10.1002/cctc.201701262
ChemCatChem
COMMUNICATION
(Table 1). Biotransformations with NDO WT resulted in 74 ± 3 % of
3b (Table 1). NDO WT and all variants were R-selective with F202A
showing the highest stereoselectivity of 87 ± 1 % ee for R-3b (Table
2). The introduction of an additional methyl group (4) led to a mixture
of mono- (4a) and dihydroxylated (4b) products (Table 1, Figure S10
and S11). Conversions of α-methylstyrene (4) with the single variant
F352V resulted in a high selectivity for production of the diol 4b
(93 %) (Table 1). In contrary, with the double variant A206V_V260I
the monohydroxylated product 4a was obtained as main product
(83 % 4a, Figure S11). In addition to the high regioselectivity, the
variants F202A and F352V notably affected the stereoselectivity of
specificity favoring the dihydroxylation of the C=C double bond over
the O-demethylation of the para-methoxy group (Figure S13).
Moreover, we expanded the substrate scope of NDO variants
towards the sterically demanding monoterpene R-limonene (11).
Despite an overall strongly decreased activity, we identified variants
H295A, A206G and V260A that affected the regioselectivity and
activity in conversions of 11 (Figure S14). The monoterpene was
monohydroxylated to carveol (-OH at C6) and mentha-1,8-dien-10-ol
(-OH at C10) (Figure S14 and S15). In biotransformations with NDO
WT, the hydroxylation mostly occurred at position C10, while the
single variant H295A favored the hydroxylation at position C6 (92 %
of the formed carveol product, Figure S14 and S15). Applying the
single variants A206G and V260A led to hydroxylation of position
C10 and the activity towards the monoterpene was increased by a
factor of five compared to the NDO WT (Figure S14). The double
variant H295A_V260A combined the selectivity of H295A for carveol
and the overall increased activity for the monoterpene of V260A by
a factor of 1.5 (Figure S15). The introduction of one or two
mutations improved the conversion of the terpene, however,
compared to other ROs like CDO or toluene dioxygenase the yields
are still very low.[13,18]
4a (Table 2). F202A exhibited
a good stereoselectivity of
80 ± 2 % ee for S-4b, while the single variant F352V gave a slightly
enantiomer enriched excess of 4b (14 ± 0 % ee (S-4b)) (Table 2).
Table 2. Stereoselectivities in whole cell biotransformations with NDO WT and
variants thereof.
Stereoselectivity ee [%][a]
These results allow us to draw some general conclusions: (1)
elongation of the alkyl substituent led to a decrease of activity, (2)
high degree of branching of the alkyl substituent forced the enzyme
to address the aromatic ring, (3) alkenyl residues with a conjugated
C=C double bond were fairly well addressed and favoured over the
dealkylation of incorporated methoxy groups, (4) gem-di-substituted
alkenyl substituents resulted in mono- as well as dihydroxylated
products, (5) trans-di-substituted alkenyl substituents directed the
reactivity towards allylic monohydroxylation and (6) isolated C=C
double bonds in alkenyl residues led to diols on absence of a
methoxy group. These observations were supported by in silico
analysis of the wild type enzyme with various substrates where
docking experiments indicated a non-flexible active site of the NDO
WT. An alignment of 19 crystal structures of NDO WT with different
co-crystallized substrates confirmed the in silico analyses as all
substrates were located in the same coordination plane (Figure
S16). However, in docking experiments with NDO variants, the
substrates were often found to be stabilized in another productive
coordination (Figure S17 and S18). A clear trend was observed for
the seven positions that affected catalysis most. In this respect, the
position 206 seems to play an important role as a set screw in the
upper part of the active site (Figure 1A and 1B).
98 ± 0
95 ± 0
82 ± 0
94 ± 0
42 ± 2
32 ± 0
34 ± 0
91 ± 0
2
3
4
(S-2a)
(S-2a)
(S-2a)
(S-2a)
(R-2a)
(S-2a)
(S-2a)
(S-2a)
78 ± 1
87 ± 1
83 ± 1
79 ± 1
46 ± 0
84 ± 0
84 ± 1
81 ± 1
(R-3b)
(R-3b)
(R-3b)
(R-3b)
(S-3b)
(R-3b)
(R-3b)
(R-3b)
55 ± 2
80 ± 2
41 ± 0
71 ± 2
14 ± 0
28 ± 0
70 ± 1
81 ± 3
(S-4a)
(S-4a)
(S-4a)
(S-4a)
(S-4a)
(S-4a)
(S-4a)
(S-4a)
> 99
(S,S-
5b)
> 95
(S,S-
5b)
19 ± 0
(R,R-
5b)
28 ± 3
(S,S-
5b)
28 ± 0
(R,R-
5b)
2 ± 0
(S,S-
5b)
5
n. d.
n. d.
> 95
> 95
> 95
94 ± 1
83 ± 1
> 95
6
n. d.
n. d.
(R-6b)
(R-6b)
(R-6b)
(R-6b)
(R-6b)
(R-6b)
[a]
Enantiomeric excess ee values were determined by GC-FID or HPLC
analyses using a chiral stationary phase (Table S5). n. d.: not detected.
These results are in line with data from literature, where
phenylalanine 352 is described as
a key amino acid for
stereoselectivity.[17] The relevance of F352 for selectivity is further
shown by the conversion of trans-β-methylstyrene (5) (Table 1 and
2). In this case, the variant F352V inverted the selectivity from
> 99 (S,S-5b) to 19 ± 0 (R,R-5b) (Table 2). Additionally, an influence
on regioselectivity has been observed at position 352 (Table 1). By
performing biotransformations with the NDO WT only the
monohydroxylated product (23 ± 0 % 5a) was observed, while
conversions with F352V resulted in 20 ± 2 % of the dihydroxylated
product 5b (Table 1). Screening of allylbenzene (6) led mainly to the
formation of the dihydroxylated product 6b, which was also favored
by the NDO WT (Table 1). Only the position L307 showed a different
product distribution by converting the substrate into a 1:1-mixture of
allylic mono- (6a) and the dihydroxylated product (6b) (Table 1).
The introduction of an additional para-methoxy group in the
substrates 3 and 6 resulted in 4-allylanisole (9) and 4-methoxy-
styrene (10). In biotransformations with 9, containing an isolated
C=C double bond in the alkene substituent, the reaction specificity
was shifted to O-demethylation (Figure S12). Replacement of the 2-
propenyl moiety with a vinyl group in 10 reversed the reaction
The control of the reaction specificity, and the regioselectivity was
strongly affected by this position as shown in biotransformations with
5 and 11. The importance of position A206 is further emphasized by
the enhanced substrate scope of the CDO variant M232A
(corresponds to the position A206 in the NDO) towards non-natural
terpene substrates like myrcene and (–)-α-pinene.[13]
The other identified positions (202, 307, 260, 295 and 352) are
mainly located on the opposite β-sheet to the position A206 in the
active site of the NDO (Figure 1A-C). In terms of regioselectivity and
activity the positions A206, V260 and H295 had a strong impact on
product distribution and yield (Figure 1B), whereas a major impact
on stereoselectivity was observed by addressing positions V260,
H295 and F352 (Figure 1C). Furthermore, positions F202, L307 and
F352 notably affected reaction specificity (Figure 1B).
This article is protected by copyright. All rights reserved.
Halder, Julia M.
