Our hope was that the resulting
unavailable from the diol directly. This strategy also raised
an opportunity to further test the scope of our N1-C4 ring
closure protocol with allylic alcohol systems. While previous
attempts to use acyclic allylic alcohols in this fashion have
been successful,6 we recognized that the proposed cyclic
system contained additional steric demands not present in
the previous systems that could make the ring closure
difficult or impossible. In this letter, we report the successful
application of our synthetic strategy to the preparation of
two novel bicyclic â-lactam systems suitable for use in
combinatorial chemistry strategies from a common carbo-
hydrate precursor.
Scheme 2. Synthesis of â-Lactams 9a-b
Toward the goal of preparing a common novel core
structure containing a endocyclic olefin suitable for late stage
functionalizations, our initial focus was the preparation of
bicyclic â-lactam 1 (Scheme 1). We envisioned that â-lactam
1 could be derived through an oxidation of the allylic acetal
moiety of â-lactam 2. Retrosynthetic cleavage of the N1-
C4 bond provided the allylic alcohol O-benzyl hydroxamate
3, which could be accessed from the differentiated diester
4. We envisioned that diester 4 could be derived through a
diastereoselective Ferrier reaction7 of glucuronic acid derived
glucal 5. Glucal 5 can be easily prepared from D-(+)-3,6-
glucuronolactone according to a reported literature proce-
dure.8
In the synthetic direction, using the literature procedure
of Wyss et al., we were able to prepare multigram quantities
of the starting glucuronic acid glucal 5 (Scheme 2).8 Next,
we investigated the possibility of developing a diastereo-
selective Ferrier reaction7 for this compound. Ferrier reactions
of glucal 5 have been previously investigated by Wieczorek
(6) (a) Zercher, C. K.; Miller, M. J. Tetrahedron Lett. 1989, 30, 7009-
7012. (b) Sheppard, A.; Miller, M. J. J. Chem. Soc., Perkin Trans. 1 1990,
2519-2525. (c) Farouz, F.; Miller, M. J. Tetrahedron Lett. 1991, 32, 3305-
3308.
(7) For discussion and lead references, see: (a) Ferrier, R. J. AdV.
Carbohydr. Chem. 1965, 20, 67-137. (b) Ferrier, R. J. AdV. Carbohydr.
Chem. 1969, 24, 199-266.
(8) Wyss, P. C.; Kiss, J.; Arnold, W. HelV. Chim. Acta 1975, 58, 1847-
1864.
136
Org. Lett., Vol. 4, No. 1, 2002

acetal could eventually be deprotected via acid-catalyzed
hydrolysis,10 a Zn-mediated reduction protocol (TCE),11 or
a selective oxidation or reduction (benzyl group). Toward
this end, glucal 5 was separately reacted with TCE and
benzyl alcohol under a variety of conditions9,12 to give
glycosides 7a-b (Scheme 2). These experiments led to
optimized reaction conditions allowing moderate yields of
both glycosides (Table 1). Of note was the isolation of a
methodology. The allylic nature of the activated alcohol
provided the possibility of forming either the desired four-
membered or an undesired six-membered heterocycle. While
our group has shown that open chain allylic alcohols can be
successfully used as â-lactam precursors,6 the cyclic nature
of allylic alcohols 8a-b imposes additional steric demands
on the proposed cyclization reaction. For example, to allow
the formation of the â-lactam ring via an SN process requires
the reactive intermediate to adopt a 5HO conformation
(Scheme 2) in which the ester and alcohol functionalities
are in pseudo-axial positions and the aglycon is in a pseudo-
equitorial position preventing anomeric and allylic stabiliza-
tion.14 One would expect that such a conformation would
be higher in energy than the corresponding OH5 conformation
(Scheme 2). This hypothesis is supported by NMR studies
of 2,3-unsaturated pyranosides.14,15 These studies have found
that R-erythro-2,3-unsaturated glycosides prefer the OH5
conformation displaying J4,5 coupling constants of ∼9 Hz,
while the corresponding â-anomers display J4,5 coupling
Table 1. Conditions for Ferrier Reaction of Glucal 5
entry ROH Lewis acid
temp./time
R/âa
yieldb
1
2
3
4
TCEc
TCE
BnOH InCl3
BF3‚OEt2
BF3‚OEt2
0 °C/4 h
-78 °C/12 h
room temp./30 min
-10 °C/4 days
5:1 53%
5:1 69%
4:1 39%
BnOH BF3‚OEt2
>99:1 53%
1
a Determined by H NMR. b Isolated. c TCE ) Trichloroethanol.
5
constants of ∼2-3 Hz indicating the adoption of a HO
conformation.14,15 The observation of a J4,5 coupling constant
of 9.50 Hz for R-7b suggests the preference for this system
single anomer of the benzyl glycoside using a catalytic
amount of BF3‚OEt2 at -10 °C (entry 4, Table 1). This
procedure proved to be especially appealing since the product
could be isolated directly from the reaction mixture via
recrystallization!
O
to also adopt a H5 conformation.
Given these considerations, our success in this endeavor
seemed far from certain. Therefore, we were delighted to
find that subjection of hydroxamates 8a-b to Mitsunobu
reaction conditions5 provided the desired novel bicyclic
â-lactams 9a-b as stable crystalline solids in good yields.
This result serves to further underscore the robustness of our
N1-C4 ring closure strategy toward the construction of
â-lactams. The structure and absolute stereochemistry of
â-lactam 9b were confirmed by X-ray diffraction.
With the successful preparation of â-lactams 9a-b, we
targeted the selective deprotection of the acetal moiety as
the next synthetic goal. For â-lactam 9a, we were disap-
pointed to find that a number of conditions failed to provide
the desired hemiacetal10,11,16 or lactone.17 However, we
hypothesized that the two benzyl moieties of â-lactam 9b
should have differing reactivities. We were encouraged by
a report in which a benzyl aglycon had been successfully
oxidized to a benzoate ester using in situ generated Collin’s
reagent.18 Model studies using diester 7b showed that these
conditions did indeed provide the targeted benzoyl aglycon
product 10 (Scheme 3). Further, treatment of a model N1-
O-benzyl-â-lactam under the same conditions showed that
no oxidation occurred at the hydroxamate benzyl group.
Encouraged by these results, we subjected â-lactam 9b to
the same conditions hoping to receive the benzoyl aglycon.
To our surprise, under these conditions, â-lactam 9b showed
With the two glycosides in hand, we set out to determine
the set of conditions necessary for the formation of the
requisite â-hydroxy hydroxamates 8a-b. While the trans-
formation of esters 7a-b to the hydroxamates 8a-b would
seem to be straightforward, we discovered that it was actually
quite challenging and required the development of a novel
strategy. For instance, after removal of the acetate group,
attempts to saponify the methyl ester with LiOH followed
by an aqueous EDAC‚HCl coupling with O-benzylhydroxy-
lamine hydrochloride (OBHA‚HCl)5 proved to be disastrous
and provided only decomposition products. This and related
attempts showed that these glycosides appear to be highly
sensitive to the reaction conditions. We were, therefore,
pressed to develop a strategy to access hydroxamates 8a-b
using neutral or near-neutral reaction conditions. To our
delight, we found that deprotection of the methyl ester could
be carried out effectively using KOTMS in THF.13 After
removal of the solvent and dissolution of the salt in dry DMF,
a mixture of EDAC‚HCl and OBHA‚HCl was added result-
ing in the formation of the hydroxamates 8a-b, which in
the case of 8b could be isolated directly from the reaction
mixture by recrystallization.
We next focused our attention on the formation of the
â-lactam ring. As mentioned earlier, this transformation
represented an interesting test of our N1-C4 ring closure
(14) Angerbauer, R.; Schmidt, R. R. Carbohydr. Res. 1981, 89, 193-
201.
(15) Dyong, I.; Schulte, G.; Lam-Chi, Q.; Friege, H. Carbohydr. Res.
1979, 68, 257-273.
(16) (a) Ates, A.; Gautier, A.; Leroy, B.; Plancher, J.-M.; Quesnel, Y.;
Marko´, I. E. Tetrahedron Lett. 1999, 40, 1799-1802. (b) Nair, V.; Nair,
L. G.; Balagopal, L.; Rajan, R. Indian J. Chem., Sect. B 1999, 38B, 1234-
1236.
(17) Grieco, P. A.; Oguri, T.; Yokoyama, Y. Tetrahedron Lett. 1978,
419-420.
(18) Hasegawa, A.; Seki, E.; Hioki, Y.; Kiso, M.; Azuma, I. Carbohydr.
Res. 1984, 131, 61-9.
(9) Wieczorek, E.; Thiem, J. Synth. Lett. 1998, 5, 467-468.
(10) (a) Isidor, J. L.; Carlson, R. M. J. Org. Chem. 1973, 38, 554-556.
(b) Tanemura, K.; Suzuki, T.; Horaguchi, T. J. Chem. Soc., Chem. Commun.
1992, 979-980. (c) Oku, A.; Kinugasa, M.; Kamada, T. Chem. Lett. 1993,
165-168.
(11) (a) Just, G.; Grozinger, K. Synthesis 1976, 457-458. (b) Dobosze-
wski, B.; Zamojski, A. Carbohydr. Res. 1984, 132, 29-38.
(12) Babu, B. S.; Balasubramanian, K. K. Tetrahedron Lett. 2000, 41,
1271-1274.
(13) Laganis, E. D.; Chenard, B. L. Tetrahedron Lett. 1984, 25, 5831-
5834.
Org. Lett., Vol. 4, No. 1, 2002
137

â-lactam 9b requires only one chromatographic separation
and that the sequence can be carried out on a multigram scale.
The future directions of this research involve targeting
specific â-lactam structures accessible from the highly
functionalized intermediates 9a-b and 1 and carrying out
investigations into the chemical reactivity of these unusual
substrates. Also of current interest is the possibility of
unmasking the C3 hydroxyl, which would allow the prepara-
tion of unusual R-amino acids via N-carboxy anhydrides.20
The results of these and other studies will be published as
appropriate.
Scheme 3. Selective Aglycon Oxidation
Acknowledgment. The authors thank Eli Lilly and
Company and the University of Notre Dame for financial
support of this research. T.B.D. is very grateful for support
in the form of both a George Wolf Fellowship and a Searle
Fellowship. The authors gratefully acknowledge Professor
Alicia Beatty for solving the crystal structure of compound
9b. The authors also thank the Lizzadro Nuclear Magnetic
Resonance Research Center for NMR resources and technical
support and the mass spectral facilities of the University of
Notre Dame for analytical services. The authors are very
grateful to Ms. Maureen Metcalf for her assistance in the
preparation of this manuscript.
no reaction. Anticipating that the reduced reativity of the
CrO3 was attributable to the pyridine ligands, we subjected
â-lactam 9b to CrO3 and acetic acid.19 We hypothesized that
under these reaction conditions, not only would the benzylic
oxidation occur but trace amounts of water in the reaction
would set the stage for an acid-catalyzed anomeric exchange
to provide the hemiacetal, which would then be oxidized to
the desired lactone. To our delight, we found that subjection
of 9b to these reaction conditions provided lactone 1 in good
overall yield (Scheme 3).
The synthesis of bicyclic â-lactams 9a-b and 1 represents
the enantioselective synthesis of two novel â-lactam tem-
plates suitable for transformation to a plethora of unique
substrate classes. The robustness of the developed chemistry
is exemplified by the fact that the conversion of glucal 5 to
Supporting Information Available: Experimental pro-
cedures and characterization for all new compounds reported,
including copies of 1H and 13C NMR spectra as well as X-ray
data for compound 9b. This material is available free of
charge via the Internet at http://pubs.acs.org.
OL017026V
(19) Araki, K.; Kawa, M.; Saito, Y.; Hashimoto, H.; Yoshimura, J. Bull.
Chem. Soc. Jpn. 1986, 59, 3137-3143.
(20) Palomo, C.; Oiarbide, M.; Landa, A.; Esnal, A.; Linden, A. J. Org.
Chem. 2001, 66, 4180-4186.
138
Org. Lett., Vol. 4, No. 1, 2002