3-phenylpropan-1-ol, or 4-phenylbutan-2-ol, respectively, with
diethylaminosulfur trifluoride (DAST), then oxidising the resultant
1
3
corresponding monofluoroalkyl benzenes. Whilst it was possible
to use this route to generate small quantities of 3-fluoropropionate
and 4-fluoropentanoate; 4-fluorobutyrate was more readily
synthesised on a multigram scale through an alternate route
1
4
starting by mono protecting 1,4-butanediol as its acetyl ester. The
unprotected hydroxyl group was then displaced by reaction with
4
DAST, the ester reduced using LiAlH , and the exposed alcohol
1
5
oxidised under Sharpless conditions (Scheme 1).
Fluorinated starter acids were fed in triplicate to 45 ml shake
cultures of both wild-type S. erythraea and ERMD1. The cultures
were supplemented with the potassium salts to a final concentra-
tion of 6 and 12 mM.§ Three control cultures of S. erythraea wild-
type and ERMD1 were grown in parallel, these were pulse fed
with deionised water. The crude extracts from the cultures were
analysed by liquid chromatography-tandem mass spectrometry
Fig. 2 LC-MS analysis of crude extract of ERMD1 strain fed with
(
LC-MS/MS)." Erythromycin A 1 was produced in the unfed
12 mM potassium 4-fluorobutyrate. (a) Total ion current; (b) Ion trace of
+
+
control cultures at a level comparable to that of the cultures fed to
a concentration of 6 and 12 mM with the potassium salt of the
alternative starter unit, suggesting that substrate toxicity was not a
problem.
[M + H] at m/z 734.2 for erythromycin A; (c) Ion trace of [M + H] at m/z
766.2 for fluoroerythromycin A. Erythromycin A and fluoroerythromycin
A co-eluted at 8.03 min. The MS/MS of the 12.38 min component reveals
that it is not fluoroerythromycin A; see ESI.{ This unidentified compound
is also present in the unfed control.
Potassium 4-fluorobutyrate, when administered at a final
concentration of 12 mM, was incorporated by the ERMD1 strain
to yield 16-fluoroerythromycin A 5 at approximately 10% of the
total erythromycins produced. The LC-MS trace exhibits peaks
corresponding to erythromycins 1, 2, 3, 3a, 4 and 4a resulting from
the incorporation of the natural starter acids (Fig. 2). The trace
shows that 16-fluoroerythromycin A 5 co-elutes with eythromycin
A 1. An as yet unidentified compound with mass of 765.2 Da and
retention time of 12.38 min was observed in the fed and control
cultures of S. erythraea wild-type and ERMD1, from the MS/MS
fragmentation patterns it is possible that this compound is an
analogue of erythromycin A bearing two additional hydroxyl
moieties in the aglycone core. No incorporation of potassium
In conclusion we have demonstrated the generation of a novel
‘
natural’ fluorinated erythromycin. Though the avermectin loading
module has been shown to be unable to accept starter acids
bearing polar functional groups such as hydroxyl moieties, we
have shown the surprising incorporation of a starter acid bearing a
highly polar carbon–fluorine bond by the modified erythromycin
PKS of ERMD1. This implies that it is not the polarity of the
functional group, but its ability to hydrogen bond that prevents
incorporation of these starter acids. It is evident that not only is the
4
-fluorobutyrate starter unit accepted but it is fully processed (no
partially processed fluoroerythromycin precursors can be observed)
by post PKS enzymes to yield a novel fluorinated erythromycin A
analogue.
4-fluorobutyrate by the wild-type strain was observed, nor were
any partially processed aglycones bearing the fluorinated moiety
observed in extracts of either S. erythraea wild-type or ERMD1.
Potassium 3-fluoropropionate, 4-fluoropentanoate and 2-fluoro-
isobutyrate were not incorporated by either S. erythraea wild-type
or ERMD1.
We thank Professors J. Staunton and P. F. Leadlay for their
helpful discussions relating to this work, their support and
encouragement and for their provision of the two bacterial strains.
The Royal Society is gratefully acknowledged for a RS BP
Dorothy Hodgkin Fellowship (R. J. M. G.).
The formula of fluoroerythromycin 5 was revealed by high-
+
resolution mass spectrometric analysis" ([M + H] : 766.4785 (obs.)
Notes and references
and 766.4753 (calc.)). The structure of this molecule was further
confirmed by comparison of its protonated and sodiated MS/MS
spectra" with the corresponding spectra of erythromycin A. The
fragmentation patterns of both the protonated and sodiated ions
are consistent with those for erythromycin A, and demonstrate the
presence of 4-fluorobutyrate as the starter unit.
{ 2-Fluoroisobutyric acid (95%) was purchased from Manchester Organics
Limited.
1
6
§
Cultures of S. erythraea WT and ERMD1 were grown in SSDM (45 ml)
and incubated in sprung flasks in an orbital shaker with a 32 mm throw at
7 uC, 190 rpm. Aqueous solutions of the potassium salts of the fluorinated
2
acids were pulse fed to resting cell cultures (48-h old cultures) at 3 6 12 h
intervals to final concentrations of 6 and 12 mM. 96 h after innoculation
the cultures were harvested by centrifugation. The supernatant was
adjusted to pH 9 with potassium hydroxide, and extracted with ethyl
4
acetate (3 6 50 ml). The organic phase was dried over MgSO and the
solvent removed under reduced pressure. The crude extract containing
erythromycins was analysed by LC-MS/MS.
"
High-resolution mass analyses were performed by electrospray ionization
on an API QSTAR pulsar (Applied Biosystems). The LC-MS/MS was
performed on a Finnigan LCQ instrument. (ThermoFinnigan, San Jose,
USA) The crude extract was eluted on a reverse phase C18 column
Scheme 1 Synthesis of potassium 4-fluorobutyrate from 1,4-butanediol.
Reagents and conditions: (i) DAST, CH
Et O, 0–20 uC, 4 h, 98%; (iii) RuCl , NaIO
RT, 24 h, 70%.
2 2
Cl , 278 uC, 3 h, 82%; (ii) LiAlH
4
,
2
3
4
, CCl –H O–MeCN (1:1:1),
2
4
(Phenomenex, 5 mm, 4.6 6 250 mm) with 20 mM ammonium acetate and
acetonitrile. The anotated spectra are included in the ESI.{
3
984 | Chem. Commun., 2005, 3983–3985
This journal is ß The Royal Society of Chemistry 2005