The first syntheses of the 1-oxo-2-oxa-5-azaspiro[3.4]octane ring system found in oxazolomycin

Article abstract of DOI:10.1021/ol0058792

matrix presented L-Proline was utilized to prepare an optically active 1-oxo-2-oxa-5-azaspiro[3.4]octane for the first time. The synthesis of the racemic system, using a tandem aldol-lactonization reaction, is also described. Ruthenium tetroxide oxidation of these compounds afforded the corresponding spiro β-lactone γ-lactams.

Full text of DOI:10.1021/ol0058792

ORGANIC
LETTERS
2000
Vol. 2, No. 14
1987-1990
The First Syntheses of the
1-Oxo-2-oxa-5-azaspiro[3.4]octane Ring
System Found in Oxazolomycin
Julien P. N. Papillon and Richard J. K. Taylor*
Department of Chemistry, UniVersity of York, Heslington, York YO10 5DD, U.K.
rjkt1@york.ac.uk
Received March 31, 2000
ABSTRACT
L-Proline was utilized to prepare an optically active 1-oxo-2-oxa-5-azaspiro[3.4]octane for the first time. The synthesis of the racemic system,
using a tandem aldol−lactonization reaction, is also described. Ruthenium tetroxide oxidation of these compounds afforded the corresponding
spiro â-lactone γ-lactams.
Oxazolomycin (1) was isolated in 1985 by Uemura and co-
workers from a strain of Streptomyces and was the first
member of a new class of antibiotics containing a spiro
â-lactone γ-lactam ring system to be discovered.1 It was
no total synthesis of these natural products has been
published, in 1990 Kende and co-workers5 disclosed an
enantioselective total synthesis of neooxazolomycin (3), the
γ-lactone congener of 1, whose isolation and structure was
reported by Uemura and co-workers.6 Whiting and He´naff
also recently reported the synthesis of racemic phthoxazolin
A, the oxazole-triene “half” of oxazolomycin.7 In addition,
the related compound lactacystin has received considerable
attention.8
found to exhibit activity against P-388 leukaemia cells,
Ehrlich ascites tumor, and Gram positive bacteria and also
to prevent crown gall formation.2 It was later found to
suppress the replication of vaccinia, herpes simplex type 1,
and influenza A viruses during one-step growth cycle
experiments in both human and chicken cells.3 Five ana-
logues of 1 have since been isolated, including 16-methyl-
oxazolomycin (2), which was described in 1997.4 Although
Herein we disclose our preliminary studies toward the spiro
lactone moiety of oxazolomycin and the synthesis of the
(4) (a) Ryu, G.; Hwang, S.; Kim, S. J. Antibiot. 1997, 50, 1064-1066.
(b) Ryu, G.; Kim, S. J. Antibot. 1999, 52, 193-197.
(5) Kende, A. S.; Kawamura, K.; De Vita, R. J. J. Am. Chem. Soc. 1990,
112, 4070-4072.
(6) Takahashi, K.; Kawabata, M.; Uemara, D. Tetrahedron Lett. 1985,
26, 1077-1078.
(7) He´naff, N.; Whiting, A. Org. Lett. 1999, 1, 1137-1139.
(8) Corey, E. J.; Li, W.-D. Z.; Nagamitsu, T.; Fenteany, G. Tetrahedron
1999, 55, 3305-3316 and references therein.
(1) Mori, T.; Takahashi, K.; Kashiwabara, M.; Uemara D. Tetrahedron
Lett. 1985, 26, 1073-1076.
(2) Kawazu, K.; Kanzaki, H.; Kawabata, G.; Kawai, S.; Kobayashi, A.
Agric. Biol. Chem. 1989, 53, 1127-1133.
(3) Tonew, E.; Tonew, M.; Gra¨fe, U.; Zo¨pel, P. Acta Virol. 1992, 36,
166-172.
10.1021/ol0058792 CCC: $19.00 © 2000 American Chemical Society
Published on Web 06/17/2000

1-oxo-2-oxa-5-azaspiro[3.4]octane system in both racemic
and optically active forms.
Racemic Route. Tandem Aldol-Lactonization. Tandem
aldol-lactonization reactions have been reported on acyclic
enolate precursors,9 and we envisioned producing the spiro-
fused ring systems 11-14 by reaction between a suitable
N-protected activated ester of proline and a carbonyl
compound. The initial study was carried out on N-carboben-
zyloxy and N-tert-butoxycarbonyl L-proline esters 4-10. The
results of this investigation are summarized in Table 1.
Table 1. Tandem Aldol-Lactonization Using
N-Alkoxycarbonyl Proline Estersa
Figure 1. ORTEP drawing of 13 (50% probability thermal
ellipsoids).
An extensive study of the conditions of the reaction using
the PfP esters 4 and 5 was carried out but offered no
significant improvements of yield. Addition of LDA to the
ester and generally poor starting material recovery prompted
us to investigate the use of less reactive esters and different
bases. As shown in Table 1, the phenyl ester 6 in combination
with LDA was clearly the best but still afforded no major
improvement.
substr
base
LDA
LDA
LDAd
NaHe
LDA
LHMDS -78 °C, THF
LICA -90 °C, THF
TMPLi -78 °C, THF
LDA
LDA
LDA
LDA
LDA
conditionsb electrophile product yield (%)
4
4
4
4
5
5
5
5
6
7
8
9
10
-90 °C, THF
-90 °C, THF
-78 °C, THF
-78 °C, DMF
-90 °C, THF
Me2CO
PhCHO
PhCHO
PhCHO
Me2CO
Me2CO
Me2CO
PhCHO
Me2CO
Me2CO
Me2CO
Me2CO
Me2CO
11
12
12
12
13
13
13
14
13
13
13
13
13
9
30 (8:7)c
no prod.
16 (3:1)c
9, 19f
11
We then turned our attention from alkoxycarbonyl to alkyl
N-protecting groups. The results using N-benzyl proline
phenyl ester (15) are shown in Table 2.
9
20g
-78 °C, THF
-78 °C, THF
-78 °C, THF
-78 °C, THF
-78 °C, THF
34
9
traces
decomp.
no prod.
Table 2. Tandem Aldol-Lactonization Using N-Benzyl Phenyl
Proline Estera
a All reactions were carried out on a 1 mmol scale using 1.1 mmol of
base unless otherwise noted. b Temperatures for deprotonation and addition
of the electrophile. c Ratio of isolated diastereomers. d 2.2 equiv of LDA
was used. e 2 equiv of NaH was used. f Yield obtained when reaction was
carried out on a 4 mmol scale. g Only one diastereomer could be detected
by 1H NMR spectroscopy. Abbreviations: PfP ) pentafluorophenyl, Succ
) succinimide.
electrophile
reaction
entry
R1
R2
time (h)
yield (%)
Deprotonation of the N-Cbz pentafluorophenyl ester 4 with
lithium diisopropylamide (LDA) at -90 °C followed by
trapping with acetone gave the spiro â-lactone 11, albeit in
extremely low yield. This system was characterized by the
typical 1824 cm-1 carbonyl stretching frequency of â-lac-
tones and a 13C NMR δCO of 170 ppm. The corresponding
reaction using benzaldehyde gave â-lactone 12 in 30% yield
as a mixture of diastereomers. The N-Boc proline 5 gave
the â-lactone 13 on treatment with LDA and acetone; an
X-ray crystallographic analysis of a single crystal confirmed
the structure of 13 (see Figure 1).10
1
2
3
4
Me
H
H
TMS
Me
Me
TMS
TMS
0.25
62, 16
33, 17b
65, 18
no product
2
2
2
a The reactions were performed at -78 °C in THF using 1.1 equiv of
LDA. b 39% of starting material 15 was recovered.
The enolate of N-benzyl proline phenyl ester 15 reacted
rapidly at -78 °C with acetone to afford the â-lactone 16 in
a gratifying 62% yield. In a attempt to prepare a methyl-
substituted â-lactone (as found in 2), we utilized ethanal as
the electrophile and only isolated 17 (33% yield). The fact
(9) See for example: Wedler, C.; Kunath, A.; Schick, H. J. Org. Chem.
1995, 60, 758-760.
1988
Org. Lett., Vol. 2, No. 14, 2000

that the â-lactone formation was observed with acetone but
not with ethanal suggested the importance of the Thorpe-
Ingold effect.11 In an attempt to exploit this observation,
reaction with trimethylsilylformaldehyde12 was carried out
but gave exclusively the uncyclized product 18 in 65% yield.
Bis(trimethylsilyl)ketone13 was also prepared and utilized in
this process, but no adduct was obtained.14
dialkyl effect. The trapping was successful, but, disappoint-
ingly, all attempts to hydrolyze 22 failed, even under the
harshest conditions (48% HBr, reflux).
Assuming that the presence of a quaternary center was
responsible for this hydrolytic stability, we turned to ben-
zyloxymethyl chloride (BOMCl) as a highly reactive and
easily available alkylating agent. Alkylation of 21 yielded a
1
single diastereomer (as judged by H NMR spectroscopy)
which underwent smooth hydrolysis (3 M HCl) to give the
R-alkylated amino acid 25 in 59% yield from L-proline
(Scheme 3). Boc protection using Johnson’s conditions16 and
Scheme 3a
Compound 16 could be easily converted into 1,1-dimethyl-
3-oxo-2-oxa-5-azaspiro[3.4]octane 19 as a stable white
powder using H2, 10% Pd/C. N-Methylation was successfully
carried out using methyl iodide and potassium carbonate in
dry DMF to give 20 (Scheme 1).
Scheme 1a
a (a) H2, Pd/C 10%, EtOAc; (b) MeI, K2CO3, DMF.
a (a) (i) LDA, THF, -78 °C, (ii) BOMCl, -78 °C to rt; (b) 3 M
HCl, reflux; (c) BOC2O, Me4NOH‚5H2O, CH3CN; (d) H2, 10%
Pd/C, MeOH; (e) DMAD, PPh3, THF, -78 °C.
Enantioselective Route. Sequential Trapping-Lacton-
ization. Having designed the first synthesis of the 1-oxo-2-
oxa-5-azaspiro[3.4]octane system in racemic form, we next
investigated an enantioselective route. The method of choice
utilized the “self-reproduction of chirality” concept intro-
duced by Seebach et al. for the asymmetric alkylation of
proline derivatives (Scheme 2).15 Our early studies focused
quantitative debenzylation (H2, Pd/C) afforded the alcohol
27, ready for lactonization, in 83% yield. Cyclization of the
Boc derivative was attempted using Vederas’ chemistry.17
Hydroxyl group activation using modified Mitsunobu condi-
tions [dimethyl azodicarboxylate (DMAD)/PPh3] afforded the
N-Boc (4S) 1-oxo-2-oxa-5-azaspiro[3.4]octane 28 in a modest
31% yield (54% based on recovered starting material). Use
of a larger excess of PPh3 (1.3 equiv) and DMAD (1.35
equiv) gave total conversion in 10 min at -78 °C and
afforded 28 in 86% yield.
Scheme 2a
(10) Crystallographic data for 13 can be obtained on request from the
Director, Cambridge Crystallographic Data Centre, University Chemical
Laboratory, Lensfield Road, Cambridge CB2 1EW, U.K.
(11) For an earlier account on the gem-substitution effect on â-lactone
formation, see: Adam, W.; Encarnacio´n, L. A. A. Chem Ber. 1982, 115,
2592-2605. See also: Jung, M. E. Synlett 1999, S1, 843-846 and
references therein.
(12) Linderman, R. J.; Suhr, Y. J. Org. Chem. 1988, 53, 1569-1572.
(13) Pan, M.; Benneche, T. Synth. Commun. 1998, 1415-1419.
(14) Reaction of N-benzyl proline methyl ester with acetone was then
carried out to check the importance of the activated ester in this reaction.
It gave the alkylated uncyclised product in 70% yield.
a (a) Pivalaldehyde, TFA, pentane, reflux; (b) (i) LDA, THF,
-78 °C, (ii) acetone, -78 °C to rt.
(15) Seebach, D.; Boes, M.; Naef, R.; Schweiser, W. B. J. Am. Chem.
Soc. 1983, 105, 5390-5398.
(16) Khalil, E. M.; Subasinghe, N. L.; Johnson R. L Tetrahedron Lett.
1996, 37, 3441-3444.
on the use of acetone as trapping agent, which would lead
to an unprotected tertiary alcohol 22. We hoped this alcohol
would later undergo efficient cyclization owing to the gem-
(17) (a) Pansare, S. V.; Huyer, G.; Arnold, L. D.; Vederas, J. C. Org.
Synth. 1991, 70, 1-9. (b) Arnold, L. D.; Kalantar, T. H.; Vederas, J. C. J.
Am. Chem. Soc. 1985, 107, 7105-7109.
Org. Lett., Vol. 2, No. 14, 2000
1989

The enantiomeric excess of 28 was determined by regi-
oselective ring opening of the â-lactone with sodium
methoxide and esterification with (R)- and (S)-2-methoxy-
2-(trifluoromethyl)phenylacetate (MTPA), to give (R,R)-29
and (R,S)-29, respectively (Scheme 4).18 Determination of
Scheme 5a
Scheme 4a
a (a) RuCl3‚xH2O, NaIO4, CH3CN/CCl4/H2O 1:1:1.5, rt.
a (a) MeONa, MeOH; (b) (R)- or (S)-MTPA, DCC, DMAP,
DCM.
â-carbon of the â-lactone, afforded the γ-lactam 31 in a more
modest yield, due to decomposition.
In summary, we have demonstrated that the 1-oxo-2-oxa-
5-azaspiro[3.4]octane ring system can be synthesized using
either tandem aldol-lactonization methodology or â-hydroxy
acid ring closure. Ruthenium tetroxide oxidation afforded,
for the first time, spiro â-lactone γ-lactam systems. Inves-
tigations are underway to utilize this methodology for the
synthesis of oxazolomycin and related compounds.
the diastereomeric excess was accomplished by comparing
selected 1H and 19F NMR data of (R,R)-29 and (R,S)-29. The
method of choice was 1H NMR spectroscopy using the H6-
methylene AB system (δ 4.57 and 4.92 for (R,R)-29, δ 4.60
and 4.87 for (R,S)-29). Both diastereomers could be detected
in both spectra in a ratio g 95:5 (i.e., ee g 90%).
Syntheses of Spiro â-Lactone γ-Lactam. Following the
successful preparation of the N-protected 1-oxo-2-oxa-5-
azaspiro[3.4]octane, attempts were made to functionalize it
and complete the first synthesis of a spiro-fused â-lactone
γ-lactam. Using Sharpless methodology,19 oxidation of the
dimethyl â-lactone 13 occurred smoothly to afford a
quantitative yield of analytically pure 30 (Scheme 5).
Oxidation of 28, which bears no substituent on the
Acknowledgment. We thank Pharmacia and Upjohn for
financial support and Drs. Eric Desarbre and Roger Crossley
for their advice and encouragement. Dr. Steve Archibald is
gratefully acknowledged for assistance with X-ray crystal-
lography.
Supporting Information Available: Experimental pro-
cedures and analytical data for compounds 13, 16, 19, 20,
28, 30, and 31. This material is available free of charge via
the Internet at http://pubs.acs.org.
(18) Dale, J. A.; Dull, D. L.; Mosher, H. S. J. Org. Chem. 1969, 34,
2543-2549.
(19) Carlsen, P. H. G.; Katsuki, T.; Martin, V. S.; Sharpless, K. B. J.
Org. Chem. 1981, 46, 3936-3938.
OL0058792
1990
Org. Lett., Vol. 2, No. 14, 2000