ACS Catalysis
Research Article
general, the data in Table 1 show that all substrates were
converted approximately equally well by OYE1, 2, and 3 and
FOYE. Although XenA led to only moderate results for 1a−g,
it showed best performance for the ethyl ketone-derived
substrate 1h in comparison to all other enzymes tested.
Preparative biotransformations were performed at the 200
mL scale using the enzyme that gave the highest product
formation on analytical scale. Products 2a and 2d gave the best
isolated yields with 61 and 62%, respectively, whereas the
isolated yield of pyrazines of 2g and 2h was 40%. Control
experiments indicated that the formed products are stable in an
aqueous solution under the conditions used and can be
recovered quantitatively.
Scheme 1. Reductive Transformation of Activated Oximes
by ERs
Next, we turned our attention to the possible reaction
pathway. As described for pyrazine formation employing
transaminases,20 an α-amino keto motif can be expected to
be an intermediate; thus, for the reaction of 1a, α-amino-β-
keto ester 3 might be the intermediate (Scheme 2).
Table 1. Transformation of the Oximes with Purified ERs
a
enzyme
Scheme 2. Proposed Reaction Pathway from 1a to Pyrazine
2a through ER-Catalyzed Oxime Reduction and
Spontaneous Oxidation
bcd
, ,
product formation 2 [%]
#
sub OYE1
OYE2
OYE3
66
OPR3
XenA
FOYE
61
64 (42)
53 (53)
51
77
53
62 (35)
14
1
2
3
4
5
6
7
8
1a
1b
1c
1d
1e
1f
64
64
48
50
61
56
59
18
68 (61)
58
49
51
63
59
56
17
40
43
38
41
49
45
49
16
15
<1
9
17
<1
<1
29
38 (39)
62
39
53 (62)
77 (57)
59 (50)
51
1g
1h
28
a
The enzymes originate from Saccharomyces pastorianus (OYE1),15
Saccharomyces cerevisiae (OYE2 and OYE3),16 Arabidopsis thaliana
(OPR3),17 Pseudomonas putida (XenA),18 and Ferrovum sp. JA12
(FOYE).19 Reaction conditions: ER (200 μg/mL, 4 μM), 10 mM
b
substrate, 0.5 mM NADPH, 50 mM glucose, 4 mg/mL GDH, 5%
DMSO (v/v), 50 mM phosphate buffer, pH 7.5, 30 °C, 24 h, and 120
c
rpm. Product formation is defined by the amount of substrate
The formation of compound 3 from the oxime (in
equilibrium with the corresponding nitroso compound)
formally requires two hydrides originating from NAD(P)H.
Either an imine or a hydroxylamine might be formed after the
first hydride transfer. Dimerization of 3 to X, followed by
spontaneous oxidation, leads to the final product. As a first
attempt to detect the intermediates, the biotransformation of
1a was followed via NMR over a period of 30 min. When
analyzing the obtained spectra, the peaks could be assigned to
the oxime substrate, buffer components, glucose, and the
S48); however, no intermediate was detected with this
method. In addition, following the reaction by HPLC−
HRMS did not lead to any hint either.
In an attempt to get hold of the proposed intermediate 3, a
cascade reaction was envisioned to reduce the carbonyl group
of the possible formed intermediate 3 by an alcohol
dehydrogenase to obtain the threonine ester 4 (Scheme 3).
Consequently, the cyclization of 3 would be circumvented and
the formation of 4 would give an indirect proof for the
formation of compound 3. For this purpose, an ADH is
required, which does not transform oxime 1a but may
transform any derivative. Fortunately, it turned out that
ADH-A from Rhodococcus ruber21 does not transform oxime
1a. Therefore, any new product observed in the cascade must
be due to ADH-catalyzed transformation of a compound
formed by reduction with the ER.
transformed to 2 as deduced from GC analysis using calibration
d
curves. Numbers in brackets represent isolated yield.
Subsequently, a library of β-keto α-oximo ester substrates
and one β-keto α-oximo amide were investigated by varying
the size of the alcohol part of the ester (R′) and the substituent
R at the ketone side (substrates 1b−1h, Table 1). Increasing
the size of the ester (R′) moiety from ethyl (1a) to bigger
groups such as benzyl (1f) led to a comparable amount of
product formation for most enzymes, except XenA, which
behaved differently and led to very low product formation for
some esters. Substrate 1g, a N-phenylamide, was converted in a
comparable fashion like the related ester 1f. Again XenA was
different because for this enzyme, 1g was the substrate, leading
to the second best value for product formation. In general, it
can be stated that the accepted scope of ester moieties is rather
broad, as far as tested.
In contrast, extending the methyl group at the ketone moiety
to an ethyl group (substrate 1h) led to low product formation
with almost all enzymes investigated (entry 8). The only
exception was the ER XenA, which gave decent product
formation of 2h for this substrate (38%). However, when the
size of the substituent at the ketone moiety (R) was extended
further to an n-butyl or phenyl group (substrates 1i and 1j, see
or any other product was detected with any enzyme. In
13378
ACS Catal. 2020, 10, 13377−13382