Approach to substituted methylcarbapenems and benzocarbacephems by radical cyclization using Cp2TiCl

Article abstract of DOI:10.1055/s-2007-980336

The reductive radical cyclization of 4-(l-methyl-2-phenyloxiranyl)-β- lactams has been achieved using titanocene monochloride. The reaction was regioselective and diastereoselective to afford carbapenems and benzocarbacephems. A rearrangement of β-hydroxy-β-phenylketones to give benzaldehyde was observed when the nitrile function was used as radical acceptor. Georg Thieme Verlag Stuttgart.

Full text of DOI:10.1055/s-2007-980336

LETTER
1243
Approach to Substituted Methylcarbapenems and Benzocarbacephems by
Radical Cyclization Using Cp TiCl
2
S
L
ubstituted Met
a
hyl Carbap
u
enems
a
nd B
r
enzocarb
a
acephems M. Monleón, Manuel Grande,* Josefa Anaya
Departamento de Química Orgánica, Universidad de Salamanca, 37008 Salamanca, Spain
E-mail: mgrande@usal.es
Received 22 February 2007
H
H
O
H
H
H
Abstract: The reductive radical cyclization of 4-(1-methyl-2-
phenyloxiranyl)-b-lactams has been achieved using titanocene
monochloride. The reaction was regioselective and diastereoselec-
tive to afford carbapenems and benzocarbacephems. A rearrange-
ment of b-hydroxy-b-phenylketones to give benzaldehyde was
observed when the nitrile function was used as radical acceptor.
MeO
O
MeO
O
Ph
N
5
1) Cp2TiCl
) H3O
4
2
3
4
6
6
5
3
O + PhCHO
2
7
1N
1N
2
1
a/1b (2:3)
2 (45%)
O
H
H
5
H
H
N
Ph
MeO
Key words: radical cyclization, titanocene chloride, methylcarbap-
enem, benzocarbacephem antibiotics, lactams
MeO
O
1) Cp2TiCl
7
6
4
3
CN
8
+ PhCHO
N
2
2) H O
3
O
1
O
Titanium(III) chloride has been extensively used as a mild
and useful reagent for various chemical transformations
3b: 3,4-cis
a: 3,4-trans
5: 6,7-cis; 5α-Me (79%)
: 6,7-trans; 5β-Me (29%)
4
6
1
such as reduction of aromatic aldehydes, glycosyl
Scheme 1 Reaction of epoxynitriles 1a/1b, 3b, and 4a with Cp TiCl
halides, vicinal dihalides, and sulfoxides. In particular, (a = 5a,6a-epoxy, b = 5b,6b-epoxy)
2
2
3
4
bis(cyclopentadienyl)titanium(III) chloride is a well-
known titanocene(III) reagent that is useful in promoting process from each one of the pure epoxybenzonitrile-2-
radical carbon–carbon bond formation. Many of the azetidinones 3 and 4 (Scheme 1), which afforded benzal-
5
reported reactions involve either carbonyl compounds,
dehyde and the 5-methylbenzocarbacephems 5 and 6,
6
7
epoxides or alkyl halides as radical precursors. Alkenes, respectively.
6
8
9
alkynes, carbonyl compounds and the cyano group
have been reported as radical acceptors.
Benzaldehyde elimination was also observed when the
green solution of Cp TiCl was added to the epoxides
2
Nowadays, radical cyclization of epoxides using ti- solutions (normal way), but in this case bigger amounts of
tanocene(III) species leading to the synthesis of a number untransformed starting materials were recovered.
of naturally occurring compounds and related products
has attracted much attention.
The structures proposed for the reaction products 2, 5 and
1
13
6
were rigorously established by IR, H NMR and
C
1
0
In previous articles on the Cp TiCl-promoted reductive NMR (including 2D experiments) and MS spectros-
2
ring opening of enantiomerically pure 4-epoxy-2-azetidi- copy.14
nones via single-electron transfer, we described the
The evolution of the epoxides 3 and 4 can be explained as
cyclization of benzyl and tertiary alkyl radicals to conju-
shown in Scheme 2 (a similar scheme can be depicted for
gated esters and aldehydes as a new approach to poly-
the transformation of 1 into the bicyclic lactam 2).
cyclic b-lactams.
Although the reductive opening of the oxiranyl ring with
In order to extend the application of these reactions to po-
titanocene monochloride (Scheme 2) should take place
lar multiple bonds, we have also applied this methodology
with the formation of the best stabilized benzylic radical
1
1
to d- and e-epoxynitrile-2-azetidinones and, surprising-
intermediate I (or the equivalent from 1) the evolution to
the lactams 2, 5, and 6 reflects the reversibility of the ox-
irane ring opening because these cyclic products should
come from the less stable tertiary homobenzylic radical
ly, benzaldehyde elimination was observed during our
study on the cyclization of the epoxynitriles with Cp TiCl
2
1
2
(
Scheme 1).
When a 2:3 diastereomeric mixture of racemic epoxyni- II. This radical should be more reactive and have better
triles 1a/1b in THF was slowly added to a green solution accessibility to the cyano group than the benzylic radical
of Cp TiCl in THF, 4a-methylcarbapenem 2 and benzal- I, driving the equilibrium towards the observed products.
2
1
3
dehyde were obtained as reaction products. Similar be- However, this behavior was not observed when the
havior was observed when we explored the 6-exo radical homobenzylic radical was secondary as we have recently
11
observed, perhaps because it is not enough stabilized to
be in an appreciable amount in equilibrium with the
benzylic radical. In that case, the cyclization products
came only from the addition of the benzylic radical to the
cyano group.
SYNLETT 2007, No. 8, pp 1243–1246
Advanced online publication: 08.05.2007
DOI: 10.1055/s-2007-980336; Art ID: G06607ST
Georg Thieme Verlag Stuttgart · New York
1
6.
0
5.
2
0
0
7
©

1
244
L. M. Monleón et al.
LETTER
The radical II cyclizes to the radical III, followed by cou-
pling to the titanocene(III) chloride present in the reaction
medium [IV, pathway (a)] and then, after hydrolysis,
gaves the b-hydroxyketones 7 which rearrange to the
products 2, 5, or 6 with loss of benzaldehyde. Otherwise,
the radical III, in equilibrium with the alkoxy radical V
TiCp Cl
2
O
Ph
MeO
5
6
3
4
2
1N
CN
O
3, 4
(or 1)
[
pathway (b)], may lose benzaldehyde through a typical b-
TiCp2Cl
Ph
Ph
cleavage generating a new radical VI. This stabilized
radical can further be coupled with titanocene(III)
chloride available in the reaction medium to give the
organometallic species VII. Finally, decomposition with
aqueous KH PO and work-up can provide the b-lactams
O
5
MeO
MeO
TiCp2Cl
5
O
6
3
4
3
4
2
2
1N
CN
1N
C N
O
O
2
4
I
II
2
, 5, and 6.
It is noteworthy that, contrary to the results we obtained in
the reactions of other 4-(1-methyl-2-phenyloxiranyl)-2-
Ph
Ph
O
O
TiCp Cl
2
MeO
O
MeO
O
N
N
b
azetidinones with Cp TiCl (not yet published), we have
TiCp Cl
2
2
(b)
not observed in the above reactions the b-lactam ring
opening by evolution of the homobenzylic radical II as
shown in Scheme 2.
N
N
V
III
The thermodynamic stability can explain the diastereo-
selectivity observed in the cyclization products through
pathways (a) or (b) so we carried out a molecular model-
ing study to determine the difference of energy minima
between each pair of diastereomers. It showed that the
energy differences are 0.6 (for 2), 2.5 (for 5) and 1.2 (for
(a)
TiCp2Cl
Ph
PhCHO
N
O
TiCp2Cl
MeO
O
.
MeO
O
N
TiCp Cl
2
1
5
TiCp Cl
2
N
N
6
) kcal/mol, with the products with trans-arrangement of
VI
TiCp2Cl
IV
H3O
Ph
H4 and H5 (or H5 and H6) being more stable. We pre-
sume that the benzaldehyde elimination should proceed
via the ionic pathway (a) because in pathway (b) the pres-
ence of the titanium(III) chloride in the reaction medium
should give rise to the reduction or to the intramolecular
O
TiCp2Cl
H
MeO
O
N
MeO
O
TiCp2Cl
O
1
N
coupling of benzaldehyde.
N
Further evidence for this hypothesis was obtained from
VII
H3O
7
the reaction of epoxyaldehydes 8 with Cp TiCl (Scheme
2
3
).
PhCHO
OH
H
The reaction of a 1:2 diastereomeric mixture of epoxyal-
MeO
O
MeO
O
O
dehydes 8a/8b with Cp TiCl, under the reported reaction
2
conditions, gave a mixture of the bicyclic b-lactams 9a/
N
N
9
b, from which the pure compound 9b could be isolated
5, 6
(or 2)
1
4
by column chromatography.
Epoxyaldehydes 8a and 8b upon treatment with Cp TiCl Scheme 2 Proposed mechanisms explaining the formation of com-
2
could also progress through radical intermediates related pounds 5, 6, and benzaldehyde
to III–V (Scheme 4) but the presence of the hydroxyben-
zyl group in b-lactams 9a and 9b rules out the alternative
pathway (b).
Finally, in order to investigate the flexibility in the rear-
rangement of hydroxybenzylketones in carbapenem sys-
tems we carried out the reaction of the epoxynitriles 10a,b
H
H
O
MeO
R1 10a/10b (1:2)
8
a/8b (1:2)
5
1
) Cp2TiCl
3
4
6
1) Cp2TiCl
2) H3O
2
R2
1N
2
) H3O
OH
O
with Cp TiCl (Scheme 3).
2
H
H
H H
MeO
8
Ph
a,b: R1 = Ph; R = CHO
10a,b: R = H; R = CN
2
MeO
O
OH
8
In this case, the cyclization of the tertiary radical at C5
generated from a 1:2 mixture of epoxynitriles 10a,b with
4
2
1
2
6
5
3
OH
O
7
1N
N
O
1
Cp TiCl, afforded a mixture of the corresponding carba-
2
penems 11a/11b, from which the pure compound 11b
9a: 3α-OH,4β-Me, 8β-OH (14%)
9b: 3β-OH,4α-Me, 8α-OH (18%)
11a: 4α-Me (34%)
11b: 4β-Me (93%)
could be isolated by column chromatography.14 The
hydroxyl IR absorption bands as well as the presence of a
Scheme 3 Reaction of epoxides 8a/8b and 10a/10b with Cp TiCl
2
typical AB system for an hydroxymethylene group in the (a = 5a,6a-epoxy, b = 5b,6b-epoxy)
Synlett 2007, No. 8, 1243–1246 © Thieme Stuttgart · New York

LETTER
Substituted Methyl Carbapenems and Benzocarbacephems
1245
Ph
Ph
(7) (a) Jana, S.; Guin, C.; Roy, S. C. Tetrahedron Lett. 2005, 46,
1155. (b) Hersant, G.; Ferjani, M. B. S.; Bennett, S. M.
Tetrahedron Lett. 2004, 45, 8123.
(8) (a) Fernández-Mateos, A.; Mateos Burón, L.; Martín de la
Nava, E. M.; Rabanedo Clemente, R.; Rubio González, R.;
Sanz González, F. Synlett 2004, 2553. (b) Fernández-
Mateos, A.; Martín de la Nava, E. M.; Pascual Coca, G.;
Ramos Silvo, A.; Rubio González, R. Org. Lett. 1999, 1,
607.
O
O
O
MeO
O
TiCp Cl
MeO
O
TiCp2Cl
2
OH
TiCp2Cl
N
N
TiCp Cl
2
(a)
III
IV
(
b)
H3O
Ph
Ph
O
OH
OH
MeO
O
MeO
O
TiCp2Cl
(9) Fernández-Mateos, A.; Mateos Burón, L.; Rabanedo
Clemente, R.; Ramos Silvo, A. I.; Rubio González, R.
Synlett 2004, 1011.
O
N
N
9a,b
V
(10) (a) Ruano, G.; Grande, M.; Anaya, J. J. Org. Chem. 2002,
6
7, 8243. (b) Ruano, G.; Martiáñez, J.; Grande, M.; Anaya,
Scheme 4 Proposed mechanism explaining the formation of com-
pounds 9a,b
J. J. Org. Chem. 2003, 68, 2024. (c) Anaya, J.; Fernández-
Mateos, A.; Grande, M.; Martiáñez, J.; Ruano, G.; Rubio-
González, M. R. Tetrahedron 2003, 59, 241.
1H NMR spectra of compounds 11a and 11b, indicates
that the rearrangement of these b-hydroxyketones has not
occurred.
(11) Monleón, L. M.; Grande, M.; Anaya, J. Tetrahedron 2007,
63, 3017.
(
12) The stereochemistry depicted in Schemes 1 and 3 for the
oxirane ring in epoxides 1a,b, 8a,b, and 10a,b were
tentatively proposed by comparison of the respective
The presence of the benzyl group in the hydroxy-
benzylketones 7 (Scheme 2) may probably be the reason
for the elimination of benzaldehyde in the radical cycliza-
tion of epoxynitriles 1, 3, and 4. The C4–C8 bond energies
in compounds 11a and 11b with a hydroxymethyl group
at C4 are larger than those of the b-hydroxy-b-phenyl-
ketones 7.
1
polarities and H NMR data with those of pure 4-(1-methyl-
10b
2
-phenyloxiranyl)-b-lactams Ia and Ib (Figure 1). This
will be described elsewhere.
R
O
Ph
H
H
MeO
O
O
5
6
3
4
O
N
CO2Me
In summary, we have analyzed the reactivity of 4-(1-
O
O
methyl-2-phenyloxiranyl)-b-lactams with Cp TiCl using
2
R
Ia: 5α,6α-epoxy
Ib: 5β,6β-epoxy
cyano and formyl groups as radical aceptors and the
shared aspect for these reactions is the regioselectivity in
the homolytic cleavage C5–O of the oxirane ring. The
benzaldehyde elimination observed in the above exam-
ples can be exploited as a new route to 4-methylcarbapen-
Figure 1
13) Typical Procedure
(
A solution of the specific epoxide (1.0 mmol) in THF (17.0
mL) was added dropwise to a green suspension of Cp TiCl,
16
ems (stable antibiotics to kidney dehydropeptidase) and
-substituted benzocarbacephems (b-lactamase inhibi-
2
5
generated from titanocene dichloride (548 mg, 2.2 mmol)
and activated zinc granules (262 mg, 4.0 mmol), in anhyd
and strictly deoxygenated THF (12.5 mL). The reaction
mixture was stirred at r.t. until a color change from green to
orange was observed, and then the reaction was quenched
with 10% v/v aq KH PO (30.0 mL). The aqueous phase was
1
7
tors). Further studies on the mechanistic and stereo-
chemical implications of this reaction are underway and
will be reported in due course.
2
4
Acknowledgment
extracted with EtOAc and the organic combined extracts
®
were filtered through Celite , dried (anhyd Na SO ) and
2
4
Financial support for this work from the Ministerio de Educación y
Ciencia of Spain (CTQ2005-05026/BQU) and the Junta de Castilla
y León (SA070/03) is gratefully acknowledged. We would also like
to thank the Ministerio de Educación y Ciencia of Spain for a grant
to L.M.M.
concentrated in vacuo. The crude material obtained was
purified by column chromatography on silica gel.
14) All these compounds are racemic mixtures but only one
stereoisomer is depicted for simplicity. The C3-, C4- or C5-
configuration for bi- or tricyclic b-lactams is based on
spectroscopic data and the configuration proposed for the
starting material. This will be described elsewhere.
Selected Data for Cyclization Products
(
References and Notes
(
(
1) Gansäuer, A.; Bauer, D. J. Org. Chem. 1998, 63, 2070.
2) Spencer, R. P.; Cavalloro, C. L.; Schwartz, J. J. Org. Chem.
Carbapenem 2: R
f
= 0.50 (7:3 benzene–EtOAc). IR (neat):
–
1 1
n = 3500, 1774, 1755 cm . H NMR (400 MHz, CDCl
):
3
1999, 64, 3987.
d = 1.12 (3 H, s), 1.53 (3 H, s), 1.21 (3 H, d, J = 7.1 Hz), 2.73
(3) Qian, Y.; Li, G.; Zheng, X.; Huang, Y. Z. Synlett 1991, 489.
(4) Zhang, Y.; Yu, Y.; Bao, W. Synth. Commun. 1995, 25, 1825.
(5) Jana, S.; Roy, S. C. Tetrahedron Lett. 2006, 47, 5949.
(6) (a) Gansäuer, A.; Lauterbach, T.; Narayan, S. Angew. Chem.
Int. Ed. 2003, 42, 3687. (b) Gansäuer, A.; Bluhm, H. Chem.
Rev. 2000, 100, 2771. (c) Gansäuer, A.; Bluhm, H.;
(1 H, dq, J = 7.1, 8.7 Hz), 3.52 (3 H, s), 3.62 (1 H, dd,
1
3
J = 4.0, 8.7 Hz), 4.68 (1 H, d, J = 4.0 Hz). C NMR (100
MHz, CDCl ): d = 12.5, 21.0, 23.4, 41.0, 58.9, 59.1, 64.4,
83.4, 170.9, 218.0. HRMS–FAB: m/z calcd for
3
+
C
H15NO
Na [M + 23]: 220.0944; found: 220.0949.
10
3
Benzocarbacephem 5: R = 0.28 (95:5 benzene–EtOAc). IR
f
–
1 1
Pierobon, M. J. Am. Chem. Soc. 1998, 120, 12849.
(neat): n = 1759, 1685 cm . H NMR (400 MHz, CDCl ):
3
d = 1.26 (3 H, d, J = 6.5 Hz), 3.01 (1 H, dq, J = 6.5, 12.8 Hz),
3.68 (3 H, s), 3.98 (1 H, dd, J = 4.5, 12.8 Hz), 4.79 (1 H, d,
(d) Nugent, W. A.; RajanBabu, T. V. J. Am. Chem. Soc.
1994, 116, 986.
Synlett 2007, No. 8, 1243–1246 © Thieme Stuttgart · New York

1
246
L. M. Monleón et al.
LETTER
J = 4.5 Hz), 7.21 (1 H, dt, J = 1.2, 7.9 Hz), 7.52 (1 H, dt,
J = 1.4, 7.9 Hz), 7.55 (1 H, dd, J = 1.2, 7.9 Hz), 7.95 (1 H,
d = 6.4, 19.1, 29.7, 52.7, 53.4, 58.7, 59.1, 59.2, 78.9, 83.6,
126.8, 128.3, 141.1, 170.6. HRMS–FAB: m/z calcd for
1
3
+
dd, J = 1.4, 7.9 Hz). C NMR (100 MHz, CDCl ): d = 10.5,
C H NO [M + 1]: 306.1700; found: 306.1722.
3
17 24
4
1
40.5, 57.7, 59.5, 84.9, 119.4, 123.2, 124.7, 127.7, 134.9,
Carbapenem 11a (from enriched mixtures): H NMR (400
1
38.0, 163.7, 194.8. HRMS–FAB: m/z calcd for
MHz, CDCl ): d = 0.97 (3 H, s), 1.10 (3 H, s), 1.64 (3 H, s),
3
+
C H NO Na [M + Na]: 254.0788; found: 254.0795.
Benzocarbacephem 6: R = 0.25 (95:5 benzene–EtOAc). IR
2.12 (1 H, br s), 3.51 (3 H, s), 4.07 (1 H, d, J = 4.5 Hz), 4.29
13
13
3
1
3
(2 H, d, J = 1.9 Hz), 4.75 (1 H, d, J = 4.5 Hz). C NMR (100
f
–
1 1
(
neat): n = 1761, 1684 cm . H NMR (400 MHz, CDCl ):
MHz, CDCl ): d = 16.3, 21.8, 23.3, 54.6, 55.7, 59.2, 65.0,
3
3
d = 1.36 (3 H, d, J = 6.6 Hz), 2.76 (1 H, dq, J = 6.6, 12.9 Hz),
3
66.9, 84.7, 171.0, 221.4.
.60 (3 H, s), 3.88 (1 H, dd, J = 2.0, 12.9 Hz), 4.66 (1 H, d,
Carbapenem 11b: R = 0.34 (1:1 benzene–EtOAc). IR (neat):
f
1
–
1
J = 2.0 Hz), 7.19 (1 H, dt, J = 1.0, 7.9 Hz), 7.57 (1 H, dt,
J = 1.5, 7.9 Hz), 7.65 (1 H, dd, J = 1.0, 7.9 Hz), 7.91 (1 H,
n = 3473, 1746 cm . H NMR (400 MHz, CDCl ): d = 1.12
3
(3 H, s), 1.17 (3 H, s), 1.62 (3 H, s), 2.12 (1 H, br s), 3.51 (3
H, s), 3.61 (1 H, d, J = 11.1 Hz), 3.85 (1 H, d, J = 11.1 Hz),
4.12 (1 H, d, J = 4.5 Hz), 4.77 (1 H, d, J = 4.5 Hz). 1 C NMR
1
3
dd, J = 1.5, 7.9 Hz). C NMR (100 MHz, CDCl ): d = 10.1,
3
3
43.9, 58.1, 61.4, 90.3, 118.7, 121.9, 124.5, 127.9, 135.2,
1
38.3, 161.3, 193.3. HRMS–FAB: m/z calcd for
(100 MHz, CDCl ): d = 16.3, 21.7, 23.1, 54.6, 55.7, 59.2,
3
+
C H NO Na [M + Na]: 254.0788; found: 254.0800.
Carbapenem 9a (from enriched mixtures): H NMR (400
64.0, 65.2, 84.7, 171.0, 221.4. HRMS–FAB: m/z calcd for
13
13
3
1
+
C H NO Na [M + 23]: 250.1055; found: 250.1050.
1
1
17
4
MHz, CDCl ): d = 1.02 (3 H, s), 1.23 (3 H, s), 1.47 (3 H, s),
(15) Conformational minima derived from analysis with CS
Chem3D Pro 5.0(1999) software (CambridgeSoft Co.
Cambridge, MA) using MM2 calculations.
(16) Neu, H. C.; Novelli, A.; Chin, N. Antimicrob. Agents
Chemother. 1989, 33, 1009.
3
2
.63 (3 H, s), 2.79 (1 H, s), 3.56 (1 H, d, J = 4.0 Hz), 4.21 (1
1
3
H, d, J = 4.0 Hz), 4.76 (1 H, s), 7.24–7.39 (5 H, m). C NMR
100 MHz, CDCl ): d = 9.1, 19.3, 29.5, 58.5, 58.9, 59.1,
(
3
62.9, 82.4, 83.9, 127.5, 127.9, 141.3, 171.3.
Carbapenem 9b: R = 0.37 (1:1 hexane–EtOAc). IR (neat):
(17) (a) Alcaide, B.; Almendros, P. Curr. Org. Chem. 2002, 6,
245. (b) Gómez-Gallego, M.; Manchedo, M. J.; Sierra, M.
A. Tetrahedron 2000, 56, 5743. (c) Setti, E. L.; Micetich, R.
G. Curr. Med. Chem. 1998, 5, 101.
f
–
1 1
n = 3411, 1731 cm . H NMR (400 MHz, CDCl ): d = 1.06
3
(3 H, s), 1.28 (3 H, s), 1.42 (3 H, s), 2.55 (3 H, s), 3.25 (1 H,
s), 3.71 (1 H, d, J = 4.3 Hz), 4.39 (1 H, d, J = 4.3 Hz), 4.60
1
3
(
1 H, s), 7.24–7.39 (5 H, m). C NMR (100 MHz, CDCl ):
3
Synlett 2007, No. 8, 1243–1246 © Thieme Stuttgart · New York

Copyright of Synlett is the property of Georg Thieme Verlag Stuttgart and its content may not
be copied or emailed to multiple sites or posted to a listserv without the copyright holder's
express written permission. However, users may print, download, or email articles for
individual use.