Journal of the American Chemical Society
Page 2 of 5
OYE-1 providing product in the highest yield (Table 1,
FIGURE 1. a. Previous studies using ketoreductases. b. The
proposed enantioselective radical dehalogenation using an
‘ene’-reductase.
1
2
3
4
5
6
7
8
entry 4). However, the ERED from G. oxydans (GluER)
afforded the highest levels of enantioselectivity favoring
the (S)-enantiomer (Table 1, entry 1).18 Importantly, the
unreacted starting material from the GluER catalyzed
reaction was found to be racemic, suggesting that the
enzyme does not preferentially react with one
enantiomer over the other. It is also interesting to note
that an engineered Baeyer-Villiger monooxygenase,
which uses the FAD cofactor to perform flavin-catalyzed
oxidations, is also capable of this reactivity, providing
product in modest yield and enantioselectivity (Table 1,
entry 10).19
In an effort to improve the enantioselectivity of the
observed dehalogenation, we targeted the conserved
tyrosine residue position at 177 for mutagenesis. Given
the role of tyrosine in the natural reaction, we
hypothesized that it could also serve as a hydrogen
source, which would diminish the overall selectivity of
the transformation. To our delight, mutation to
phenylalanine (Y177F) provided product in 69% yield
and 97:3 er (Table 1, entry 11) with the remaining mass
balance being unreacted starting material. Interestingly,
while mutation of the conserved tyrosine residues in
PETNr and YqjM provided a similar increase in yield, no
change in the enantioselectivity was observed (see
supplemental information). The reaction yield could be
further improved by increasing the catalyst loading to
We targeted the enantioselective dehalogenation of
α-bromoesters as an ideal model reaction to investigate
the ability of flavoenzymes to catalyze radical reactions.
Enantioselective hydrogen atom transfer to this class of
substrate is a challenging reaction for small molecule
catalysts, typically providing low levels of
enantioselectivity with stoichiometric reagents.13 EREDs
were selected as an ideal enzyme platform because
they are known to be highly evolvable, broadly
substrate permissive, and capable of providing high
levels of enantioselectivity for enoate reduction.
Furthermore, the mechanism of this reduction has been
extensively studied and is well understood to proceed
via a hydride transfer mechanism.14 However, early
studies by Miura found that old yellow enzyme (OYE1) is
able to reduce menadione to the corresponding radical
anion via a single electron pathway.15 Mechanistically,
we hypothesized that a dehalogenation could occur via
electron transfer from FMNHhq to the substrate,
followed by rapid mesolytic cleavage to form α-acyl
radical (Figure 1b). This species could abstract a
hydrogen atom from FMNHsq (calc N–H BDE = 59.9
kcal/mol) to form the reduced product and FMN.16
Mechanistically, this proposed reactivity is reminiscent
of flavin-dependent iodotyrosine deiodinase where
FMNHhq is proposed to reduce the keto tautomer of
iodotyrosine to form a phenoxy radical and FMNHsq.17
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
0.75 mol
% with no observed change in the
enantioselectivity (Table 1, entry 12). The reaction is
tolerant of cell free lysates, furnishing product with
good enantioselectivity albeit in modest yield (see
supplemental information).
With ideal conditions in hand, we explored the scope
and limitation of this reaction with the evolved G.
oxydans variant GluER-Y177F. The reaction is broadly
tolerant of substituents appended to the meta position
of the arene, with both electron rich (Table 2, 4 and 14)
and electron poor (Table 2, 6, 8, 10, 12) substituents
TABLE 1. Screen of Structurally Diverse Flavin-Dependent
Oxidoreductases
providing product with
good
to excellent
enantioselectivity. Ortho substituents are also accepted,
albeit in modest yield and enantioselectivity (Table 2,
16). Intriguingly, para substituents are poorly tolerated
with ethyl esters, providing less than 10% of the desired
product with the remaining mass balance being
unreacted starting material. The para substituted
methyl ester variants afford the product in higher yield
but again with low levels of enantioselectivity (Table 2,
18). Testing other variants failed to provide product in
superior
enantioselectivity
(see
supplemental
information). Heterocyclic arenes are reactive
substrates, but unfortunately provide low levels of
enantioselectivity presumably due to an altered binding
mode afforded by the Lewis basic nitrogen (Table 2, 20).
We began our investigation by subjecting α-bromo-α-
aryl ester 1 to a panel of nine structurally diverse EREDs
(Table 1). To our surprise, all of the enzymes tested
provided the desired dehalogenated product 2 with
ACS Paragon Plus Environment