Organic Letters
Letter
Next step was to implement the feasibility of reduction by
KRED in this one-pot protocol. Carbonyl reductase from
Kluyveromyces thermotolerans (KtCR) was first chosen as a
model enzyme for the optimization of the cascade reaction,
which was identified with satisfactory (R)-selectivity for
acetophenone.12 The initial study was performed by adding
KtCR and other cofactors (including NADP+) in one pot after
decarboxylative fluorination. Glucose dehydrogenase (GDH)
was also added to recycle the cofactor. To our delight, the
cascade system in one-pot was proceeded successfully with
high total conversion (92%) and excellent (S)-stereoselectivity
(ee = 99%, the R/S assignment was changed in accordance
with the Cahn−Ingold−Prelog priority rules).
Scheme 3. Substrate Scope of the Cascade Process for the
Production of (S)-Vicinal Fluoro Alcohols
a
However, the high costs of the cofactors motivated us to
seek a more economical way to carry out this transformation.
The whole-cell reaction system was chosen to overcome this
limitation by regenerating the NADPH intracellular.13 Thus,
we performed this whole-cell transformation (Figure 2) by
a
Reaction conditions: 1 mmol of β-ketonic ester (1) and Selectfluor
(30 mM, 1.2 equiv) were dissolved in 40 mL of buffer (200 mM PBS,
pH = 7.4) containing 4 mL of Candida antarctica lipase B (CAL-B,
5000 U/mL). The reaction mixture was shaken vigorously at 50 °C
for 12 h. Then 400 mL of whole-cell culture of KtCR (resuspended in
50 mM sodium phosphate buffer, pH 6.5) with glucose (100 mmol)
was added into the above mixture and shaken at 30 °C overnight. All
yields are isolated yields of products. The ee values were determined
by chiral GC.
Figure 2. Whole-cell reaction system to produce chiral vicinal fluoro
alcohols.
adding 10 mL of the culture medium of KtCR (OD600 ≈ 6)
into the above-mentioned mixture of 1a. As expected, 2a was
able to permeate the cell membrane, and the whole-cell one-
pot reaction successfully provided the product (R)-3a in
excellent total conversion (92%) and selectivity (99%).
Based on the successful one-pot process of 1a, a range of β-
ketonic esters were subjected to the hydrolysis, fluorination
and reduction step by step to yield chiral vicinal fluoro
alcohols. To assess the scalability of these cascade process, we
performed 200 mg scale reactions for all tested substrates (1a−
1l), and isolated yields were obtained for comparison (Scheme
3, Table S1). It can be found that both electron-donating and
electron-withdrawing group substituted substrates afforded the
corresponding products in acceptable isolated yields (20−
92%) and high stereoselectivity (99%). Among them, 3i gave
the best result with 92% isolated yields and 99% ee, while 3e
with the lowest yield. Actually, for all tested substrates, the
conversions of the second step (enzymatic reduction) were
first step of the lipase-catalyzed hydrolysis and fluorination
were completely different. For example, the conversion of 1e in
the lipase-catalyzed hydrolysis and fluorination was only 23%,
resulting in the lowest isolated yield (20%) for 3e.
Interestingly, the bulky β-ketonic esters (1k) and thienyl-
substituted substrate (1l) were also tolerated with this method
to give the (S)-isomer of 3k and 3l with excellent
enantioselectivity (ee = 99%).
structures on the CAL-B/YtBE-catalyzed cascade process was
also similar to that of the CAL-B/KtCR system. Compound 1i
provided a better result than other substrates, and the isolated
yield of 3e was still the lowest because of the poor conversion
in the first step of lipase-catalyzed hydrolysis and fluorination.
Notably, YtbE showed the same selectivity preference for 3f as
that of KtCR, while they were opposite in the cases of other
products, implying the distinct binding or recognition
mechanism for 3f. For thienyl-substituted substrate (1l),
YtbE showed good activity, but the selectivity is very poor.
These results shown in both Scheme 3 and 4 documented the
synthetic utility of double enzyme-catalyzed one-pot process
for the preparation of enantiocomplementary vicinal fluoro
alcohols.
In summary, we have developed a one-pot, three-step
cascade process combining enzymatic hydrolysis, decarbox-
ylative fluorination, and KRED-catalyzed reduction to achieve
enantiocomplementary vicinal fluoro alcohols. A series of β-
ketonic esters can be successfully converted into the
corresponding chiral vicinal fluoro alcohols with either (R)-
or (S)-configuration in accordance with the selectivity of the
KRED used. This one-pot cascade process was performed in
the aqueous phase under moderate conditions without the
separation of intermediates. A whole-cell system was
implemented to regenerate NADPH in situ to make the
transformation proceed in an economical way. All these
advantages should enable chemoenzymatic transformations
that display great potential in green synthetic chemistry and
sustainable development. However, in the current scale-up
reactions of double enzyme-catalyzed one-pot process, the
The (S)-selective KRED could be replaced by an (R)-
selective KRED; thus, chiral vicinal fluoro alcohols with tailor-
made steteoselectivity were achieved. By using the NADPH-
dependent aldo-keto reductase from Bacillus sp. ECU0013
(YtbE)14 as the KRED, most of the tested aromatic carboxylic
esters could also be converted into corresponding (R)-alcohols
with acceptable isolated yields and high ee values through the
above one-pot, three-step process. The influence of substrate
C
Org. Lett. XXXX, XXX, XXX−XXX