Stereodivergent Synthesis of 3-Hydroxyprolines and 3-Hydroxypipecolic Acids via Ketoreductase-Catalyzed Dynamic Kinetic Reduction

Article abstract of DOI:10.1002/adsc.201900871

We report a practical enzymatic approach for the stereoselective synthesis of hydroxylated cyclic amino acids. Using ketoreductases, cyclic ketoesters are converted with high diastereo- and enantioselectivity to all isomers of 3-hydroxyproline and 3-hydroxypipecolic acid via a dynamic kinetic reduction reaction. This work highlights the ability of enzymes to provide solutions to challenges in stereoselective synthesis. (Figure presented.).

Full text of DOI:10.1002/adsc.201900871

COMMUNICATIONS
DOI: 10.1002/adsc.201900871
Stereodivergent Synthesis of 3-Hydroxyprolines and
3-Hydroxypipecolic Acids via Ketoreductase-Catalyzed Dynamic
Kinetic Reduction
Christopher K. Prier,a,* Michael M.-C. Lo,a Hongming Li,a and
Nobuyoshi Yasudaa
a
Merck & Co., Inc., 126 E Lincoln Avenue, Rahway, NJ 07065, USA
phone: (732) 594-3856
E-mail: christopher_prier@merck.com
Manuscript received: July 27, 2019; Revised manuscript received: September 13, 2019;
Version of record online: ■■■, ■■■■
Supporting information for this article is available on the WWW under https://doi.org/10.1002/adsc.201900871
employed to access diverse chiral building blocks.[9]
When reductive enzymes are employed, such processes
Abstract: We report a practical enzymatic approach
for the stereoselective synthesis of hydroxylated
have been termed dynamic reductive kinetic resolution
cyclic amino acids. Using ketoreductases, cyclic
(DYRKR);[10] these reactions have the notable potential
ketoesters are converted with high diastereo- and
to set epimerizable stereocenters in addition to the
enantioselectivity to all isomers of 3-hydroxyproline
center undergoing reduction. As an alternative to
and 3-hydroxypipecolic acid via a dynamic kinetic
existing routes to hydroxyproline and hydroxypipecolic
reduction reaction. This work highlights the ability
acid structures, we envisioned the dynamic reductive
of enzymes to provide solutions to challenges in
kinetic reduction of a β-ketoester with the generic
stereoselective synthesis.
structure 1 (Figure 1). The synthesis of one hydrox-
yproline isomer, 3-cis-D-hydroxyproline, has been
demonstrated via such a DYRKR approach using
Dynamic kinetic resolution
Keywords: Biocatalysis; Enzymes; Amino acids;
yeasts.[11] However, we anticipated that employing
single ketoreductase (KRED) enzymes as catalysts,
rather than whole organisms, could provide access to
Hydroxyprolines and hydroxypipecolic acids are priv-
ileged pharmacophores, appearing as structural compo-
nents of many drugs and bioactive compounds (Fig-
ure 1).[1] These structures are also useful as synthetic
intermediates,[2] and their incorporation into proteins
can alter peptide conformations, stability, and aggrega-
tion behavior.[3] Insufficient levels of proline hydrox-
ylation in the biopolymer collagen is associated with a
number of disease states.[4] We required a robust
approach to access significant quantities of 3-hydrox-
yprolines with high stereochemical purity to support
ongoing drug development programs. In particular,
access to the trans-3-hydroxy stereochemistry pre-
sented a significant challenge. Diverse routes to this
structure have been reported using chiral pool starting
materials,[5] asymmetric catalysis,[6] and enzymatic
kinetic resolution,[7,8] but in general these reported
approaches require many synthetic steps and provide
Figure 1. 3-Hydroxyproline and 3-hydroxypipecolic acid struc-
tures in bioactive compounds, and the dynamic kinetic
resolution approach envisioned for their synthesis.
low overall yield.
Dynamic kinetic resolution (DKR) is a well-
established strategy in biocatalysis, and has been
Adv. Synth. Catal. 2019, 361, 1–5
1
© 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
��
These are not the final page numbers!

COMMUNICATIONS
asc.wiley-vch.de
different stereoisomers of 3-hydroxyproline, including ketoester 2 having different protecting groups were
those with the more difficult to access trans-stereo- converted to the desired trans products with inferior
chemistry. We recognized as well that CÀ H hydrox- levels of activity and/or selectivity (see Figure S4).
ylation using proline hydroxylases can provide rapid
Having identified selective variants for the produc-
access to hydroxyprolines.[12] These enzymes, however, tion of each desired isomer, we then performed these
are typically active only on L-amino acids. We were reactions on preparative scale, using at least 1.0 g
attracted to a DKR approach as it could provide access ketoester (and up to 30 g ketoester) (Figure 2). The
to both the L- and D-series of hydroxyproline analogs, reactions could be run at moderately high ketoester
both of which were required for our drug discovery loadings (up to 50 g/L) using isopropanol (IPA) as the
efforts.
stoichiometric reductant. In cases where the KRED did
The ketoester 2 was prepared in three steps from N- not tolerate high levels of isopropanol, glucose
Boc-β-alanine, following literature precedent.[13] Initial dehydrogenase (GHD) was employed for NADP
experiments demonstrated that the α-stereocenter read- recycling, using glucose as the stoichiometric reductant
ily undergoes racemization in phosphate buffer at and DMSO as the organic co-solvent. In all cases, high
°
pH 7.0 and 30 C. Screening a commercial library of yields of product could be obtained after a simple
KREDs under these conditions against ketoester 2, we extraction into organic solvent (see Supporting Infor-
then gratifyingly identified KREDs generating all four mation).
possible stereoisomers with high levels of diastereo-
To provide access to the 3-hydroxypipecolic acid
and enantiopurity (Figure 2). Using KRED P1B2, series, we then explored dynamic kinetic resolution
ketoester 2 was transformed selectively to the 3-cis-L- using the ketoester 4 (Figure 3). Many of the ketor-
proline analog 3b, while KRED 119 delivered the 3- eductases identified for selective reduction of 2 were
cis-D-proline analog 3d. More significantly, the also effective for reduction of ketoester 4, but in some
challenging 3-trans stereochemistry of isomer 3a cases superior selectivities could be obtained using
could be accessed using the enzyme KRED NADH different KRED enzymes. In general, higher catalyst
101 (an NADH-dependent reductase). The enantiomer- loadings were required to achieve full conversion of
ic trans isomer 3c could be accessed in excellent
diastereoselectivity and good enantioselectivity using
KRED 264, an enzyme previously evolved for the
DKR reduction of a bicyclic γ-ketoester.[14] Analogs of
Figure 3. Ketoreductase-catalyzed dynamic kinetic reduction
provides access to all four stereoisomers of 3-hydroxypipecolic
acid. Reactions performed on at least 800 mg scale. KREDs are
lyophilized powders of clarified E. coli lysate. All reactions
a
employ 1 g/L NAD(P). After separation of the diastereomers
Figure 2. Ketoreductase-catalyzed dynamic kinetic reduction
reactions provide access to all four stereoisomers of 3-
hydroxyproline. Reactions performed on at least 1.0 g scale.
KREDs are lyophilized powders of clarified E. coli lysate. All
reactions employ 1 g/L NAD(P). IPA=isopropanol; GDH=
glucose dehydrogenase.
by column chromatography, the major (2S,3S) isomer 5a was
obtained in 61% yield, >99% de, 98% ee. The minor isomer is
the (2R,3S) compound 5d (>99% ee). After separation of the
diastereomers by column chromatography, the major (2R,3R)
isomer 5c was obtained in 54% yield, >99% de, 99% ee. The
minor isomer is the (2S,3R) compound 5b (94% ee).
b
Adv. Synth. Catal. 2019, 361, 1–5
2
© 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
��
These are not the final page numbers!

COMMUNICATIONS
asc.wiley-vch.de
over MgSO4, filtered, and concentrated. The product 1-(tert-
butyl) 2-methyl (2R,3S)-3-hydroxypyrrolidine-1,2-dicarboxylate
(3d) was obtained as a pale yellow oil (740 mg, 3.02 mmol,
73% yield, 93% de, >99% ee).
ketoester 4 compared to 2. Thus, KRED P1D3 and
KRED 119 were employed to generate the cis-3-
hydroxypipecolic acid isomers 5b and 5d with high
diastereo- and enantioselectivity (Figure 3). In the case
of the trans-hydroxy isomers, the best KREDs identi-
fied were highly enantioselective but only moderately
diastereoselective. However, the diastereomers could
be readily separated by silica gel chromatography,
providing straightforward access to the trans-hydrox-
ypipecolic acid compounds 5a and 5c in high
diastereo- and enantiopurity (>99% de, �98% ee).
In conclusion, we have demonstrated the enzymatic
and stereoselective synthesis of all isomers of 3-
hydroxyproline and 3-hydroxypipecolic acid from β-
ketoesters via a dynamic kinetic reduction approach.
This chemistry employs readily available ketoreductase
enzymes and cofactors, takes place under mild con-
ditions, is readily scalable, and allows for simple
product isolation. The ability to generate all possible
stereoisomers emphasizes the utility of enzymes for
stereodivergent synthesis. Furthermore, while we have
not performed protein engineering in this study, we
expect that the ketoreductases identified here may be
readily evolved for higher levels of selectivity and
activity toward any of these transformations.
References
[1] a) S. Harper, J. A. McCauley, M. T. Rudd, M. Ferrara,
M. DiFilippo, B. Crescenzi, U. Koch, A. Petrocchi,
M. K. Holloway, J. W. Butcher, J. J. Romano, K. J. Bush,
K. F. Gilbert, C. J. McIntyre, K. T. Nguyen, E. Nizi, S. S.
Carroll, S. W. Ludmerer, C. Burlein, J. M. DiMuzio,
D. J. Graham, C. M. McHale, M. W. Stahlhut, D. B.
Olsen, E. Monteagudo, S. Cianetti, C. Giuliano, V. Pucci,
N. Trainor, C. M. Fandozzi, M. Rowley, P. J. Coleman,
J. P. Vacca, V. Summa, N. J. Liverton, ACS Med. Chem.
Lett. 2012, 3, 332–336; b) J. Gillard, A. Abraham, P. C.
Anderson, P. L. Beaulieu, T. Bogri, Y. Bousquet, L.
Grenier, I. Guse, P. Lavallée, J. Org. Chem. 1996, 61,
2226–2231; c) J. M. Balkovec, D. L. Hughes, P. S.
Masurekar, C. A. Sable, R. E. Schwartz, S. B. Singh,
Nat. Prod. Rep. 2014, 31, 15–34; d) S. Suksamrarn, N.
Suwannapoch, N. Aunchai, M. Kuno, P. Ratananukul, R.
Haritakun, C. Jansakul, S. Ruchirawat, Tetrahedron
2005, 61, 1175–1180; e) J. D. Scott, R. M. Williams,
Angew. Chem. Int. Ed. 2001, 40, 1463–1465.
[2] S. P. Miller, Y.-L. Zhong, Z. Liu, M. Simeone, N.
Yasuda, J. Limanto, Z. Chen, J. Lynch, V. Capodanno,
Org. Lett. 2014, 16, 174–177.
Experimental Section
Example protocol using isopropanol as terminal reductant:
In a 500 mL jacketed flask were combined with stirring 220 mg
NADP and 198 mL 0.1 M phosphate buffer pH 7.0. Then,
without stirring and at room temperature, 550 mg KRED P1B2
were added and let dissolve over 30 min. The reaction mixture
[3] a) C. M. Taylor, R. Hardré, P. J. B. Edwards, J. Org.
Chem. 2005, 70, 1306–1315; b) C. L. Jenkins, L. E.
Bretscher, I. A. Guzei, R. T. Raines, J. Am. Chem. Soc.
2003, 125, 6422–6427; c) S. A. Lieblich, K. Y. Fang,
J. K. B. Cahn, J. Rawson, J. LeBon, H. T. Ku, D. A.
Tirrell, J. Am. Chem. Soc. 2017, 139, 8384–8387.
[4] D. M. Hudson, D. R. Eyre, Connect. Tissue Res. 2013,
54, 245–251.
[5] a) J.-O. Durand, M. Larchevêque, Y. Petit, Tetrahedron
Lett. 1998, 39, 5743–5746; b) J. H. Lee, J. E. Kang,
M. S. Yang, K. Y. Kang, K. H. Park, Tetrahedron 2001,
57, 10071–10076; c) J. Mulzer, A. Meier, J. Buschmann,
P. Luger, J. Org. Chem. 1996, 61, 566–572; d) C.
Herdeis, H. P. Hubmann, H. Lotter, Tetrahedron: Asym-
metry 1994, 5, 119–128.
°
was then heated to 30 C, and a solution of ketoester 2 (11.0 g,
45.2 mmol) in 22 mL IPA was then added slowly with stirring.
After 17 h, the reaction was cooled to room temperature, and
60 mL of 3:1 IPAC (isopropyl acetate):IPA were added with
slow stirring. The layers were separated, and the aqueous layer
was extracted twice more with 3:1 IPAC:IPA. The organic
layers were combined, washed with brine, dried over MgSO4,
filtered, and concentrated. The product 1-(tert-butyl) 2-methyl
(2S,3R)-3-hydroxypyrrolidine-1,2-dicarboxylate (3b) was iso-
lated as a yellow oil (10.3 g, 42.0 mmol, 93% yield, 99% de,
>99% ee).
[6] a) P. Hughes, J. Clardy, J. Org. Chem. 1989, 54, 3260–
3264; b) O. Poupardin, C. Greck, J.-P. Genêt, Synlett
1998, 1279–1281; c) T. P. Kumar, S. Chandrasekhar,
Synthesis 2012, 44, 2889–2894.
[7] J. C. Sheehan, J. G. Whitney, J. Am. Chem. Soc. 1963,
85, 3863–3865.
[8] For a review on the synthesis of 3-hydroxypipecolic
acids, see: A. Cochi, D. G. Pardo, J. Cossy, Eur. J. Org.
Chem. 2013, 809–829.
[9] a) O. Verho, J.-E. Bäckvall, J. Am. Chem. Soc. 2015,
137, 3996–4009; b) F. Rebolledo, J. González-Sabín, V.
Gotor, in: Stereoselective Synthesis of Drugs and Natural
Products (Eds.: V. Andrushko, N. Andrushko), John
Wiley & Sons, 2013, 1683–1711.
Example protocol using glucose as terminal reductant: In a
100 mL vessel, glucose (888 mg), NADP (40 mg), glucose
dehydrogenase (GDH, 80 mg), and KRED 119 (250 mg) were
dissolved in 36 mL 0.1 M sodium phosphate buffer pH 7.0. A
solution of ketoester 2 (1.0 g, 4.1 mmol) in DMSO (3 mL) was
°
then added, and the reaction was heated to 30 C with stirring.
Continuous pH adjustment was performed by automated
addition of 1 N NaOH to maintain the pH at 7.0. After 17 h,
reaction was cooled to room temperature, and 10 mL of 3:1
IPAC:IPA were added with slow stirring. The layers were
separated, and the aqueous layer was extracted three more times
with 3:1 IPAC:IPA. The organic layers were combined, washed
with brine, and concentrated. The resulting residue was re-
dissolved in EtOAc, washed with water (2×) and brine, dried
Adv. Synth. Catal. 2019, 361, 1–5
3
© 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
��
These are not the final page numbers!

COMMUNICATIONS
asc.wiley-vch.de
[10] G. A. Applegate, D. B. Berkowitz, Adv. Synth. Catal. [13] a) R. J. Moreau, E. J. Sorensen, Tetrahedron 2007, 63,
2015, 357, 1619–1632.
[11] a) J. Cooper, P. T. Gallagher, D. W. Knight, J. Chem.
6446–6453; b) J. Wang, Y. Hou, P. Wu, J. Chem. Soc.
Perkin Trans. 1, 1999, 2277–2280.
Soc. Chem. Commun. 1988, 8, 509–510; b) R. Bhide, R. [14] A. M. Hyde, Z. Liu, B. Kosjek, L. Tan, A. Klapars, E. R.
Mortezaei, A. Scilimati, C. J. Sih, Tetrahedron Lett.
1990, 31, 4827–4830; c) D. W. Knight, N. Lewis, A. C.
Share, D. Haigh, Tetrahedron: Asymmetry 1993, 4, 625–
628.
Ashley, Y.-L. Zhong, O. Alvizo, N. J. Agard, G. Liu, X.
Gu, N. Yasuda, J. Limanto, M. A. Huffman, D. M.
Tschaen, Org. Lett. 2016, 18, 5888–5891.
[12] W. Hüttel, Chem. Ing. Tech. 2013, 85, 809–817.
Adv. Synth. Catal. 2019, 361, 1–5
4
© 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
��
These are not the final page numbers!

COMMUNICATIONS
Stereodivergent Synthesis of 3-Hydroxyprolines and 3-Hy-
droxypipecolic Acids via Ketoreductase-Catalyzed Dynamic
Kinetic Reduction
Adv. Synth. Catal. 2019, 361, 1–5
C. K. Prier*, M. M.-C. Lo, H. Li, N. Yasuda