Radical rearrangement: A strategy for conversion of cephalosporin to 1-carba(dethia)cephalosporin

Article abstract of DOI:10.1016/S0040-4020(00)00417-8

A radical rearrangement approach to semisynthesis of the carbacephem class of β-lactam antibiotics is reported. The crucial bond construction in assembly of the carbacephem framework was accomplished by intramolecular C-C bond formation between an azetidi

Full text of DOI:10.1016/S0040-4020(00)00417-8

TETRAHEDRON
Pergamon
Tetrahedron 56 (2000) 5679±5685
Radical Rearrangement: A Strategy for Conversion of
Cephalosporin to 1-Carba(dethia)cephalosporin
²
Noreen G. Halligan, Raymond F. Brown, Douglas O. Spry and Larry C. Blaszczak
*
Lilly Research Laboratories, Discovery Chemistry Research, Eli Lilly & Company, Indianapolis, IN 46285, USA
Received 24 November 1999; accepted 7 February 2000
AbstractÐA radical rearrangement approach to semisynthesis of the carbacephem class of b-lactam antibiotics is reported. The crucial
bond construction in assembly of the carbacephem framework was accomplished by intramolecular C±C bond formation between an
azetidin-2-one-4-yl and a pendant diene ester. This reactive intermediate was generated by fragmentation of a cephem derived radical
followed by loss of sulfur dioxide. The radical precursor was prepared in 63% yield over four steps and the subsequent rearrangement
reaction provided carbacephem products in a single step at 23% yield. q 2000 Elsevier Science Ltd. All rights reserved.
An investigation of the 1-carba(dethia)cephem (carba-
cephem) class of antibiotics identi®ed loracarbef (Lorabidw),
a compound of suf®cient interest to support development to
commercialization. Loracarbef is a potent antibacterial
agent,1 is well absorbed following oral administration,2
and exhibits an astonishing degree of chemical stability.
The seemingly conservative atom interchange that differ-
entiates cephems and carbacephems, substitution of carbon
for sulfur at position 1 adjacent to the bridgehead,
profoundly in¯uences the chemical properties of the
bicyclic b-lactam nucleus without adversely affecting the
antibacterial characteristics.3
Because loracarbef and other members of the carbacephem
class of compounds differ from their fermentation derived
analogs (e.g. cefaclor) by the carbon substitution, they are
available only through lengthy syntheses.4
One well-established approach to enantiospeci®c synthesis
of new bicyclic b-lactams is through structural modi®cation
of inexpensive fermentation products such as penicillin and
cephalosporin.5 Even though all classes of b-lactam anti-
biotics have been synthesized in this way, the routes from
penicillin to carbacephems are cumbersome.6 Nonetheless,
obvious structural similarity between the bicyclic core
structures of cephems and carbacephems strongly suggests
the possibility of ®nding an atom ef®cient chemical trans-
formation that would connect the two b-lactam ring
systems. As an extension of our program to investigate
new radical based synthetic methodology for installation
of carbon±carbon bonds at C4 of azetidin-2-one deriva-
tives,7 we envisaged a homolytic retrosynthetic dis-
connection in a generalized target structure such as 1
(Scheme 1). Suf®ciently nucleophilic carbon-centered
radicals react readily with electron de®cient ole®ns.8
Furthermore, inspection of molecular models suggests that
cyclization of 2 would be facilitated by very favorable
orbital overlap in the proposed transition state. However,
consideration of synthetic schemes for generation of 2
immediately reveals two signi®cant challenges. First,
generation of a radical such as 2 from a monocyclic
precursor by the usual reductive processes would likely
be precluded by preferential reduction of the diene ester.
Secondly, stereoselective introduction of the diene ester
would require multistep transformations of known
penicillin-derived azetidinone intermediates and thus
compromise the potential ef®ciency of the overall
process.
Keywords: cephalosporins; radicals; rearrangements.
* Corresponding author.
Deceased.
²
0040±4020/00/$ - see front matter q 2000 Elsevier Science Ltd. All rights reserved.
PII: S0040-4020(00)00417-8

5680
N. G. Halligan et al. / Tetrahedron 56 (2000) 5679±5685
Scheme 1.
The propensity of b-sulfenyl, sul®nyl and sulfonyl carbon
radicals toward fragmentation9 offers the possibility of a
very ef®cient approach to 2 wherein a potential method
for stereoselective synthesis of the required diene ester
substituent is embedded in methodology for generation of
2 (Scheme 2). Presumably 2 could be formed on loss of SO2
from sul®nyl radical10 3 which, in turn, could result from
fragmentation of cephem sulfone radical 4.
nucleus or penicillin. The straightforward process for acti-
vation of the cephem is outlined below (Scheme 3). Only a
standard tributyltin hydride reduction of 8 seemed to be
required for generation of 4 and initiation of the proposed
rearrangement.
When 8 was subjected to variations on the standard radical
ring closure conditions12,13 (tributyltin hydride, 2,20-azo-
bisisobutyronitrile [AIBN], benzene, 808C), 9 was observed
as the major product (60±70%) accompanied by the desired
carbacephem compounds 10 (15±25%), numerous minor
products and, curiously, gas evolution. The major features
of the reaction outcome are illustrated in Scheme 4. As
expected, the carbacephem products were formed as a
mixture of all possible ring junction and double bond isomer
combinations. The 6,7-trans ring junction was favored and
the D2 (non-conjugated) double bond isomer predominated
regardless of ring junction stereochemistry. Furthermore,
the C2 epimers of 9 appeared to interconvert upon chroma-
tography. This phenomenon caused substantial dif®culty in
chromatographic detection of the carbacephem reaction
products and resulted in tedious puri®cation.
Thus, a carbon-centered radical exocyclic at C2 in a cepha-
losporin sulfone contains all of the structural and electronic
elements necessary to drive a rearrangement to the corre-
sponding carbacephem radical. Synthesis of a precursor to 4
presents no synthetic dif®culties of consequence. We now
report demonstration of a pathway for the direct conversion
of a cephalosporin to the corresponding carbacephem.
We chose 7-phenoxyacetamido-10-desacetoxycephalosporin
allyl ester 5 as a model substrate for investigation of the
proposed activation±rearrangement sequence. This simple
cephalosporin is easily synthesized in high yield by classical
procedures11 from either the 10-desacetoxy cephalosporin
Scheme 2.
Scheme 3. Conditions: (a) 2.2 equiv. m-CPBA; DMF±EtOAcˆ1:1, 08C; 94%. (b) 37% aq. formaldehyde; Me2NH2Cl; THF; 258C; 92%. (c) PhSeH; CH3CN;
258C; 83% RˆPhOCH2.
Scheme 4.

N. G. Halligan et al. / Tetrahedron 56 (2000) 5679±5685
5681
Scheme 5.
A large matrix of reaction conditions was explored includ-
ing variations on solvent, reaction temperature, concentra-
tion and energy source with no signi®cant variation in the
results. The simplest explanation that could account for the
major observed products and for the large amount of direct
reduction product 9 would be to assume that quenching of
radical 4 by tributyltin hydride is faster than the proposed
fragmentation and rearrangement. Premature reduction is
often a shortcoming of synthetic routes based on radical
cyclizations. However, very slow addition of the tributyltin
hydride and high dilution failed to change the product
distribution signi®cantly. Furthermore, when the reaction
was conducted with or without AIBN in the presence of a
chain-terminating 10% (v/v) nitrobenzene, 9 was still
produced. Finally, the gas evolution phenomenon is
unexplained by simple premature hydrogen atom transfer.
Together these results suggest that other chemistry is
operative in the rearrangement reaction mixture at a rate
or rates competitive with the radical chain process.
Tributyltin hydride reacts with carboxylic acids in the
absence of radical initiators to form tributyltin carboxylates
and hydrogen.14 The C2 methine proton of 8 and seleno-
phenol both represent acids that might exhibit similar
reactivity with tributyltin hydride. As a test of this hypothe-
sis, cephalosporin sulfone 9 was heated with 1 equiv. of
tributyltin hydride in toluene at 958C for 1 h. No gas evolu-
tion was observed and no evidence for enolstannyl ether
1
formation could be found in the H NMR spectrum of an
aliquot concentrate. Upon addition of one equivalent of
selenophenol, gas evolution commenced and (n-Bu)3Sn-
SePh was formed. Thus, a mechanistic rationale emerges
to account for the inescapably large amount of direct reduc-
tion product (Scheme 5). Presumably when 8 is heated in
Scheme 6.

5682
N. G. Halligan et al. / Tetrahedron 56 (2000) 5679±5685
solution, an equilibrium is established between 8 and 7 plus
selenophenol. Tributyltin hydride can then react with the
free selenophenol by chemistry that is not part of the desired
radical chain pathway. This process could account for the
(hydrogen) gas evolution and a large fraction of the
(n-Bu)3SnSePh. Tributyltin hydride can react with 7 as
well to afford 9, the same product that results from
premature hydrogen atom transfer to 4 in the radical chain
manifold. Tributyltin radical itself can react with 7 through
the radical chain process. When the exomethylene cephem 7
was treated independently with tributyltin hydride in
toluene at 1108C, 9 was the exclusive product; in the
presence of AIBN both reduction product 9 and the
Bu3Snz addition product 11 were formed. Indeed 11 was
always isolated in varying amounts from rearrangement
reaction mixtures.
initially confounding product isolation. A simple expedient
rendered the carbacephem products of the reaction more
readily available for isolation and analysis. All cephem
sulfones were eliminated from the crude reaction mixtures
by reductive opening of their dihydrothiazine rings through
the use of electrochemical methods16 followed by extractive
removal of the resulting azetidin-2-one-4-sul®nic acid salts
(Scheme 7).
The carbacephem product is a mixture of ring junction and
double bond isomers that largely favors the trans ring junc-
tion (86:14) and the D2 double bond isomer (88:12). The
trans preference in the ring junction stereochemistry is
easily rationalized by assuming preferential approach of
the diene ester terminus from the less hindered face of the
azetidin-2-one-4-yl 2. Presumably the distribution of the
double bond isomers is re¯ective of a spin density ratio
that favors the capto-dative radical stabilization17 at C4
(10d) over carbonyl conjugation stabilization at C2 (10c).
Radical 4 is most likely involved in a fragmentation/
readdition equilibrium with 3 that supports both direct
reduction and carbacephem formation (Scheme 6). The
latter product can only result from displacement of the equi-
librium by loss of SO2 from 3. Considering the non-radical
chain pathway for degradation of 8, the rate constant15 of ca.
2£106 M21 s21 for depletion of 4 by (n-Bu)3Sn±H, and our
observed formation of carbacephem product at 15±25%, we
conclude that loss of SO2 from 3 is quite an ef®cient process.
We expected that closure of the sulfonyl radical 3 onto the
terminus of the diene ester would be unfavorable based on
steric as well as radical philicity considerations. However,
the reaction mixture produced a very small amount of
material that is isomeric with 9 and exhibits an AA0BB0
pattern centered around 2.9d in place of the diastereotopic
methyl doublets of 9 at 1.5d in the 1H NMR spectrum. We
tentatively assign the structure 3a to this substance. We
were unable to ®nd any product attributable either to
5-exo ring closure or to hydrogen atom transfer to 2 by
(n-Bu)3Sn-H, suggesting that 6-endo ring closure of 2 to
the carbacephem radicals 10c and 10d is very fast.
Finally, we required a transformation of the major carbace-
phem product, the D2 isomers, to the biologically active D3
isomers. Isomerization of double bond mixtures of cephalo-
sporin to the biologically active D3 form is accomplished by
a two step process of oxidation at sulfur (D2 cephalosporin
sulfoxides are unknown) followed by a return to the sul®de
oxidation state under neutral or acidic conditions.18
Obviously this sequence cannot be applied to the car-
bacephem ring system. We found, in contrast to the cephalo-
sporin system, that treatment of the carbacephem isomer
mixture with catalytic 1,8-diazabicyclo[5.4.0]undec-7-ene
(DBU) in dichloromethane results in a simpli®cation of
10a/10b to the pair of ring junction diastereomers at C6
(10b), each with the double bond migrated completely to
the D3 position (Scheme 8).
We have illustrated a straightforward approach to car-
bacephem synthesis that is based on the close structural
relationship between the target molecule and the naturally
occurring molecules that are readily available from fermen-
tation as starting materials. In concept the cephem-to-
carbacephem pathway involves only regioselective addition
The C2 epimers of 9 appeared to interconvert under all
chromatographic conditions examined, effectively obscur-
ing chromatographic observation of reaction progress and
Scheme 7.
Scheme 8.

N. G. Halligan et al. / Tetrahedron 56 (2000) 5679±5685
5683
of a single carbon atom to the cephalosporin nucleus,
removal of the sulfur atom and speci®c reconnection of
the remaining components. There are strategic advantages
of this approach to synthesis of carbacephems over others
that have been recorded. High yield fermentation is the
source for several important aspects of the carbacephem
product: ®fteen of sixteen atoms in the nucleus core
structure, the absolute stereochemistry and the legendary
b-lactam that is responsible for biological activity. Follow-
ing some simple adjustments to the natural product, trans-
formation of cephem framework to the corresponding
carbacephem framework can be accomplished in a single
step. Thus, strategic validation of the radical rearrangement
approach to carbacephem synthesis has been fully realized.
Certain tactical issues, most notably radical generation
methodology and control of ring junction diastereo-
selection, remain to be addressed and are currently under
investigation.
analytical sample was prepared by recrystallization from
isopropyl ether/dichloromethane which afforded 5 as a
1
white crystalline solid: H NMR (DMSO-d6, 300 MHz) d
9.05 (d, Jˆ8.4 Hz, 1H), 7.29 (td, Jˆ1.1, 7.3 Hz, 2H), 6.98±
6.92 (m, 3H), 6.01±5.88 (m, 1H), 5.67 (dd, Jˆ4.4, 8.1 Hz,
1H), 5.38 (dd, Jˆ1.5, 17.2 Hz, 1H), 5.24 (dd, Jˆ1.5,
10.2 Hz, 1H), 5.12 (d, Jˆ4.8 Hz, 1H), 4.71 (dd, Jˆ1.5,
4.0 Hz, 2H), 4.63 (d, Jˆ3.7 Hz, 2H), 3.52 (q, Jˆ17.9 Hz,
2H), 2.05 (s, 3H); 13C NMR (DMSO-d6, 75 MHz) d 168.5,
164.0, 161.6, 157.7; MS (FAB) m/z 389.1 (97%, M1H); IR
(KBr) nmax 3273 (m), 1762 (s), 1719 (s), 1671 (s), 1543 (m),
1496 (m), 1385 (m), 1221 (s) cm21; [a]D20ˆ11205.1 (cˆ1
DMSO); UV±vis (EtOH, nm) 268 (6643).
Allyl (6R,7R) 7-phenoxyacetamido-1,1-dioxo-10-desace-
toxyceph-3-em-4-carboxylate (6). To a solution of 5
(15.0 g, 38.6 mmol) in DMF (160 mL) at ice bath tempera-
ture was added m-chloroperoxybenzoic acid (85% tech-
nical, 17.7 g, ca. 85 mmol) in ethyl acetate (160 mL)
dropwise at such a rate as to maintain a reaction temperature
of 10±158C. After the addition was complete, the cooling
bath was removed and stirring continued at ambient
temperature for 3 h. The reaction mixture was diluted with
ethyl acetate and extracted three times with N HCl/saturated
brineˆ1:1 (v/v), once with N sodium thiosulfate, twice with
saturated aqueous sodium bicarbonate once with saturated
brine. The organic phase was dried (MgSO4) and concen-
trated to a tan solid. Recrystallization from isopropyl ether/
dichloromethane provided 13.1 g (81%) of 6 as an off-white
crystalline solid: 1H NMR (DMSO-d6, 300 MHz) d 8.57 (d,
Jˆ9.5 Hz, 1H), 7.29 (td, Jˆ1.1, 5.5 Hz, 2H), 7.00±6.90 (m,
3H), 6.07 (dd, Jˆ4.8, 9.1 Hz, 1H), 6.00±5.88 (m, 1H), 5.39
(dd, Jˆ1.5, 17.6 Hz, 1H), 5.26 (dd, Jˆ1.5, 10.2 Hz, 1H),
4.74 (dd, Jˆ1.5, 5.5 Hz, 2H), 4.66 (q, Jˆ8.78 Hz, 2H), 4.32
(q, Jˆ18.3 Hz, 2H), 2.01 (s, 3H); 13C NMR (DMSO-d6,
75 MHz) d 168.1, 163.9, 160.7, 157.3; MS (FAB) m/z
421.1 (74%, M1H); IR (KBr) nmax 3394 (m), 1764 (s),
Experimental
Allyl (6R,7R) 7-phenoxyacetamido-10-desacetoxyceph-
a slurry of 7-amino-
3-em-4-carboxylate (5). To
desacetoxycephalosporanic acid (42.8 g, 200 mmol) in a
mixture of dioxane (400 mL) and water (200 mL) was
added dropwise with stirring 2N aqueous sodium hydroxide
(100 mL, 200 mmol). The reaction mixture became a brown
homogeneous solution after ca. 30 min. One dropping
funnel was charged with phenoxyacetyl chloride (29 mL,
210 mmol) in dioxane (130 mL); another was charged
with 2N aqueous sodium hydroxide (110 mL, 220 mmol).
The reaction vessel was cooled in an ice bath and simul-
taneous additions of acid chloride solution and hydroxide
solution were rate adjusted to maintain pHˆ8±9 (meter).
When the acid chloride addition was complete, the pH was
adjusted to 8.5, the cooling bath was removed, and stirring
continued for 1 h at ambient temperature. After removal of
most of the dioxane in vacuo, the aqueous solution was
extracted once with ether and taken directly to the esteri®-
cation reaction.
1726 (s), 1525 (s), 1329 (s), 1227 (s), 1128 (s) cm21
;
[a]2D0ˆ2279.4 (cˆ1 DMSO); UV±vis (EtOH, nm) 261
(8638).
Allyl (6R,7R) 7-phenoxyacetamido-1,1-dioxo-2-methyl-
ene-10-desacetoxyceph-3-em-4-carboxylate (7). To
a
Tetrabutylammonium hydrogen sulfate (71.3 g, 210 mmol)
was dissolved in water (300 mL) and the pH adjusted to 7.5
with 2N aqueous sodium hydroxide (meter). The aqueous
acylation reaction mixture and the tetrabutylammonium sul-
fate solution were combined in a separatory funnel and
extracted with three portions of dichloromethane. The
combined organic extracts were dried (MgSO4) and concen-
trated to a brown oil. The oily salt was dissolved in acetone
(150 mL) and the solution cooled in an ice bath. Allyl
bromide (19.5 mL, 225 mmol, puri®ed by ®ltration through
activity I neutral alumina immediately prior to use) was
added dropwise. Upon completion of the addition, the cool-
ing bath was removed and the reaction mixture stirred at
ambient temperature for 2 h. After removal in vacuo of most
of the acetone, the thick brown oil was diluted with ether/
dichloromethane (3:1ˆv/v) and extracted twice with pH 7
buffer and once with saturated brine. The organic phase was
dried (MgSO4), treated with decolorizing carbon, and
concentrated to 65.2 g (84%) of 5 as a yellow solid. The
crude material, contaminated by several percent of D2
isomer, was taken directly to the oxidation step. The
solution of 6 (43.0 g, 102 mmol) in dioxane (600 mL) was
added an excess of aqueous formaldehyde (37% solution,
150 mL) and dimethylamine hydrochloride (12.6 g,
155 mmol). After stirring at ambient temperature 1.5 h, tlc
analysis (1:1ˆEtOAc±hexanes, double elution) indicated
complete reaction. After approximately half of the dioxane
had been removed in vacuo, (some precipitate deposited)
the remaining mixture was diluted with ethyl acetate and
extracted with N HCl and brine. The aqueous phase was
back extracted with two portions of ethyl acetate. The
combined organic extracts were dried (MgSO4) and concen-
trated to a solid. Crystallization from isopropyl ether/
dichloromethane (two crops) provided a total of 40.3 g
(91%) of 7 as an off-white crystalline solid: 1H NMR
(DMSO-d6, 300 MHz) d 8.77 (dd, Jˆ3.3, 9.5 Hz, 1H),
7.33±7.26 (m, 2H), 7.00±6.89 (m, 3H), 6.58 (d,
Jˆ2.6 Hz, 1H), 6.50 (d, Jˆ2.6 Hz, 1H), 6.19±6.13 (m,
1H), 6.03±5.09 (m, 1H), 5.62 (t, Jˆ4.8 Hz, 1H), 5.41 (dt,
Jˆ1.8, 17.2 Hz, 1H), 5.29 (dt, Jˆ1.5, 10.6 Hz, 1H), 4.80±
4.74 (m, 2H), 4.68 (dd, Jˆ3.7, 9.5 Hz, 2H), 2.16 (s, 3H); 13C

5684
N. G. Halligan et al. / Tetrahedron 56 (2000) 5679±5685
NMR (DMSO-d6, 75 MHz) d 168.2, 163.1, 161.0, 157.4;
MS (FAB) m/z 587.1 (100%, M1matrix), 433.1 (26%,
M1H); IR (KBr) nmax 3398 (m), 1772 (s), 1732 (s), 1715
7.32±7.26 (m, 2H), 7.00±6.90 (m, 3H), 6.16 (dd, Jˆ4.8,
9.5 Hz, 1H), 5.98±5.87 (m, 1H), 5.57 (d, Jˆ4.8 Hz, 1H),
5.39 (dd, Jˆ1.5, 17.2 Hz, 1H), 5.26 (dd, Jˆ1.5, 10.2 Hz,
1H), 4.74 (dd, Jˆ1.5, 5.5 Hz, 2H), 4.65 (s, 2H), 4.12 (q,
Jˆ7.0 Hz, 1H), 2.06 (s, 3H), 1.55 (d, Jˆ7.0 Hz, 3H); 13C
NMR (DMSO-d6, 75 MHz) d 168.2, 168.0, 164.1, 163.5,
161.0, 161.0, 157.4; MS (ESI) m/z 435.1 (100%, M1H);
IR (CHCL3) nmax 3410 (m), 1789 (m), 1722 (m), 1518
(m), 1328 (m), 1230 (m) 759 (m) cm21; UV±vis (EtOH,
nm) 263 (8537).
(s), 1534 (s), 1322 (s), 1312 (s), 1143 (s) cm21
;
[a]2D0ˆ2132.4 (c 1 DMSO); UV±vis (EtOH, nm) 261
(8038), 305 (2216).
Allyl (6R,7R) 7-phenoxyacetamido-1,1-dioxo-2-phenyl-
selenomethyl-10-desacetoxyceph-3-em-4-carboxylate (8).
To a stirred solution of 7 (1.30 g, 3.0 mmol) in dry aceto-
nitrile was added selenophenol (0.35 mL, 3.3 mmol) in one
portion. After ca. 30 min, a colorless crystalline solid
formed and was collected. The ®ltrate was concentrated to
a yellow solid. The yellow solid was triturated with ether,
collected and combined with the crystalline material to give,
after drying in vacuo, a total of 1.47 g (83%) of white solid:
1H NMR (DMSO-d6, 300 MHz) d 8.61 (d, Jˆ9.5 Hz, 1H),
7.65±7.56 (m, 2H), 7.37±7.21 (m, 5H), 7.02±6.88 (m, 3H),
6.24 (dd, Jˆ4.8, 9.5 Hz, 1H), 6.01±5.86 (m, 1H), 5.76 (d,
Jˆ4.8 Hz, 2H), 5.44±5.24 (m, 2H), 4.82±4.60 (m, 4H), 3.78
(dd, Jˆ5.1, 14.3 Hz, 1H), 3.36 (dd, 4.4, 14.6 Hz, 1H), 1.74
(s, 3H); 13C NMR (DMSO-d6, 75 MHz) d 168.1, 164.0,
160.7, 157.4; MS (FAB) m/z 591.1 (70%, M1H); IR
(KBr) nmax 3384 (m), 1767 (s), 1726 (s), 1600 (m), 1528
(m), 1496 (m), 1324 (m), 1231 (m), 1109 (m), 1060 (m)
Electrochemical removal of sulfones and double bond
isomerization. The partially puri®ed rearrangement product
(1.74 g) and tetrabutylammonium tosylate (1.80 g) were
dissolved in a mixture of DMF (50 mL) and methanol
(30 mL). The solution was placed in an electrochemical
cell19 consisting of a Hg pool cathode and a Pt wire
anode. The system was cooled to 0±58C and thoroughly
purged of oxygen by conducting an Ar stream through the
reaction mixture; purging was continued during the course
of the reaction. The reduction was performed at a constant
potential of 21.4 V. When 9 had been consumed as judged
by tlc analysis (60:40ˆEtOAc±hexanes) the reaction
mixture was concentrated in vacuo, redissolved in dichloro-
methane and then extracted with N HCl, saturated aqueous
NaHCO3, and saturated brine. The organic phase was dried
(Na2SO4) and concentrated to an oil which was taken
directly to the next step. The crude reduction product
(498 mg) was dissolved in dichloromethane (20 mL) and
treated at room temperature with DBU (42 mg, 0.3 mmol).
After 1.5 h, the reaction mixture was extracted with N HCl,
dried (Na2SO4), and concentrated to an oil. The crude
product was puri®ed by ¯ash chromatography to afford
377 mg of 6R (trans) carbacephem and 61 mg of 6S (cis)
carbacephem (23% of 10 based on 8).
cm21
.
Radical Rearrangement of 8. A slurry of 8 (3.04 g,
5.14 mmol) and AIBN (0.851 g, 5.19 mmol) in benzene
(190 mL) was heated to re¯ux. Tributyltin hydride
(3.3 mL, 11.9 mmol) was added dropwise over ca. 5 min.
After 1.5 h, tlc analysis (60:40ˆEtOAc±hexanes) indicated
complete consumption of 8. The reaction mixture was
cooled, concentrated in vacuo, and partitioned between
acetonitrile and hexanes. The acetonitrile solution was
concentrated to an orange oil and submitted to ¯ash chro-
matography on silica gel using an elution gradient of
hexanes to ethyl acetate. The early elution product, 11,
and representative fractions containing chromatographi-
cally pure 9 were collected for characterization. The
remainder of the material (1.74 g) was taken directly
to electrochemical reduction for removal of cephalosporin
sulfones.
Allyl (6R,7R) 7-phenoxyacetamido-10-desacetoxy-1-carba-
(dethia)ceph-3-em-4-carboxylate (10). Yellow oil; 1H
NMR (DMSO-d6, 300 MHz) d 8.88 (d, Jˆ8.8 Hz, 1H),
7.30 (td, Jˆ1.8, 8.4 Hz, 2H), 6.96 (td, Jˆ0.73, 7.3 Hz,
3H), 5.99±5.87 (m, 1H), 5.38 (dd, Jˆ1.5, 17.2 Hz, 1H),
5.38 (dd, Jˆ3.7, 5.1 Hz, 1H), 5.22 (dd, Jˆ1.5, 10.2 Hz,
1H), 4.68 (d, Jˆ5.5 Hz, 2H), 4.58 (s, 2H), 3.78 (ddd,
Jˆ5.1, 5.1, 10.2 Hz, 1H), 2.32±2.27 (m, 2H), 1.94 (s,
3H), 1.79±1.63 (m, 2H); 13C NMR (DMSO-d6, 75 MHz)
d 168.2, 165.0, 161.8, 157.6; MS (FAB) m/z 371.1 (93%,
M1H); IR (CHCl3) nmax 3416 (m), 3026 (m), 1764 (s), 1689
(s), 1523 (m), 1496 (m), 1389 (m) cm21; UV±vis (EtOH,
nm) 268 (9191).
Allyl (6R,7R) 7-phenoxyacetamido-1,1-dioxo-2-tributyl-
stannylmethyl-10-desacetoxyceph-3-em-4-carboxylate
(11). Yellow oil; 1H NMR (DMSO-d6, 300 MHz) d 8.56 (d,
Jˆ10.2 Hz, 1H), 7.32±7.26 (m, 2H), 7.00±6.92 (m, 3H),
6.12 (dd, Jˆ4.4, 9.9 Hz, 1H), 6.02±5.89 (m, 1H), 5.39±
5.25 (m, 2H), 5.04 (d, Jˆ4.4 Hz, 1H), 4.68±4.67 (m, 3H),
1.74 (s, 2H), 1.64±1.38 (m, 8H), 1.37±1.21 (m, 9H), 0.95±
0.84 (m, 15H); 13C NMR (DMSO-d6, 75 MHz) d 167.9,
165.9, 163.7, 157.2, 139.4, 132.7, 129.5, 128.5, 121.4,
118.8, 64.2, 59.8, 53.9, 28.3, 26.6, 16.9, 13.5, 10.2, 6.6;
MS (ESI) m/z 725.2 (100%, M1H), 723.3 (63%, M2H);
IR (CHCl3) nmax 2959 (m), 1797 (m), 1746 (m), 1697 (m),
Allyl (6S,7R) 7-phenoxyacetamido-10-desacetoxy-1-carba-
(dethia)ceph-3-em-4-carboxylate (10). Yellow oil; 1H
NMR (CDCl3, 300 MHz) d 8.25 (d, Jˆ7.3 Hz, 1H), 7.41±
7.29 (m, 2H), 7.12±7.00 (td, Jˆ1.1, 7.3 Hz, 1H), 6.97±6.85
(dd, Jˆ1.1, 8.8 Hz, 2H), 6.09±5.88 (m, 1H), 5.41 (dd,
Jˆ1.5, 17.2 Hz, 1H), 5.27 (dd, Jˆ1.1, 10.2 Hz, 1H), 4.82
(dd, Jˆ1.83, 7.3 Hz, 1H), 4.76 (d, Jˆ1.5, 5.9 Hz, 1H), 4.52
(s, 2H), 3.50 (ddd, Jˆ2.2, 3.7, 11.3 Hz, 1H), 2.44±2.47 (m,
2H), 2.00 (s, 3H), 1.82±1.65 (m, 2H); 13C NMR (CDCl3,
75 MHz) d 168.7, 161.2, 160.2, 156.9; MS (FD) m/z 371
(100%, M1H); IR (CHCl3) nmax 3019 (s), 1763 (s), 1688
1521 (m), 1495 (m), 1315 (m), 1241 (m), 1136 (m) cm21
UV±vis (EtOH, nm) 247 (0.2768).
;
Allyl (6R,7R) 7-phenoxyacetamido-1,1-dioxo-2-methyl-
10-desacetoxyceph-3-em-4-carboxylate (9). Waxy off-
white solid (data reported for the major diastereomer); H
1
(s), 1524 ((m), 1495 (s), 1441 (m), 1390 (m), 1233 (s) cm21
UV±vis (EtOH, nm) 268 (7620).
;
NMR (DMSO-d6, 300 MHz) d 8.68 (d, Jˆ9.1 Hz, 1H),

N. G. Halligan et al. / Tetrahedron 56 (2000) 5679±5685
5685
6. (a) Deeter, J. B.; Hall, D. A.; Jordan, C. L.; Justice, R. M.;
Kinnick, M. D.; Morin, J. M. Jr.; Paschal, J. W.; Ternansky, R. J.
Tetrahedron Lett. 1993, 34, 3051. (b) Halligan, N. G.; Blaszczak,
L.C. Unpublished results.
Acknowledgements
The authors wish to thank Dr Jacqueline S. Larew and her
colleagues in the Physical Chemistry Department of
Discovery Chemistry Research at Lilly Research
Laboratories for spectral data; Professor Bruce Branchaud,
Department of Chemistry, University of Oregon, for
numerous stimulating discussions on aspects of radical
chemistry and Ms Elizabeth Dingess-Hammond for
assistance in preparation of the manuscript.
7. Blaszczak, L. C.; Armour, H. K.; Halligan, N. G. Tetrahedron
Lett. 1990, 31, 5693.
8. Giese, B. Angew. Chem., Intl. Ed. Engl. 1983, 22, 753.
9. Ueno, Y.; Aoki, S.; Okawara, M. J. Am. Chem. Soc. 1979, 101,
5414.
È
10. Kohler, J. J.; Speckamp, W. N. Tetrahedron Lett. 1977, 18,
635.
11. Penicillins and CephalosporinsÐChemistry and Biology,
Flynn, E. H., Ed.; Academic Press: New York, 1972; Chapter 16.
12. For an excellent review of the important issues associated with
radical chemistry applied to synthesis, see Jasperse, C. P.; Curran,
D. P.; Fevig, T. L. Chem. Rev. 1991, 91, 1237.
13. Giese, B. Radicals in Organic Synthesis: Formation of
Carbon±Carbon Bonds; Baldwin, J. E., Ed.; Pergamon: New
York, 1986; Chapter 4.
References
1. Counter, F. T.; Ensminger, P. W.; Alborn, W. E.; Preston, D. A.;
Turner, J. A. Abstracts of Papers, 27th Interscience Conference on
Antimicrobial Agents and Chemotherapy, New York, NY;
American Society for Microbiology: Washington, DC, Abstr.
1187.
2. Fassbender, M.; Lode, H.; Schaberg, T.; Borner, K.; Koeppe, P.
Clin. Infect. Dis. 1993, 16, 646.
14. Poller, R. C. Chemistry of Organotin Compounds; Academic
Press: New York, 1970; p 174.
3. Blaszczak, L. C.; Brown, R. F.; Cook, G. K.; Hornback, W. J.;
Hoying, R. C.; Indelicato, J. M.; Jordan, C. L.; Katner, A. S.;
Kinnick, M. D.; McDonald III, J. H.; Morin Jr., J. M.; Munroe,
J. E.; Pasini, C. E. J. Med. Chem. 1990, 33, 1656.
15. Chatgilialogu, C.; Ingold, K. U.; Sciano, J. C. J. Am. Chem.
Soc. 1981, 103, 7739.
16. Hall, D. A. U.S. Patent 4,436,596, 1984.
17. Viehe, H. G.; Janousek, Z.; Merenyi, R. Acc. Chem. Res. 1985,
18, 148.
4. (a) Guthikonda, R. N.; Cama, L. D.; Christensen, B. G. J. Am.
Chem. Soc. 1974, 96, 7584. (b) Evans, D. A.; Sjogren, E. B.
Tetrahedron Lett. 1985, 32, 3787. (c) Guzzo, P. R.; Miller, M. J.
J. Org. Chem. 1994, 59, 4862. (d) Metais, E.; Overman, L. E.;
Rodriguez, M. I.; Stearns, B. A. J. Org. Chem. 1997, 62, 9210.
5. Chemistry and Biology of b-Lactam Antibiotics; Morin, R. B.;
Gorman, M., Eds.; Academic Press: New York, 1982; Vol. 2.
18. Spry, D. O. Tetrahedron Lett. 1980, 21, 1289.
19. Technique of Electroorganic Synthesis, Weinberg, N. L., Ed.;
Wiley: New York, 1974.