Isolation of 3' -O-acetylchloramphenicol: a possible intermediate in chloramphenicol biosynthesis.

Article abstract of DOI:10.1016/S0960-894X(01)00739-9

3' -O-acetylchloramphenicol, commonly formed from chloramphenicol by resistant bacteria, has been isolated from the antibiotic-producing organism. Biosynthetic experiments suggest that it is a protected intermediate in chloramphenicol biosynthesis, implicating acetylation as a self-resistance mechanism in the producing organism.

Full text of DOI:10.1016/S0960-894X(01)00739-9

Bioorganic & Medicinal Chemistry Letters 12 (2002) 283–286
0
Isolation of 3 -O-Acetylchloramphenicol: A Possible Intermediate
in Chloramphenicol Biosynthesis
a,b,y
Elizabeth A. Lewis, Mahmood Piraee, Karl-Heinz van Pe
´
e,b
c
a
Frank Groß,
a
c,
Leo C. Vining and Robert L. White *
aDepartment of Biology, Dalhousie University, Halifax, Nova Scotia, Canada B3H 4J1
bInstitut fu¨r Biochemie, Technische Universita¨t Dresden, D-01062 Dresden, Germany
Department of Chemistry, Dalhousie University, Halifax, Nova Scotia, Canada B3H 4J3
c
Received 9 August 2001; accepted 25 October 2001
0
Abstract—3 -O-Acetylchloramphenicol, commonly formed from chloramphenicol by resistant bacteria, has been isolated from the
antibiotic-producing organism. Biosynthetic experiments suggest that it is a protected intermediate in chloramphenicol biosynthesis,
implicating acetylation as a self-resistance mechanism in the producing organism. # 2002 Elsevier Science Ltd. All rights reserved.
5
Acetylation, adenylation, and phosphorylation are
common modifications conferring resistance to anti-
Cultures of S. venezuelae strain 13s were grown at
ꢀ
6
30 C in glucose–isoleucine–salts solution with aeration
on a rotary shaker, and harvested after 7 days by filtra-
tion. The culture filtrate was extracted with three half-
volumes of ethyl acetate, and the extract after evapora-
1
,2
biotics in pathogenic microorganisms. Many bacteria
3
inactivate chloramphenicol (1) by acetylation, which
prevents the modified antibiotic from binding to pro-
karyotic ribosomes and inhibiting peptidyl transferase
7
tion in vacuo, was assayed by reversed-phase HPLC.
4
during protein biosynthesis. Chloramphenicol acetyl-
3
Chloramphenicol and an unknown metabolite with a
longer retention time were detected. To obtain sufficient
material for NMR analysis, the evaporated extract from
500 mL of culture filtrate was redissolved in water and
fractionated on a silica C18 column (95Â0.9 cm, Baker-
bond, 40 mm) by gradient elution with methanol–water.
The NMR spectra of the second metabolite were similar
to those of chloramphenicol except for additional reso-
transferase (CAT, EC 2.3.1.28), catalyzes acetylation of
the primary hydroxyl group by acetyl-CoA, yielding 3 -
0
O-acetylchloramphenicol (3). In a slower, nonenzymatic
reaction, the acetyl group migrates to the secondary
hydroxyl group. 1 -O-Acetylchloramphenicol (2) is also
0
substrate for CAT, forming 1 ,3 -O-diacetyl-
0
0
a
chloramphenicol (4), although the reaction is slower
than for chloramphenicol itself. None of the acetylated
chloramphenicol derivatives exhibits antibiotic activity,
and the action of CAT is the most common mechanism
1
3
nances at 172.5 and 20.6 ppm in the C spectrum, and
1
at 2.02 ppm (singlet, 3H) in the H spectrum, which
were consistent with an O-acetyl group.
3
for acquiring resistance to chloramphenicol in bacteria.
Although detected in Streptomyces species, remarkably,
however, CAT has never been reported in a strain pro-
1
,4
ducing the antibiotic. The present report describes the
0
isolation of 3 -O-acetylchloramphenicol from chlor-
amphenicol-producing cultures of Streptomyces vene-
zuelae, and experiments to determine the biosynthetic
relationship between the antibiotic and its acetylated
derivative.
*Corresponding author. Tel.:+1-902-494-6403; fax:+1-902-494-1310;
e-mail: robert.white@dal.ca
Present address: Syngenta Biotechnology, Inc., PO Box 12257, 3054
Cornwallis Road, Research Triangle Park, NC 27709-2257, USA.
y
To obtain a precise structural assignment, O-acetyl
8
,9
derivatives were prepared from chloramphenicol (1)
0
960-894X/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(01)00739-9

284
F. Groß et al. / Bioorg. Med. Chem. Lett. 12 (2002) 283–286
and acetic anhydride10 and separated by preparative
To investigate the biosynthesis of the acetyl and
dichloroacetyl groups, S. venezuelae cultures (500 mL)
were supplemented at 72, 96 and 120 h after inoculation
9
TLC (silica gel; CHCl /CH OH, 19:1). Extraction of
3
3
the separated, fluorescence-quenching bands into cold,
dry CH CN yielded chloramphenicol (1, R 0.15), 1 -O-
9
0
acetylchloramphenicol (2, Rf 0.36), 3 -O-acetylchlor-
13
with sterile aqueous sodium [2- C]acetate (3Â0.5 mol,
13
3
f
0
amphenicol (3, Rf 0.45) and 1 ,3 -di-O-acetylchlor-
13
13
99.7% C, MSD Isotopes) or ethyl [2,4- C ]aceto-
2
0
amphenicol (4, R 0.70); no isomerization of 2 or 3 in
0
acetate (3Â0.25 mol, 99% C, Cambridge Isotope
Laboratories). The metabolites were isolated as descri-
bed above and C enrichments were determined by
f
1
3
acetonitrile solution was apparent after several weeks
storage at À20 C. On the other hand, two TLC spots
ꢀ
were detected, and by H NMR spectroscopy 1:4 and
19
13
NMR spectroscopy (Table 2). Label from [2- C]ace-
tate was incorporated into the dichloroacetyl moieties of
1
1
1
0
C enrichment of the dichloromethine carbon was
1
:5 mixtures, close to the equilibrium ratio of 2 and 3,
chloramphenicol and 3 -O-acetylchloramphenicol; the
13
0
0
were observed when the separated bands of 1 - and 3 -
O-acetylchloramphenicol were extracted from silica gel
into ethyl acetate.
more than 5 times that of the amide carbonyl, support-
ing previous studies that showed labelling of both
1
4
20
carbons from [2- C]acetate and incorporation of
13
The NMR spectra12 of chloramphenicol and 1 ,3 -di-O-
0
0
[1,2- C ]acetate with considerable bond scission.
2
16
0
chloramphenicol was observed. Administration of ethyl
acetylchloramphenicol were consistent with published
1
Negligible enrichment of the acetyl group in 3 -O-acetyl-
13,14
13 15,16
H
and
mers were distinguished in H NMR spectra by chemi-
C
NMR data. The monoacetyl iso-
1
13
[2,4- C ]acetoacetate to S. venezuelae yielded com-
2
0
.0 and 3.5 ppm for 1 -O-acetylchloramphenicol
0
cal shift differences for H-2 and H-3 (i.e., resonances at
6
plementary labelling results (Table 2); the acetyl group
0
0
at 5.1 and 4.2 ppm for 3 -O-acetylchloramphenicol )
13,17
and
13
of 3 -O-acetylchloramphenicol was highly enriched,
whereas the dichloroacetyl moieties of both metabolites
were enriched to a much smaller extent.
0
and by 1-2 ppm downfield shifts of the C resonance of
1
3
0
0
C-1 or C-3 upon acetylation (Table 1). The position of
the acetyl group in 3 was further confirmed by an
HMBC experiment that showed long range coupling
between the carbon of the acetyl carbonyl and the
While a common origin for the dichloroacetyl groups in
0
chloramphenicol and 3 -O-acetylchloramphenicol is
indicated by parallel labelling (in both experiments) of
their dichloromethine and amide carbonyl carbon
0
0
1
13
H-2 ,H-3 spin system. Comparison of the H and
1
C
8
13
NMR data with those of the synthetic monoacetylated
samples revealed that the second metabolite had been
atoms, the different levels of C incorporation into the
0
dichloroacetyl and acetyl components of 3 -O-acetyl-
0
0
isolated as a mixture of 1 - and 3 -O-acetylchlor-
amphenicol. This is consistent with the relative ease
with which mono-acetyl chloramphenicols inter-
convert, but the approximately 1:6 and 1:9 ratios of
:3 found in the product indicate that equilibrium was
not established. Since the proportion of 3 in the isolated
chloramphenicol point to different routes of biosynthesis.
Assuming the acetyl group is derived from acetyl–CoA,
the labelling results imply that ethyl acetoacetate, which
could generate acetyl–CoA from acetoacetate (or its ester)
by a retro Claisen condensation without consumption
of energy, is more efficient than acetate as a precursor of
acetyl–CoA. They also suggest that acetate is not incor-
porated into the dichloroacetyl unit via acetyl–CoA.
1
1
2
0
material exceeded that in the equilibrium mixture, 3 -O-
acetylchloramphenicol is probably the initial acetylated
chloramphenicol formed in the S. venezuelae culture,
yielding 1 -O-acetylchloramphenicol by nonenzymatic
0
isomerization.
The isotopic results do not define the order in which
0
chloramphenicol and 3 -O-acetylchloramphenicol are
13
1
0 0
a,b
Table 1.
C and H NMR data, including observed HMBC correlations, for 1 - and 3 -O-acetylchloramphenicol
0
0
Position
1 -O-Acetylchloramphenicol (2)
3 -O-Acetylchloramphenicol (3)
13C
d H
1
c
d
13C
d H
1
HMBC
d
0
1
2
3
4
5
6
148.8
128.7
124.5
146.5
124.5
128.7
73.8
56.5
61.4
67.6
150.2
128.3
124.2
148.5
124.2
128.3
71.2
55.2
64.0
67.5
H-3, H-5, H-1
0
d
d
d
7.58 (2H)
8.18 (2H)
7.60 (2H)
8.16 (2H)
H-6, H-1
H-2, H-5
d
H-2, H-3, H-5, H-6,
H-3, H-6
H-2, H-1
H-2, H-6, H-2 , H-3
d
d
d
8.18 (2H)
7.58 (2H)
8.16 (2H)
7.60 (2H)
d
0
0
0
0
1
6.02 (d, J=5 Hz, 1H)
4.16–4.31 (m, 1H)
3.43–3.63 (m, 2H)
6.10 (s, 1H)
5.07 (br s, 1H)
4.15–4.35 (m, 3H)
4.15–4.35 (m, 3H)
6.04 (s, 1H)
0
0
0
2
0
H-3
H-1
3
–
–
–
–
–
CHCl
C¼O (amide)
CH
2
165.1
21.1
170.9
165.1
21.0
171.6
–CHCl
2
3
2.12 (s, 3H)
2.01 (s, 3H)
0
0
C¼O (ester)
H-2 , H-3 , –CH
3
NH
7.24 (br d, J=9 Hz, 1H)
7.11 (br d, J=8 Hz, 1H)
a
CD
3
CN, Bruker AMX 400 spectrometer, chemical shifts refer to CHD
2 H 3 C
CN (d 1.94 ppm) and CD CN (d 1.39 ppm).
b
c
Assignments and one-bond C–H connections were determined in JMOD and HETCOR experiments.
50.1 MHz.
Apparent d, 9 Hz splitting.
2
d

F. Groß et al. / Bioorg. Med. Chem. Lett. 12 (2002) 283–286
Table 2. Incorporation of 13C above natural abundance (%)a
Substrate Chloramphenicol (1)
285
0
3 -O-Acetylchloramphenicol (3)
Cl
2
CH–
–C¼O
Cl
2
CH–
–C¼O (amide)
CH
3
–
–C¼O (ester)
À0.1
Sodium [2- 3C]acetate
1
4.8 (5.2)b
0.5
0.6
0.2
4.9 (7.1)b
0.5
0.7
d
0.2 (0.9)c
16.2 (15.0)c
13
d
Ethyl [2,4-
C
2
]acetoacetate
—
—
a
The H and 13C NMR spectra of samples in CD
OD were acquired on a Bruker AMX-400 spectrometer. A 3.7 ms 30 pulse with an acquisition
ꢀ
3
1 13
1
time of 3.4 s and a 1-s delay were used to collect the H NMR spectra. The C NMR spectra (800 or 1000 scans) were acquired using 64 K data
points in a 30 kHz spectral window, a 7.25 ms 90 pulse, a 1.08-s acquisition time, and a 4-s delay. A line broadening of 2 Hz was used to process the
free-induction-decay data. Spectra for the natural-abundance reference and C-enriched samples were acquired under identical conditions. The
ꢀ
13
13
19
percentage C enrichment was calculated for individual carbons of the acetyl and dichloroacetyl units by normalizing peak intensities to the
average peak intensity of the phenylpropanoid carbon atoms. Error limits are about ꢁ0.2%. Numbers in parentheses are incorporations calculated
b1
13
1
by integrating the C satellite peaks in the H NMR spectrum.
J
J
CH=180 Hz.
CH=130 Hz.
Signal not observed.
c1
d
released by the biosynthetic process. To determine whe-
0
with the late growth phase and with the particulate
fraction of sonicated mycelium, together with evidence
for export of chloramphenicol from S. venezuelae
mycelium producing the antibiotic, suggest that hydro-
ther 3 -O-acetylchloramphenicol might be formed from
chloramphenicol, cell-free extracts prepared by sonica-
tion of 7-day mycelium were incubated with
chloramphenicol (1 mM) and acetyl–CoA (1 mM). No
O-acetylchloramphenicol was detected by HPLC, and a
0
lysis of 3 -O-acetylchloramphenicol is linked to excre-
tion of chloramphenicol. Metabolic shielding with
acetylated intermediates to protect the producing
organism is used during biosynthesis of the antibiotics
bialaphos and kanamycin, and a series of phosphory-
lated compounds are intermediates in streptomycin
2
1
colorimetric assay showed no CAT activity. Mixing
the S. venezuelae extract with an active extract from
Escherichia coli cc118 did not diminish the E. coli CAT
activity. The lack of CAT activity and the failure to
detect genes homologous to known CAT genes in S.
1
0
biosynthesis. Formation of 3 -O-acetylchloramphenicol
and esterase-mediated generation of the metabolically
toxic chloramphenicol during excretion may constitute
a self-resistance mechanism in S. venezuelae.
venezuelae (Wynands and van Pe
´
that 3 -O-acetylchloramphenicol is not formed from
e, unpublished) suggest
0
chloramphenicol in the producing organism.
1
,4,22
On the other hand, it was observed by HPLC that
O-acetylchloramphenicol added to washed mycelium
prepared from 5- to 7-day (late growth phase), but not
from 3- or 4-day (exponential phase) cultures, was lar-
gely hydrolyzed to chloramphenicol within 60 min.
Since O-acetylchloramphenicol remained unaltered in
culture fluid passed through a 0.2 mm filter, and in con-
trol incubations with boiled mycelium, the results
implicate O-acetylchloramphenicol esterase activity in
S. venezuelae mycelium. After sonication and cen-
trifugation (10 min at 5000g) of 7-day mycelium, the
esterase was located mainly in the insoluble pellet.
Addition of the esterase inhibitors acetylsalicylic acid
Without coupling to chloramphenicol export, esterase
catalyzed hydrolysis of O-acetylchloramphenicol in the
last step of chloramphenicol biosynthesis would defeat
acetylation of the antibiotic as a resistance mechanism
in S. venezuelae. Not only would the active antibiotic be
regenerated, but a futile cycle of acetylation/deacetyl-
ation would deplete the cell of acetyl–CoA and
ultimately ATP. To evade chloramphenicol toxicity
antibiotic-producing strains of Streptomyces have
evolved alternative protective mechanisms. Under
physiological conditions in which genes for biosynthesis
of the antibiotic are not expressed, chloramphenicol
7
4
can be phosphorylated or degraded. When the bio-
synthesis genes are expressed, export of chlor-
amphenicol from the mycelium prevents access to
12 mM) and sodium fluoride (0.5 mM)23 to S. vene-
2
2
(
zuelae cultures during the antibiotic production phase
increased accumulation of O-acetylchloramphenicol in
4
susceptible target sites in the producing organism, and
0
the broth. An esterase hydrolyzing 3 -O-acetyl-
a complementary self-resistance mechanism is provided
by the use of acetylated intermediates in the biosyn-
thetic pathway.
2
2
chloramphenicol has been reported to be present in
cultures of most chloramphenicol-producing stretomy-
cetes, but was located in the soluble fraction after soni-
cation. It was not responsible for failure of the cultures
to exhibit CAT activity.
Acknowledgements
0
The enzyme-catalyzed conversion of 3 -O-acetylchlor-
We thank the Natural Sciences and Engineering
Research Council of Canada (NSERC) and the Deutsche
Forschungsgemeinschaft (DFG, grant Pe 348/5-2) for
financial support, the Atlantic Region Magnetic Reso-
nance Centre for NMR spectra, and MedMira Labora-
tories Ltd. for support to R.L.W. during the
preparation of this manuscript.
amphenicol to chloramphenicol and the accumulation
of acetylated antibiotic on addition of inhibitors impli-
cate O-acetylation as a biosynthetic step en route to
chloramphenicol, with the ester as a biosynthetic inter-
mediate, and its hydrolysis as the last step in the path-
way. Association of esterase activity in S. venezuelae

2
86
F. Groß et al. / Bioorg. Med. Chem. Lett. 12 (2002) 283–286
1H), 4.09–4.03 (m, 1H), 3.74–3.60 (m, 2H), 3.20 (dd, J=6 and
References and Notes
1
3
5
Hz, 1H), C NMR d 164.9, 150.9, 148.4, 128.2, 124.2, 71.8,
1
1
2
. Cundliffe, E. Annu. Rev. Microbiol. 1989, 43, 207.
. Shaw, K. J.; Rather, P. N.; Hare, R. S.; Miller, G. H.
67.6, 62.7, 57.9. 4: H NMR d 8.18 (apparent d, splitting of 9
Hz, 2H), 7.58 (apparent d, splitting of 9 Hz, 2H), 7.27 (br d,
J=9 Hz, 1H), 6.06 (s, 1H), 5.99 (d, J=4 Hz, 1H), 4.56–4.50
Microbiol. Rev. 1993, 57, 138.
. (a) Shaw, W. V.; Leslie, A. G. W. Annu. Rev. Biophys.
1
3
3
(m, 1H), 4.21–4.05 (m, 2H), 2.14 (s, 3H), 2.00 (s, 3H),
C
Biophys. Chem. 1991, 20, 363. (b) Murray, I. A.; Shaw, W. V.
Antimicrob. Agents Chemother. 1997, 41, 1.
NMR d 171.4, 170.7, 165.2, 148.5, 145.8, 128.7, 124.6, 73.9,
67.4, 63.4, 53.6, 21.1, 21.0.
4. Vining, L. C.; Stuttard, C. In Genetics and Biochemistry of
Antibiotic Production; Vining, L. C., Stuttard, C., Eds.; But-
terworth-Heinemann: Boston, 1995; p 505.
13. Argoudelis, A. D.; Coats, J. H. J. Antibiot. 1971, 24, 206.
14. Hazra, B. G.; Pore, V. S.; Maybhate, S. P.; Natekar,
M. V.; Rao, A. S. Synth. Commun. 1989, 19, 1763.
15. Munro, M. H. G.; Taniguchi, M.; Rinehart, K. L., Jr.;
Gottlieb, D. Tetrahedron Lett. 1975, 2659.
5. Malik, V. S.; Vining, L. C. Can. J. Microbiol. 1970, 16, 173.
. Doull, J.; Ahmed, Z.; Stuttard, C.; Vining, L. C. J. Gen.
6
Microbiol. 1985, 131, 97.
. Mosher, R. H.; Camp, D. J.; Yang, K.; Brown, M. P.;
Shaw, W. V.; Vining, L. C. J. Biol. Chem. 1995, 270, 27000.
16. Simonsen, J. N.; Paramasigamani, K.; Vining, L. C.;
McInnes, A. G.; Walter, J. A.; Wright, J. L. C. Can. J.
Microbiol. 1978, 24, 136.
7
8
9
. Kleanthous, C.; Shaw, W. V. Biochem. J. 1984, 223, 211.
. Murray, I. A.; Derrick, J. P.; White, A. J.; Drabble, K.;
17. Murray, I. A.; Lewendon, A.; Williams, J. A.; Cullis, P. M.;
Shaw, W. V.; Leslie, A. G. W. Biochemistry 1991, 30, 3763.
1
13
Wharton, C. W.; Shaw, W. V. Biochemistry 1994, 33, 9826.
0. Procedure for the acetylation of chloramphenicol: Acetic
anhydride (0.047 mL, 0.5 mmol) was added dropwise via a
18.
H
(400.1 MHz) and
C (100.6 MHz) NMR data
1
1
(CD OD). 2 (minor species): H NMR d 8.20 (apparent d),
7.60 (apparent d, partially obscured), 6.24 (s), 6.10 (d), H-2
obscured, 3.45–3.65 (m), 2.14 (s). Only one enriched carbon of
2 at either 67.4 and 20.7 ppm was observed in the C NMR
spectra of O-acetylchoramphenicol isolated in the feeding
experiments. 3 (major species): H NMR d 8.17 (apparent d,
splitting of 9 Hz, 2H), 7.64 (apparent d, splitting of 8 Hz, 2H),
3
0
syringe to a stirred solution of 1 (162 mg, 0.5 mmol) and
4
idine (1 mL, 12.5 mmol) at 0 C. After stirring for 20 min at
1
3
-(dimethylamino)pyridine (DMAP, 10 mg, 15% mol) in pyr-
ꢀ
ꢀ
1
0
C, ice-cold water (3 mL) was added, and the mixture was
extracted with ethyl acetate (3Â20 mL). The combined
extracts were dried over anhydrous MgSO and concentrated
4
6.19 (s, 1H), 5.06 (d, J=3 Hz, 1H), 4.40–4.35 (m, 2H), 4.23–
1
3
in vacuo to a brown oil from which 1 (17 mg), 2 (13 mg), 3 (34
mg) and 4 (17 mg) were isolated by preparative TLC. The
relative yields of the acetylated chloramphenicol derivatives
were sensitive to the acetylation reaction conditions. When
4.21 (m, 1H), 2.02 (s, 3H), C NMR d 172.5, 166.7, 150.8,
128.4, 124.2, 71.7, 67.3, 64.8, 55.7, 20.6.
19. White, R. L.; Ramezani, M.; Gharbia, S. E.; Seth, R.;
Doherty-Kirby, A. L.; Shah, H. Biotechnol. Appl. Biochem.
1995, 22, 385.
20. Gottlieb, D.; Carter, H. E.; Robbins, P. W.; Burg, R. W.
J. Bacteriol. 1962, 84, 888.
21. Shaw, W. V. Methods Enzymol. 1975, 43, 737.
22. Nakano, H.; Matsuhashi, Y.; Takeuchi, T.; Umezawa, H.
J. Antibiot. 1977, 30, 76.
23. (a) Riefler, J. F., III; Higerd, T. B. Biochim. Biophys. Acta
1976, 429, 191. (b) Dulac, R. W.; Yang, T. J. Exp. Hematol.
1991, 19, 59.
DMAP was omitted from the reaction mixture, equal amounts
ꢀ
3
and 4 were obtained at 0 or À20 C, whereas excess acetic
anhydride (0.75 mmol) produced 4 as the major product.
1. Nakagawa, Y.; Nitahara, Y.; Miyamura, S. Antimicrob.
Agents Chemother. 1979, 16, 719.
1
1
13
1
(
2
2.
CD
H), 7.60 (apparent d, splitting of 9 Hz, 2H), 7.04 (br d, J=8
Hz, 1H), 6.05 (s, 1H), 5.15 (t, J=3 Hz, 1H), 4.11 (d, J=4 Hz,
H
(400.1 MHz) and
C
(100.6 MHz) NMR data
1
3
CN). 1: H NMR d 8.16 (apparent d, splitting of 9 Hz,