Scalable Synthesis of the Potent HIV Inhibitor BMS-986001 by Non-enzymatic Dynamic Kinetic Asymmetric Transformation (DYKAT)

Article abstract of DOI:10.1002/anie.201502290

Abstract Described herein is the synthesis of BMS-986001 by employing two novel organocatalytic transformations: 1)a highly selective pyranose to furanose ring tautomerization to access an advanced intermediate, and 2)an unprecedented small-molecule-mediated dynamic kinetic resolution to access a variety of enantiopure pyranones, one of which served as a versatile building block for the multigram, stereoselective, and chromatography-free synthesis of BMS-986001. The synthesis required five chemical transformations and resulted in a 44 % overall yield. Good dynamic: Described is the synthesis of BMS-986001 by employing two novel organocatalytic transformations: a highly selective pyranose to furanose ring tautomerization, and an unprecedented small-molecule-mediated dynamic kinetic asymmetric transformation (DYKAT) to access enantiopure pyranones. BMS-986001 was synthesized in five steps in an overall yield of 44 %. Bz=benzoyl.

Full text of DOI:10.1002/anie.201502290

Angewandte
Chemie
International Edition: DOI: 10.1002/anie.201502290
German Edition: DOI: 10.1002/ange.201502290
Organocatalysis
Scalable Synthesis of the Potent HIV Inhibitor BMS-986001 by Non-
Enzymatic Dynamic Kinetic Asymmetric Transformation (DYKAT)**
Adrian Ortiz,* Tamas Benkovics, Gregory L. Beutner, Zhongping Shi, Michael Bultman,
Jeffrey Nye, Chris Sfouggatakis, and David R. Kronenthal
Abstract: Described herein is the synthesis of BMS-986001 by
employing two novel organocatalytic transformations: 1) a
highly selective pyranose to furanose ring tautomerization to
access an advanced intermediate, and 2) an unprecedented
small-molecule-mediated dynamic kinetic resolution to access
a variety of enantiopure pyranones, one of which served as
a versatile building block for the multigram, stereoselective,
and chromatography-free synthesis of BMS-986001. The syn-
thesis required five chemical transformations and resulted in
a 44% overall yield.
S
ince the FDA approval of azidothymidine (AZT) in 1987
as the first NRTI (nucleoside reverse transcriptase inhibitor)
treatment of the HIV virus, the scientific community has been
continuously searching for safer and more efficacious thera-
pies. The last 20 years of research in this area has resulted in
vastly improved therapeutics and treatment strategies.[1]
Despite these improvements, viral drug resistance[2] and
side-effects to the prescribed therapies remain outstanding
issues.[3] BMS-986001 (1) is a thymidine NRTI which has been
developed to maintain the in vivo antiviral activity demon-
strated by other NRTIꢀs, but lacks the associated toxicity side
effects. Recent clinical data has shown this investigational
therapy to be effective in reducing viral load while exhibiting
a significantly improved safety profile, when compared to the
standard of care.[4] To aid the development of this compound,
a unique, expedient, and scalable synthesis of 1 was required.
The development of this new route resulted in several
interesting observations, and the development of two organo-
catalytic transformations to set key structural and stereo-
chemical elements as described herein.
Figure 1. Retrosynthetic analysis of BMS-986001 (1). Bz=benzoyl.
quent 1,2-addition of the alkyne moiety would provide the
pyranose 6b. Next, a ring tautomerization/acylation sequence
and a subsequent Vorbrꢁggen reaction could be employed to
convert the pyranose ring into the desired furanose nucleo-
side 12. Finally, oxidation of the thioether and thermolysis of
ꢀ
the resulting sulfilimine could install the required C2 C3
dehydrofuranose moiety present in 1. The success of this
strategy hinged on the accessibility of optically enriched (S)-3.
Similar structural pyranone derivatives have demonstrated
broad utility as versatile building blocks in organic synthesis,[5]
and as key components in the development of new synthetic
methods.[6] However, all previous approaches to (S)-3 and
similar derivatives suffered from unsatisfactory yields,[7] and
required the use of either chiral chromatography, derivatiza-
tion, or enzyme-mediated resolution to impart high enantio-
purity. Our previously published work (Scheme 1a) employed
an enzymatic resolution by destructive transesterification to
deliver (S)-3a in high purity and enantioselectivity, but in
moderate overall yield (26% from 2a).[8]
Retrosynthetic analysis of the targeted structure 1 led us
to define pyranone (S)-3 as the key enantioenriched building
block from which a substrate-controlled, diastereoselective
synthesis was envisioned (Figure 1). In the forward sense,
a diastereoselective 1,4 addition of an arylthiol and subse-
To increase both efficiency and overall yield, a dynamic
kinetic asymmetric transformation (DYKAT) was considered
for the acylation of the racemic lactol 2a (Scheme 1b).
Limited precedence existed for this transformation biocata-
lytically, and most reports achieved only low to moderate
enantioselectivity.[9] In fact, in our hands, screening of more
than 100 enzymes led to either the undesired isomer [(R)-
3a][10] or (S)-3a with low selectivity. Despite the emergence of
a number of catalysts shown to resolve secondary alcohols by
way of non-enzymatic selective acylations,[11] to the best of
our knowledge, there have been no reports of a small
molecule facilitating this important transformation on
a lactol. An initial screen of organocatalysts resulted in low
levels of conversion and/or selectivity. Surprisingly, the best
result was achieved using levamisole (A), an inexpensive and
[*] Dr. A. Ortiz, Dr. T. Benkovics, Dr. G. L. Beutner, Dr. Z. Shi,
Dr. M. Bultman, Dr. J. Nye, Dr. C. Sfouggatakis, Dr. D. R. Kronenthal
Chemical Development, Bristol-Myers Squibb
1 Squibb Drive, New Brunswick, NJ 08903 (USA)
E-mail: Adrian.Ortiz@bms.com
[**] We thank Dr.’s R. Parsons, R. Waltermire, and M. D. Eastgate for
supporting this work, Dr.’s Charles Pathirana and David Ayers for
assistance with structural elucidation, Merrill Davies for help with
chiral separation, and Jonathan Marshall for MS analysis. We would
also like to thank Prof. Phil Baran for helpful discussions in the
drafting of this manuscript.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201502290.
Angew. Chem. Int. Ed. 2015, 54, 1 – 5
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1
These are not the final page numbers!

.
Angewandte
Communications
Table 2: Levamisole-mediated DYKATof the selected pyranose lactols 2a
and 4a–d.[a]
Starting material Product R1
R2
R3
Yield [%][b] ee [%][c]
2a
4a
4b
4c
4d
3c
5a
5b
5c
5d
H
Me
H
H
H
H
H
Me
H
H
H
H
H
91
91
90
88
86
82
72
55
Me 96
Ph 96
[a] Reaction conditions: levamisole (0.05 equiv), isobutyric anhydride
(1.2 equiv), 20 mL toluene/g substrate, 1 h, RT. [b] Determined after
isolation by chromatography. [c] Determined by HPLC analysis on
a chiral stationary phase.
Scheme 1. Synthetic routes to (S)-3 by enzymatic resolution (a) and by
non-enzymatic DYKAT (b). DMAP=4-(N,N-dimethylamino)pyridine.
materials in numerous total syntheses.[5] Application of the
optimized reaction conditions to a series of substituted
pyranose lactols (4a–d) demonstrated good tolerance for
alkyl substitution at C2 (4c!5c), C3 (4b!5b), and C5 (4a!
5a). While aryl substitution at C2 (4d!5d) led to reduced
enantioselectivity (55% ee), the high yield and crystallinity of
the product still enabled a practical alternative (> 99% ee
after recrystallization from methyl tert-butyl ether) to enrich-
ment by chiral chromatography. Further studies and mecha-
nistic understanding of this transformation are currently
under investigation.
commercially available small molecule (Scheme 1b; for
details see Table 1), which currently has found only limited
utility in kinetic resolutions.[12]
Optimization of this transformation ensued with the use
of Bz2O (Table 1, entry 1), which we found to be suboptimal,
thus confirming previous reports.[13] This shortcoming
Table 1: Optimization of the levamisole-mediated DYKAT of 2a (see
Scheme 1 for reaction equation).[a]
Entry
R
3
Solvent
Conv. [%][b]
ee [%][c]
With a preparation of enantiopure (S)-3d secured, we
proceeded with its utilization in the synthesis of 1 as outlined
in Scheme 2. Diastereoselective thioconjugate addition of p-
thiocresol with (S)-3d resulted in the C2 sulfide 6a (> 50:1),
which was immediately reacted with lithio-TMS-acetylene
(ꢀ788C) in a single-pot operation to afford the crystalline
pyranose 6b in high yield and selectivity (70% yield, > 20:1
d.r.). Installation of the aryl sulfide at this juncture was
designed to play a critical role throughout the synthesis, thus
acting as a relay for stereochemical information. With the
crucial remote C4 stereocenter successfully installed, we
proceeded to confront what would be one of the most
challenging transformations in the synthesis: selective pyr-
anose to furanose ring-chain tautomerization (6b!11).
Carbohydrate ring-chain tautomerization is known to be
substrate-dependent and effected by a number of external
variables (i.e., solvent, pH, temperature, and pressure).[14]
Although initially discouraged by literature accounts describ-
ing the general preference of carbohydrates to exist in the
pyranose form,[15] we were hopeful a solution would be
discovered. Initial hydrolysis of 6b (aq. HCl, CH3CN, 208C,
10 h) provided access to the furanose form 8a/b, but also
vastly increased the complexity of the system by generating
a rapidly equilibrating mixture of four lactol isomers (7a/
1
2
3
Bz
Ac
CH2Ph
iPr
CH2Ph
3a
3b
3d
3c
3d
t-amyl alcohol
t-amyl alcohol
t-amyl alcohol
toluene
60
>95
>80
>95
>95
33
60
60
88
79
4
5[a]
toluene
[a] Reaction conditions: levamisole (0.05 equiv), anhydride (1.2 equiv),
10 mL solvent/g substrate, 1 h, RT. [b] Determined by HPLC analysis.
[c] Determined by HPLC analysis on a chiral stationary phase.
prompted us to explore alternative alkyl anhydrides, resulting
in increased reaction rates, albeit with modest enantioselec-
tivities (entries 2 and 3; 60–70% ee). Employment of a non-
polar solvent (toluene) and isobutyric anhydride allowed us to
achieve our highest level of enantioenrichment (entry 4).
However, the acylated product 3c was noncrystalline, thus
preventing further enantioenrichment by crystallization. For-
tunately, the crystalline phenylacetate derivative 3d provided
similar initial selectivity (79% ee crude) with the benefit of
further enrichment by crystallization (99% ee), a small
compromise which enabled its preparation on multigram
scale (entry 5) from racemic 2a.
While extremely pleased with the DKR results on 2a, we
were curious about the scope of this important and previously
unprecedented transformation (Table 2). Similar acylated
pyranose lactols have been broadly utilized as chiral building
blocks in organic chemistry and have functioned as starting
1
b$8a/b). H NMR studies conducted on this lactol mixture
revealed a mild preference for the furanose form in nonpolar
solvents ([D6]benzene = 60% versus CD3CN = 5%), presum-
2
www.angewandte.org
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2015, 54, 1 – 5
These are not the final page numbers!

Angewandte
Chemie
Scheme 2. Synthesis of BMS-986001 (1) from (S)-3d. Reagents and conditions: a) p-thiocresol (1.1 equiv), iPr2EtN (0.05 equiv), toluene, 208C,
30 min, then lithio-TMS-acetylene (2.5 equiv), THF (3.0 equiv), ꢀ788C, 1 h, 70%; b) 1n HCl (0.5 equiv), CH3CN/H2O (5:1), 208C, 10 h, then Bz2O
(3.0 equiv), levamisole (0.1 equiv), toluene, 208C, 48 h, then K2HPO4 (3.0 equiv), nBu4HSO4 (0.2 equiv), H2O (3.0 equiv), toluene, 08C, 1.5 h,
91%; c) bis(TMS)thymine (1.4 equiv), TMSOTf (1.4 equiv), CH3CN, 10 to 208C, 4.5 h, 86%, (13:1 d.r.); d) Chloramine-T (1.2 equiv), AcOH
(0.1 equiv), CH3CN, 208C, 2.5 h, then nBuOH, 958C, 3 h, 87%; e) DBU (0.05 equiv), MeOH, 608C, 15 h, 92%. DBU=1,8-diazabicyclo[5-4-
0]undec-7-ene, Tf=trifluoromethanesulfonyl, THF=tetrahydrofuran, TMS=trimethylsilyl, Ts=4-toluenesulfonyl.
ably because of a key hydrogen bond between the ring oxygen
atom and the C5 hydroxy group.[16] The discovery of this
solvent effect on the tautomer equilibrium prompted us to
replace CH3CN with toluene prior to advancing. Extensive
optimization of the selective benzoylation, led to the discov-
ery that 8a/b could be trapped as its bis(benzoate) derivative
(10a) using levamisole (10 mol%, Bz2O, toluene, 208C, 48 h)
as a nucleophilic organocatalyst. These reaction conditions
resulted in remarkable selectivity (98:2, furanose/pyranose,
> 50:1 a:b), and facilitated a highly successful ring-chain
tautomerization transformation. After mild and selective
hydrolytic removal of the TMS group (K2HPO4, nBu4NHSO4,
H2O, toluene, 08C), the crystalline a-Bz-furanose 11 was
isolated in 91% overall yield from 6b. Next, a Vorbrꢁggen
reaction with bis(TMS)thymine furnished the nucleoside 12 in
86% yield and with high b-selectivity (13:1 d.r.), presumably
influenced by the presence of the C2 aryl sulfide.[17] The
nucleoside 12 was then converted into the dehydrofuranose
14 through a one-pot, two-step procedure involving selective
Chloramine T[18] mediated sulfur oxidation to give the
corresponding sulfilimine 13 followed by thermolysis under
relatively mild reaction conditions (958C, 3 h, 87% yield). At
this point, the crucial role of the C2 aryl sulfide becomes
clear: initially, it provided conformational control and relayed
facial selectivity guiding the introduction of the C4 acetylene
(6a!6b), later it provided assistance governing the b-
selective nucleoside addition (11!12), and now it has
served as a handle for the installation of the dehydrofuranose
backbone (12!14). Finally, HIV-inhibitor BMS-986001 (1)
was obtained in excellent yield (92%) through a DBU-
catalyzed transesterification.
In summary, we have demonstrated a novel and highly
efficient synthesis of BMS-986001 (1) in five steps and 44%
overall yield from (S)-3d. Highlighted in this work is the
unexpected utility of the inexpensive and readily available
organocatalyst, levamisole (A), first as a means to prepare
enantiomerically enriched building block (S)-3d (by an
unprecedented non-enzymatic DYKAT of lactol 2a), and
later as a highly effective catalyst facilitating the ring-chain
tautomerization of a pyranose into a furanose (6b!11).
Complete studies regarding the development of these leva-
misole-mediated transformations, along with an account
covering the application of this chemistry in the preparation
more than 250 kg of 1 will be disclosed in due course.[19]
Keywords: inhibitors · kinetic resolution · organocatalysis ·
synthesis design · tautomerism
[1] For a recent review, see: S. Broder, Antiviral Res. 2010, 85, 1 – 18.
[2] E. L. Asahchop, M. A. Wainberg, R. D. Sloan, C. L. Tremblay,
Antimicrob. Agents Chemother. 2012, 56, 5000 – 5008, and
references therein.
[3] Panel on Antiretroviral Guidelines for Adults and Adolescents.
Guidelines for the Use of Antiretroviral Agents in HIV-1-
Angew. Chem. Int. Ed. 2015, 54, 1 – 5
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3
These are not the final page numbers!

.
Angewandte
Communications
infected Adults and Adolescents: 2012 Update. Washington,
DC: U.S. Department of Health and Human Services, 2012.
[4] a) L. Cotte, P. Dellamonica, F. Raffi, Y. Yazdanpanah, J.-M.
Michel Molina, F. Bouꢂ, Y. Urata, H. P. Chan, L. Zhu, I. Chang,
R. Bertz, G. J. Hanna, D. M. Grasela, C. Hwang, J. Acquired
Immune Defic. Syndr. 2013, 63, 346 – 354; b) F. Wang, O. P. Flint,
Antimicrob. Agents Chemother. 2013, 57, 6205 – 6212; c) G.
Yang, G. E. Dutschman, C. J. Wang, H. Tanaka, M. Baba, K. S.
Anderson, Y. C. Cheng, Antiviral Res. 2007, 73, 185 – 191;
d) G. E. Dutschman, S. P. Grill, E. A. Gullen, K. Haraguchi, S.
Takeda, H. Tanaka, M. Baba, Y. C. Cheng, Antimicrob. Agents
Chemother. 2004, 48, 1640 – 1646.
[5] a) H. Takayama, Z. Jia, L. Kremer, J. O. Bauer, C. Strohmann, S.
Ziegler, A. P. Antonchock, H. Waldman, Angew. Chem. Int. Ed.
2013, 52, 12404 – 12408; Angew. Chem. 2013, 125, 12630 – 12634;
b) M. A. Ali, N. Bhogal, J. B. C. Findlay, C. W. G. Fishwick, J.
Med. Chem. 2005, 48, 5655 – 5658; c) K. L. Jackson, J. A.
Henderson, J. C. Morris, H. Motoyoshi, A. J. Phillips, Tetrahe-
dron Lett. 2008, 49, 2939 – 2941; d) K. C. Nicolaou, M. O.
Fredrick, A. C. B. Burtuloso, R. M. Denton, F. Rivas, K. P.
Cole, R. J. Aversa, R. Gibe, T. Umezawa, T. Suzuki, J. Am.
Chem. Soc. 2008, 130, 7466 – 7476; e) R. Jones, M. J. Kriche, Org.
Lett. 2009, 11, 1849 – 1851.
[8] T. Benkovics, A. Ortiz, Z. Guo, A. Goswami, P. Deshpande, Org.
Synth. 2014, 91, 293 – 306.
[9] M. Van den Heuvel, A. D. Cuiper, H. Van der Deen, R. M.
Kellogg, B. L. Feringa, Tetrahedron Lett. 1997, 38, 1655 – 1658.
[10] A. Yamazaki, Y. Iriyama, Y. Ootsuka, H. Kurihara (Nissan
Chemical Industries), WO201126082, 2011.
[11] G. C. Fu, Acc. Chem. Res. 2004, 37, 542 – 547.
[12] X. Li, V. Birman, Org. Lett. 2006, 8, 1351 – 1354.
[13] See Table 3 in Ref. [12].
[14] a) For a review see: P. R. Jones, Chem. Rev. 1963, 63, 461 – 487;
b) L. Guasch, M. Sitzmann, M. C. Nicklaus, J. Chem. Inf. Model.
2014, 54, 2423 – 2432; Y. C. Martin, J. Comput.-Aided Mol. Des.
2009, 23, 693 – 704.
[15] Advances in Carbohydrate Chemistry and Biochemistry, Vol. 24
(Eds.: M. L. Wolfrom, R. S. Tipson), Academic Press, New York,
1969, pp. 43 – 63.
[16] Calculated most stable forms of 7a/b and 8a/b lactol mixture in
the gas phase, [B3LYP/6-31G(d)]. 8a < 8b (2.0 kcalmolꢀ1) < 7b
(2.4 kcalmolꢀ1) < 7a (2.8 kcalmolꢀ1).
[17] L. J. Wilson, D. Liotta, Tetrahedron Lett. 1990, 31, 1815 – 1818.
[18] A. L. Marzinzik, K. B. Sharpless, J. Org. Chem. 2001, 66, 594 –
596.
[19] “Sulfilimine and Sulfoxide Methods for the Preparation of
Festinavir”: A. Ortiz, T. Benkovics, Z. Shi, P. P. Deshpande, Z.
Guo, D. R. Kronenthal, C. Sfouggatakis, WO2013177243A1,
2013.
[6] a) N. Z. Burns, M. R. Whitten, E. N. Jacobsen, J. Am. Chem. Soc.
2011, 133, 14578 – 14581; b) T. C. Coombs, M. D. Lee, H. Wong,
M. Armstrong, B. Cheng, W. Chen, A. F. Moretto, L. S.
Liebeskind, J. Org. Chem. 2008, 73, 882 – 888; c) A. Orue, E.
Reyes, J. L. Vicario, L. Carillo, U. Uria, Org. Lett. 2012, 14,
3740 – 3743.
[7] M. P. Georgiadis, K. E. Albizati, T. M. Georgiadis, Org. Prep.
Proced. 1992, 24, 95 – 118.
Received: March 11, 2015
Published online: && &&, &&&&
4
www.angewandte.org
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2015, 54, 1 – 5
These are not the final page numbers!

Angewandte
Chemie
Communications
Organocatalysis
Good dynamic: Described is the synthe-
sis of BMS-986001 by employing two
novel organocatalytic transformations:
a highly selective pyranose to furanose
ring tautomerization, and an unprece-
dented small-molecule-mediated
dynamic kinetic asymmetric transforma-
tion (DYKAT) to access enantiopure pyr-
anones. BMS-986001 was synthesized in
five steps in an overall yield of 44%. Bz=
benzoyl.
A. Ortiz,* T. Benkovics, G. L. Beutner,
Z. Shi, M. Bultman, J. Nye,
C. Sfouggatakis,
D. R. Kronenthal
&&&& — &&&&
Scalable Synthesis of the Potent HIV
Inhibitor BMS-986001 by Non-Enzymatic
Dynamic Kinetic Asymmetric
Transformation (DYKAT)
Angew. Chem. Int. Ed. 2015, 54, 1 – 5
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5
These are not the final page numbers!