Biosynthetic Studies of ω-Cycloheptyl Fatty Acids in Alicyclobacillus cycloheptanicus. Formation of Cycloheptanecarboxylic Acid from Phenylacetic Acid

Article abstract of DOI:10.1021/jo962402o

The formation of the structurally novel, mono-substituted cycloheptane ring in ω-cycloheptyl fatty acids in Alicyclobacillus cycloheptanicus (formerly Bacillus cycloheptanicus) has been examined. Feeding experiments with ¹³C- and ²H-labeled intermediates demonstrated that cycloheptanecarboxylic acid (3), probably as its CoA thioester, is the starter unit for ω-cycloheptyl fatty acid biosynthesis. Analysis of the resultant labeling pattern from a feeding experiment with [U-¹³C₆]-glucose suggested a shikimate pathway origin of 3 via aromatic amino acids. [1,2-¹³C₂]Phenylacetic acid (6) was efficiently metabolized into the 3-derived moiety in a manner reminiscent of the seven-membered ring Pseudomonas metabolite thiotropocin. The fates of the aromatic and benzylic hydrogens of 6 were determined; these dictated various boundary conditions for the biosynthetic pathway from 6 to 3. Taken together with the results from feeding experiments with postulated cycloheptenylcarboxylate biosynthetic intermediates, the data lead us to propose a pathway which involves an oxidative ring-expansion of 6 to a hydroxynorcaradiene intermediate followed by a series of double bond reductions and dehydrations to the saturated 3.

Full text of DOI:10.1021/jo962402o

J . Org. Chem. 1997, 62, 2173-2185
2173
Biosyn th etic Stu d ies of ω-Cycloh ep tyl F a tty Acid s in
Alicycloba cillu s cycloh epta n icu s. F or m a tion of
Cycloh ep ta n eca r boxylic Acid fr om P h en yla cetic Acid
Bradley S. Moore,* Kevin Walker, Ingo Tornus, Sandeep Handa,† Karl Poralla,‡ and
Heinz G. Floss
The Department of Chemistry BG-10, University of Washington, Seattle, Washington 98195, and the
Botanisches Institut, Mikrobiologie, Universita¨t Tu¨bingen, Auf der Morgenstelle 1,
D-72076 Tu¨bingen, Germany
Received December 30, 1996X
The formation of the structurally novel, mono-substituted cycloheptane ring in ω-cycloheptyl fatty
acids in Alicyclobacillus cycloheptanicus (formerly Bacillus cycloheptanicus) has been examined.
2
Feeding experiments with 13C- and H-labeled intermediates demonstrated that cycloheptanecar-
boxylic acid (3), probably as its CoA thioester, is the starter unit for ω-cycloheptyl fatty acid
biosynthesis. Analysis of the resultant labeling pattern from a feeding experiment with [U-13C6]-
glucose suggested a shikimate pathway origin of 3 via aromatic amino acids. [1,2-13C2]Phenylacetic
acid (6) was efficiently metabolized into the 3-derived moiety in a manner reminiscent of the seven-
membered ring Pseudomonas metabolite thiotropocin. The fates of the aromatic and benzylic
hydrogens of 6 were determined; these dictated various boundary conditions for the biosynthetic
pathway from 6 to 3. Taken together with the results from feeding experiments with postulated
cycloheptenylcarboxylate biosynthetic intermediates, the data lead us to propose a pathway which
involves an oxidative ring-expansion of 6 to a hydroxynorcaradiene intermediate followed by a series
of double bond reductions and dehydrations to the saturated 3.
Most bacilli predominantly possess branched chain
fatty acids which are biosynthetically derived from
branched chain amino acids.1 Three thermoacidophilic
bacilli alternatively produce ω-cycloalkyl fatty acids as
their main fatty acid components.2 As they form a
distinct phylogenetic group, they have recently been
reclassified into the new genus Alicyclobacillus based on
this common phenotypic trait.3 Comparative 16S rRNA
sequence data for Alicyclobacillus are markedly different
from those of the traditional Bacillus species, including
the thermophiles Bacillus stearothermophilus and Bacil-
lus coagulans. Alicyclobacillus acidocaldarius4 and Ali-
cyclobacills acidoterrestris5 both contain 70-90% ω-cy-
clohexyl fatty acids, whereas Alicyclobacillus cyclo-
heptanicus6 possesses ω-cycloheptyl fatty acids. Consid-
erable evidence7-9 suggests that these compounds serve
a membrane-stabilizing function and are essential to the
growth of these organisms at high temperature and low
pH.
origin of the cyclohexanecarboxylic acid starter unit in
ω-cyclohexyl fatty acid synthesis.10-12 We recently re-
ported the complete sequence of metabolic steps for the
conversion of shikimic acid to cyclohexanecarboxylic acid
in A. acidocaldarius13,14 and in the ansatrienin
A
producer Streptomyces collinus.14,15 The pathway in-
volves a series of dehydrations and double bond reduc-
tions interspersed in such a way that no intermediate is
ever aromatic.
Unlike the occurrence of a cyclohexane ring in ω-cy-
clohexyl fatty acids of A. acidocaldarius, A. acidoterris-
tris, Curtobacterium pusillum,16 and Achyranthes aspera17
and in several Streptomyces antibiotics,18-25 the occur-
rence of a fully saturated, monosubstituted cycloheptane
ring is unique to the ω-cycloheptyl fatty acids of A.
cycloheptanicus. ω-Cycloheptylundecanoic (1a ), -trideca-
noic, and -R-hydroxyundecanoic acids comprise nearly
(10) DeRosa, M.; Gambacorta, A.; BuÅLock, J . D. Phytochemistry
1974, 13, 1793. DeRosa, M.; Gambacorta, A.; Minale, L.; BuÅLock, J .
D. Biochem. J . 1972, 128, 751.
(11) Oshima, M.; Ariga, T. J . Biol. Chem. 1975, 250, 6963.
(12) Furukawa, J .; Tsuyuki, T.; Morisaka, N.; Uemura, N.; Koiso,
Y.; Umezawa, B.; Kawaguchi, A.; Iwasaki, S.; Okuda, S. Chem. Pharm.
Bull. 1986, 34, 5176.
(13) Moore, B. S.; Poralla, K.; Floss, H. G. J . Am. Chem. Soc. 1993,
115, 5267.
(14) Moore, B. S.; Floss, H. G. J . Nat. Prod. 1994, 57, 382.
(15) Moore, B. S.; Cho, H.; Casati, R.; Kennedy, E.; Reynolds, K. A.;
Mocek, U.; Beale, J . M.; Floss, H. G. J . Am. Chem. Soc.1993, 115, 5254.
(16) Suzuki, K.; Saito, K.; Kawaguchi, A.; Okuda, S.; Komagata, K.
J . Gen. Appl. Microbiol. 1981, 27, 261.
Several groups have determined the shikimic acid
* Address correspondence to this author. Tel: (206) 543-9950. Fax:
(206) 543-8318. Email: bmoore@chem.washington.edu.
† Present address: Department of Chemistry, University of Not-
tingham, University Park, Nottingham NG7 2RD, England.
‡ Universita¨t Tu¨bingen.
X Abstract published in Advance ACS Abstracts, March 15, 1997.
(1) Kaneda, T. Microbiol. Rev. 1991, 55, 288.
(2) Deinhard, G.; Poralla, K. Biospektrum 1996, 2, 40.
(3) Wisotzkey, J . D.; J urtshuk, P.; Fox, G. E.; Deinhard, G.; Poralla,
K. Int. J . Syst. Bacteriol. 1992, 42, 263.
(4) Darland, G.; Brock, T. D. J . Gen. Microbiol. 1971, 67, 9.
(5) Deinhard, G.; Blanz, P.; Poralla, K.; Altan, E. Syst. Appl.
Microbiol. 1987, 10, 47.
(17) Misra, T. N.; Singh, R. S.; Pandey, H. S.; Prasad, C.; Singh, B.
P. Phytochemistry 1993, 33, 221.
(6) (a) Deinhard, G.; Saar, J .; Krischke, W.; Poralla, K. Syst. Appl.
Microbiol. 1987, 10, 68. (b) Allgaier, H.; Poralla, K.; J ung, G. Liebigs.
Ann. Chem. 1985, 378. (c) Poralla, K.; Ko¨nig, W. A. FEMS Microbiol.
Lett. 1983, 16, 303.
(7) DeRosa, M.; Gambacorta, A.; Minale, L.; BuÅLock, J . D. J . Chem.
Soc., Chem. Comm. 1971, 1334.
(8) Krischke, W.; Poralla, K. Arch. Microbiol. 1990, 153, 463.
(9) Mantsch, H. H.; Madec, C.; Lewis, R. N. A. H.; McElhaney, R.
N. Biochim. Biophys. Acta. 1989, 980, 42. Kannenberg, E.; Blume, A.;
Poralla, K. FEBS Lett. 1984, 172, 331.
(18) Ansatrienins: Damberg, M.; Russ, P.; Zeeck, A. Tetrahedron
Lett. 1982, 59. Weber, W.; Za¨hner, H.; Damberg, M.; Russ, P.; Zeeck,
A. Zbl. Bakt. Hyg., I. Abt. Orig. 1981, C2, 122.
(19) Mycotrienins: Hosokawa, N.; Naganawa, H.; Hamada, M.;
Takeuchi, T.; Ikeno, S.; Hori, M. J . Antibiot. 1996, 49, 425. Hiramoto,
S.; Sugita, M.; Ando, C.; Sasaki, T.; Furihata, K.; Seto, H.; Otake, N.
J . Antibiot. 1985, 38, 1103. Sugita, M.; Sasaki, T.; Furihata, K.; Seto,
H.; Otake, N. J . Antibiot. 1982, 35, 1467. Sugita, M.; Natori, Y.; Sasaki,
T.; Furihata, K.; Shimazu, A.; Seto, H.; Otake, N. J . Antibiot. 1982,
35, 1460.
S0022-3263(96)02402-4 CCC: $14.00 © 1997 American Chemical Society

2174 J . Org. Chem., Vol. 62, No. 7, 1997
Moore et al.
Sch em e 1. Ea r ly Hyp oth esis for th e Biosyn th esis
of Cycloh ep ta n eca r boxylic Acid via
Sch em e 2. Biosyn th esis of Cycloh exa n eca r boxylic
Acid in A. a cid oca ld a r iu s a n d S. collin u s.13,15
3-Deoxy-D-m a n n o-2-octu loson a te 8-P h osp h a te
(KDO) in A. cycloh epta n icu s
arabinose 5-phosphate,27 and may be metabolized to the
homoshikimate analog 2 (Scheme 1) in a fashion similar
to the conversion of 3-deoxy-D-arabino-2-heptulosonate
7-phosphate (DAHP) into shikimic acid (4) (Scheme 2).28
A series of dehydrations and double bond reductions of
2, akin to those in the conversion of 4 to 5,13,15 would then
yield 3. In fact, although 3 (probably as its CoA thioester)
is the starter unit of ω-cycloheptyl fatty acid synthesis
in A. cycloheptanicus, it is rather derived via a ring
expansion of phenylacetic acid.
80% of the fatty acids of A. cycloheptanicus.6a,26 Three
additional ω-cycloheptyl fatty acids, ω-cycloheptylnonano-
ic, -decanoic and -R-hydroxytridecanoic acids, were re-
cently identified in the minor fatty acid components.26
The biologically novel structural feature of the saturated
cycloheptane ring is interesting from a biosynthetic point
of view and prompted us to carry out a study of its mode
of formation.
Resu lts
Isola tion a n d Ch a r a cter iza tion of ω-Cycloh ep tyl
F a tty Acid s. Biosynthesized ω-cycloheptyl fatty acids
were typically analyzed by a combination of GC-MS (as
methyl esters such as 1b) and NMR (as phenacyl esters
such as 1c). Derivatization of the fatty acid extract to
the UV-absorbing phenacyl ester derivatives allowed for
the isolation of 1c by reverse phase HPLC29 and subse-
quent analysis by NMR.30
On the basis of the biosynthesis of the cyclohexanecar-
boxylic acid (5) starter unit of ω-cyclohexyl fatty acids,
we initially contemplated an analogous pathway to the
hypothetical starter unit cycloheptanecarboxylic acid (3)
(Scheme 1). 3-Deoxy-D-manno-2-octulosonate 8-phos-
phate (KDO), an integral constituent of the lipopolysac-
charide in most Gram-negative bacteria, is derived from
the condensation of phosphoenolpyruvate (PEP) and
The unequivocal 13C-NMR signal assignment of 1c, the
most abundant ω-cycloheptyl fatty acid, was neccessary
in order to conduct labeling experiments. The ring
methylene carbon signals, which each integrate to two
due to the symmetry-induced degeneracy of the ring,
were initially assigned on the basis of the known 13C
assignments of cycloheptanecarboxylic acid.31 The com-
plete assignments were eventually determined through
coupling analyses of 13C-labeled samples of 1c from
various feeding experiments. These assignments will be
discussed throughout the manuscript in conjunction with
the appropriate feeding experiment. The 1H signals were
then assigned by correlation to their respective carbons
via a heteronuclear mutliple quantum coherence (HMQC)
experiment.
(20) Trienomycins: Smith, A. B., III; Wood, J . L.; Wong, W.; Gould,
A. E.; Rizzo, C. J .; Funayama, S.; Oh mura, S. J . Am. Chem. Soc. 1990,
112, 7425. Funayama, S.; Okada, K.; Iwasaki, K.; Komiyama, K.;
Umezawa, I. J . Antibiot. 1985, 38, 1677. Funayama, S.; Okada, K.;
Komiyama, K.; Umezawa, I. J . Antibiot. 1985, 38, 1107. Umezawa, I.;
Funayama, S.; Okada, K.; Iwasaki, K.; Satoh, J .; Masuda, K.; Ko-
miyama, K. J . Antibiot. 1985, 38, 699.
(21) Thiazinotrienomycins: Hosokawa, N.; Naganawa, H.; Iinuma,
N.; Hamada, M.; Takeuchi, T.; Kanbe, T.; Hori, M. J . Antibiot. 1995,
48, 471.
Aceta te F eed in g Exp er im en ts. Compound 1 may
be biosynthesized from a functionalized straight chain
fatty acid analogous to the formation of dictyotene from
(22) Asukamycin: Kakinuma, K.; Ikekama, N.; Nakagawa, A.;
Oh mura, S. J . Am. Chem. Soc. 1979, 101, 3402.
(23) Phospholine: Ozasa, T.; Tanaka, K.; Sasamata, M.; Kaniwa,
H.; Shimizu, M.; Matsumoto, H.; Iwanami, M. J . Antibiot. 1989, 42,
1339. Ozasa, T.; Suzuki, K.; Sasamata, M.; Tanaka, K.; Kobori, M.;
Kadota, S.; Nagai, K.; Saito, T.; Watanabe, S.; Iwanami, M. J . Antibiot.
1989, 42, 1331.
(24) Phoslactomycins: Fushimi, S.; Nishikawa, S.; Shimazu, A.;
Seto, H. J . Antibiot. 1989, 42, 1019. Fushimi, S.; Furihata, K.; Seto,
H. J . Antibiot. 1989, 42, 1026.
(25) 2-Pyranone derivative I-i: Kohama, T.; Nakamura, T.; Ki-
noshita, T.; Kaneko, I.; Shiraishi, A. J . Antibiot. 1993, 46, 1512.
Kohama, T.; Enokita, R.; Okazaki, T.; Miyaoka, H.; Torikata, A.;
Inukai, M.; Kaneko, I.; Kagasaki, T.; Sakaida, Y.; Satoh, A.; Shiraishi,
A. J . Antibiot. 1993, 46, 1503. Maeda, M.; Kodama, T.; Asami, S. J pn.
Kokai 304893 (′89), Dec 8, 1989.
(27) Unger, F. M. Adv. Carbohydr. Chem. Biochem. 1981, 38, 323
and references therein.
(28) Haslam, E. Shikimic Acid: Metabolim and Metabolites; J ohn
Wiley & Sons, Chichester, 1993.
(29) Borch, R. F. Anal. Chem. 1975, 47, 2437.
(30) Aside from this method of derivatization being rapid and
convenient, the presence of the phenacyl moiety permits (1) the direct
quantitation of the molar ratios of the fatty acid derivatives based on
HPLC peak areas29 and (2) the determination of 13C abundances in
the biosynthesized 1c by integration of the 13C-NMR signals of the
fatty acid carbons relative to that of the phenacyl ester methylene
carbon.
(31) The Stadtler Standard Carbon-13 NMR Spectra, Stadtler
Research Laboratories, Division of Bio-Rad Laboratories Inc.: Phili-
delphia, 1980.
(26) Moore, B. S.; Poralla, K.; Floss, H. G. J . Nat. Prod. 1995, 58,
590.

Biosynthetic Studies of ω-Cycloheptyl Fatty Acids
Ta ble 1. 13C Distr ibu tion in 1c Biosyn th esized fr om Sin gly La beled Gen er a l P r ecu r sor s
relative 13C abundancesa
[1-13C]glucose [2-13C]glucose
ndc
0.4
J . Org. Chem., Vol. 62, No. 7, 1997 2175
carbon no.
δc (ppmb)
[1-13C]acetate
[2-13C]acetate
[3-13C]glucose
[6-13C]glucose
1
173.2
33.9
24.9
29.1
29.3
29.5
29.6
29.7
30.0
27.4
38.3
39.3
34.7
26.6
28.6
13.1
0.7
0.9
18.1
2.6
nd
nd
2
2.3
1.2
2.1
1.3
2.1
1.4
2.2
1.2
2.1
1.1
3.4
4.7
4.7
2.3
1.0
2.1
0.9
1.5
1.4
1.8
1.4
1.6
0.7
4.4
0.6
4.6
2.9
3.1
1.0
0.9
0.9
1.0
0.8
0.9
0.9
0.9
1.4
1.0
0.6
1.5
4.1
5.8
1.5
0.6
1.5
0.8
1.6
0.8
1.6
0.7
1.3
0.6
1.9
4.0
3.7
1.1
3
4
28.0
nd
17.8
nd
18.9
nd
26.6
nd
17.4
1.3
5
6
7
8
9
30.2
nd
32.3
nd
28.7
0.7
1.0
10
11
12
0.9
1.2
13/18
14/17
15/16
1.8
1.7
2.1
2.5
2.6
2.1
a
b
Relative to the abundance of the phenacyl ester methylene carbon (65.5 ppm) ) 1.0. Referenced to CDCl3. c nd ) not detected.
Sch em e 3. P r ecu r sor s of
ω-Cycloh ep tylu n d eca n oic Acid (1a )
was next administered to A. cycloheptanicus (Scheme 3).
[8-13C]3 was synthesized from cycloheptyl magnesium
bromide and 13CO2 in analogy to the procedure of Gil-
man.34 Concentrations of [8-13C]3 greater than 1 mg per
100 mL culture inhibited the growth of A. cycloheptani-
cus. This inhibition may be due to 3 acting like a
detergent. 13C-NMR analysis of the resultant 1c indi-
cated 13C-enrichment at 38.3 ppm (C-11) of 50% (Table
2). Due to the high incorporation, the signals of the
carbons adjacent to C-11 were split into doublets roughly
centered around the natural abundance singlet, thus
providing further 13C-chemical shift assignments. From
the coupling constants, the two [2-13C]acetate-derived
1
carbons resonating at 27.4 ppm (AM doublet, J ) 34.9
arachidonic acid.32 Conversely, cycloheptanecarboxylic
acid may be the starter unit for fatty acid synthesis
similar to ω-cyclohexyl fatty acid formation in which
cyclohexanecarboxylic acid initiates chain elongation.13
Feeding experiments with 13C-labeled acetate differenti-
ated between these two scenarios.
Hz) and 29.7 (3J ) 2.1 Hz) must be C-10 and C-8,
respectively. By difference, the remaining unassigned
[2-13C]acetate-derived carbon at 29.5 ppm should be C-6.
In addition, two ring carbon signals at 39.3 ppm (AB
1
doublet, J ) 34.9 Hz) and 26.6 ppm (3J ) 2.6 Hz) were
assigned to C-12 and C-14/17, respectively.
Sodium [1-13C]- and [2-13C]acetate both highly enriched
(20-30%) two different sets of five chain carbons in 1c
(Table 1). Absolutely no label was detected in the ring
carbons, thus eliminating the fatty acid cyclization
hypothesis. The data rather indicate that five acetate
units are added to some non-acetate-derived precursor
which provides C-11 through C-18. The one remaining
chain methylene carbon (38.3 ppm) not labeled by
acetate, by difference, must be C-11.
[7-13C]Cyclohexanecarboxylic acid13 was also fed to A.
cycloheptanicus to probe whether ω-cycloheptyl fatty
acids are enriched and/or if ω-cyclohexyl fatty acids are
formed.35 13C-NMR analysis of the resultant 1c showed
no incorporation into this compound. However, ω-cyclo-
hexyl-R-hydroxyundecanoate and ω-cyclohexylundecanoate
were produced in amounts comparable to the correspond-
ing ω-cycloheptyl fatty acids. Both were enriched 100%
at C-11 (37.6 ppm). Phenacyl ω-cyclohexylundecanoate
isolated from A. acidocaldarius was identical chromato-
graphically (GC, HPLC) and spectroscopically (MS, NMR)
to that formed in this experiment by A. cycloheptanicus.
The biosynthesis of 3 was next probed in feeding
experiments with [1,7-13C2]shikimic acid36 and L-[methyl-
13C]methionine. A shikimic acid-derived precursor con-
ceivably may ring expand with the addition of a carbon
from the one-carbon pool to a cycloheptyl intermediate,
which, in turn, is converted into 3. The arrangement of
13C in positions 1 and 7 of shikimic acid allows for the
distinction of biosynthetic pathways in which shikimic
acid is incorporated intact with all seven carbons from a
pathway in which it is incorporated via the aromatic
The 13C NMR chemical shift assignments of the chain
carbons of 1c were in part determined from 1c biosyn-
thesized from sodium [1,2-13C2]acetate (Scheme 3). Intact
incorporation of acetate was evident from the presence
of satellites in the 1D 13C-NMR spectrum, and the
presence of several C-C connectivities was confirmed by
2D INADEQUATE33 NMR. The high acetate incorpora-
tion allowed assignment of carbons 1-4. A third pair
involving carbons resonating at 30.0 and 27.4 ppm was
also observed. However, the remaining two acetate-
derived pairs of carbons involving signals at 29.3, 29.5,
29.6, and 29.7 were not detected due to near chemical
shift equivalence of these carbons and consequent strong
second order coupling.
Or igin of th e Cycloh ep ta n e Rin g. Based on the
(34) Gilman, H.; Kirby, R. H. Organic Syntheses; Wiley: New York,
1941; Collect. Vol. I, p 361.
results of the [13C]acetate experiments with 1, [8-13C]3
(35) In the ω-cyclohexyl fatty acid producer A. acidocaldarius,
various cycloalkylcarboxylic acids, when exogenously added, were
converted to the corresponding ω-cycloalkyl fatty acids.8
(36) Cho, H.; Heide, L.; Floss, H. G. J . Labelled Compd. Radiopharm.
1992, 31, 589.
(32) Stratmann, K.; Boland, W.; Mu¨ller, D. G. Tetrahedron 1993,
49, 3755. Angew. Chem. Int. Ed. Engl. 1992, 31, 1246.
(33) Bax, A.; Freeman, R.; Kempsell, S. P. J . Am. Chem. Soc. 1980,
102, 4849-4851.

2176 J . Org. Chem., Vol. 62, No. 7, 1997
Moore et al.
Ta ble 2. 13C-NMR An a lysis of 1c Obta in ed fr om F eed in g Exp er im en ts w ith 13C-La beled P r ecu r sor s.
precursor
amount (mg) fed per 1 L culture
% enrichment at C-11 (38.3 ppm) in 1c
[8-13C]cycloheptanecarboxylic acid (3)
10
20
100
20
15
17
24
5.5
5
50a
20b,c
27b,c
3a
[1,2-13C2]phenylacetic acid (6)
[1,2-13C2]phenylacetic acid (6)
[8-13C]cyclohepta-2,4,6-trienecarboxylic acid (14)
[1-2H,813C]cyclohepta-2,4,6-trienecarboxylic acid (14)
[8-13C]cyclohepta-1,4,6-trienecarboxylic acid (15)
[3-2H,813C]cyclohepta-1,4,6-trienecarboxylic acid (15)
[8-13C]cyclohepta-1,3-dienecarboxylic acid (17)
[8-13C]cyclohepta-2-enecarboxylic acid (18)
[8-13C]cyclohepta-2-enecarboxylic acid (18)
[8-13C]cyclohept-1-enecarboxylic acid (19)
4a
8a
7a
8a
26a
48a
12a
10
10d
a
% Specific incorporation ) (A - B)/B, where A ) intergrated NMR signal of enriched carbon and B ) intergrated NMR signal of the
natural abundance carbon. (doublet, J C11-C12 ) 34.6 Hz). c % specific incorporation ) C/B, where C ) summation of the integrated
NMR signals of the enriched carbon doublet. 10 mg fed to 300 mL culture.
b
1
d
F igu r e 1. Partial 125.8 MHz 13C-NMR spectrum in CDCl3 of phenacyl ω-cycloheptylundecanoate (1c) enriched in a feeding
experiment with [U-13C6]glucose. Individual resonances for the cycloheptanecarboxylic acid derived carbons are shown.
amino acids with loss of C-7. Analyses of the 13C-NMR
spectrum of 1c and the GC-MS spectrum of 1b from the
experiment revealed no 13C incorporation.37 Secondly, no
label was detected in the resultant 1c from L-[methyl-
13C]methionine, even when A. cycloheptanicus was grown
in a defined medium (medium B) in which the unlabeled
methionine was entirely replaced with labeled methion-
ine.
Glu cose F eed in g Exp er im en ts. Since none of the
previous experiments had provided labeling of the cyclo-
heptanecarboxylic acid-derived carbons (C-11 to C-18),
[U-13C3]glycerol was fed as a more general precursor to
probe the biosynthesis of the starter unit 3. By the
known metabolism of glycerol via glycolysis and the
Krebs cycle, a general acetate labeling pattern should at
least be observed from C-1 to C-10 in addition to any
labeling in the cycloheptanecarboxylic acid-derived car-
bons C-11 to C-18. However, no incorporation of label
from glycerol was observed, even when the feeding
experiment was conducted in the defined medium in
which the glucose was entirely replaced with labeled
glycerol. Clearly, glycerol is not being assimilated or
metabolized by A. cycloheptanicus.
Alternatively, feeding experiments were performed
with 13C-labeled glucose, another general precursor which,
if metabolized, will also produce metabolites via glycolysis
and the Krebs cycle. In the normal sporulation medium
(medium A), [U-13C6]glucose did not significantly enrich
1. However, [U-13C6]glucose extensively labeled 1 when
A. cycloheptanicus was fermented in the defined medium
from which unlabeled glucose and L-alanine were ex-
cluded. A drawback of the feeding experiments in the
defined medium is that considerably more branched
chain fatty acids, and consequently less ω-cycloheptyl
fatty acids, are produced.
Analysis of the 1D 13C-NMR spectrum (Figure 1)
revealed that each carbon is coupled to at least one of its
neighboring carbons. The spectrum is rather complex
due to overlapping multiplets and to the symmetry-
induced degeneracy of the ring methylene carbons.
Initial assignments of labeling patterns were made from
the analysis of the coupling patterns and spectral simu-
lation with the Bruker PANIC program. The 13C-NMR
signals for C-11 and C-12 appear as a pair of enhanced
and coupled AB doublets (J ) 34.6 Hz) approximately
centered around the natural abundance singlet. For
C-13/18, the AX-type doublet (J ) 34.2 Hz) is less intense
than the enhanced central resonance, indicative of an
additional single enrichment. Small triplets are also
apparent at C-12 and C-13/18 due to the high overall
(37) The nonincorporation of [1,7-13C2]shikimic acid turned out to
be a false negative result, as was later demonstrated in a feeding
experiment with [U-13C6]glucose. We have observed in some organisms
that labeled-shikimic acid failed to be incorporated into shikimate
pathway-derived metabolites, probably because of poor uptake into the
cells. For example, 13C-labeled shikimate was not incorporated into
the cyclohexanecarboxylic acid moiety of asukamycin.38 However, the
expected shikimate type labeling pattern from [U-13C3]glycerol was
observed.38
(38) Thiericke, R.; Zeeck, A.; Nakagawa, A.; Oh mura, S.; Herrold, R.
E.; Wu, S. T. S.; Beale, J . M.; Floss, H. G. J . Am. Chem. Soc. 1990,
112, 3979.

Biosynthetic Studies of ω-Cycloheptyl Fatty Acids
J . Org. Chem., Vol. 62, No. 7, 1997 2177
Sch em e 4. 13C-La belin g a n d 13C-13C Cou p lin g P a tter n s in ω-Cycloh ep tyl F a tty Acid s Biosyn th esized fr om
[U-13C6]Glu cose
incorporation of glucose (approximately 30%). The sig-
nals at C-14/17 and C-15/16 are much more complex due
to overlapping first and second order couplings. First,
both the C-14/17 and C-15/16 signals contain doublets
of 34.8 and 33.4 Hz, respectively. The doublet at C-14/
17 actually represents two independent two-bond cou-
plings with C-15/16 and C-13/18; this became very
apparent when the 13C-NMR spectrum was run at lower
field (75.1 MHz), as two doublets of types AB and AX of
32.9 and 34.6 Hz, respectively, were detected. The
remaining lines of the signals for C-14/17 and C-15/16
are part of an ABX spin system. The originally degener-
ate carbons C-15 and C-16 have become nondegenerate
and thus couple due to a nonsymmetrical alignment of
the 13C-enriched four-carbon unit within the cycloheptane
ring.
symmetrical arrangement of the unit within the cyclo-
heptane ring, it is impossible to distinguish, from cou-
pling alone, between a four-carbon unit (1e) and two two-
carbon units.
Singly 13C-labeled glucose samples were also fed to
probe the orientation of the various glucose-derived
fragments (Table 1). [1-13C]- and [6-13C]glucose both
labeled 1c in roughly the same manner. Through gly-
colysis, carbons at positions 1 and 6 of glucose become
equivalent. In addition to labeling the [2-13C]acetate-
derived carbons, carbons 12, 13/18, and 14/17 were
enriched at 2.8, 1.6, and 1.8%, respectively. [2-13C]-
Glucose enriched the [1-13C]acetate-derived carbons and
carbons 11, 13/18, 14/17, and 15/16 at 3.4, 1.6, 0.7, and
0.7%. And lastly, [3-13C]glucose labeled only carbons 14/
17 and 15/16 at 1.4 and 2.2%, respectively.
The labeling patterns of the cycloheptanecarboxylic
acid moiety are not consistent with the KDO model in
which PEP and arabinose 5-phosphate condense (Scheme
1). In this scenario, intact three- and five-carbon units
would label C-11 through C-13 and C-14 through C-18,
respectively.40 Rather, the labeling pattern is identical
to that recently reported for the shikimate-derived
thiotropocin, which was shown to be formed via pheny-
lacetic acid (6).41 “Oxidative ring expansion [of pheny-
lacetic acid] would involve either of the ortho carbon
atoms of the aromatic ring with an equal probability of
either cleaving the bond between the two-carbon and
four-carbon units [route a] or scission of the two-carbon
unit itself [route b], resulting in the generation of a pair
of carbon atoms without labeled neighbors [Scheme 4].” 41
F eed in g Exp er im en ts w ith Ar om a tic P r ecu r sor s.
To test whether the formation of the cycloheptane ring
of 1 is similar to that of the tropolone ring of thiotropo-
cin,41 20 mg of [1,2-13C2]6 was administered to a 1 L
culture of A. cycloheptanicus. The 13C-NMR spectrum of
the resultant 1c showed two enhanced AB doublets for
C-11 and C-12 (J ) 34.6 Hz, 20% specific incorporation)
(Table 2), indicating intact incorporation (Scheme 4),
which is consistent with the pathway proposed by Cane
and co-workers.41 The phenylacetic acid presumably
arises from L-phenylalanine (7), and the following experi-
ments bear this out.
Routine double quantum (DQ) and triple quantum (TQ)
INADEQUATE experiments were attempted to help
clarify this coupling system. However, low sample
concentrations caused these experiments to be unsuc-
cessful. To overcome the low sensitivity of the INAD-
EQUATE experiment, an inverse-detected version of this
experiment was developed. Details of this experiment
have been published.39
The results from this successful experiment corrobo-
rated the interpretation of the 1D coupling pattern. In
addition, the assignment of each carbon resonance was
unambiguously determined. Five signals appeared at the
DQ freqencies of the 10 chain carbons (C-1 through C-10)
which were acetate-derived. Four additional DQ transi-
tions were detected which correspond to the cyclohep-
tanecarboxylic acid-derived portion; one of these comes
from two degenerate signals on equivalent sides of the
ring (C-15 and C-16). Additionally, two nonadjacent DQ
signals are present, indicative of intact three-carbon
units. The TQ spectrum yielded cross peaks correspond-
ing to those for the nonadjacent DQ signals. The two
overlapping three-carbon units (C-13,14,15 and C-14,15,-
16) are part of an overall four-carbon unit.
In addition to the four-carbon unit from C-13 to C-16
(or likewise C-15 to C-18), there are two two-carbon units
at C-11 to C-12 and C-17 to C-18 (or likewise C-13 to
C-14) to complete the labeling of the ring (1d , Scheme
4). However, there are two signals which are not
addressed in this pattern: the strong single enrichment
at C-13/18 and the AB-type doublet between C-14/17 and
C-15/16. There must therefore be a second labeling
pattern present in C-13 through C-18. As there are only
these two remaining NMR signals, the labeling pattern
must be symmetrical. There may be an intact four-
carbon unit extending from C-14 to C-17, but due to the
(39) Pratum, T.; Moore, B. S. J . Mag. Res. Series B, 1993, 102, 91.
(40) Depending upon the mode and degree of glucose metabolism,
the arabinose 5-phosphate moiety may not appear solely as an intact
five-carbon unit. In addition, the labeling pattern may appear as intact
two- and three-carbon units corresponding to C-1 f C-2 and C-3 f
C-5 of the pentose, respectively. Catabolism of glucose to glyceralde-
hyde 3-phosphate and dihydroxyacetone phosphate during glycolysis
would account for this.
(41) Cane, D. E.; Wu, Z.; Van Epp, J . E. J . Am. Chem. Soc. 1992,
114, 8479.

2178 J . Org. Chem., Vol. 62, No. 7, 1997
Moore et al.
Ta ble 3. Ma ss Sp ectr a l An a lysis of 1b Obta in ed fr om F eed in g Exp er im en ts w ith 2H-La beled P oten tia l P r ecu r sor s
m/z (relative intensity)
precursor
296
297
298
299
300
301
302
303
unlabeled control
100
100
100
100
100
100
100
21.4
35.5
34.0
42.3
20.8
25.0
21.6
2.7
9.4
10.0
7.5
2.7
3.8
0.3
17.4
6.4
1.1
0.3
[2H7]phenylacetic acid (6)
51.6
52.0
0.1
74.2
64.0
45.2
20.8
13.9
3.9
L-[ring-2H5]phenylalanine (7)
D,L-[3,3-2H2]phenylalanine (7)
(()-[2-2H1]mandelic acid (12)
[1-2H1,8-13C]cyclohepta-2,4,6-trienecarboxylic acid (14)
[3-2H1,8-13C]cyclohepta-1,4,6-trienecarboxylic acid (15)
0.4
2.2
8.4
a
Sch em e 5. Syn th esis of [3′,5′-2H2]- a n d
[2′,6′-2H2]P h en yla cetic Acid s
a
(a) NaCN, H2O/EtOH; (b) KOH/H2O, ∆; (c) (i) Zn, NaOD/D2O,
∆, (ii) HCl; (d) (i) KOH/H2O, ∆, (ii) HCl; (e) Raney Ni, NaOD/D2O.
during the metabolism and if there is no deuterium
migration, this ratio should theoretically be 1:2:2. From
the MS and NMR data, it thus appears that one-fourth
of the deuterium from C-14/17 is lost. Therefore at least
two tetradeuterio species (1f and 1g) are required to
account for the equal deuterium loss from either C-14/
17 methylene hydrogen. In addition, two pentadeuterio
species (1h and 1i) are likewise needed to account for
the equal distribution of deuterium at H-13/18ax and
H-13/18eq. Consequently, the four deuterated products
are present in a 1:1:1:1 ratio. C-12 is epimeric in each
pair of tetra- and pentadeuterated 1. Thus, a biosyn-
thetic pathway proceeding through a C2 symmetrical
intermediate would account for the equal distribution of
deuterium between the axial and equatorial positions at
each ring methylene carbon.
Phenylacetic acids with deuterium exclusively at the
ortho-, meta-, or para-positions were synthesized and fed
in order to corroborate the above interpretation and
determine which aromatic hydrogen is partially lost
during the biosynthetic transformation from 6 to 1. 4′-
Bromophenylacetic acid was dehalogenated with zinc
powder in 30% NaOD/D2O to give [4′-2H]phenylacetic acid
as previously described.42 This approach was also uti-
lized for the synthesis of [3′,5′-2H2]phenylacetic acid from
3′,5′-dibromophenylacetic acid (9), which in turn was
prepared from 3′,5′-dibromobenzyl bromide via the cya-
nide 8 (Scheme 5). Dehalogenation of 2′,6′-dichlorophe-
nylacetic acid to [2′,6′-2H2]phenylacetic acid was alter-
nately achieved with Raney Ni-Al alloy in 30% NaOD/
D2O in a fashion similar to the dehalogenation of 2,6-
dichlorophenol.43 In each synthesis, deuterium (99%)
was also detected at the benzylic position of the resultant
phenylacetic acid; this deuterium was completely ex-
changed with hydrogen by refluxing in 30% KOH/H2O.
F igu r e 2. 2H-NMR spectra of 1c biosynthesized from L-[ring-
2H5]7, [2′,6′-2H2]6, [3′,5′-2H2]6, and [4′-2H]6.
The fate of the aromatic hydrogens of phenylacetic acid
was followed in a feeding experiment with L-[ring-2H5]7.
Phenylalanine should be converted to phenylacetic acid
by successive transamination and R-oxidation. Ten
percent of the aromatic deuterium atoms were lost in the
conversion of L-[ring-2H5]7 to deuterated 1. Equal
amounts of tetra- and pentadeuterio species (21.7% and
22.7%, respectively) (Table 3) were measured for the
resultant 1b by mass spectrometry. In addition, minor
amounts of mono-, di-, and trideuterio species were
detected (5%, 2%, and 2%, respectively), which may
reflect reduction(s) of an unlabeled cycloheptanecarboxy-
lic acid pathway intermediate by NAD(P)H that has
acquired deuterium from the catabolism of L-[ring-2H5]-
7. This rationale also explains the occurrence of a
hexadeuterio species (3.5%), reflecting the reduction of
an L-[ring-2H5]7-derived intermediate by labeled NAD-
(P)H.
The 2H-NMR spectrum (Figure 2) of 1c from this
feeding experiment displayed six signals (Table 4). Each
unique ring methylene hydrogen (six total) contained
deuterium, and furthermore, each methylene hydrogen
equatorial/axial pair (three total) was equally enriched.
The ratio of enrichment at H-13/18, H-14/17, and H-15/
16 was approximately 1:1.5:2. If no deuterium is lost
(42) Langhals, H.; Fischer, H. Chem. Ber. 1978, 111, 543.
(43) Tashiro, M.; Fukata, G. J . Org. Chem. 1977, 42, 835.

Biosynthetic Studies of ω-Cycloheptyl Fatty Acids
J . Org. Chem., Vol. 62, No. 7, 1997 2179
Ta ble 4. 2H-NMR An a lysis of 1c Obta in ed fr om F eed in g Exp er im en ts w ith 2H-La beled P oten tia l P r ecu r sor s
product (phenacyl ω-cycloheptylundecanoate)
δ 2H-NMR signal(s)a
precursor
(relative abundance)
labeled hydrogen(s)
[2′,6′-2H2]phenylacetic acid (6)
[3′,5′-2H2]phenylacetic acid (6)
[4′-2H1]phenylacetic acid (6)
[2,2,4′-2H3]phenylacetic acid (6)
[2H7]phenylacetic acid (6)
1.64 (1), 1.58 (1), 1.36 (1), 1.11 (1)
H-13/18eq, H-14/17eq, H-14/17ax, H-13/18ax
1.58 (1), 1.52 (1.6), 1.42 (1.7), 1.36 (1) H-14/17eq, H-15/16eq, H-15/16ax, H-14/17ax
1.52 (1), 1.42 (1) H-15/16eq, H-15/16ax
1.58 (1), 1.52 (1.8), 1.42 (1.8), 1.36 (1) H-14/17eq, H-15/16eq, H-15/16ax, H-14/17ax
1.64 (1), 1.58 (1.5), 1.52 (2), 1.42
(2.1), 1.35 (1.4), 1.11 (1)
1.64 (1), 1.58 (1.7), 1.52 (2.4), 1.42
(2.5), 1.36 (1.8), 1.11 (1)
1.60 (1), 1.38 (1)
H-13/18eq, H-14/17eq, H-15/16eq,H-15/16ax
H-14/17ax, H-13/18ax
,
L-[ring-2H5]phenylalanine (7)
H-13/18eq, H-14/17eq, H-15/16eq, H-15/16ax
H-14/17ax, H-13/18ax
H-14/17eq, H-14/17ax
,
D,L-[3,3-2H2]phenylalanine (7)
L-[2,3S-2H2]phenylalanine (7)
L-[3R-2H1]phenylalanine (7)
(()-[2-2H1]mandelic acid (12)
1.60 (1), 1.39 (1.2)
no incorporation
no incorporation
H-14/17eq, H-14/17ax
[1-2H1,8-13C]cyclohepta-2,4,6-trienecarboxylic acid (14) 1.38
H-14/17ax
H-14/17eq, H-14/17ax
[3-2H1,8-13C]cyclohepta-1,4,6-trienecarboxylic acid (15) 1.59 (1), 1.37 (1)
a
CHCl3 used as reference.
a
Sch em e 6. Syn th esis of L-[3R-2H1]- a n d
[2′,6′-2H2]Phenylacetic acid was fed to A. cycloheptani-
cus, and the resultant 1 was doubly enriched. The
corresponding 2H-NMR spectrum (Figure 2) indicated
that both ortho hydrogens are fully retained, as it showed
four peaks with identical intensities at H-13/18 and H-14/
17 (Table 4). In contrast, 1 biosynthesized from [3′,5′-
2H2]6 was both singly and doubly enriched. The ratio of
deuterium at the equatorial and axial positions of C-14/
17 to that at C-15/16 was 1:1.7 (Figure 2). Therefore,
approximately 50% of one of the meta-hydrogens of 6,
the one that is ultimately positioned at C-14/17 in 1, is
lost during the biosynthetic conversion from 6 to 1. And
thirdly, deuterium from [4′-2H1]6 was incorporated equally
at both H-15/16 positions of 1 as expected. The NMR
data corroborate the above interpretation that the bio-
synthetic pathway must proceed through a symmetrical
intermediate.
To determine whether the benzylic hydrogens are
retained, a feeding experiment with D,L-[3-2H2]phenyl-
alanine was performed. GC-MS analysis of the result-
ant 1b revealed the presence of 17.3% monodeuterated
species (Table 3). No dideuterio species was detected.
The location of the deuterium atom was determined by
2H-NMR of the corresponding 1c. Two deuterium signals
of equal intensity were observed at 1.36 ppm (H-14/17ax)
and 1.58 ppm (H-14/17eq) (Table 4).
L-[2,3S-2H2]P h en yla la n in es
a
(a) Rh((R)-prophos)(NBD)ClO4, D2 (H2), THF; (b) 6 N HCl, ∆.
Sch em e 7. Con ver sion of L-[3R-2H1]- a n d
L-[2,3S-2H2]7 by A. cycloh epta n icu s in to 1
In order to ascertain the individual fate of the pro-3S
and pro-3R hydrogens of 7, L-[3R-2H]7 and L-[2,3S-2H2]7
were prepared by adaptations of procedures in the
literature (Scheme 6).44 2-Benzamido[3-2H]cinnamic acid44
([3-2H]10) was hydrogenated in the presence of the chiral
catalyst [Rh((R)-prophos)(NBD)]ClO445 to produce (2S,3R)-
N-benzoyl[3R-2H]phenylalanine ([3R-2H]11) in 90% ee
(Scheme 6). Alternatively, unlabeled 10 was hydroge-
nated in D2 with the rhodium catalyst to (2S,3S)-N-
benzoyl[2,3-2H2]phenylalanine ([2,3S-2H2]11). Deprotec-
tion yielded L-[3R-2H]7 and L-[2,3S-2H2]7, respectively.
The resultant 1c from the feeding experiment with
L-[2,3S-2H2]7 (Scheme 7) was singly enriched and con-
tained an equal amount of deuterium at δ 1.39 and 1.60
in a regiospecific manner identical to that observed with
1c from D,L-[3-2H2]7. In contrast, 1c derived from L-[3R-
2H]7 was not enriched with deuterium. Thus, the pro-
3S hydrogen of phenylalanine is stereoselectively re-
tained during the conversion to 1 at C-14/17.
The incorporation of the deuterium from the benzylic
position relative to that of the aromatic ring was assessed
by feeding [2,2,4′-2H3]phenylacetic acid. In addition to
the two 2H-NMR signals present in the resultant 1c at δ
1.42 and 1.52 in a 1:1 ratio, a second pair of equivalent
singlets resided at δ 1.36 and 1.58, but in a 1:1.8 ratio
relative to the first set of signals (Table 4). The relative
incorporation of the deuterium from the benzylic position
to that from the aromatic ring was 55%, indicating that
approximately half of the pro-3S hydrogen of 7 is lost
during the conversion.
Mandelic acid (12) was next tested as a possible
intermediate as the ring expansion may proceed via an
R-hydroxylation of phenylacetic acid. Introduction of
deuterium at C-2 of methyl mandelate was achieved by
the reduction of methyl benzoylformate with boron deu-
teride. Treatment with NaBD4, unfortunately, resulted
in the reduction of both carbonyl groups. However,
(44) Wightman, R. H.; Staunton, J .; Battersby, A. R.; Hanson, K.
R. J . Chem. Soc., Perkin Trans. 1 1972, 2355.
(45) Fryzuk, M. D.; Bosnich, B. J . Am. Chem. Soc. 1978, 100, 5491.

2180 J . Org. Chem., Vol. 62, No. 7, 1997
Moore et al.
Sch em e 8. Hyp oth etica l F or m a tion of
Cycloh ep ta n eca r boxylic Acid (3) via P h en yla cetic
Acid (6)
Sch em e 9.a Con ver sion of [1-2H,8-13C]14 a n d
[3-2H1,8-13C]15 by A. cycloh epta n icu s in to 1
a
The 13C-label is denoted with a dot and is incorporated into
C-11 of 1 with the respective percentages. The percentage deute-
rium retention is relative to that of 13C. The implied absolute
stereochemistry of the cycloheptane moiety is not known, but is
shown to illustrate the relative stereochemistry.
reduction with Zn(BD4)2 (formed from NaBD4 and ZnCl2)46
in Et2O cleanly gave methyl [2-2H]mandelate in 80%
yield. Base hydrolysis of the methyl ester completed the
synthesis.
[8-13C]14 (20 mg) was fed to a 1 L culture of A.
cycloheptanicus. The resultant 1c was labeled with 13C
at C-11 (3%) (Table 2), indicating that 14 was completely
reduced to the saturated cycloheptane. However, the
percent incorporation is low, especially relative to that
of 6 (20%). [8-13C]14 is not labile in the warm, acidic
culture medium, as first thought, as it was detected
unaltered in the spent culture medium after the 24 h
fermentation. No other triene isomer was detected. A
competition experiment with [1,2-13C2]6 was conducted
to probe whether the incorporation of 6 would diminish
that of 14. Ten milligrams of [8-13C]14 and 5 mg of [1,2-
13C2]6 were fed to a 500 mL culture of A. cycloheptanicus,
and the resultant 1c was singly enriched at C-11 (4.5%)
and doubly enriched at C-11 and C-12 (10%).
The resultant 1 from the feeding of [2-2H]12 was
analyzed by 2H-NMR and MS. The 2H-NMR spectra
failed to show any deuterium signals at H-14/17 nor at
any other positions. In addition, the MS did not reveal
any enhancement of the (M + 1)+ peak. It thus appears
that mandelic acid is not a precursor and that the
rearrangement may proceed directly from phenylacetic
acid.
Cycloh ep t a t r ien eca r b oxylic Acid F eed in g E x-
p er im en ts. Ring expansion of phenylacetic acid in A.
cycloheptanicus may proceed via a coenzyme-B12-medi-
ated or similar rearrangement to the norcaradiene 13
(Buchner’s acid47 ), which is in equilibrium with cyclo-
hepta-2,4,6-trienecarboxylic acid48 (14). A suprafacial
1,3-allylic isomerization of the ∆2-double bond of 14
catalyzed by a single enzyme base49 to give cyclohepta-
1,4,6-trienecarboxylic acid (15) would then rationalize the
retention of the pro-3S hydrogen of phenylalanine.
Subsequent double bond reductions and isomerizations
through cyclohepta-1,3-trienecarboxylic acid (17) would
yield the saturated 3. To test whether such a pathway
(Scheme 8) is operative, labeled 14 and 15 were synthe-
sized and fed to A. cycloheptanicus.
Betz and Daub have reported the synthesis of 14
through methanolysis of cyclohepta-2,4,6-trienylcyanide
(20) to the methyl ester 21 followed by saponification.50
[8-13C]20 was prepared analogously to the procedure of
Doering and Knox,51 instead using tropylium tetrafluo-
roborate52 and potassium [13C]cyanide. We then followed
the Betz and Daub procedure50 to convert [8-13C]20 to
[8-13C]14, although we modified the methanolysis step
(20 f 21). The cyano group was converted to the methyl
ester by treatment of 20 with chlorotrimethylsilane in
methanol for 48 h instead of saturating the methanol
solution with HCl gas. [1-2H,8-13C]14 was additionally
prepared by the treatment of [8-13C]14 with 2 N NaOD
in D2O.53
The fate of the C-1 hydrogen of 14 was examined in a
second feeding experiment with [1-2H,8-13C]14 (Scheme
9). The resultant 1c was enriched once again with 13C
2
at C-11 (4%). The H-NMR spectrum indicated a single
signal at 1.38 ppm (H-14/17ax); no deuterium was evident
at 1.60 ppm (H-14/17eq). Recall that both methylene
hydrogens at C-14/17 of 1c biosynthesized from L-[2,3S-
2H2]phenylalanine were equally enriched with deuterium,
which is attributed to the pathway from phenylalanine
proceeding through a symmetrical intermediate. Yet in
this case, only one position is labeled, indicating that if
14 is a pathway intermediate, a C2 symmetrical inter-
mediate must occur before and not after 14. The total
incorporation of [1-2H,8-13C]14 was measured by MS on
the resultant 1b and found to be 3.9% [(M + 1)+ ) 3.5%
enrichment and (M + 2)+ ) 0.4% enrichment], while the
relative deuterium incorporation was only 10% (Table 3),
corresponding to a greater deuterium washout than that
observed from the benzylic position in 7. No deuterium
was lost from the recovered [1-2H,8-13C]14.
We followed the Takahashi et al. syntheses of 1553
when preparing [8-13C]15 from [8-13C]20 and [3-2H1,
8-13C]15 from [8-13C]14. In both feeding experiments, 15
was recovered from the spent culture medium unaltered
and, in the second case, without loss of deuterium. As
in the feeding experiment with 14, no other triene isomer
was detected by GC-MS. The resultant 1c from the
feeding of [8-13C]15 was more highly enriched at C-11
(8%) than that from [8-13C]14, which may indicate that
15 is a more advanced precursor than 14. A competition
experiment with [1,2-13C2]6 was also conducted. Ten
milligrams of [8-13C]15 and 5 mg of [1,2-13C2]6 were fed
to 500 mL of culture, and the resultant 1c was singly
(46) Gensler, W. J .; J ohnson, F.; Sloan, D. B. J . Am. Chem. Soc.
1960, 82, 6074.
(47) Braren, W.; Buchner, E. Ber. Dtsch. Chem. Ges. 1901, 34, 982,
and references therein; for a review, see Maier, G. Angew. Chem. 1967,
79, 446.
(48) Wehner, R.; Gu¨nther, H. J . Am. Chem. Soc. 1975, 97, 923.
(49) Hanson, K. R.; Rose, I. A. Acc. Chem. Res. 1975, 8, 1.
(50) Betz, W.; Daub, J . Chem. Ber. 1972, 105, 1778.
(51) Doering, W. von E.; Knox, L. H. J . Am. Chem. Soc. 1957, 79,
352.
(52) Conrow, K. Org. Syntheses 1963, 43, 101.
(53) Takahashi, K.; Yamamoto, H.; Nozoe, T. Bull. Chem. Soc. J pn.
1970, 43, 200.

Biosynthetic Studies of ω-Cycloheptyl Fatty Acids
J . Org. Chem., Vol. 62, No. 7, 1997 2181
elimination55 followed by ester hydrolysis yielded the R,â-
unsaturated [8-13C]19. The ∆2-monoene [8-13C]18 was
synthesized from 3-bromocycloheptene56 and sodium [13C]-
cyanide based on the synthesis of 24.57 anti-Addition of
bromine to [8-13C]18 followed by two base-induced dehy-
drobrominations of the crude stereoisomeric dibromides
afforded a 3.6:1 mixture of the 1,3- and 1,6-regioisomeric
dienes. Resultant chromatography and recrystallization
yielded pure [8-13C]17. 13C-Label from each of the [8-13C]-
cycloheptenylcarboxylic acids was efficiently incorporated
into the C-11 position of the resultant 1c (Table 2),
lending support for the conversion of 17 to 3 via 18 and
19.
Sch em e 10. Con ver sion of Cycloh exylca r boxylic
Acid Der iva tives by A. cycloh epta n icu s in to
ω-Cycloh exyl F a tty Acid s
enriched at C-11 (8%) and doubly enriched at C-11 and
C-12 (10%).
Discu ssion
The fate of the C-3 hydrogens of 15 was examined in
a feeding experiment with [3-2H1,8-13C]15 (Scheme 9).
Once again, the resultant 1c was enriched with 13C at
C-11 to the extent of 7%. The 2H-NMR spectrum showed
the expected two signals in equal proportions at 1.59 and
1.37 ppm corresponding to H-14/17eq and H-14/17ax,
respectively (Table 4). The relative deuterium incorpora-
tion was measured by MS on the corresponding 1b. The
(M + 1)+ peak was not increased, whereas the (M + 2)+
peak was enhanced 7.8%, indicating that the deuterium
retention relative to 13C is 100% (Table 3).
F eed in g Exp er im en ts w ith Cycloh ep tyl Mim ics.
Intermediates late in the metabolic pathway from shikimic
acid to cyclohexanecarboxylic acid13,15 were administered
to A. cycloheptanicus to determine whether these inter-
mediates mimic their corresponding cycloheptyl coun-
terparts and, as a consequence, are inadvertently con-
verted into ω-cyclohexyl fatty acids. The resultant fatty
acid methyl esters from these experiments were analyzed
by GC-MS.
Initial experiments on the biosynthesis of the unique
ring structure in ω-cycloheptyl fatty acids have been
presented. As in the case of the ω-cyclohexyl fatty acids
from A. acidocaldarius,13 the corresponding cycloheptane-
carboxylic acid (probably as its CoA thioester) serves as
the starter unit for ω-cycloheptyl fatty acid biosynthesis
in A. cycloheptanicus. The [U-13C6]glucose feeding ex-
periment demonstrated that all eight carbons of the
cycloheptanecarboxylic acid moiety are derived from
glucose. Analysis of the resultant labeling pattern sug-
gested a shikimate pathway origin, but via aromatic
amino acids rather than directly from the seven carbons
of shikimate.
The labeling pattern of the cycloheptanecarboxylic acid
moiety in 1 biosynthesized from [U-13C6]glucose was
reminiscent of that recently published by Cane et al. for
the seven-membered ring Pseudomonas metabolite thiotro-
pocin.41 A feeding experiment with [1,2-13C2]phenylacetic
acid in A. cycloheptanicus demonstrated that the cyclo-
heptanecarboxylic acid-derived moiety of 1 was also
elaborated from phenylacetic acid. The fates of the
aromatic and benzylic hydrogens of phenylacetic acid
were then determined through feeding experiments with
deuterated phenylalanine and phenylacetic acid. The
results from these experiments dictate the following
boundary conditions for the biosynthetic pathway from
phenylacetic acid to cycloheptanecarboxylic acid:
Cyclohex-1- and -2-enecarboxylic acids54 (23, 24) were
both converted by A. cycloheptanicus into fully reduced
ω-cyclohexyl fatty acids. The ratio of methyl ω-cyclo-
hexylundecanoate to methyl ω-cycloheptylundecanoate in
the two experiments was 1:3.5 and 1:2.5, respectively.
These ratios roughly equate to relative incorporations of
the cyclohexyl mimics, giving values of 22 and 29%,
respectively. Conversely, 5-hydroxycyclohex-1-ene- and
cyclohexa-1,5-dienecarboxylic acids (25, 26) were not
metabolized to ω-cyclohexyl fatty acids. Cyclohexa-1,3-
dienecarboxylic acid (27), which is not an intermediate
in the A. acidocaldarius cyclohexanecarboxylate pathway
but was instead converted into ω-cyclohex-3-enyl fatty
acids,13 was also tested as a mimic. GC-MS analysis
revealed that the fully reduced methyl ω-cyclohexylun-
decanoate was formed in addition to methyl ω-cyclohep-
tylundecanoate in a 1:11 ratio (8% relative incorporation).
On the basis of these results, it is suggested that the final
steps in the formation of 3 may involve two double bond
reductions of 17, which is consistent with the proposed
pathway outlined in Scheme 8, and similar to that
proposed for the reduction of the mimic 27 (Scheme 10).
Cycloh ep ten ylca r boxylic Acid F eed in g Exp er i-
m en ts. The corresponding cycloheptenylcarboxylic acids
were synthesized with a 13C-label at the carboxyl carbon
and fed to A. cycloheptanicus to substantiate the above
results with the cyclohexyl mimics. Direct dehydroge-
nation of the methyl ester of [8-13C]3 by selenoxide syn
(1) The biosynthetic pathway proceeds through a C2
symmetrical intermediate; this would account for the
equal distribution of deuterium between the equatorial
and axial positions at each ring methylene carbon.
(2) Fifty percent of the aromatic meta-hydrogen that
ultimately resides equally at H-14/17ax and H-14/17eq of
1 is lost during the biosynthetic transformation. Other-
wise, all of the other aromatic hydrogens are retained.
(3) The pro-3S hydrogen of 7 is retained at 50% in the
resultant 1 equally at H-14/17ax and H-14/17eq, whereas
the pro-3R hydrogen is entirely lost. The partial reten-
tion and migration of the pro-3S hydrogen is suggestive
of a suprafacial 1,3-allylic isomerization.49
Evidence in support of the final stages in the sequence
(17 f 3) comes from the experiments with the cyclohexyl
mimics and the incorporations of 17-19. In fact, incuba-
tions of A. cycloheptanicus cell-free extracts with the
(55) Reich, H. J .; Renga, J . M.; Reich, I. L. J . Am. Chem. Soc. 1975,
97, 5434.
(56) Hatch, L. F.; Bachmann, G. Chem. Ber. 1964, 97, 132.
(57) Davies, S. G.; Whitman, G. H. J . Chem. Soc., Perkin Trans. 1
1976, 2279.
(54) All of the cyclohexyl intermediates were previously synthesized
and are described in ref 15.

2182 J . Org. Chem., Vol. 62, No. 7, 1997
Moore et al.
Sch em e 11. P r op osed Oxid a tive Rin g Exp a n sion of P h en yla cetic a cid (6) by A. cycloh epta n icu s
Exp er im en ta l Section
coenzyme A thioesters of monoenes 18 and 19 with
NADPH resulted in the formation of 3.58
1
Gen er a l P r oced u r es. The H-, 2H-, and 13C-NMR spectra
The hypothetical pathway outlined in Scheme 8 does
not allow for the symmetrization of the labeling patterns
nor for the observed partial losses of deuterium from 6
as dictated above. Likewise, various features from the
feeding experiment with the hypothetical intermediate
14 raise questions as to whether it is an intermediates
on the pathway from 6 to 3 or whether its incorporation
represents nonphysiological channeling of added com-
pounds onto the pathway. For instance, it is incorporated
into 1 much less efficiently than 6, and the fate and
percent transfer of the deuterium did not match that from
the corresponding position in precursors 6 and 7.
were recorded on an IBM AF-300 spectrometer operating at a
field strength of 7.1 T. Additional
13C-NMR measurements
on D-[U-13C6]-glucose-derived 1c were performed on a Bruker
AM-500 spectrometer. Chemical shifts are given in parts per
million (ppm) and are adjusted to the TMS scale by reference
to the solvent signal. Coupling constants (J ) are given in Hertz
(Hz). GC-MS was carried out on a Kratos Profile mass
spectrometer (Manchester, U.K.). Analytical TLC was ex-
ecuted on precoated silica gel 60F-254 plates. Compounds on
the plates were visualized under UV light and/or by spraying
with a KMnO4 solution (1.0 g of KMnO4, 100 mL of 1 N NaOH)
and heating at 120 °C. Mobilities are quoted relative to the
solvent front (Rf). Column chromatography was performed on
230-400 mesh silica gel from Aldrich.
For these reasons an alternate pathway is proposed
which invokes an oxidative ring-expansion reaction which
could be mediated by a cytochrome P450-like enzyme59
(Scheme 11). Enzymatic or nonenzymatic ring expansion
of the hydroxynorcaradiene intermediate 28 followed by
tautomerization would yield the ketodiene 30. A su-
prafacial 1,3-allylic isomerization of the methine hydro-
gen (HS), which originated from the pro-3S hydrogen in
7, gives 31. A subsequent nonenzymatic enolization to
32 would result in the observed 50% losses of hydrogen
originating from both the aromatic meta (Hm) and pro-S
benzylic (HS) positions in 6. Enol reduction followed by
dehydration gives the triene 15, which is then processed
to the saturated 3 as illustrated in Scheme 8.
Ma ter ia ls. A. cycloheptanicus (formerly B. cycloheptani-
cus)3 was obtained from the Deutsche Sammlung fu¨r Mikro-
organismen (DSM 4006) and A. acidocaldarius (formerly B.
acidocaldarius)3 from the American Type Culture Collection,
Rockville, MD (ATCC 27009). All chemicals were of reagent
grade and were used without further purification. Ingredients
for fermentations were purchased from Difco and Sigma. The
following companies supplied the stable-isotope labeled com-
pounds (all 13C or 2H enrichments were at least 98 atom %):
sodium [1-13C]acetate, sodium [2-13C]acetate, sodium [1,2-13C2]-
acetate, [1-2H]benzaldehyde, D2, 30% NaOD in D2O, D2O,
Cambridge Isotope Laboratories; D-[U-13C6]glucose, D-[1-13C]-
glucose, D-[2-13C]-glucose, D-[3-13C]glucose, D-[6-13C]glucose,
D-[1,2-13C2]glucose, L-[methyl-13C]methionine, 13CO2, K13CN,
[1,2-13C2]phenylacetic acid, L-[ring-2H5]phenylalanine, D,L-[3-
2H2]phenylalanine, Isotec Inc; [U-13C3]glycerol, Los Alamos
Stable Isotope Resource; [2H7]phenylacetic acid, C/D/N Isotopes
Inc; NaBD4, Aldrich.
F er m en ta tion . A loopful of A. cycloheptanicus from a plate
culture was inoculated into 100 mL of the sporulation medium
(medium A) in a 500 mL baffled Erlenmeyer flask. The
medium contained (NH4)2SO4 (0.2 g), CaCl2‚2H2O (0.25 g),
MgSO4‚7H2O (0.5 g), KH2PO4 (3.0 g), glucose (3.0 g), yeast
extract (3.0 g), and 1 mL of trace element solution per liter.
The trace element solution consisted of FeSO4‚7H2O (0.28 g),
MnCl2‚4H2O (1.25 g), and ZnSO4‚7H2O (0.48 g) per liter. The
pH of the sporulation medium was adjusted to 4.0 with 6 M
H2SO4. The inoculated culture was incubated at 50 °C for 24
h on a rotary shaker at 150 rpm. The inoculum (5-10 mL)
was then transferred to 100 mL of the same sporulation
medium or to 100 mL of a defined medium (medium B) and
incubated at 50 °C for 24 h with shaking at 150 rpm. Medium
B consisted of the sporulation medium elements minus yeast
extract, plus an additional 2 g of glucose and the following
amino acids and vitamins per liter: L-methionine (50 mg),
L-alanine (320 mg), L-isoleucine (280 mg), L-valine (160 mg),
The pathway presented in Scheme 11 accounts for the
observed losses of deuterium as described above but is
limited by the fact that it does not advance through a
symmetrical intermediate to allow for the observed
scrambling of deuterium atoms from the feeding experi-
ments with 6 and 7. We have not been able to reconcile
these observations to our satisfaction without jeopardiz-
ing the other pathway constraints. However, if the final
reduction of the ∆1-monoene 19 were to proceed without
overall stereocontrol, then the resultant 3 would be
epimeric at C-1, resulting in the observed “scrambling”
of deuterium atoms. Cell-free experiments are in progress
to probe the early stages in the pathway in order to define
the complete sequence of reactions.
(58) Tornus, I.; Floss, H. G. Unpublished results.
(59) Ortiz de Montellano, P. R. Cytochrome P-450: Structure,
Mechanism, and Biochemistry; Plenum Press, New York, 1986.

Biosynthetic Studies of ω-Cycloheptyl Fatty Acids
J . Org. Chem., Vol. 62, No. 7, 1997 2183
L-leucine (180 mg), biotin (100 µg), niacin (2 mg), thiamine
(150 µg), vitamin B12 (100 µg), and pantothenic acid (5 mg)
per liter. The trace element solution for this medium (1 mL
per liter) is CaCl2‚2H2O (0.66 g), ZnSO4‚7H2O (0.18 g), CuSO4‚-
5H2O (0.16 g), MnSO4‚H2O (0.15 g), CoCl2‚6H2O (0.18 g), H3-
BO3 (0.1 g), Na2MoO4‚2H2O (0.3 g), and FeSO4‚H2O (1.0 g) per
liter. All media were sterilized for 20 min at 121 °C in a steam
autoclave.
133.8, 134.4, 173.2 (C-1), 192.4 (C-2′); GC-MS m/ z (relative
intensity) 400 (M+, 0.1), 281 [(M - CH2C(O)Ph)+, 2], 265 [(M
- OCH2C(O)Ph)+, 2], 105 (PhCO+, 100), 97 (C7H13+, 7).
P h en acyl ω-cycloh eptyl-r-h ydr oxyu n decan oate: Rf 0.64
(hexane:EtOAc 2:1); 13C-NMR (CDCl3) δ 24.8, 26.6, 27.4, 28.6,
29.4, 29.5, 29.6, 29.7, 30.0, 34.6, 34.7, 38.3, 39.3, 66.5, 70.7,
127.8, 129.0, 134.0, 134.1, 175.1, 191.4.
Isola tion of P h en a cyl ω-Cycloh exylu n d eca n oa te fr om
A. a cid oca ld a r iu s. The fermentation of A. acidocaldarius
and isolation of the fatty acids was done as described in ref
13. The fatty acids were derivatized to their phenacyl esters
and purified by C-18 HPLC as described above. Phenacyl
ω-cyclohexylundecanoate (3 mg/1 L culture) eluted at 44 min:
13C-NMR (CDCl3) δ 24.9, 26.5, 26.8, 26.9, 29.1, 29.3, 29.5, 29.6,
29.7, 30.0, 33.5, 34.0, 37.6, 37.7, 65.8, 127.8, 128.8, 133.8, 134.4,
173.2, 192.3.
F eed in g Exp er im en ts w ith La beled P r ecu r sor s. Feed-
ing experiments were carried out in either medium A or B. In
general, single doses of precursor were administered to the
fermentation at the time of inoculation and the cultures were
harvested 24 h later. Precursors were added as sterile
solutions in either water or a 3% NaHCO3 solution to the
following media in the amounts indicated per culture volume.
Medium A: sodium [1-13C]acetate (450 mg/900 mL), sodium
[2-13C]acetate (250 mg/1 L), sodium [1,2-13C2]acetate (250 mg/1
L), [U-13C3]glycerol (100 mg/1 L), L-[methyl-13C]methionine (100
mg/1 L), (-)-[1,7-13C2]shikimic acid (5 mg/100 mL; 100 mg/
1L), [8-13C]cycloheptanecarboxylic acid (10 mg/1 L), [7-13C]-
cyclohexanecarboxylic acid (20 mg/1 L), cyclohex-1-enecarbox-
ylic acid (1.5 mg/100 mL), cyclohex-2-enecarboxylic acid (1.9
mg/100 mL), 5-hydroxycyclohex-1-enecarboxylic acid (1.2 mg/
100 mL), cyclohexa-1,5-dienecarboxylic acid (1.4 mg/100 mL),
cyclohexa-1,3-dienecarboxylic acid (1.4 mg/100 mL), [1,2-13C2]-
phenylacetic acid (20 and 100 mg/1 L), [2′,6′-2H2]phenylacetic
acid (100 mg/1 L), [3′,5′-2H2]phenylacetic acid (100 mg/1 L),
[4′-2H]phenylacetic acid (100 mg/1 L), [2H7]phenylacetic acid
(100 mg/1 L), L-[ring-2H5]phenylalanine (100 mg/1 L), D,L-[3-
2H2]phenylalanine (100 mg/1 L); L-[3R-2H1]phenylalanine (100
mg/1 L), L-[2,3S-2H2]phenylalanine (100 mg/1 L), (()-[2-2H]-
mandelic acid (90 mg/900 mL), [8-13C]cyclohepta-2,4,6-trien-
ecarboxylic acid (20 mg/1 L), [1-2H,8-13C]cyclohepta-2,4,6-
trienecarboxylic acid (15 mg/1 L), [8-13C]cyclohepta-1,4,6-
trienecarboxylic acid (17 mg/1 L), [3-2H1,8-13C]cyclohepta-1,4,6-
trienecarboxylic acid (24 mg/1 L), [8-13C]cyclohepta-1,3-
dienecarboxylic acid (5.5 mg/1 L), [8-13C]cyclohept-2-ene-
carboxylic acid (5 and 10 mg/1 L), [8-13C]cyclohept-1-enecar-
boxylic acid (10 mg/300 mL). Medium A without unlabeled
Syn th esis of La beled P r ecu r sor s. [8-13C]Cycloh ep ta n -
eca r boxylic a cid ([8-13C]3) was prepared from cycloheptyl
bromide (1.5 g, 8.5 mmol) and 13CO2 (206 mL, 9.2 mmol) to
give [8-13C]3 (662 mg, 55%) in an analogous manner to the
3
synthesis of [7-13C]5:13 13C-NMR δ (CDCl3) 26.3 (d, J C3,6-C8
)
)
4.2 Hz, C-3,6), 28.3 (s, C-4,5), 30.6 (s, C-2,7), 44.9 (d, 1J C1-C8
54.0 Hz, C-1), 183.6 (enhanced signal with small satellites
1
symmetrically arranged, J C8-C1 ) 54.0 Hz, C-8).
3′,5′-Dibr om oben zyl cya n id e (8) was prepared from 3′,5′-
dibromobenzyl bromide (2 g, 6.1 mmol) and NaCN (377 mg,
7.7 mmol) in 80% yield analogous to the synthesis of benzyl
cyanide:60 EIMS m/ z (relative intensity) 273 (M+, 0.50), 275
[(M + 2)+, 1.00], 277 [(M + 4)+, 0.48]; High-resolution EIMS
m/ z 272.87897 (calcd for C8H5Br2N, 272.87887); 1H-NMR
(acetone-d6) δ 4.04 (s, H-2), 7.63-7.75; 13C-NMR (acetone-d6)
δ 22.7 (C-2), 118.2 (C-1), 123.8 (C-4′), 131.1 (C-3′,5′), 134.0 (C-
2′,6′), 136.9 (C-1′). Anal. Calcd for C8H5Br2N: C, 35.03; H,
1.84; N, 5.11. Found: C, 34.90; H, 1.83; N, 5.04.
3′,5′-Dibr om op h en yla cetic Acid (9). Compound 8 (1 g,
3.6 mmol) was hydrolyzed in KOH to give 9 in 90% yield:60
EIMS m/ z (relative intensity) 292 (M+, 0.50), 294 [(M + 2)+,
1.00], 296 [(M + 4)+, 0.49]; high-resolution EIMS m/ z 291.87324
(calcd for C8H6Br2O2, 291.87345); 1H-NMR (acetone-d6) δ 3.76
(s, H-2), 7.55-7.65. 13C-NMR (acetone-d6) δ 39.9 (C-2), 123.0
(C-4′), 132.5 (C-3′,5′), 132.8 (C-2′,6′), 140.4 (C-1′), 171.8 (C-1).
Anal. Calcd. for C8H6Br2O2: C, 32.66; H, 2.06; O, 10.88.
Found: C, 32.48; H, 1.97; O, 10.89.
glucose: D-[U-13C6]glucose (250 mg/500 mL). Medium
B
without unlabeled L-alanine and glucose: sodium [1,2-13C2]-
acetate (100 mg/1 L), D-[1-13C]glucose (50 mg/500 mL), D-[2-
13C]glucose (50 mg/500 mL), D-[3-13C]glucose (100 mg/1 L), D-[6-
13C]glucose (100 mg/1 L), D-[1,2-13C2]glucose (100 mg/1 L), D-[U-
13C6]glucose (200 mg/2 L), [U-13C3]glycerol (50 mg/500 mL).
Medium B without unlabeled L-methionine: L-[methyl-13C]-
methionine (25 mg/500 mL).
[3′,5′-2H2]P h en yla cetic Acid ([3′,5′-2H2]6). Compound 9
(755 mg, 2.57 mmol) was stirred for 2 h at 80 °C with Na2CO3
(286 mg, 2.70 mmol) in D2O (2 mL). The D2O was removed
under vacuum at 100 °C, and the sodium salt was dissolved
in 7.5% NaOD in D2O (4 mL). Zinc dust (2 g, 30.6 mmol) was
added in one portion, and the mixture stirred at reflux for 36
h. The reaction mixture was then cooled to room temperature,
diluted with H2O (40 mL), and acidified by dropwise addition
of concentrated HCl. The phenylacetic acid was extracted into
Et2O (3 × 40 mL), dried (MgSO4), and concentrated to give 6
(264 mg, 73% yield). The product was refluxed overnight in
30% KOH/H2O to exchange the deuterated benzylic position
with hydrogen. Workup as described above gave [3′,5′-2H2]6
as a white solid (255 mg): mp 76-77 °C (lit.61 mp 76.5 °C);
high-resolution EIMS m/ z 138.06502 (calcd for C8H6D2O2,
138.06498); 1H-NMR (acetone-d6) δ 3.62 (s, 2H, H-2), 7.24, 7.31
(m, 3H); 13C-NMR (acetone-d6) δ 40.9 (C-2), 127.3 (C-4′), 128.8
(t, J ) 24.4 Hz, C-3′,5′), 130.0 (C-2′,6′), 135.7 (C-1′), 172.8 (C-
1).
Isola tion of P h en a cyl ω-Cycloh ep tyl F a tty Acid s. The
cultures were centrifuged at 14900g for 15 min. The wet cells
were suspended in 10-20 mL of 2.5% KOH in 1:3 MeOH:H2O,
sealed in 2-4 10 mL capped vials and heated in an oven at
100 °C overnight. The basic solution was then extracted with
3 × 20 mL n-hexane and acidified with 6 M HCl, and the fatty
acids were then extracted with 3 × 20 mL of 1:2 n-hexane:
EtOAc. The extract was evaporated, and per 10 mg of residue,
12 mg phenacyl bromide and 10 mg triethylamine in 2 mL
acetone were added and the mixture was stirred overnight at
room temperature.29 The fatty acid derivatives were purified
by reverse-phase HPLC on a 10 × 300 mm C-18 (10 µm)
column with 9:1 MeCN:H2O (4 mL/min, 254 nm detection) to
give phenacyl ω-cycloheptyl-R-hydroxyundecanoate (tr 28 min)
and phenacyl ω-cycloheptylundecanoate (tr 56 min). The
average amount of phenacyl ω-cycloheptylundecanoate varied
according to the medium [medium A (6.0 mg/L), medium B
without unlabeled glucose and L-alanine (1.0 mg/L)].
P h en a cyl ω-Cycloh ep tylu n d eca n oa te (1c): Rf 0.90 (hex-
ane:EtOAc 2:1). 1H-NMR (CDCl3) δ 1.08 (m, H-13/18ax, 1H),
1.14 (m, H-11, 2H), 1.18-1.28 (m, H-4 to H-10, 14H), 1.32 (m,
H-14/17ax, 1H), 1.36 (m, H-12, 1H), 1.41 (m, H-15/16ax, 1H),
1.49 (m, H-15/16eq, 1H), 1.55 (m, H-14/17eq, 1H), 1.62 (m, H-13/
18eq, 1H), 1.66 (m, H-3, 2H), 2.44 (t, J ) 7.6 Hz, H-2, 2H),
5.31 (s, 2H), 7.47 (t, J ) 7.6 Hz, 2H), 7.58 (tm, J ) 7.3 Hz,
1H), 7.90 (dm, J ) 7.8 Hz, 2H); 13C-NMR (CDCl3) δ 24.9 (C-
3), 26.6 (C-14/17), 27.4 (C-10), 28.6 (C-15/16), 29.1 (C-4), 29.3
(C-5), 29.5 (C-6), 29.6 (C-7), 29.7 (C-8), 30.0 (C-9), 33.9 (C-2),
34.7 (C-13/18), 38.3 (C-11), 39.3 (C-12), 65.5 (C-1′), 127.8, 128.8,
[2′,6′-2H2]P h en yla cetic Acid ([2′,6′-2H2]6). 2′,6′-Dichlo-
rophenylacetic acid (1.5 g, 7.3 mmol) was stirred with Na2-
CO3 (848 mg, 8.0 mmol) in D2O at 80 °C for 3 h. The D2O was
removed, and the sodium salt was stirred with 30% NaOD in
D2O (5 mL) at 80 °C for 36 h. After heating, the D2O was
removed until a wet paste remained. D2O (10 mL) was added
to the mixture in an ice bath, and Ni-Al alloy (1.4 g) was
added in one portion. After the Al dissolved and the evolution
of D2 ceased, the mixture was warmed to room temperature
and stirred for 30 min. The reaction was worked up as
(60) Vogel, A. Textbook of Practical Organic Chemistry, 4th ed.;
Longman: London, 1978.
(61) The Merck Index; Merck & Co., Inc.: Rahway, N.J ., 1989.

2184 J . Org. Chem., Vol. 62, No. 7, 1997
Moore et al.
described above to give [2′,6′-2H2]6 (800 mg) in 80% yield: mp
EtOAc 6:1) gave pure [8-13C]21 as an oil (909 mg, 6.0 mmol)
and a mixture of [8-13C]20 and -21 (172 mg): Rf 0.70 (hexane:
EtOAc 5:1); GC-MS m/ z (relative intensity) 151 (M+, 7), 119
(7), 91 (100). 1H-NMR (CDCl3) δ 2.52 (m, H-1, 1H), 3.76 (d, J
) 3.9 Hz, COOCH3, 3H), 5.40 (m, H-2,7, 2H), 6.23 (dm, J )
8.8 Hz, H-3,6, 2H), 6.62 (t, J ) 3.1 Hz, H-4,5, 2H); 13C-NMR
(CDCl3) δ 43.8 (d, J C1-C8 ) 61.9 Hz, C-1), 52.1 (d, J ) 3.9 Hz,
COOCH3, 3H), 116.9, 125.5 (d, J ) 4.9 Hz), 130.8, 173.4
(enhanced signal with small satellites symmetrically arranged,
J C8-C1 ) 62.0 Hz, C-8).
1
74.5-76 °C; H-NMR (acetone-d6) δ 3.63 (s, 2H, H-2), 7.20-
7.36 (m, 3H); 13C-NMR (acetone-d6) δ 41.1 (C-2), 127.5 (C-4′),
129.0 (C-3′,5′), 129.9 (t, J ) 24.3 Hz, C-2′,6′), 135.6 (C-1′), 172.8
(C-1).
L-[3R-2H]P h en yla la n in e ([3R-2H]7). [3R-2H]11 was pre-
pared by hydrogenation of [3-2H]1044 with [Rh((R)-prophos)-
(NBD)]ClO4 as described.45 Deprotection yielded [3R-2H]7:44
[R]25 ) -26.5 (c ) 0.9 in H2O), [lit.62 [R]25 ) -29.4 (c ) 1.0
D
D
in H2O)]; 1H-NMR (D2O) δ 3.09 (d, J ) 5.5 Hz, H-3S), 3.80 (d,
J ) 5.5 Hz, H-2), 7.14-7.29 (m, 5H); 13C-NMR (D2O) δ 36.4 (t,
J C3-2H ) 19.4 Hz, C-3), 56.3 (C-2), 127.8, 129.3, 129.7, 135.6,
173.5 (C-1).
[8-13C]Cycloh ep ta -2,4,6-tr ien eca r boxylic Acid ([8-13C]-
14). [8-13C]21 (909 mg, 6.0 mmol) was hydrolyzed as described
by Betz and Daub50 to give [8-13C]14 (680 mg, 5.0 mmol, 92%
yield based on consumed 21): GC-MS m/ z (relative intensity)
137 (M+, 15), 119 (12), 91 (C7H7+, 100); 1H-NMR (CDCl3) δ
2.58 (m, H-1, 1H), 5.36 (m, H-2,7, 2H), 6.29 (dm, J ) 8.8 Hz,
H-3,6, 2H), 6.64 (t, J ) 3.2 Hz, H-4,5, 2H), 11.27 (br s, COOH,
1H); 13C-NMR (CDCl3) δ 42.9 (d, J C1-C8 ) 59.8 Hz, C-1), 113.3,
125.9 (d, J ) 4.7 Hz), 130.7, 179.9 (enhanced signal with small
satellites symmetrically arranged, J C8-C1 ) 59.8 Hz, C-8).
[1-2H,8-13C]Cycloh ep ta -2,4,6-tr ien eca r boxylic a cid ([1-
2H,8-13C]14) was prepared from [8-13C]14 (133 mg, 0.97 mmol)
and 2 N NaOD in D2O as described53 to give [1-2H,8-13C]14
(127 mg): GC-MS m/ z (relative intensity) 138 (M+, 14), 120
(12), 92 (C7H6D+, 100); 1H-NMR (CDCl3) δ 2.58 (m, H-1, 0.01H),
5.35 (dd, J ) 8.7, 4.0 Hz, H-2,7, 2H), 6.29 (ddd, J ) 8.6, 3.4,
L-[2,3S-2H2]P h en yla la n in e ([2,3S-2H2]7). [2,3S-2H2]11
was prepared by hydrogenation of 10 with D2.45 Deprotection
yielded [2,3S-2H2]7:44 [R]25D ) -26.5 (c ) 1.0 in H2O); 1H-NMR
(D2O) δ 2.93 (s, H-3R), 7.11-7.24 (m, 5H); 13C-NMR (D2O) δ
36.4 (t, J C3-2H ) 19.7 Hz, C-3), 55.3 (t, J C2-2H ) 18.5 Hz, C-2),
127.9, 129.4, 129.6, 135.4, 173.0 (C-1).
Zin c Bor od eu ter id e.46 A mixture of ZnCl2 (1.25 g, 9.2
mmol) and anhydrous Et2O (50 mL) was refluxed for 2 h in a
100 mL flask. The mixture was cooled to room temperature,
and the supernatant, a saturated solution of ZnCl2 in Et2O
(33 mL, ca. 6 mmol), was slowly added to a stirred solution of
NaBD4 (0.5 g, 11.9 mmol) in 17 mL of Et2O. This mixture
was stirred for 2 d at room temperature. The solids were
allowed to settle, and the solution of Zn(BD4)2 (approximately
238 mM in Et2O) was removed and stored in a stoppered bottle
at 0 °C.
2
3.1 Hz, H-3,6, 2H), 6.64 (dd, J ) 3.3, 3.1 Hz, H-4,5, 2H); H-
NMR (CHCl3) δ 2.53 (br s, H-1); 13C-NMR (CDCl3) δ 42.5 (dt,
J C1-C8 ) 59.6 Hz, J C1-2H ) 20.6 Hz, C-1), 113.1, 125.9 (d, J )
4.8 Hz), 130.6, 179.8 (enhanced signal with small satellites
symmetrically arranged, J C8-C1 ) 59.7 Hz, C-8).
Meth yl [2-2H]Ma n d ela te. Zn(BD4)2 (2.6 mL of a 238 mM
solution in Et2O) was added dropwise to a stirred solution of
methyl benzoylformate (200 mg, 1.22 mmol) in anhydrous Et2O
(5 mL) in an ice bath. After stirring for 1 h, the reaction was
quenched by the addition of H2O (10 mL) followed by 1 N HCl
(2 mL). The Et2O layer was decanted, and the aqueous
solution was extracted again with Et2O (2 × 10 mL). The Et2O
fractions were dried (MgSO4) and evaporated to give an oil
(163 mg, 0.98 mmol, 80% yield): Rf 0.46 (hexane:EtOAc 2:1);
IR (neat) 3500, 1740 cm-1; 1H-NMR (CDCl3) δ 3.65 (br s, 1H),
3.74 (s, 3H), 5.17 (s, 0.01H), 7.32-7.42 (m, 5H); 2H-NMR
[8-13C]Cycloh ep t a -1,4,6-t r ien yl Cya n id e ([8-13C]22).
[8-13C]20 (614 mg, 4.80 mmol) was isomerized to [8-13C]22 (250
mg) as described:53 Rf 0.60 (hexane:EtOAc 5:1); 1H-NMR
(CDCl3) δ 2.37 (t, J ) 7.0 Hz, H-3), 5.38 (dt, J ) 9.3, 6.7 Hz,
H-4), 6.03 (q, J ) 7.3 Hz, H-2), 6.23 (dd, J ) 9.3, 5.5 Hz, H-5),
6.55 (dd, J ) 11.2, 3.3 Hz, H-7), 6.72 (dd, J ) 11.2, 5.6 Hz,
H-6); 13C-NMR (CDCl3) δ 27.8 (d, J ) 6.1 Hz, C-3), 112.1 (d,
J C1-C8 ) 78.3 Hz, C-1), 118.2 (enhanced signal with small
satellites symmetrically arranged, J C8-C1 ) 78.3 Hz, C-8), 121.5
(s), 127.3 (d, J ) 2.8 Hz), 127.5 (s), 133.8 (d, J ) 2.0 Hz), 133.9
(d, J ) 5.9 Hz).
(CHCl3) δ 5.16; 13C-NMR (CDCl3) δ 53.0, 72.5 (t, J
Hz, C-2), 126.6, 128.5, 128.6, 138.1, 174.1.
) 22.6
13
2
C- H
[8-13C]Cycloh ep ta -1,4,6-tr ien eca r boxylic a cid ([8-13C]-
15) was prepared from [8-13C]22 (250 mg, 1.82 mmol) to give
[8-13C]15 (214 mg) as described:53 GC-MS m/ z (relative
[2-2H]Ma n d elic Acid ([2-2H]12). To a solution of methyl
[2-2H]mandelate (100 mg, 0.6 mmol) in H2O (5 mL) was added
one KOH pellet. The mixture was stirred at room temperature
for 1 h. The resulting solution was washed with Et2O (3 × 10
mL), and the aqueous layer was acidified with 6 N HCl
(approximately 0.5 mL). The liberated acid was extracted with
Et2O (5 × 10 mL), and the Et2O fractions were combined, dried
(MgSO4), and concentrated to give [2-2H]12 (83.5 mg, 91%):
1H-NMR (acetone-d6) δ 5.20 (s, 0.02H), 7.29-7.51 (m, 5H); 2H-
intensity) 137 (M+, 18), 91 (C7H7+, 100); H-NMR (CDCl3) δ
1
2.37 (t, J ) 7.3 Hz, 2H, H-3), 5.41 (dt, J ) 9.2, 6.7 Hz, 1H,
H-4), 6.23 (dd, J ) 9.3, 5.5 Hz, 1H, H-5), 6.59 (q, J ) 7.0 Hz,
1H, H-2), 6.73 (dd, J )11.4, 5.5 Hz, 1H, H-6), 7.13 (br d, J )
11.5 Hz, 1H, H-7), 11.29 (br s, COOH); 13C-NMR (CDCl3) δ
27.8 (d, J ) 5.0 Hz, C-3), 120.5 (s), 127.7 (d, J ) 2.7 Hz), 127.8
(s), 129.4 (d, J C1-C8 ) 70.8 Hz, C-1), 131.8 (d, J ) 2.4 Hz), 132.2
(d, J ) 4.9 Hz), 172.4 (enhanced signal with small satellites
symmetrically arranged, J C8-C1 ) 70.8 Hz, C-8).
13
13
2
NMR (acetone) δ 5.14; C-NMR (CD3OD) δ 73.7 (t, J
)
C- H
22.6 Hz, C-2), 127.8, 129.2, 129.4, 140.6, 176.1.
[8-13C]Cycloh ep ta -2,4,6-tr ien yl Cya n id e ([8-13C]20) was
prepared from K13CN (742 mg, 11.2 mmol) and tropylium
tetrafluoroborate52 (2.00 g, 11.2 mmol) according to Doering
and Knox51 to give [8-13C]20 (1.237 g, 10.5 mmol, 94%): Rf 0.60
(hexane:EtOAc 5:1); 1H-NMR (CDCl3) δ 2.97 (m, H-1, 1H), 5.34
(m, H-2,7, 2H), 6.29 (dt, J ) 8.9, 3.0 Hz, H-3,6, 2H), 6.70 (t, J
) 3.2 Hz, H-4,5, 2H); 13C-NMR (CDCl3) δ 29.6 (d, J C1-C8 ) 61.5
Hz, C-1), 115.9 (d, J ) 2.9 Hz), 119.6 (enhanced signal with
small satellites symmetrically arranged, J C8-C1 ) 61.1 Hz, C-8),
127.1 (d, J ) 5.5 Hz), 131.2 (s).
[3-2H1,8-13C]Cycloh ep ta -1,4,6-tr ien eca r boxylic a cid ([3-
2H1,8-13C]15) was prepared from [8-13C]14 (135 mg, 0.98 mmol)
and 2 N NaOD in D2O to give [3-2H1,8-13C]15 (125 mg):53 GC-
MS m/ z (relative intensity) 138 (M+, 14), 120 (11), 92 (C7H6D+,
100). 1H-NMR (CDCl3) δ 2.32 (t, J ) 7.1 Hz, 1H, H-3), 5.40
(dd, J ) 9.2, 6.7 Hz, 1H, H-4), 6.23 (dd, J ) 9.3, 5.5 Hz, 1H,
H-5), 6.59 (t, J ) 6.8 Hz, 1H, H-2), 6.73 (dd, J )11.4, 5.5 Hz,
1H, H-6), 7.13 (dd, J ) 11.5, 1.2 Hz, 1H, H-7). 13C-NMR
(CDCl3) δ 27.4 (dt, J C3-C8 ) 5.0 Hz, J C3-2H ) 20.2 Hz, C-3),
Meth yl [8-13C]Cycloh ep ta -2,4,6-tr ien eca r boxyla te ([8-
13C]21). [8-13C]20 (1.237 g, 10.5 mmol) was dissolved in
anhydrous Et2O (15 mL) and MeOH (20 mL). Chlorotrimeth-
ylsilane (5 mL, 39.4 mmol) was next added slowly to the above
stirred solution in an ice bath. The solution was stirred at
room temperature for 48 h, at which time it turned pinkish.
Water (30 mL) was slowly added over 30 min. The solution
was then extracted with Et2O layers (4 × 40 mL), and the Et2O
layers were combined, dried (MgSO4), and evaporated to give
a dark oil (1.197 g). Flash chromatography (silica gel, hexane:
120.4 (s), 127.7 (d, J ) 2.3 Hz), 127.8 (s), 129.3 (d, J C1-C8
)
70.6 Hz, C-1), 131.7 (d, J ) 2.7 Hz), 132.1 (d, J ) 4.2 Hz),
172.5 (enhanced signal, C-8).
Meth yl [8-13C]cycloh ep t-1-en eca r boxyla te63 was pre-
pared from [8-13C]3 (50 mg, 0.35 mmol) in 63% yield in analogy
to the synthesis of methyl cyclohex-1-enecarboxylate from 5:55
1
Rf 0.15 (Et2O:hexane 1:9); H-NMR (CDCl3) δ 1.44-1.49 (m,
4H), 1.65-1.69 (m, 2H), 2.22 (dd, J ) 11.1, 6.7 Hz, 2H), 2.44
(dd, J ) 5.4, 5.3 Hz, 2H), 3.65 (s, 3H), 7.11 (t, J ) 6.7 Hz, 1H);
13C-NMR (CDCl3) δ 25.7, 26.1, 27.3, 28.7, 31.9, 51.6, 136.4 (d,
J C1-C8 ) 59.2 Hz, C-1), 144.4, 168.5 (enhanced signal, C-8).
(62) The optical purity was confirmed by comparison of the product’s
optical rotation with that of authentic material from Sigma.
(63) Pawlak, J . L.; Berchtold, G. A. J . Org. Chem. 1988, 53, 4063.

Biosynthetic Studies of ω-Cycloheptyl Fatty Acids
J . Org. Chem., Vol. 62, No. 7, 1997 2185
[8-13C]Cycloh ep t-1-en eca r boxylic Acid ([8-13C]19). To
a solution of methyl [8-13C]cyclohept-1-enecarboxylate (31 mg,
0.2 mmol) in MeOH (0.5 mL) at 0 °C was added a solution of
lithium hydroxide monohydrate (21 mg, 0.5 mmol) in H2O (0.5
mL). The mixture was stirred at 0 °C for 2 h and then the
MeOH was removed in vacuo. The aqueous solution was
diluted with H2O (5 mL), extracted with Et2O (2 × 10 mL),
acidified to pH 1 with dilute HCl, and extracted with Et2O (3
× 15 mL). The latter extracts were combined and dried (Na2-
SO4), and the solvent was removed in vacuo to give [8-13C]19
(26 mg, 92% yield) as a colorless oil: 1H-NMR (CDCl3) δ 1.48-
1.53 (m, 4H), 1.69-1.79 (m, 2H), 2.30 (dd, J ) 11.1, 6.7 Hz,
2H), 2.49 (dd, J ) 5.5, 5.3 Hz, 2H), 7.33 (t, J ) 6.7 Hz, 1H);
13C-NMR (CDCl3) δ 25.6, 26.1, 27.0, 29.0, 32.0, 135.9 (d, J C1-C8
) 59.1 Hz, C-1), 147.3, 173.5 (enhanced signal, C-8).
[8-13C]Cycloh ep t-1,3-d ien eca r boxylic Acid ([8-13C]17).
Bromine (53 µL, 1.03 mmol) in chloroform (60 µL) was slowly
added to a stirred solution of [8-13C]18 (141 mg, 1 mmol) in
chloroform (0.5 mL mL) at -10 °C. After stirring at room
temperature for 10 min, the solvent was removed under
vacuum (1 Torr) to quantitatively furnish a 2.8:1 (cis:trans)
mixture of diastereomeric 2,3-dibromo[8-13C]cycloheptanecar-
boxylic acids as a yellow-brownish solid. The bromides were
dissolved in THF (5 mL), treated at -78 °C with slow addition
of potassium tert-butoxide (95%, 354 mg, 3 mmol), and stirred
at -10 °C. After 1 h, water (10 mL) was added, and the
aqueous phase was washed with pentane (20 mL) before
adjusting to pH 2 (concentrated HCl). The collected organic
layers from the Et2O extraction (3 × 20 mL) were dried
(MgSO4) and concentrated to give 98 mg of a yellowish oil.
Silica flash column chromatography (hexane:EtOAc 5:1) af-
forded a 3.6:1 mixture of [8-13C]cyclohepta-1,3- and -1,6-
dienecarboxylic acids (colorless crystals, 63 mg, 45%) according
to 1H-NMR (ratio of the olefinic protons). Recrystallization
(30 mg in 1 mL hexane, -78 °C) afforded pure [8-13C]17 (5.5
mg): 1H-NMR (CDCl3) δ 1.86 (m, 2H, H-6), 2.42 (m, 2H, H-5),
2.63 (m, 2H, H-7), 5.93 (m, 1H, H-4), 6.21 (dt, J ) 11.7, 4.9
Hz, 1H, H-3), 7.14 (t, J ) 7.4, 1H, H-2); 13C-NMR (CDCl3) δ
24.5 (d, J ) 2.9 Hz), 29.2 (d, J ) 2.9 Hz), 33.0, 123.6 (d, J )
7.6 Hz), 123.9, 133.3 (d, J ) 69.5 Hz, C-1), 136.3 (d, J ) 2.0
Hz), 173.8 (enhanced signal, C-8). Anal. Calcd. for C8H10O2:
C, 69.55; H, 7.30. Found: C, 69.49; H, 7.11 (for unlabeled 17).
[8-13C]Cycloh ep t -2-en eca r b oxylic Acid ([8-13C]18).
Freshly distilled 3-bromocycloheptene56 (350 mg, 2 mmol) was
added dropwise at room temperature to a stirred solution of
sodium [13C]cyanide (200 mg, 4 mmol) in 1-methyl-2-pyrroli-
dinone (1.5 mL). After stirring for 2.5 h, the mixture was
diluted with Et2O (40 mL) and extracted with H2O (30 mL).
The aqueous phase was extracted with Et2O (2 × 40 mL), and
the collected Et2O extracts were washed with water (10 mL),
dried (MgSO4), and evaporated. Chlorotrimethylsilane (1.3
mL, 10 mmol) was added to the crude nitrile (brown oil, 13C-
NMR in benzene-d6: δCN 120.7 ppm) in absolute Et2O (3 mL)
and anhydrous MeOH (4 mL) at 0 °C and stirred at room
temperature for 4 d. The reaction mixture was poured into
crushed ice (10 mL) and the product extracted with Et2O (4 ×
20 mL). Drying (MgSO4) and evaporation of the solvent gave
the crude methyl ester as a yellowish oil (175 mg) that was
directly subjected to a solution of lithium hydroxide monohy-
drate (95 mg, 2.26 mmol) in 40% aqueous MeOH (2.5 mL).
After stirring for 10 h at room temperature, the alkaline
solution was washed with Et2O (3 × 5 mL) and acidified to
pH 2 with concentrated HCl. The solution was extracted with
Et2O (3 × 30 mL), the extract dried (MgSO4), and the solvent
removed in vacuum to yield [8-13C]18 as an oil (94 mg, 33%
yield): 1H-NMR (CDCl3) δ 1.30-2.20 (m, 8H), 3.29 (m, 1H,
H-1), 5.89 (m, 2H); 13C-NMR (CDCl3) δ 26.5, 28.5, 30.3 (d, J )
4.6 Hz), 30.4, 45.8 (d, J ) 55 Hz, C-1), 129.6, 133.7 (d, J ) 5.2
Hz), 181.6 (enhanced signal, C-8). Anal. Calcd. for C8H12O2:
C, 68.55; H, 8.63. Found: C, 68.12; H, 8.57 (for unlabeled 18).
Ack n ow led gm en t. We thank Dr. Thomas Pratum,
University of Washington, for carrying out the HMQC
experiment. This work was supported by grants from
NIH (AI 20264) and NATO (CRG 910456) and fellow-
ships to B.S.M. (NIH, NRSA # GM 08437), K.D.W (NIH,
GM 32333), and I.T. (Deutsche Forschungsgemein-
schaft, To 179/1-1). We also acknowledge the supply
of 13C-labeled compounds provided by the Los Alamos
Stable Isotope Resource, supported by NIH grant RR
02231 and USDOE/OHER.
J O962402O