S. Qiu, et al.
BioorganicChemistry103(2020)104228
efficiency towards (5R)-1 was 36.31 s−1·mM−1. Under the optimized
conditions, it completely reduced (5R)-1 at load of up to 80 g·L-1 in
1.5 h, giving (3R,5R)-3 in d.e.p > 99.5% and STY of 660.0 g·L-1·d-1. In
present work, to engineer KmAKR, we used the combinational screening
strategy to screen hot spots around the intersection of the neighbor-
hoods of the active center and surface loop. A synchronously im-
provement in both activity and thermostability was achieved. The
Variant VI is much more active and thermo-stable than KlAKR-Y295W-
W296L-I125V-S30P-Q212R-I63W [31].
3.7. Environmental factor analysis and economic analysis
Asymmetric bioreduction of (5R)-1 and (5S)-2 to (3R,5R)-3,
(3R,5S)-4 by Variant VI is greener and more economic than the che-
mical reduction. As for the conditions of reaction, a variety of organic
reagents consumption is zero for Variant VI. In contrast to the chemical
asymmetric reduction using metal catalysts and NaBH4, this enzymatic
process used glucose as sustainable co-substrates, affording water as a
clean byproduct, making this method a promising green chemistry
approach to manufacture atorvastatin intermediate (3R,5R)-3 and
“Super statins” chiral intermediate (3R,5S)-4. Biocatalytic process em-
ploying Variant VI avoids the use of potentially hazardous hydrogen
and heavy metal catalysts throughout the process, thus obviating con-
cerns for their removal from waste streams and/or contamination of the
product. More importantly, the enzyme catalysts and the co-substrate
glucose are derived from renewable raw materials and are completely
biodegradable. The by-products of the reaction are gluconate, a tiny
amount of residual glucose, enzyme, cell debris and minerals, and the
waste water is directly suitable for biotreatment. In term of economy, in
this bioprocess, no exogenous NADPH or NADP+ was added and the
atom efficiency is 56.3%, attributing to that the use of glucose as the
reductant for cofactor regeneration is cost effective but not particularly
atom efficient.[30] Fortunately, glucose is a renewable resource and the
Fig. 5. Conversion profile of Variant V and Variant VI catalyzed reduction of
(5R)-1 and (5S)-2 to (3R,5R)-3 and (3R,5S)-4.
to pharmacologically active (3R,5R)-3, (3R,5S)-4, accompanied by the
consumption of NADPH. Therefore, a glucose dehydrogenase from
Exiguobacterium sibiricum was introduced to regenerate the desired co-
factor NADPH from NADP+ using glucose as the co-substrate (Fig. 4).
The enzymatic asymmetric reduction of (5R)-1 at 200 g L-1 was
conducted using the Variant V, the reaction was accomplished com-
pletely within 4.5 h using 15.0 g DCW L−1 of E. coli cells expressing
Variant V and 5.0 g DCW L−1 of E. coli cells expressing EsGDH on
100 mL scale. Under the same scale, 450 g L-1 (5R)-1, 400 g L-1 (5S)-2
were completely converted by 4.5 g DCW L-1 Variant VI. Due to the
outstanding thermal stability of the Variant VI, we tried to raise the
reaction temperature from 35 °C to 40 °C to accelerate reaction rate. As
shown in Fig. 5, the reaction catalyzed 200 g L-1 (5R)-1 by Variant V
achieved > 99% conversion in 4.5 h, and the corresponding STY
reached 670.5 g L-1 d-1 with d.e.p > 99.5%. The “best” mutant Variant
VI catalyzed 450 g L-1 (5R)-1 for 7.0 h, giving 99% conversion and STY
of 1.24 kg L-1 d-1. 400 g L-1 (5S)-2 was converted into (3R,5S)-4 within
5 h, STY reaching 1.34 kg L-1 d-1. The “best” mutant Variant VI sig-
nificantly decreased the catalyst load, resulting in a dramatic increase
of substrate/catalyst ratio of from 10.0 g g−1 to 75.0 g g−1 (Table S5).
Reviewing the entire evolution campaign from WT to Variant VI,
five mutation sites were acquired to enhance the stereoselectivity, ac-
tivity and thermostability. As shown in Figure S13, the location of five
mutation site can be divided into 3 portions, core shell, middle shell and
surface shell. Core shell is the active center, NADPH binding region and
substrate binding region, and site 63 is located in the core shell. Surface
shell is the enzyme surface region, where site 296 and 297 are located.
Between the core shell and surface shell is the middle shell, where site
28 and 29 are located. The sites surrounding active center is the
common hot spot for activity enhancement via rational and semi-ra-
tional engineering. However, the sites surrounding active center re-
presents only a small portion of the whole enzyme. Identifying more hot
spots from surface shell and middle shell via other methods, such as
error prone PCR, to discover further reinforcing mutations other than
only core shell for enzyme activity and thermostability engineering is
helpful to develop ideal industrial biocatalysts with better catalytic
performance.
4. Conclusion
In summary, KmAKR mutants Variant VI and VII that exhibit si-
multaneously improved thermostability and activity compared with the
parent enzyme Variant III were constructed using the combinational
screening strategy and ISM. It was worth noting that the trade-off be-
tween activity and thermostability was avoided. Analysis of Docking,
protein interaction calculator and MD stimulation, mutations of Y28A
and T63M or T63L shortened the distance between the H atom of cat-
alytic site Tyr64-OH and the carbonyl oxygen atom of (5R)-1, accel-
erating the transfer rate of protons. As for thermostability, flexible loops
around the active site were rigidified through introducing additional
hydrogen bonds and hydrophobic interactions in mutants Variant VI
and Variant VII, leading to improvements in t1/2, Tm, T5015. Most im-
portantly, Variant VI allowed for asymmetric bioreduction of (5R)-1
and (5S)-2 to (3R,5R)-3 and (3R,5S)-4 at elevated temperatures (e.g.,
40 °C), dramatically shortening the reaction times and enhancing the
productivity significantly. Finally,
a STY of (3R,5R)-3 up to
1.24 kg·L−1·d−1 was achieved with a 75.0 g·g−1 S/C mass ratio, and a
STY of (3R,5S)-4 up to 1.34 kg·L−1·d−1 was achieved with a 66.6 g·g−1
S/C mass ratio. Hence, Variant VI is a very promising biocatalyst for
large scale production of statins chiral intermediates (3R,5R)-3 and
(3R,5S)-4.
To date, only limited alcohol dehydrogenases/reductases have been
reported for chirl diol (3R,5R)-3 and (3R,5S)-4 biocatalysis. In our
previous work, an aldo–keto reductase KlAKR-WT was cloned from the
Kluyveromyces lactis (Table 5), which had excellent R-stereoselectivity
towards (5R)-1, a sequence identity of 89% and the same catalytic
tetrad with the KmAKR [58]. Our previous work on KlAKR was mainly
focused on activity improvement through modification on amino acid
residues that were situated in the neighborhood of tunnels of substrate-
entrance and product-release, (5R)-1 and NADPH binding pockets, and
within 4 Å distance from the bound (5R)-1, yielding a positive variant
Declaration of Competing Interest
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influ-
ence the work reported in this paper.
KlAKR-Y295W-W296L-I125V-S30P-Q212R-I63W,
whose
catalytic
9