.
Angewandte
Communications
+
[15]
nicotinamide cofactor. Hence, the NAD(P) required in the
first oxidative step is regenerated in the subsequent step.
More specifically, cyclohexanol (1) is oxidized by an alcohol
dehydrogenase (ADH) to cyclohexanone (2) at the expense
coccus sp. for the hydrolytic ring-opening of e-caprolactone
[
16]
(3), a primary ADH from E. coli
for the oxidation of
alcohol 4, an alanine dehydrogenase (AlaDH) from Bacillus
subtilis, and various w-transaminases for the amination. In
this system, substrate 3 was transformed only into alcohol 4
with a high conversion (99%), but formation of amine 6 was
insignificant (1%), irrespective of the transaminase employed
(see Table S4, entries 1–4 in the Supporting Information).
Testing module 2.0 with alternative ADHs (horse liver ADH
E-isoenzyme, ADH from Bacillus stearothermophilus, two
ADHs from Candida tropicalis) led to the same result. An
investigation of the amination of aldehyde 5 exclusively
showed that all the transaminases converted this substrate
with high conversion (see Table S5 in the Supporting Infor-
mation) and also that the AlaDH was not inhibited.
Subsequently, it turned out that all the tested ADHs were
unable to oxidize the primary alcohol moiety of the 6-
hydroxycarboxylic acid (4). Although an ADH from Acine-
+
of NADP to give NADPH; the latter is consumed during the
following NADPH-dependent Baeyer–Villiger monooxyge-
nase (BVMO) step to give e-caprolactone (3) and thereby
regenerating the NADP for the first step. It is important to
note that since the first module is NADP-dependent, the
second module has to be NAD-dependent to allow both
modules to run simultaneously and independently without
interference. Hence, the second module starts with the
hydrolysis of 3 to liberate the alcohol moiety to give 6-
hydroxyhexanoic acid (4). For the amination of the primary
+
[
3b,10]
alcohol,
alcohol 4 is oxidized to 6-oxohexanoic acid (5)
and finally aminated by an w-transaminase (w-TA) in the
presence of an alanine dehydrogenase (AlaDH) to give 6-
aminohexanoic acid (6). The latter reaction requires l-alanine
as an amine donor to give pyruvate as a co-product, which is
recycled by the AlaDH to l-alanine, thereby consuming the
ammonia and NADH. Since NADH is generated in the
oxidation step, module 2 again represents a self-sufficient
redox sequence.
In each module, an alcohol moiety is oxidized: this is
a secondary alcohol in module 1 and a primary alcohol in
module 2.0. To avoid interference of the two modules, the
alcohol dehydrogenases for each step need to be carefully
selected to preferentially transform the intended alcohol
exclusively. Similarly, the w-transaminase must aminate
exclusively compound 5 but not cyclohexanone (2). There-
fore, it has to exhibit perfect chemoselectivity to distinguish
between a ketone and an aldehyde.
[17]
tobacter NCIB 9871
Brevibacterium epidermitis strain HCU has been reported
as well as an isoenzyme from
[18]
to perform the oxidation of 6-hydroxycarboxylic acid (4),
these enzymes turned out to be unsuitable for our reaction at
elevated substrate concentrations because of their poor
stability and problems with expression. Interestingly, related
substrates such as 1-hexanol and the corresponding ethyl ester
of 4 (ethyl 6-hydroxyhexanoate) were readily oxidized by all
the primary ADHs employed and were efficiently converted
into the corresponding amines with an ADH/w-TA/AlaDH
system (see Table S4, entries 5–12 in the Supporting Infor-
mation). Thus, it was concluded that the carboxylic acid
moiety of 6-hydroxyhexanoic acid (4) inhibits the oxidation
by the ADHs tested. For example, the K value of 4 for the
prim-ADH from Bacillus stearothermophilus (ADH-ht)
i
[19]
For the first module, the BVMO originating from
[
11]
Acinetobacter calcoaceticus turned out to be most suitable
for our purpose, with the C376 LM400I double mutant
applied because of its higher oxidative stability compared to
Since the BVMO was NADPH-
dependent, an NADP -dependent ADH was selected (ADH
was determined to be 98 mm. Since the ethyl and methyl
esters of 4 were substrates for oxidation as well as for the
overall amination, an alternative reaction sequence was
envisioned in which an ester of 4 was generated as an
intermediate. To achieve this, lactone 3 has to be opened in
aqueous buffer with an alcohol (ethanol/methanol) rather
than water as the nucleophile. This should afford the
corresponding ester and thereby introduce capping of the
carboxylic acid moiety. Methanol was the nucleophile of
choice over ethanol, because the primary ADHs employed
oxidize ethanol, thus leading to an unwanted side reaction,
while the oxidation of methanol was negligible. Enzymes
catalyzing the ring opening of e-caprolactone (3) have in
general been described for polymerization in organic sol-
vents, since hydrolysis to the corresponding carboxylic acid
is expected in aqueous solution. In fact, to the best of our
knowledge, the ring opening of a lactone in water with an
alcoholic nucleophile has not been described before.
On testing various hydrolytic enzymes for the unprece-
dented transformation of e-caprolactone (3) into the corre-
sponding methyl ester in aqueous buffer, it turned out that
neither the lactonase from Rhodococcus sp. nor any lipase (12
tested) led to useful ester formation (see Table S6, entries 1–
[
12]
the wild-type enzyme.
+
[
13]
from Lactobacillus brevis).
Although the ADH/BVMO
[14]
cascade has been investigated by others, it turned out that
the selection of the BVMO variant was crucial to achieve
excellent conversion (up to 98%) into e-caprolactone (3),
even at a relatively high substrate loading (96% at 20 gl ,
00 mm; Table 1). This result could be confirmed on a prep-
arative scale (100 mg substrate 1), which resulted in 99%
conversion, with 96% formation of e-caprolactone (3) and
a 75% yield of the isolated product.
The more demanding module 2.0 was tested first by
employing the following enzymes: a lactonase from Rhodo-
À1
2
[20]
Table 1: Synthesis of e-caprolactone from cyclohexanol through a redox
self-sufficient cascade reaction (module 1, Scheme 1).
[
a]
Entry
Conc. 1 [mm]
Conv. [%]
Ketone 2 [%]
Lactone 3 [%]
1
2
50
200
>99
99
2
3
98
96
[
b]
1
3 in the Supporting Information). However, out of the 18
[
a] Reaction conditions: Na HPO /KH PO buffer pH 8.0, 2 mm MgCl ,
2
4
2
4
2
esterases tested, four showed activity, namely an esterase
from Bacillus subtilis (esterase 008-SD), two esterases from
2
18C, 170 rpm orbital shaker, 20 h, 1 bar O ; 0.2 U ADH and 0.2 U
2
+
+
BVMO, 0.3 mg NADP . [b] 0.8 U ADH and 0.8 U BVMO, 1.2 mg NADP .
1
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 14153 –14157