Precursor-directed biosynthesis of 12-ethyl erythromycin

Article abstract of DOI:10.1016/S0968-0896(98)00081-9

A precursor-directed method for the biosynthesis of novel 6-deoxyerythronolide B derivatives has been extended to allow alteration of the functionality at C-12. We recently described a simple and practical method for harnessing the biosynthetic potential

Full text of DOI:10.1016/S0968-0896(98)00081-9

BIOORGANIC &
MEDICINAL
CHEMISTRY
Bioorganic & Medicinal Chemistry 6 (1998) 1171±1177
Precursor-Directed Biosynthesis of 12-Ethyl Erythromycin
John R. Jacobsen,a Adrian T. Keatinge-Clay,a David E. Caneb
and Chaitan Khoslaa, c, d,
*
aDepartment of Chemical Engineering, Stanford University, Stanford, CA 94305-5025, USA
bDepartment of Chemistry, Box H, Brown University, Providence RI 02912, USA
cDepartment of Chemistry, Stanford University, Stanford, CA 94305-5025, USA
dDepartment of Biochemistry, Stanford University, Stanford, CA 94305-5025, USA
Received 5 February 1998; accepted 13 April 1998
AbstractÐA precursor-directed method for the biosynthesis of novel 6-deoxyerythronolide B derivatives has been
extended to allow alteration of the functionality at C-12. We recently described a simple and practical method for
harnessing the biosynthetic potential of the erythromycin pathway to generate novel molecules of natural product-like
complexity by feeding designed synthetic molecules to an engineered mutant strain having an altered 6-deoxy-
erythronolide B synthase (DEBS). Our initial applications of this technique focused on alteration of the ethyl side chain of
6-dEB (C14-C15). We now report the extension of this approach to modi®cation of the C-12 substituent. An appro-
priately designed substrate is shown to incorporate into 6-dEB biosynthesis, yielding a 6-dEB analogue bearing a 12-
ethyl group. This 6-dEB analogue is a substrate for post-polyketide tailoring enzymes, and is converted into the cor-
responding analogue of erythromycin C. These results show that many of the downstream active sites are tolerant of
this unnatural functionality and suggest that a wide variety of erythromycin derivatives should be accessible by this
approach or by total biosynthesis via genetic engineering. # 1998 Elsevier Science Ltd. All rights reserved.
Introduction
Modular polyketide synthases such as 6-deoxyerythro-
Polyketides comprise a large and diverse class of natural
products, many of which possess important biological
and medicinal properties.1 A tremendous diversity of
structures, many of which are highly complex, are pro-
duced via a common biosynthetic pathway which clo-
sely parallels the biosynthesis of fatty acids.2 Like fatty
acid biosynthesis, polyketide biosynthesis proceeds by
repetitive condensation of simple carboxylic acid
monomers. Structural complexity is introduced by var-
iation in the degree of reduction after each condensa-
tion, by the generation of stereocenters, and by
variation of monomers. Additional complexity is intro-
duced by `post-polyketide' enzymes which carry out
oxidations, alkylations, glycosylations and other trans-
formations. The abundance of e€ective, polyketide-
derived pharmaceuticals makes this class of compounds
an attractive target for drug discovery.1
nolide B synthase (DEBS) utilize a separate set of
active sites for each condensation step. The biosynthesis
of 6-deoxyerythronolide B (6dEB) involves six con-
densation steps, and consequently DEBS is organized
into six groups of active sites, called `modules', each of
which is responsible for one cycle of extension and pro-
cessing (Fig. 1(A)).3 Each module possesses domains
analogous to the ketosynthase (KS), acyl transferase
(AT), and acyl carrier protein (ACP) enzymes of fatty
acid biosynthesis. Other domains analogous to keto-
reductase (KR), dehydratase (DH), enoyl reductase
(ER), and thioesterase (TE) enzymes are present in var-
ious combinations, giving rise to the complex ®nal
structure of 6-dEB (1). The six modules of DEBS are
organized into three large polypeptides (each
>300 kDa) with two modules contained in each of the
DEBS1, DEBS2, and DEBS3 proteins.
The modular structure of these enzymes has made them
attractive targets for genetic engineering despite a lack
of protein structural information. The feasibility of
Key words: Polyketide synthase; erythromycin; diketide; muta-
tional biosynthesis.
*Corresponding author. Fax: 650 723 9780.
0968-0896/98/$Ðsee front matter # 1998 Elsevier Science Ltd. All rights reserved
PII: S0968-0896(98)00081-9

1172
J. R. Jacobsen et al./Bioorg. Med. Chem. 6 (1998) 1171±1177
Figure 1. (A) Genetic model of DEBS. DEBS consists of three large polypeptides, each of which contains two `modules'. Each module
catalyzes the addition of a single methylmalonyl-C0A extender unit, along with any reductive steps. Each module contains a variety of
domains including a ketosynase (KS), acyltransferase (AT), and acyl carrier protein (ACP). All modules except Module 3 have active
ketoreductase (KR) domains, and Module 4 contains dehydratase (DH) and enoyl reductase (ER) domains. Module 6 bears a
thioesterase (TE) domain which catalyzes the cyclization of the heptaketide intermediate to form 6-dEB (1). Proposed enzyme-bound
intermediates are shown for each module. (B) KS10 mutant of DEBS (pJRJ2). Mutation of an active site cysteine residue in the
Module 1 ketosynthase domain (KS1) inactivities this domain. No 6-dEB-like products are observed in fermentations of strains
expressing this DEBS mutant.
generating new polyketides has been demonstrated
via loss-of-function mutagenesis within reductive
domains,3±5 replacement of acyltransferase domains to
alter starter or extender unit speci®city,6±9 replacement
of a reductase domain to alter product stereochemistry,10
and gain-of-function mutagenesis to introduce novel
catalytic activities within modules.11,12 However, these
approaches are time consuming and limited to metabo-
lically available precursors.
precursor-directed biosynthesis of a C12 derivative of
6-dEB in which the methyl side chain found on C12
of 6-dEB has been substituted with an ethyl side
chain to give 12-desmethyl-12-ethyl-6-deoxyerythronolide
B (4).
Results
We recently reported the development of a generally
applicable, fermentation-based strategy in which chemi-
cally synthesized, cell-permeable, non-natural pre-
cursors are transformed into novel, natural product-like
molecules by a genetically engineered PKS.13 DEBS
KS10 (Fig. 1(B)) carries an inactivating mutation in
the KS domain of module 1. Because it is unable to
synthesize the ®rst (diketide) intermediate, no 6-dEB (1)
is produced. However, when provided with a synthetic
analogue of this intermediate (2), DEBS KS10 eciently
converts the diketide substrate to 6-dEB.
Construction of the strain CH999/pJRJ2, which expres-
ses the DEBS KS10 mutant, has been described pre-
viously.13 When grown on R2YE medium, CH999/
pJRJ2 produces no detectable 6-dEB-like products.
However, when substrate 2 was added to growing
cultures of this strain, 6-dEB was eciently produced
(Fig. 2). Our previous studies with unnatural sub-
strates have focused primarily on alterations to the ethyl
side chain of 6-dEB (C14-C15). In order to test the
generality of this method we prepared substrate 3 which
contains a C2-ethyl group. We expected that this sub-
strate would be taken up by module 2 of the DEBS KS10
mutant and converted into 12-desmethyl-12-ethyl-6-dEB.
In addition to converting diketide 2 to 6-dEB, DEBS
KS10 was shown to utilize `unnatural' substrates and
convert these to corresponding analogues of 6-dEB.
Initial studies focused primarily on introduction of
novel functionality at C15 of 6-dEB. We now report the
Substrate 3 (Scheme 1) was prepared by methods ana-
logous to those described previously for the synthesis of
substrate 2.13±15 The Evans chiral oxazolidinone was

J. R. Jacobsen et al./Bioorg. Med. Chem. 6 (1998) 1171±1177
1173
Scheme 1. Synthesis of 3. Reagents for transformations: (a) Bu2BOTf, EtN(iPr)2, EtCHO; (b) TBDMSOTf, EtN(iPr)2; (c) LiOH,
H2O2; (d) (PhO)2P(O)N3, Et3N, N-acetylcysteamine; (e) HF.
acylated with butyryl chloride to provide compound 5.
The asymmetric aldol condensation of 5 with propion-
aldehyde a€orded the desired diketide as the oxazo-
lidinone imide 6. Imide 6 was protected with tert-
butyldimethylsilyl tri¯ate to give imide 7, and the chiral
auxiliary was removed with lithium hydroperoxide to
give carboxylic acid 8. The free acid was coupled to N-
acetylcysteamine (NAC) to a€ord thioester 9. NAC
thioesters have been used as acyl donors in a variety of
studies of fatty acid and polyketide synthases. Finally,
the silyl protecting group was removed with hydro-
¯uoric acid to give substrate 3. Substrate 3 was admi-
nistered to growing cultures of CH999/pJRJ2 on R2YE
agar plates. Polyketide products were obtained by
extraction of the agar medium with ethyl acetate. Silica
gel chromatography gave 5 mg of 12-desmethyl-12-
ethyl-6-dEB (4) from 2 L of agar plates. The product
structure was identi®ed by 1H NMR, 13C NMR and
1H±1H COSY NMR. All spectra are very similar to
those of 6-dEB (1). Notable di€erences include the
appearance of a new methyl triplet (0.88 ppm) and a
shift of the signal for the proton attached to C11. The
high-resolution mass spectrum con®rms a molecular
formula of C22H40O6, which di€ers from 6-dEB by a
single methylene unit.
The successful processing of these unnatural inter-
mediates by the `downstream' modules of DEBS led us
to investigate whether the post-PKS enzymes in the
erythromycin biosynthetic pathway might also process
Figure 3. Processing of 6-dEB and 4 to erythromycins. 6-dEB
is converted to erythromycin D (10) by oxidation at C6 and
glycosylation at C3 and C5. Hydoxylation of C12 a€ords ery-
thromycin C (12), and subsequent methylation of the mycarose,
sugar a€ords erythromycin A (13). Methylation of mycarose
without C12 oxidation gives erythromycin B (11). S. erythrea
A34 (which expresses all postpolyketide processing enzymes)
converts the 12-ethyl derivative of 6-dEB (4) to its ery-
thromycin C analog (14).
Figure 2. Conversion of diketide substrates by DEBS KS10.
Substrate 2 is utilized by DEBS KS10 and eciently converted
to 6-dEB. Substrate 3 is converted to 4, an analog of 6-dEB in
which the methyl group attached to C12 has been replaced by
an ethyl group.

1174
J. R. Jacobsen et al./Bioorg. Med. Chem. 6 (1998) 1171±1177
6-dEB analogue 4. In the natural producer organism,
Saccharopolyspora erythrea, 6-dEB undergoes several
enzyme-catalyzed transformations. Oxidation at C6 and
glycosylations at C3 and C5 a€ord erythromycin D (10)
and subsequent transformations a€ord erythromycins B
(11), C (12), and A (13) (Fig. 3). In order to test the
ability of the modi®cation enzymes to act on 4, we uti-
lized an S. erythrea mutant (A34)16 which is unable to
synthesize 6-dEB. This strain produces no erythromycin
when grown on R2YE plates (as judged by the ability of
extracts to inhibit growth of the erythromycin-suscep-
tible bacterium Bacillus cereus). However, when 6-dEB
(which has no antibacterial activity) is added to the cul-
ture medium, extracts exhibited potent antibacterial
activity. A sample of 6-dEB derivative 4 was assayed for
conversion by this strain. A crude extract demonstrated
inhibition of B. cereus growth, and mass spectrometry
was used to identify the major component of the extract
as 14 (Fig. 3), an analogue of erythromycin C.
used to introduce `handles' for the synthetic modi®cation
of products. A similar approach was recently reported in
which a biosynthetic pathway was used to incorporate a
reactive ketone group into a cell surface glycoside, allow-
ing speci®c modi®cations to the surfaces of intact cells.23
It is somewhat surprising that the post-polyketide
enzymes of S. erythrea convert 4 to an erythromycin C
analogue. Earlier studies with unnatural 6-dEB deriva-
tives showed that the derivatives examined were con-
verted to erythromycin D analogues.13 Despite the fact
that it di€ers from 6-dEB at C12, the 12-ethyl derivative
is apparently a substrate for the hydroxylase which acts
on C12 (ery K gene product). The erythromycin C ana-
logue (14) is, however, apparently not a substrate for the
methylase (ery G gene product) which converts erythro-
mycin C to erythromycin A.
In addition to being a source of potentially useful
molecules, chemo-biosynthesis can also provide valu-
able information to guide genetic engineering. Ethyl side
chains, derived from incorporation of ethylmalonyl-
CoA extender units, are observed in some macrolides
such as tylosin. Although DEBS normally utilizes only
methylmalonyl-CoA extender units, the successful
incorporation of substrate 3 demonstrates tolerance of
intermediates bearing ethyl side chains by DEBS
(modules 2±6) as well as other enzymes involved in
erythromycin biosynthesis. This suggests that substitu-
tion of the module 1 AT domain for a heterologous
ethylmalonyl transferase domain should allow for pro-
duction of 4 by total biosynthesis. Importantly, the use
of substrate 3 as a chemical probe allowed this determi-
nation to be made without additional genetic manip-
ulations and without addressing the availability of
ethylmalonyl-CoA. Ethylmalonyl-CoA is an unusual
metabolite which requires a dedicated pathway for its
biosynthesis and is absent from many bacteria including
Saccharopolyspora coelicolor. The facile, precursor-
directed synthesis of derivatives such as 4 and 14 should
be useful in assessing which engineered mutants of
DEBS or other polyketide synthases would be most
likely to yield interesting products.
Discussion
The importance of erythromycin as an antibiotic makes
it an attractive target for medicinal chemistry, however
the challenges involved in total synthesis of macrolides17
make this approach impractical for the preparation of
derivatives. Recent advances have demonstrated that
new derivatives of erythromycin and other macrolides
are likely to yield valuable new pharmaceuticals.18±20
While the synthetic modi®cation of natural products is a
proven method for generating derivatives, engineered
biosynthesis o€ers a complementary approach which
allows the exploration of di€erent types of modi®ca-
tions which might be dicult or impossible to achieve
by semisynthesis.
Precursor-directed biosynthesis combines the power of
synthetic chemistry to generate a wide variety of sub-
structures with the power of biosynthesis to selectively
and rapidly incorporate these substructures into a com-
plex sca€old of proven pharmacological importance. In
addition to our approach which utilizes a KS1 knockout
mutation and synthetic diketide substrates, another
method of precursor-directed biosynthesis was recently
reported which utilizes an engineered PKS with relaxed
starter-unit speci®city.21 Although this engineered PKS
can eciently incorporate simple carboxylic acids into
erythromycin derivatives, this approach is limited to
alteration of the ethyl side chain of 6-dEB (C14-C15).
Experimental
General methods
Reagents and chemicals were obtained from commercial
suppliers and used without further puri®cation.
The ability to generate new derivatives at C12 of
erythromycin may provide a useful route to the investi-
gation of new pharmaceuticals since this region of
the molecule of particular medicinal interest.19,22 Sub-
strates which bear reactive functional groups might be
(4S)-4-benzyl-3-butyryl-2-oxazolidinone (5). A solution
of (4S)-4-benzyl-2-oxazolidinone (1.3 g, 7.1 mmol) in
25 mL freshly distilled THF was cooled to 78 ꢀC.
n-Butyllithium (4.5 mL of a 1.6 M solution in hexanes)
was added dropwise by syringe. After 5 min, butyryl

J. R. Jacobsen et al./Bioorg. Med. Chem. 6 (1998) 1171±1177
1175
chloride (0.83 mL, 8.0 mmol) was added dropwise by
syringe. The solution was stirred for 30 min at 78 ꢀC,
then warmed slowly to room temperature. After an
additional 30 min, the reaction was quenched by addi-
tion of 4.5 mL saturated aqueous ammonium chloride.
The mixture was concentrated in vacuo, and the residue
was extracted (2Â50 mL CH2Cl2). The combined
extracts were washed (50 mL saturated aqueous
NaHCO3 followed by 50 mL brine), dried (MgSO4) and
concentrated in vacuo to a€ord the product (1.6 g, 89%)
as a colorless oil. Rf 0.37 (20% ethyl acetate in hexanes).
1H NMR (400 MHz, CDCl3) d 1.01 (t, 3H, J=7.4 Hz,
C(40)-H3), 1.67±1.78 (m, 2H, C(30)-H2), 2.77 (dd, 1H,
J=9.5 Hz, 13.4 Hz, one of Ph-CH2), 2.83±3.01 (m, 2H,
C(20)-H2), 3.30 (dd, 1H, J=3.2 Hz, 13.3 Hz, one of
Ph-CH2), 4.14±4.22 (m, 2H, O-CH2), 4.61±4.71 (m, 1H,
N-CH), 7.20‹7.36 (m, 5H, Ph-H5).
added.25 After 2 h, 150 mL of water was added, and the
aqueous layer was extracted (3Â100 mL CH2Cl2).
The organic layers were combined, dried (MgSO4),
and concentrated in vacuo. Flash chromatography
(6Â10 cm silica gel, 40% ethyl acetate in hexanes)
a€orded product 7 as a colorless oil (7.9 g, 96%). Rf 0.86
(50% ethyl acetate in hexanes). 1H NMR (400 MHz,
CDCl3) d 0.02 (s, 3H, Si-CH3), 0.03 (s, 3H, Si-CH3),
0.84±0.96 (m, 15H, Si-C(CH3)3, C(20)CH2-CH3, C(50)-
H3), 1.50±1.90 (m, 4H, C(40)-H2, C(20)-CH2), 2.72 (dd,
1H, J=10.1 Hz, 13.2 Hz, one of Ph-CH2), 3.36 (dd, 1H,
J=3.1 Hz, 13.2 Hz, one of Ph-CH2), 3.94±3.99 (m, 2H,
C(20)-H, C(30)-H), 4.09±4.18 (m, 2H, O-CH2), 4.62±4.68
(m, 1H, N-CH), 7.21±7.36 (m, 5H, Ph-H5).
(2S,3R)-2-Ethyl-3-(tert-butyldimethylsiloxy)-pentanoic
acid (8). Oxazolidinone 7 (7.9 g, 19 mmol) was dissolved
in 145 mL THF. Water (50 mL) was added and the
solution was cooled to 0 ꢀC with vigorous stirring.
Hydrogen peroxide (18 mL of a 30% aqueous solution)
was added followed by lithium hydroxide (1.6 g,
38 mmol).26 The reaction was allowed to warm slowly to
room temperature. After 24 h, the reaction was quen-
ched by addition of sodium sul®te (115 mL of a 20%
aqueous solution). After stirring for 10 min, the reaction
was concentrated in vacuo to remove THF and then
acidi®ed to pH 2 by slow addition of 1 N HCl. The
resulting cloudy mixture was extracted (3Â30 mL
CH2Cl2). The organic layers were combined, dried
(MgSO4) and concentrated in vacuo. The residue was
puri®ed by chromatography (6Â15 cm silica gel, 15%
ethyl acetate in hexanes) to a€ord the desired product
(2.25 g, 46%). Rf 0.73 (50% ethyl acetate in hexanes).
1H NMR (400 MHz, CDCl3) d 0.10 (s, 6H, Si-(CH3)2),
0.84±0.96 (m, 15H, Si-C(CH3)3, C(2)CH2-CH3, C(5)-H3),
1.45±1.81 (m, 4H, C(2)-CH2, C(4)-H2), 2.40±2.45 (m, 1H,
C(2)-H), 3.84±3.88 (m, 1H, C(3)-H).
(4S,20S,30R)-3-(20-Ethyl-30-hydroxypentanoyl)-4-benzyl-
2-oxazolidinone (6). To 120 mL of dry CH2Cl2 was
added 8.2g (33 mmol) of (4S)-3-butyryl-4-benzyl-2-oxazo-
lidinone, and the solution was cooled to 0 ꢀC.
Bu2BOTf (39 mL of a 1.0 M solution in CH2Cl2) was
added by syringe, followed by dropwise addition of di-
isopropylethylamine (7.7 mL, 44 mmol).24 The reaction
was cooled to 78 ꢀC and freshly distilled propionalde-
hyde (2.7 mL, 37 mmol) was added. The reaction was
stirred at 78 ꢀC for 30 min then warmed to 0 ꢀC and
stirred for an additional 2 h. A mixture of 40 mL phos-
phate bu€er (1.0 M sodium phosphate, pH 7.4) and
100 mL methanol was added to the reaction followed by
dropwise addition of 100 mL 2:1 methanol:30% H2O2.
After 1.5 h, organic solvents were evaporated in vacuo
and the residual material resuspended in 100 mL of 5%
aqueous NaHCO3 and extracted (3Â100 mL CH2Cl2).
The combined extracts were washed (100 mL 5% aqu-
eous NaHCO3 followed by 100 mL brine), dried
(MgSO4) and concentrated in vacuo. Flash chromato-
graphy (6Â15 cm silica gel, gradient of 15±30% ethyl
acetate in hexanes) a€orded 6.0 g (59%) of the desired
product as a colorless oil. Rf 0.49 (50% ethyl acetate in
hexanes). 1H NMR (400 MHz, CDCl3) d 0.99 (t, 6H,
J=7.4 Hz, C(50)-H3, C(20)CH2-CH3), 1.51±1.58 (m, 2H,
C(20)-CH2), 1.70±1.77 (m, 2H, one of C(40)-CH2), 1.86±
1.92 (m, 2H, one of C(40)-CH2), 2.48 (br d, 1H, OH),
2.71 (dd, 1H, J=10.3 Hz, 13.2 Hz, one of Ph-CH2), 3.37
(dd, 1H, J=3.3 Hz, 13.2 Hz, one of Ph-CH2), 3.78±3.83
(m, 1H, C(20)-H), 3.99±4.04 (m, 1H, C(30)-H), 4.14±4.22
(m, 2H, O-CH2), 4.71±4.76 (m, 1H, N-CH), 7.21±7.36
(m, 5H, Ph-H5).
(2S,3R)-2-Ethyl-3-hydroxypentanoyl-N-acetylcysteamine
thioester (3). In 100 mL of dry dimethylformamide was
dissolved 2.1 g (8.1 mmol) of carboxylic acid 8. The
resulting solution was cooled to 0 ꢀC, and diphenylphos-
phoryl azide (5.2 mL, 24 mmol) and triethylamine
(4.5 mL, 32 mmol) were added.14 After 2 h, N-acetyl-
cysteamine (2.6 mL, 24 mmol) was added dropwise.
The reaction was allowed to warm slowly to room tem-
perature. After 24 h, 200 mL of water was added, and
the resulting mixture was extracted with ether
(3Â100 mL). The combined organics were dried
(MgSO4) and concentrated in vacuo. The residue was
partially puri®ed by chromatography (6Â15 cm silica
gel, 50% ethyl acetate in hexanes) to give intermediate
9. The protected thioester 9 (1.8 g, 5 mmol) was dis-
solved in acetonitrile (96 mL) and water (19 mL), and
hydro¯uoric acid (14 mL, 280 mmol, 48% aqueous)
was added. After 2 h, saturated aqueous NaHCO3
(4S,20S,30R)-(20-Ethyl-30-tert-butyldimethylsiloxypentan-
oyl)-4-benzyl-2-oxazolidinone (7). To 6.0 g (20 mmol) of
oxazolidinone 6 in 100 mL CH2Cl2 at 0 ꢀC, diisopropyl-
ethylamine (5.8 mL, 33 mmol) and tert-butyldimethyl-
silyl tri¯uoromethanesulfonate (6.8 mL, 30 mmol) were

1176
J. R. Jacobsen et al./Bioorg. Med. Chem. 6 (1998) 1171±1177
(approximately 300 mL) was added until the pH reached
7.5. The acetonitrile was removed in vacuo, and the
aqueous residue extracted (5Â75 mL CH2Cl2). The organic
layers were combined and dried (MgSO4). The residue
was chromatographed (3Â15 cm silica gel, 80% ethyl
acetate in hexanes) to give the desired product (540 mg,
29%, two steps). Rf 0.16 (70% ethyl acetate in hexanes).
1H NMR (400 MHz, CDCl3) d 0.95 (t, 3H, J=7.4 Hz,
C(2)CH2-CH3), 0.98 (t, 3H, J=7.3 Hz, C(5)-H3), 1.46±
1.54 (m, 2H, C(4)-H2), 1.65±1.84 (m, 2H, C(2)-CH2),
1.97 (s, 3H, NCO-CH3), 2.28 (br s, 1H, OH), 2.58±2.64
(m, 1H, C(2)-H), 3.00±3.12 (m, 2H, S-CH2), 3.39±3.52
(m, 2H, N-CH2), 3.69±3.75 (m, 1H, C(3)-H), 5.88 (br s,
1H, NH). 13C NMR (100 MHz, CDCl3) d 10.4, 12.1,
20.5, 23.2, 27.4, 28.6, 39.6, 61.4, 73.8, 170.5, 203.5.
threa A34 was then applied so as to give lawns. After 8
days of growth, the media were homogenized and
extracted three times with 98.5:1.5 ethyl acetate:-
triethylamine. Extracts were dried (MgSO4) and con-
centrated. The crude extracts were examined by TLC
and mass spectrometry. Mass analysis suggests conver-
sion of 4 to the corresponding erythromycin C analogue
(14): Rf 0.41 (streak) (80:20:0.1, CHCl3:MeOH:NH4OH).
MS (APCI+): 733 (M+).
Qualitative assay for antibacterial activity. Crude
extracts from 80 mL cultures of S. erythrea A34 fed with
either 2 mg 6-dEB, 2 mg 12-desmethyl-12-ethyl-6-dEB,
or no polyketide were each dissolved in 1.4 mL ethanol.
Filter discs were soaked in these ethanolic solutions,
dried and laid over freshly-plated lawns of Bacillus cer-
eus.27 After incubation for 12 h at 37 ꢀC, inhibition of
bacterial growth was evident for both erythromycin
derivatives but not for the control extract.
Production and puri®cation of polyketides
The construction of CH999/pJRJ2, which expresses a
KS1 mutant of DEBS has been described previously.13
0
Lawns of S. coelicolor CH999/pJRJ2 were grown on
R2YE agar plates containing 0.3 mg/mL sodium pro-
pionate. After three days, each agar plate was overlaid
with 1.5 mL of a 20 mM substrate solution in 9:1
water:DMSO. After an additional 4 days, the agar
media (2.0 L) were homogenized and extracted three
times with ethyl acetate. The combined extracts were
dried (MgSO4) and ®ltered through silica gel (1.2Â4 cm,
eluted with 25 mL of 50% ethyl acetate in hexanes). The
extract was puri®ed by silica gel chromatography
(1.2Â14 cm silica gel, gradient of 15±25% ethyl acet-
ate in hexanes) to a€ord the product (4). Rf 0.62 (60%
ethyl acetate in hexanes). 1H NMR (500 MHz, CDCl3) d
0.88 (t, 3H, J=7.4 Hz, C12CH2-Me), 0.95 (t, 3H,
J=7.3 Hz, H15), 1.03±1.09 (m, 12H, C4-Me, C6-Me,
C8-Me, C10-Me), 1.30 (d, 3H, J=6.8 Hz, C2-Me), 1.35±
1.45 (m, 2H, H7b), 1.50±1.92 (m, 7H, H4, H7a, H12,
H12-CH2, H14a, H14b), 2.02±2.10 (m, 1H, H6), 2.63±
2.83 (m, 2H, H8, H10), 2.78 (dq, 1H, J=10.5 Hz,
6.7 Hz, H2), 3.85 (ddd, 1H, J=10.2 Hz, 4.6 Hz, 2.3 Hz,
H11), 3.94 (d, 1H, J=10.4 Hz, H3), 4.02 (d, 1H,
J=4.4 Hz, H5), 5.16 (ddd, 1H, J=9.7 Hz, 4.0 Hz,
1.6 Hz, H13); 13C NMR (100 MHz, CDCl3): d 7.1 (C4-
Me), 10.8 (C15), 11.6 (C8-Me), 12.5 (C12-Me), 13.3
(C10-Me), 14.7 (C2-Me), 16.4 (C6-Me), 27.3 (C12-CH2),
25.6 (C14), 35.5 (C6), 37.4 (C7), 37.6 (C4), 39.4 (C8),
43.8 (C2), 44.0 (C10), 46.2 (C12), 69.3 (C11), 76.5 (C13),
77.2 (C5), 79.7 (C3), 178.6 (C1) , 213.8 (C9); HRMS
(FAB+) Calcd for (C22H40O6)Cs+: 533.1879. Found
533.1898.
Acknowledgements
The authors thank H. Fu for mass spectroscopy of ery-
thromycin derivatives. Research was supported by
grants from the National Institutes of Health (CA66736
to C.K. and GM22172 to D.E.C.), an NSF Young
Investigator Award (to C.K.), and a David and Lucille
Packard Fellowship for Science and Engineering (to
C.K.). J.R.J. is a recipient of a National Institute of
General Medical Sciences postdoctoral fellowship (1
F32 GM18590-01).
References
1. Katz, L.; Donadio, S. Ann. Rev. Microbiol. 1993, 47, 875.
2. O'Hagan, D. O. The Polyketide Metabolites; Ellis Horwood:
Chichester, U.K., 1991.
3. Donadio, S.; Staver, M. J.; McAlpine, J. B.; Swanson, S. J.;
Katz, L. Science 1991, 252, 675.
4. Donadio, S.; McAlpine, J. B.; Sheldon, P. J.; Jackson, M.;
Katz, L. Proc. Natl. Acad. Sci. U.S.A. 1993, 90, 7119.
5. Bedford, D.; Jacobsen, J. R.; Luo, G.; Cane, D. E.; Khosla,
C. Chem. Biol. 1996, 3, 827.
6. Oliynyk, M.; Brown, M. J. B.; Cortes, J.; Staunton, J.; Lea-
dlay, P. F. Chem. Biol. 1996, 3, 833.
7. Kuhstoss, S.; Huber, M.; Turner, J. R.; Paschal, J. W.; Rao,
R. N. Gene 1996, 183, 231.
8. Liu, L.; Thamchaipenet, A.; Fu, H.; Betlach, M.; Ashley, G.
J. Am. Chem. Soc. 1997, 119, 10553.
9. Ruan, X.; Pereda, A.; Stassi, D. L.; Zeidner, D.; Summers,
R. G.; Jackson, M.; Shivakumar, A.; Kakavas, S.; Staver, M.
J.; Donadio, S.; Katz, L. J. Bacteriol. 1997, 179, 6416.
Post-polyketide processing of 12-desmethyl-12-ethyl-6-
dEB
10. Kao, C. M.; McPherson, M.; McDaniel, R. N.; Fu, H.;
Cane, D. E.; Khosla, C. J. Am. Chem. Soc. 1998, 120, 2478.
Puri®ed 4 (2 mg dissolved in 1 mL ethanol) was layered
onto R2YE plates (80 mL) and allowed to dry. S. ery-
11. McDaniel, R. M. J. Am. Chem. Soc. 1997, 119, 4309.

J. R. Jacobsen et al./Bioorg. Med. Chem. 6 (1998) 1171±1177
1177
12. Kao, C. M.; McPherson, M.; McDaniel, R. N.; Fu, H.;
Cane, D. E.; Khosla, C. J. Am. Chem. Soc. 1997, 119,
11339.
20. Morimoto, S.; Misawa, Y.; Adachi, T.; Nagate, T.; Wata-
nabe, Y.; Omura, S. J. Antibiot. 1990, 43, 286.
21. Marsden, A. F. A.; Wilkinson, B.; Cortes, J.; Dunster, N.
J.; Staunton, J.; Leadlay, P. F. Science 1998, 279, 199.
13. Jacobsen, J. R.; Hutchinson, C. R.; Cane, D. E.; Khosla,
C. Science 1997, 277, 367.
22. Adachi, T.; Morimoto, S.; Kondoh, H.; Nagate, T.; Wata-
nabe, Y.; Sota, K. J. Antibiot. 1988, 41, 966.
14. Cane, D. E.; Yang, C.-C. J. Am. Chem. Soc. 1987, 109, 1255.
15. Cane, D. E.; Luo, G.; Khosla, C.; Kao, C. M.; Katz, L. J.
Antibiot. 1995, 48, 647.
23. Mahal, L. K.; Yarema, K. J.; Bertozzi, C. R. Science 1997,
276, 1125.
16. Weber, J. M.; Wierman, C. K.; Hutchinson, C. R. J. Bac-
teriol. 1985, 164, 425.
24. Cane, D. E.; Lambalot, R. H.; Prabhakaran, P. C.; Ott, W.
R. J. Am. Chem. Soc. 1993, 115, 522.
17. Paterson, I.; Mansuri, M. M. Tetrahedron 1985, 41, 3569.
25. Cane, D. E.; Luo, G. J. Am. Chem. Soc. 1995, 117, 6633.
18. Baker, W. R.; Clark, J. D.; Stephens, R. L.; Kim, K. H. J.
Org. Chem. 1988, 53, 2340.
26. Evans, D. A.; Britton, T. C.; Ellman, J. A. Tetrahedron
Lett. 1987, 28, 6141.
19. Griesgraber, G.; Or, Y. S.; Chu, D. T. W.; Nilius, A. M.;
Johnson, P. M.; Flamm, R. K.; Henry, R. F.; Plattner, J. J. J.
Antibiot. 1996, 49, 465.
27. Vincent, J. G.; Vincent, H. W. Proc. Soc. Exp. Biol. Med.
1944, 55, 162.