Developing a Biocascade Process: Concurrent Ketone Reduction-Nitrile Hydrolysis of 2-Oxocycloalkanecarbonitriles

Article abstract of DOI:10.1021/acs.orglett.6b01510

A stereoselective bioreduction of 2-oxocycloalkanecarbonitriles was concurrently coupled to a whole cell-catalyzed nitrile hydrolysis in one-pot. The first step, mediated by ketoreductases, involved a dynamic reductive kinetic resolution, which led to 2-hydroxycycloalkanenitriles in very high enantio- and diastereomeric ratios. Then, the simultaneous exposure to nitrile hydratase and amidase from whole cells of Rhodococcus rhodochrous provided the corresponding 2-hydroxycycloalkanecarboxylic acids with excellent overall yield and optical purity for the all-enzymatic cascade.

Full text of DOI:10.1021/acs.orglett.6b01510

Letter
pubs.acs.org/OrgLett
Developing a Biocascade Process: Concurrent Ketone Reduction-
Nitrile Hydrolysis of 2‑Oxocycloalkanecarbonitriles
Elisa Liardo,† Nicolas Ríos-Lombardía,‡ Francisco Morís,‡ Javier Gonzalez-Sabín,*,‡
́
́
and Francisca Rebolledo*,†
†Departament of Organic and Inorganic Chemistry, University of Oviedo, Julian
́
Clavería, 8, 33006 Oviedo, Spain
‡EntreChem, SL, Edificio Científico Tecnolog
́
ico, Campus El Cristo, E-33006 Oviedo, Spain
S
* Supporting Information
ABSTRACT: A stereoselective bioreduction of 2-oxocycloal-
kanecarbonitriles was concurrently coupled to a whole cell-
catalyzed nitrile hydrolysis in one-pot. The first step, mediated
by ketoreductases, involved a dynamic reductive kinetic
resolution, which led to 2-hydroxycycloalkanenitriles in very
high enantio- and diastereomeric ratios. Then, the simulta-
neous exposure to nitrile hydratase and amidase from whole
cells of Rhodococcus rhodochrous provided the corresponding 2-
hydroxycycloalkanecarboxylic acids with excellent overall yield
and optical purity for the all-enzymatic cascade.
Scheme 1. KRED and Enzymatic Cascade for Transforming 2-
Oxocycloalkanecarbonitriles
xidative−reductive transformations belong to the most
Oimportant reactions in organic synthesis. Redox-active
enzymes such as oxygenases, alcohol dehydrogenases, amine
dehydrogenases, and ene-reductases selectively catalyze the
introduction and modification of functional groups under mild
reaction conditions and can create chiral centers with excellent
stereoselectivity. Although some critical issues such as cofactor-
dependency, regeneration systems, or unfavorable reaction
equilibria are associated with these processes, different
approaches such as the use of biomimetic cofactors1 or the
coupling of several enzymatic steps have been developed.2
However, the possibility of performing multistep enzymatic
synthesis in a concurrent fashion is particularly appealing since
this strategy can reduce costs significantly. This fact, combined to
the intrinsic green features of enzymes, increase the sustainability
expectations of biocatalysis-based chemical manufacturing.
Multistep reactions in whole cell fashion were described as
early as in the 1980s for the production of amino acids.3 As for
the possibilities to combine several isolated enzyme classes
facilitating a one-pot cascade, the use of redox enzymes has been
successfully demonstrated in recent years.2
Early reports on the bioreduction of cyclic β-ketonitriles relied
on biotransformations by fungi and yeasts, leading to the
corresponding cis-(1S,2S)-β-hydroxy nitriles in high enantio- and
diastereomeric excess.4 More recently, isolated carbonyl
reductases were sequentially coupled with nitrilases to provide
enantiopure linear β-hydroxy carboxylic acids from α-unsub-
stituted β-ketonitriles.5 In this sense, an interesting alternative to
nitrilases could be the use of two enzymes, nitrile-hydratase and
amidase, which are found in different strains of Rhodococci. In the
light of this background, we envisaged to test the bioreduction of
several 2-oxocycloalkanecarbonitriles (1a−c) concurrent to the
enzymatic hydrolysis of the cyano group (Scheme 1). Thus,
taking advantage of the epimerizable stereocenter of 1a−c (pKa =
7.84 for 1a),4 our first goal was to identify highly stereoselective
ketoreductases (KREDs) to promote efficient dynamic reductive
kinetic resolutions (DYRKR).6 Then, the resulting optically
active β-hydroxynitriles 2a−c would undergo, by the successive
action of a nitrile-hydratase (NHase) and an amidase, a further
hydrolysis to the final β-hydroxyacids 3a−c. However, although
water is the natural enzyme environment, a number of issues
should be addressed to implement such enzymatic cascade.7 A
major challenge is the compatibility of the different enzymes with
the preferred pH, temperature, concentration, and cosolvents,
but also specific activities and stability should be balanced and
inhibition avoided. In addition, to set a process as shown in
Scheme 1, the tandem NHase-amidase system should ideally not
be active toward the starting β-ketonitrile, or in the worst case,
much slower than toward the intermediate β-hydroxynitrile.
The starting β-ketonitriles 1a−c were synthesized from
inexpensive alkanedinitriles following literature procedures.4
Then, the bioreduction was tested using ketoreductases from the
Codex KRED Screening Kit with isopropanol (IPA) for cofactor
recycling and as cosolvent. Initially, the screening was performed
Received: May 25, 2016
© XXXX American Chemical Society
A
DOI: 10.1021/acs.orglett.6b01510
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
under the standard conditions at pH 7.0 (Tables S1, S3, and S5),8
which should be enough to promote the desired racemization.
Alternatively, bioreductions of 1a−c at pH 5.0 and 10 were also
tested (Tables S2 and S4).8 Thus, all the KREDs led to complete
conversion after 24 h, and the resulting 2-hydroxycycloalkaneni-
triles were isolated in high yield (>95%). With regards to the
stereoselectivity, results were biocatalyst- and ketone-dependent,
the most representative are shown in Table 1.
Thus, KRED-P1-A04 rendered, by means of a highly efficacious
DYRKR, (1S,2S)-2c in >99 dr and >99% ee. It is worth
mentioning that both enantiomers of cis-2-hydroxycyclohepta-
nenitrile were prepared in this report for the first time.
Nitriles are immediate precursors of carboxylic acids, but harsh
conditions typically used for their hydrolysis are often
incompatible with most other functional groups.9 The
biotransformation of nitriles, however, either through a direct
conversion to a carboxylic acid catalyzed by a nitrilase or through
the previous NHase-catalyzed hydration followed by the
amidase-catalyzed hydrolysis of the resulting amide, is a
conveniently mild alternative.10 In this regard, several Rhodococci
catalyzed the hydrolysis of closely related cyclic N-protected-β-
aminonitriles with comparable trends: five-membered substrates
were transformed significantly faster than the six-membered
homologues, and the trans-derivatives reacted faster than the cis-
counterparts. Moreover, the enantioselectivity was higher for the
trans-isomers.11
Table 1. DYRKR of β-Ketonitriles 1a−c Catalyzed by
a b
,
KREDs
Initial screening experiments were performed with both β-
ketonitrile 1a and a cis/trans mixture of β-hydroxynitrile 2a
employing nitrilases from the Codex Nitrilase Screening Kit, but
all the attempts showed low activities. Accordingly, we focused
our attention on whole-cell biocatalysts, namely, the commer-
cially available bacterium Rhodococcus rhodochrous IFO 15564.12
Thus, biotransformations of 1a−c and 2a−c were performed
with a standard cell concentration of microorganism in the
metabolic resting phase [approximately 0.9 mg/mL of aqueous
0.1 M phosphate buffer pH 8.0, 1% EtOH v/v (equivalent to A650
= 1.0)]. TLC analysis after 24 h showed complete conversion of
cis- and trans-isomers of 2a−c into the corresponding β-
hydroxyacids 3a−c and no reaction with the β-ketonitriles 1a−
c (Scheme 2). Interestingly, both facts fulfilled the prerequisites
β-
entry ketonitrile
ef
,
ef
,
KRED
cis/trans
eecis (%)
eetrans (%)
c
1
2
3
4
5
6
7
8
9
1a
1a
1a
1a
1a
1a
1a
1b
1b
1b
1b
1c
1c
P2-D11
P2-D11
P1-B12
P1-B12
P2-G03
P2-G03
NADH101
P1-A04
P2-H07
P1-B10
P1-B12
P1- A04
P1- B10
10:90
6:94
>99 (1R,2S)
>99 (1R,2S)
d
c
83:17
88:12
96:4
98 (1R,2R)
>99 (1R,2R)
90 (1S,2S)
d
c
d
d
c
98:2
95 (1S,2S)
>99:<1
>99:<1
>99:<1
94:6
>99 (1S,2S)
>99 (1S,2S)
>99 (1S,2S)
>99 (1R,2R)
>99 (1R,2R)
>99 (1S,2S)
>99 (1R,2R)
c
c
c
c
c
10
11
12
13
>99:<1
>99:<1
98:2
Scheme 2. Preliminary Study of R. rhodochrous Activity
a
Substrate (20 mM) in KH2PO4 buffer, 125 mM (1.25 mM MgSO4, 1
mM NADP+), pH 7.0 or 5.0 (900 μL), KRED (2 mg), IPA (190 μL),
b
c
d
24 h at 250 rpm and 30 °C. Conversion >99%. pH = 7.0. pH = 5.0.
e
f
Measured by chiral GC. Absolute configuration established as
detailed in the SI.
In the case of the cyclopentanone derivative 1a (entries 1−7),
half of KREDs exhibited cis diastereoselectivity; meanwhile, the
other half afforded predominantly the trans-counterpart, the best
results being obtained at pH 5.0. Pleasantly, we could identify
biocatalysts that gave rise to three out of the four possible
stereoisomers with high diastereomeric ratio and enantiomeric
excess >99%. Remarkably, KRED-P2-D11 yielded the trans
isomer (1R,2S)-2a [(1R,2S)-2-hydroxycyclopentanecarboni-
trile] in 16:1 dr and >99% ee (entry 2). Regarding the cis-
isomer, KRED-P1-B12 provided enantiopure (1R,2R)-2a with
7:1 dr (entry 4); meanwhile, KRED-NADH101 gave its
counterpart (1S,2S)-2a with an excellent dr and >99% ee
(entry 7).
In the case of the cyclohexanone derivative 1b, all the KREDs
afforded the cis-diastereomer predominantly, with better
performances at pH 7.0. It is noteworthy the case of KRED-
P1-A04 and KRED-P2-H07, which exhibited total selectivity
toward (1S,2S)-2b, with >99 dr and >99% ee (entries 8−9).
Likewise, the enantiomer (1R,2R)-2b could also be obtained in
32:1 dr and >99% ee from the reaction catalyzed by KRED-P1-
B12 (entry 11). The best KREDs found in the bioreduction of 1b
remained the same for its seven-membered analogous 1c (entries
12−13), again favoring the formation of the cis-diastereomer.
for the implementation of the biocascade: (1) the starting
material is substrate only for the KRED; (2) the tandem NHase-
amidase is very active toward the product of the KRED,
independently of its absolute configuration. Accordingly, the
selectivity of the biocatalyst was not analyzed in depth since the
second step of the cascade will be fueled with optically active β-
hydroxynitriles in very high dr and ee. Even so, R. rhodochrous
displayed higher reactivity and enantioselectivity toward the
trans-isomers (data not shown), in accordance with the previous
background.11
Next, we took the challenge of coupling both steps into a
concurrent process with both enzymes working “hand-in hand”.
Actually, the previous report dealing with linear β-ketonitriles
was performed sequentially, and, once the carbonyl reductase-
catalyzed reaction was completed, the pH was readjusted and the
nitrilase added to the medium.5 In preliminary cascade attempts,
1b was subjected to the medium used in the bioreductions but
containing both R. rhodochrous IFO-15564 and a KRED.
Regarding the cosolvent, IPA (essential for KREDs) and EtOH
B
DOI: 10.1021/acs.orglett.6b01510
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
(typical with R. rhodochrous) were checked in different ratios (1−
15% v/v). The most significant outcomes were the following: (1)
EtOH is not accepted for KREDs; (2) 15% v/v of IPA inhibits
the activity of R. rhodochrous at A650 = 1−4; (3) both R.
rhodochrous (A650 = 3−4) and KRED are active with 5% v/v of
IPA; and (4) the activity of R. rhodochrous significantly decreased
at pH 5. Based on this parametrization, the biotransformation of
β-ketonitriles were designed using phosphate buffer (125 mM,
1.25 mM MgSO4, 1 mM NADP+) pH 7.0, IPA 5% v/v, and whole
cells of A650 = 4. Under these optimized conditions, both
biocatalysts were functional and led to the target β-hydroxyacids.
All the reactions assayed at pH 7.0 in Table 1 were submitted
to the biocascade. Thus, both cyclohexylic and cycloheptylic β-
ketonitriles 1b and 1c underwent effective bioconversion into the
corresponding β-hydroxyacids 3b and 3c with nearly the same dr
and ee (selected examples in Scheme 3).13 Regarding 1b, the
pH was the only adjustment before the addition of the
microorganism.
Finally, to exploit the synthetic utility of the enzymatic
platform developed herein, a biocascade using 1b was carried out
with whole cells of R. rhodochrous grown in the presence of
DEPA, an inhibitor of the amidase activity of the microorganism
(Scheme 5). Under these conditions, (1R,2S)-2-hydroxycyclo-
Scheme 5. Enzymatic Platform Towards Valuable Optically
Active Compounds Starting from Alkanedinitriles
Scheme 3. Biocascade Towards Stereoisomers of 3b and 3c
hexanecarboxamide (4b) was isolated from the cascade with
KRED-P1-A04. This compound is also an immediate precursor
of (1S,2R)-cis-2-aminocyclohexanol (5b), completing a spec-
trum of valuable optically active molecules from inexpensive
pimelonitrile, as exemplified in Scheme 5.
In summary, we have developed an enzymatic cascade process
in aqueous medium combining a highly selective DYRKR of 2-
oxocycloalkanenitriles, mediated by KREDs, with whole cells
containing NHase-amidase activity for a further nitrile hydrolysis.
Thus, the success of the strategy lied both in the synchronization
between a fast racemization compared to the bioreduction of the
nonpreferred enantiomer in the first step, and a microorganism
that converts the cyano group of the β-hydroxynitrile
intermediate but not the one contained in the starting β-
ketoxynitrile. Interestingly, the KREDs showed markedly cis-
diastereoselectivity, complementing the existing methodologies,
mostly aimed at trans-diastereoisomers, readily available from
epoxides and aziridines.
combination of R. rhodochrous IFO-15564 and the stereo-
complementary ketoreductases P1-A04 and P1-B12 delivered,
respectively, (1R,2S)-3b and (1S,2R)-3b in >95% yield (30 mg
scale) and very high selectivity: >99:1 dr and >99% ee (Scheme
3). Further extension to 1c provided both antipodes of the
previously unreported cis-2-hydroxycycloheptanecarboxylic acid
(3c), namely, (1R,2S)-3c and (1S,2R)-3c in nearly quantitative
yield and complete selectivity (>99 dr and >99% ee).
Although operationally viable, the cascade for the 5-membered
ring homologue 1a was challenging since the optimal selectivity
in the bioreduction step was achieved at pH 5.0 (Table 1, entries
1−7), which inhibits the activity of R. rhodochrous. Accordingly,
we turned our attention toward a stepwise process. Thus, a 30
mg-scale bioreduction of 1a was performed with KRED-P2-D11
at pH 5.0, and once the reduction was completed, pH was raised
to 7.0 and a bacterial suspension of high absorbance (A650 = 4)
added. As a result, (1S,2S)-3a was isolated in very high dr (19:1)
and excellent ee (>99%), despite a moderate yield of 52% due to
the incomplete extraction from the aqueous medium (Scheme
4).14 It is of note the simplicity of the setup, despite not being a
“true cascade” since the isolation of the β-hydroxynitrile
intermediate was not necessary. Actually, both steps were
performed under identical reaction medium, and a change on the
ASSOCIATED CONTENT
* Supporting Information
■
S
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.orglett.6b01510.
Experimental procedures, enzymatic screenings, charac-
terization data, copies of the corresponding 1H, 13C
spectra, and GC chromatograms (PDF)
AUTHOR INFORMATION
Corresponding Authors
■
*E-mail: jgsabin@entrechem.com.
*E-mail: frv@uniovi.es.
Notes
Scheme 4. One-Pot Sequential Synthesis of (1S,2S)-3a
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
E.L. acknowledges funding from the European Union’s Horizon
2020 MSCA ITN-EID program under grant agreement No
634200 (Project BIOCASCADES). N.R.-L. acknowledges
MINECO for funding under Torres-Quevedo program (PTQ-
C
DOI: 10.1021/acs.orglett.6b01510
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
12-05 407). Authors from University of Oviedo thank the
Principado de Asturias for financial support (FC-15-GRUPIN-
14-002).
REFERENCES
■
(1) (a) Knaus, T.; Paul, C. E.; Levy, C. W.; De Vries, S.; Mutti, F. G.;
Hollmann, F.; Scrutton, N. S. J. Am. Chem. Soc. 2016, 138, 1033.
(b) Paul, C. E.; Arends, I. W. C. E.; Hollmann, F. ACS Catal. 2014, 4,
788.
(2) (a) Ricca, E.; Brucher, B.; Schrittwieser, J. H. Adv. Synth. Catal.
2011, 353, 2239. (b) Simon, R. C.; Richter, N.; Busto, E.; Kroutil, W.
ACS Catal. 2014, 4, 129. (c) Gargiulo, S.; Opperman, D. J.; Hanefeld, U.;
Arends, I. W. C. E.; Hollmann, F. Chem. Commun. 2012, 48, 6630.
(d) Kalaitzakis, D.; Smonou, I. J. Org. Chem. 2010, 75, 8658. For
combinations of organocatalysts or metals with enzymes, see: (e) Baer,
K.; Krauβer, M.; Burda, E.; Hummel, W.; Berkessel, A.; Groger, H.
̈
Angew. Chem., Int. Ed. 2009, 48, 9355. (f) Ríos-Lombardía, N.; Vidal, C.;
́
́
Liardo, E.; Morís, F.; García-Alvarez, J.; Gonzalez-Sabín, J. Angew. Chem.,
Int. Ed. 2016, DOI: 10.1002/anie.201601840. (g) Dey, S.; Powell, D. R.;
Hu, C.; Berkowitz, D. B. Angew. Chem., Int. Ed. 2007, 46, 7010.
(h) Karukurichi, K. R.; Fei, X.; Swyka, R. A.; Broussy, S.; Shen, W.; Dey,
S.; Roy, S. K.; Berkowitz, D. B. Sci. Adv. 2015, 1, e1500066.
(3) Yamada, H.; Shimizu, S. Angew. Chem., Int. Ed. Engl. 1988, 27, 622.
(4) Dehli, J. R.; Gotor, V. J. Org. Chem. 2002, 67, 6816.
(5) Ankati, H.; Zhu, D.; Yang, Y.; Biehl, E. R.; Hua, L. J. Org. Chem.
2009, 74, 1658.
(6) Applegate, G. A.; Berkowitz, D. B. Adv. Synth. Catal. 2015, 357,
1619.
(7) Bornscheuer, U. T. Angew. Chem., Int. Ed. 2015, 54, 2.
(8) See Supporting Information for all the enzymatic screenings.
(9) Hann, E. C.; Sigmund, A. E.; Fager, S. K.; Cooling, F. B.; Gavagan, J.
E.; Ben-Bassat, A.; Chauhan, S.; Payne, M. S.; Hennessey, S. M.;
DiCosimo, R. Adv. Synth. Catal. 2003, 345, 775.
(10) Martínkova,
(11) Winkler, M.; Martínkova,
N. Tetrahedron 2005, 61, 4249.
́
L.; Kren, V. Biocatal. Biotransform. 2002, 20, 73.
́
L.; Knall, A. C.; Krahulec, S.; Klempier,
(12) Formerly known as Rhodococcus butanica ATCC 21197
(a) Kakeya, H.; Sakai, N.; Sugai, T.; Ohta, H. Tetrahedron Lett. 1991,
32, 1343. (b) Moran
521.
́
-Ramallal, R.; Liz, R.; Gotor, V. Org. Lett. 2007, 9,
(13) For 3a-c, the Cahn−Ingold−Prelog priority of the substituents at
the C-1 position changed with respect to 2a-c.
(14) The starting material 1a and the intermediate 2a were not
detected, which indicates that the sequential process worked efficiently.
D
DOI: 10.1021/acs.orglett.6b01510
Org. Lett. XXXX, XXX, XXX−XXX