Five-Step Enantioselective Synthesis of Islatravir via Asymmetric Ketone Alkynylation and an Ozonolysis Cascade

Article abstract of DOI:10.1002/chem.202003091

A 5-step enantioselective synthesis of the potent anti-HIV nucleoside islatravir is reported. The highly efficient route was enabled by a novel enantioselective alkynylation of an α,β-unsaturated ketone, a unique ozonolysis-dealkylation cascade in water,

Full text of DOI:10.1002/chem.202003091

Accepted Article
Accepted Article
Title: Five-Step Enantioselective Synthesis of Islatravir via Asymmetric
Ketone Alkynylation and an Ozonolysis Cascade
Authors: Aaron Whittaker, Niki R. Patel, Mark A. Huffman, Xiao Wang,
Bangwei Ding, Mark McLaughlin, Justin A. Newman, Teresa
Andreani, Kevin M. Maloney, and Heather C. Johnson
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Chem. Eur. J. 10.1002/chem.202003091
Link to VoR: https://doi.org/10.1002/chem.202003091
0
1/2020

Chemistry - A European Journal
10.1002/chem.202003091
RESEARCH ARTICLE
Five-Step Enantioselective Synthesis of Islatravir via Asymmetric
Ketone Alkynylation and an Ozonolysis Cascade
Niki R. Patel, Mark A. Huffman, Xiao Wang, Bangwei Ding, Mark McLaughlin, Justin A. Newman,
Teresa Andreani, Kevin M. Maloney, Heather C. Johnson* and Aaron M. Whittaker*
N. R. Patel, M. A. Huffman, X. Wang, B. Ding, M. McLaughlin, J. Newman, T. Andreani, K. M. Maloney, H. C. Johnson and A. M. Whittaker
Department of Process Research and Development, MRL, Merck & Co., Inc., Rahway, NJ 07065 USA
[*]
Correspondence to aaron.whittaker@merck.com and heather.johnson@merck.com
Supporting information for this article is given via a link at the end of the document.
Abstract: A 5-step enantioselective synthesis of the potent anti-HIV
We proposed a general strategy of using an olefin as a masked
aldehyde to cleanly reveal 2-ethynylglyceraldehyde-3-phosphate
nucleoside islatravir is reported. The highly efficient route was
enabled by
a novel enantioselective alkynylation of an α,β-
2
3
via ozonolysis (Scheme 1). The chiral allylic alcohol precursor
unsaturated ketone, a unique ozonolysis-dealkylation cascade in
water, and an enzymatic aldol-glycosylation cascade.
could arise from an enantioselective alkynylation of an αʹ-
phosphorylated-α,β-unsaturated ketone 4, which, in turn, can be
readily accessed from commodity chemicals. To accomplish the
proposed route, we needed to develop an unprecedented
enantioselective alkynylation, demonstrate a safe and selective
ozonolysis, and understand the handling of highly functionalized
and unstable intermediates.
Introduction
According to the World Health Organization, approximately 37
million people are currently living with HIV, with 1.8 million newly
infected in 2017 alone.[1] While nucleoside reverse transcriptase
inhibitors (NRTIs) have served as a cornerstone of highly active
antiretroviral therapy (HAART) for HIV, novel drug classes are
needed to improve drug adherence through improved safety,
long-term efficacy, and simplified dosing regimens.[2] One
particularly promising molecule is islatravir (1, EFdA or MK-
8591),
a
long-acting nucleoside reverse transcriptase
translocation inhibitor (NRTTI) that binds to the polymerase
active site of HIV reverse transcriptase.[3] This investigational
new drug has been shown to suppress HIV viral load for
extended duration at low doses, opening the possibility of
implantable formulations that have promising implications for
drug adherence.[4]
A number of approaches to the total synthesis of islatravir (1)
Scheme 1. Retrosynthetic approach to islatravir (1).
have been reported, including 22-step[5] and 18-step syntheses
[6]
from the Kuwaraha group and
a 15-step synthesis from
[7]
McLaughlin and co-workers. Considering the recent clinical
promise surrounding islatravir (1), we sought a more efficient
synthesis to enable an inexpensive manufacturing scale supply.
To this end, we recently disclosed a biocatalytic route to
islatravir involving a highly efficient aldol-glycosylation one-pot
biocatalytic cascade.[8] During the development of the
biocatalytic route, we simultaneously explored alternative routes
leveraging the highly efficient enzymatic glycosylation, but
Recognizing the challenges associated with developing an
asymmetric enone alkynylation, we prioritized the synthesis of a
library of electronically and sterically diverse substrates at the
outset. Benzylideneacetone derivatives were considered an
attractive, inexpensive starting point since the αʹ-
phosphorylated-α,β-unsaturated ketones could be accessed in
two steps from commodity chemicals.[11]
halogenation of electon-rich ketones (5),[12] or a Wittig olefination
of electron-defficient aldehydes (6) using commercially
available ylide[13] provided the α-chloroketones that were readily
Electrophilic
accessing key intermediates from chemocatalytic approaches.[
9]
a
Herein, we describe the shortest and highest yielding of these
approaches enabled through novel enantioselective
a
[10]
alkynylation of an alpha-beta unsaturated ketone.
N
converted to the α-phosphates (4) by an S 2 reaction (Scheme
2 [
).
14]
Results and Discussion
Our retrosynthetic approach intercepts the biocatalytic cascade
at the key intermediate, 2-ethynylglyceraldehyde-3-phosphate, 2.
1
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.202003091
RESEARCH ARTICLE
variety of ligands (via high-throughput experimentation) to
induce enantioselective alkynylation of enone 4b by
ethynylmagnesium chloride (7). Chiral ligand classes explored
include BINOLs and chiral diols, PyBox, salens and salans.
2
Metal additives such as LiHMDS, ZnR (R = Me, Et) and triflate
salts were also screened.
Table 2. Identification of chiral ligand and reaction conditions
Scheme 2. Two-step access to α-phosphorylated benzalacetone derivatives.
BTMACl I = benzyltrimethylammonium dichloroiodate.
2
Investigation into the racemic alkynylation of the resulting
ketones using ethynyl magnesium chloride (7) quickly revealed
that the diethyl phosphate derivatives were unstable toward
strongly nucleophilic conditions, while the bulky tert-butyl
phosphates proceeded quantitatively (Table 1, 4a vs 4b).[15]
Further advantages to employing tert-butyl phosphates are the
ease of deprotection using mild acid (vide infra), and the
unexpected observation that most of the tert-butyl phosphate
containing compounds were crystalline solids. Less rigid alkyl
phosphates (eg. 4a) often exist as oils having implications on
stability, as well as precluding bulk purification via
recrystallization. Finally, we discovered that alkynylation of
electron rich substrates provided products with improved stability
which is reflected in the isolated yields compared to the NMR
yields (3c vs 3g).
Entry
ligand
L1
e.r.
conditions
1
2
3
4
5
6
44:56
42:58
79:21
82:18
77:23
69:31
50 mol% L1, no metal additive
50 mol% L2, 50 mol% LiHMDS
50 mol% L3, 2 equiv 7
L2
L3
L3
As entry 3, with 25 mol% LiHMDS
As entry 4, ethynylMgBr instead of 7
As entry 4, toluene as solvent
Table 1. Racemic alkynylation of various α-phosphorylated ketones.
L3
L3
>90% conversion to 3b observed in all conditions shown in table.
Most ligand classes evaluated resulted in poor enantioselectivity
and/or conversion (see SI), including salen ligands, which have
previously been used in enantioselective alkynylation of ketones
SM
Ar
Ph
R
Et
allylic alcohol
yield[a]
57 (65)
96
(
Table 2, entry 1, L1).[16c, 17] While related reduced ligand L2
4
4
4
4
4
4
4
a
b
c
d
e
f
3a
3b[b]
3c[b]
3d
provided similar enantioselectivity (Table 2, entry 2), the N-
methyl salan ligand L3 was identified as a promising lead,
providing >95% conversion with 79:21 e.r., indicating the
importance of the N-Me group on the ligand (Table 2, entry 3).
Salan ligands incorporating the diaminocyclohexane backbone
have previously been used for various asymmetric synthetic
transformations, including organogallium or organozinc addition
to aldehydes, when paired with metals such as Al, V, Zn, Ti, Nb
and Zr.[ LiHMDS as an additive resulted in higher er (Table 2,
entry 4), but higher loadings led to an erosion in the product
enantioselectivity (e.g. 64:36 er at 100 mol% LiHMDS).
Additionally, the use of ethynylmagnesium bromide instead of
the corresponding chloride (7) resulted in a lower e.r. (Table 2,
entry 5). Use of 2-MeTHF or MTBE as solvent resulted in similar
enantioselectivities, while the selectivity was worse in toluene,
Ph
tBu
4-OMeC
6
H
4
tBu
93
tBu
naphthyl
93
4-FC
4-CF
4-CNC
6
H
4
tBu
tBu
tBu
3e
88
3
C
6
H
4
3f
81 (96)
77 (93)
18]
g
[
6
H
4
3g
a] Isolated yields shown with NMR assay yield in
parenthesis. [b] Halogenation, phosphorylation and alkynylation
run on > 400 g scale.
With a library of electronically diverse substrates with stability
toward alkynylation designed into the structure, we pursued an
enantioselective alkynylation. In contrast to the alkynylation of
aldehydes, enantioselective alkynylation of ketones is more
challenging, and efforts involving alkynylation of enones have
only focused on 1,4-addition.[16] Accordingly, we evaluated a
3
CF Ph and CPME (Table 1 and SI). Cooling the reaction to –
20 °C failed to further improve the er.
We next investigated the impact of the substrate phenyl
electronic properties upon the enantioselectivity of the product.
The presence of electron-donating aryl groups led to higher
enantioselectivities, with a negative Hammett correlation (Figure
2
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.202003091
RESEARCH ARTICLE
1
). The small magnitude of the ρ value indicates this is a minor
t
5
6
7
R = Bu, R’ = H (L4)
88:12
49:51
90:10
As above
As above
effect: para-methoxy substitution (substrate 4c) provided the
highest enantioselectivity of 83:17 e.r. under these conditions,
compared with 80:20 and 74:26 when
respectively.
t
t
R = Bu, R’ = Bu
R = Ph or CN,
L4
100 mol% L4; 50 mol%
LiHMDS, 3 eq. 7
We next sought a deeper mechanistic understanding of the new
alkynylation reaction. Both the salan-mediated and racemic
alkynylations of substrates 4b and 4c proceeded rapidly to
>95% conversion within 30 seconds at 0 °C. Attempts to further
cool the reaction to monitor kinetics were hampered by
precipitation of the catalyst at ca. –30 °C. Although the reaction
was too fast to monitor kinetics by conventional methods, clues
into the nature of the alkynylation catalyst were provided by
varying the amount of reagent 7 while holding 4b (1 equiv), L3
(0.5 equiv) and LiHMDS (0.25 equiv) constant. With 0.5 equiv 7,
no reaction was observed within 1 hour, but with 1 equiv, 13%
conversion was observed in the same timeframe. With 1.5 equiv,
60% conversion was observed, and 2 equiv afforded complete
conversion. These results suggest that, with 0.5 equiv L3,
approximately 1 equiv of 7 is sequestered in catalyst formation,
implying that the catalyst structure involves 2 Mg ions per salan.
Figure 1. Hammett plot depicting enantiomeric ratio vs σpara of the substrate.
For point a, aryl group = 2-naphthyl. Reactions conducted at 4 °C.
To explore the effect of the ligand upon the alkynylation of 4c,
modifications of L1 were conducted. As observed in Jacobsen’s
Mn-salen-catalyzed epoxidation,[19] the electronic properties of
the aryl group on the ligand played an important role in
controlling the product enantioselectivity. The employment of
electron-donating substituents on the 4-position of the phenol
rings generally promoted higher enantioselectivities in the
reaction (Table 3). Ortho- substituents (e.g. 2,4-di-Cl or 2,4-di-
To probe the involvement of the alkynyl group in catalyst
formation, experiments were conducted in which different
Grignard reagents were added sequentially to salan L3 followed
1
2
by exposure to the enone 4b (Table 4). When R = ethyl and R
= ethynyl (Table 4, Entry 3), the major product is alkynylation
1
2
with a similar ee as when R = R = ethynyl (Table 4, Entry 1).
This result suggests that the active catalyst is formed from the
1
addition of R MgCl whether using EtMgCl or 7. Entry 2 suggests
t
Bu substitution patterns) significantly decreased the er (Table 3
that the Grignard reagent (7) is consumed mostly for complex
formation, and then EtMgCl adds to the ketone. Importantly for
scale-up considerations, utilization of EtMgCl instead of (7) for
catalyst formation significantly improves the safety profile
(avoiding liberation of acetylene gas), volumetric efficiency and
cost of the catalyst formation step.[
and SI), which may suggest dimer formation is required for
catalysis and this process is hindered by the presence of steric
bulk proximal to the metal (vide infra). The optimal combination
of ligand L4 with substrate 4c achieved an er of 90:10 for
alkynylated product 3c (Table 3, entry 7).
20]
Table 3. Effect of ligand electronics and reaction conditions upon the
enantioselectivity of the alkynylation of 4c
Table 4. Mixed Grignard experiments
Entry
R1
R2
alkynylation[a]
%
alkylation[a]
%
alkynylation
e.r.
1
2
3
Ethynyl
Ethynyl
Ethyl
Ethynyl
Ethyl
100
29
-
71
7
79:21
ND[b]
Ethynyl
93
79:21
Entry
ligand
e.r.
conditions
[a] Relative % alkynylation vs alkylation shown, determined by SFC analysis.
1
2
3
4
R = H, R’ = H (L3)
R = Me, R’ = H
R = Cl, R’ = H
84:16
84:16
74:26
86:14
As above
As above
As above
As above
[b] Not determined due to overlapping side products in the SFC
chromatogram.
A non-linear effect was observed in the alkynylation product
ee when varying the ee of salan L4 (Figure 2). Simulations of
R = OMe, R’ = H
) model[21] fit the experimental data well,
Kagan’s ML
consistent with a dimeric catalyst species being involved in the
2 2
(or (ML)
3
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.202003091
RESEARCH ARTICLE
enantioselectivity-determining step. A similar non-linear effect
The solution state NMR spectra of complex 8 are consistent with
the solid-state structure. Two 1:1 sets of salan signals were
observed, indicating the existence of two different conformations
of salan subunits (as was observed in the solid-state structure,
see SI), and both subunits featured a heavily shielded N-
was also observed with ligand L3. This observation is in contrast
to V-salan catalyzed oxidation reactions,[
18d]
or Cozzi’s Ti-salen
catalyzed ketone alkynylation,[16c] in which monomeric catalyst
species were proposed, although many other dimeric salen
complexes have been used previously in asymmetric
1
13
methine (δ
H/ C: 2.10/55.0, 1.63/66.0 ppm). NMR
[22]
catalysis.
spectroscopy of freshly-made in situ salan/7 mixtures shows
peaks matching the predicted chemical shifts of complex 8 (see
SI). This shows that LiHDMS is not required for formation of 8.
Several other species are present within the in situ formed
catalyst mixture, assigned by DOSY experiment as dimers with
similar molecular sizes to structure 8. Upon addition of Li-based
4
additives LiBF or LiTMP (which give similar enantioselectivities
as LiHMDS, Table 2), dimeric species (including complete 8)
remain, albeit with differing distributions/chemical shifts. While
we cannot definitively rule out a monomeric active catalyst, the
available evidence points to a dimeric active catalyst in which
the coordination can be subtly influenced by lithium-based
additives to provide a small increase in the e.r. of the obtained
product. Moreover, the evidence is consistent with ethynyl
magnesium bromide producing lower enantioselectivity than the
corresponding chloride (7) for the alkynylation (vide supra) likely
due to significant alterations to the dimeric catalyst structure
when replacing chloride with bromide.
2
Figure 2. Non-linear effect using L4 and Kagan’s ML model.
Under optimized conditions (3.75 equiv 7, 0.90 equiv L4,
and 0.57 equiv LiHMDS), 1 g of α,β-unsaturated ketone 4c
afforded chiral product 3c in 96% yield and 90:10 er. With the
enantioenriched allylic alcohol in hand, dealkylation of the tert-
butyl groups and ozonolysis of the olefin would provide the
desired 2-ethynylglyceraldehyde-3-phosphate (2). An initial
investigation showed that the acidic conditions required to
catalyze dealkylation of the phosphate also catalyzed the
isomerization of the allylic alcohol to form a conjugated enyne
Consistent with the observed non-linear effect, crystallographic
evidence of dimeric species was obtained: addition of LiHMDS
and 7 in THF to ligand L4, followed by layering the solution with
hexanes at 5 °C, yielded colorless single crystals suitable for X-
ray diffraction. The solid-state structure of this complex (8,
Figure 3) shows that two Mg centers, with approximate
octahedral coordination geometry, are each coordinated to a
salan ligand, and bridged by two Cl atoms. Two additional
pentacoordinate Mg atoms (each bound to a chloride and THF
molecule) bridge the phenoxy groups on opposing salan ligands.
This structure is consistent with earlier results that implied a 2:1
Mg:salan ratio is required to form the alkynylating agent.[
(Scheme 3, 9). Gratifyingly, reversal of the steps allowed for
clean ozonolysis of the allylic alcohol to 2-
ethynylglyceraldehyde-3-di-tert-butylphosphate, thus eliminating
concerns around allylic transposition.[26] Unsurprisingly, the
aldehyde generated from ozonolysis was found to be unstable in
organic solvents, but readily formed a stabilized hydrate in
aqueous solvents. Difficulty in isolating the di-tert-
butylphosphate due to decomposition via phosphate ester
migration (Scheme 3, 10) guided us toward the goal of
developing a one-pot procedure to convert allylic alcohol 3c
directly to hydrate 2.
23]
While many catalysts utilizing
a salan ligand with the
diaminocyclohexane backbone are proposed to operate as
monomeric species (with supporting crystallographic evidence in
several cases[18h, 24]), multimetallic or dimeric catalysts have also
been reported.[18e, 25]
Drawing inspiration from Dussault and Molander,[27] we
viewed ozonolysis in water with an acetone co-solvent as
optimal for our purposes. In addition to the aqueous conditions
stabilizing hydrate 2 (see Scheme 3, 11), the safe mixed solvent
system emerged as a rare combination that served to solubilize
both starting material 3c and product 2. Indeed, a 9:1 mixture
of acetone-water enabled creation of a homogeneous reaction
mixture throughout the course of the transformation, and no
ozonides were detected at any point during the reaction.
Furthermore, we were pleased to observe that dealkylation of
the phosphate was spontaneous and autocatalytic after the
2
Me S quench. Upon revealing the free phosphate, addition of
pentanes allowed for separation of the acetone layer containing
the majority of the byproducts leaving only 0.8 equiv of DMSO
and 1.5 equiv of tBuOH in the remaining aqueous solution,
containing hydrate 2 in 93% yield.
Figure 3. Solid-state structure of complex 8. Displacement ellipsoids are
shown at 50% probability.
4
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.202003091
RESEARCH ARTICLE
accomplished, it severely affects the overall efficiency of the
route, as expected. However, the 90:10 e.r. obtained in the
alkynylation approach is not only acceptable for delivering high
ee product, but actually provides a higher overall yield (54% vs.
HO
HO
H2O:Me2CO (1:9)
(tBuO)2OPO
(HO)2OPO
OH
OH
Ar
O3, then Me2S
Ar = 4-OMeC6H4
3
c
2, 93%
4
7%) than recrystallizing at an earlier stage.[30] The unique ability
+H2O
HO
to carry through imperfect enantiomeric ratios and eliminate time
and material-consuming recrystallization steps highlights a less
obvious advantage of employing highly specific biocatalytic
cascades late in a synthesis.[31]
HO
R
Nu
OH
HO
Ar
O-
O
t
O
O
OPO(O Bu)2
O
R
P
RO OR
(10) dephosphorylation
when R
(
9) acid catalyzed
(11) oligomerization
upon concentration
isomerization
H
Conclusion
Scheme 3. Ozonolysis of allylic alcohol 3c and side reactions avoided by
using an aqueous cascade protocol.
In conclusion, we have developed
a 5-step asymmetric
synthesis of islatravir (1) from commodity chemicals. To achieve
this goal, a novel reaction involving the asymmetric alkynylation
of an αʹ-phosphorylated-α,β-unsaturated ketone and an
ozonolysis-dealkylation cascade were developed. Finally, an
enzymatic aldol-glycosylation cascade was leveraged to
construct the desired bonds with complete stereoselectivity
regardless of the enatiomeric purify of the penultimate,
highlighting the power of late stage biocatalytic transformations.
Because of the high concentration (up to 0.4 M) and high
purity of the ozonolysis reaction stream, isolation of the product
as a calcium salt was possible simply by addition of Ca(OAc)
2
i
[28]
followed by PrOH trituration. Although isolation enabled good
purity control at the penultimate step, it was deemed
unnecessary as our ozonolysis crude stream could be used
directly in the aldol-glycosylation cascade more effectively then
when contaminated by calcium salts.[29] In fact, the high purity
and concentration of intermediate 2 from the ozonolysis stream
allowed the synthesis to be completed with more favorable
conditions and higher yield than reported in our previous fully
biocatalytic approach (82% vs 76%) (Table 5).[8]
Acknowledgements
Thanks to Yafei Li (Pharmaron, Inc.) for assistance with ligand
synthesis.
Table 5. Stereospecifity of the biocatalytic aldol-glycosylation cascade.
Keywords: synthesis • asymmetric alkynylation • ozonolysis •
biocatalysis •
[1]
[2]
[3]
HIV
Surveillance
Supplemental Report 2017; 22(No. 2).
http://www.cdc.gov/hiv/library/reports/hivsurveillance.html.
July 2017. Accessed Oct 22, 2019.
Published
a) M. P. de Bethune, Antiviral Res. 2010, 85, 75-90; b) D. G. Prajapati,
R. Ramajayam, M. R. Yadav, R. Giridhar, Bioog. Med. Chem. 2009, 17,
5744-5762; c) C. Flexner, Nat. Rev. Drug Discovery 2007, 6, 959-966.
a) E. Michailidis, A. D. Huber, E. M. Ryan, Y. T. Ong, M. D. Leslie, K. B.
Matzek, K. Singh, B. Marchand, A. N. Hagedorn, K. A. Kirby, L. C.
Rohan, E. N. Kodama, H. Mitsuya, M. A. Parniak, S. G. Sarafianos, J.
Biol. Chem. 2014, 289, 24533-24548; b) H. Ohrui, Chem. Rec. 2006, 6,
e.r. of 3[a]
9:1
e.r. of API
>99:1
d.r. of API
>99:1
yield of 1[b]
overall yield [c]
9
82
44
74
47[d]
32
5
9
0:50
0:10
>99:1
>99:1
1
33-143; c) S. Kohgo, K. Yamada, K. Kitano, Y. Iwai, S. Sakata, N.
>99:1
>99:1
54
Ashida, H. Hayakawa, D. Nameki, E. Kodama, M. Matsuoka, H.
Mitsuya, H. Ohrui, Nucleosides, Nucleotides Nucleic Acids 2004, 23,
[
a] er was determined for the alkynylation products and API,
6
71-690.
but not for 2. [b] Isolated yield for the aldol-glycosylation
cascade vs 2. Yields vs. the limiting reagent (2-FA is 0.85
equiv) are > 93% in all cases (see SI). [c] Isolated yield of the
entire route from anisylidene acetone. [d] An upgrade from
[
4]
a) C. A. Stoddart, S. A. Galkina, P. Joshi, G. Kosikova, M. E. Moreno, J.
M. Rivera, B. Sloan, A. B. Reeve, S. G. Sarafianos, M. Murphey-Corb,
M. A. Parniak, Antimicrob. Agents Chemother. 2015, 59, 4190-4198; b)
S. E. Barrett, R. S. Teller, S. P. Forster, L. Li, M. A. Mackey, D.
Skomski, Z. Yang, K. L. Fillgrove, G. J. Doto, S. L. Wood, J. Lebron, J.
A. Grobler, R. I. Sanchez, Z. Liu, B. Lu, T. Niu, L. Sun, M. E. Gindy,
Antimicrob. Agents Chemother. 2018, 62, e01058-18.
9
0:10 e.r. to 99:1 e.r. of 3c can be achieved by
recrystallization from MeCN/H2O in 78% recovery. DERA =
deoxyribose aldolase, PNP purine nucleoside
=
[
5]
6]
K. Fukuyama, H. Ohrui, S. Kuwahara, Org. Lett. 2015, 17, 828-831.
a) M. Kageyama, T. Miyagi, M. Yoshida, T. Nagasawa, H. Ohrui, S.
Kuwahara, Biosci., Biotechnol., Biochem. 2012, 76, 1219-1225; b) M.
Kageyama, T. Nagasawa, M. Yoshida, H. Ohrui, S. Kuwahara, Org.
Lett. 2011, 13, 5264-5266.
phosphorylase, PPM = phosphopentamutase, SUP = sucrose
phosphorylatse, TEoA = triethanolamine.
[
The enzymatic cascade was shown to be completely
[7]
M. McLaughlin, J. Kong, K. M. Belyk, B. Chen, A. W. Gibson, S. P.
Keen, D. R. Lieberman, E. M. Milczek, J. C. Moore, D. Murray, F. Peng,
J. Qi, R. A. Reamer, Z. J. Song, L. Tan, L. Wang, M. J. Williams, Org.
Lett. 2017, 19, 926-929.
stereoselective providing islatravir as
independent of the enantiomeric purity of the incoming substrate
Table 5). A comparison of overall yields shows that, while
discarding >50% of the substrate in the final step can be
a
single isomer
(
5
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.202003091
RESEARCH ARTICLE
[
8]
Huffman, M. A.; Fryszkowska, A.; Alvizo, O; Borra-Garske, M.;
Campos, K. R.; Canada, K. A.; Devine, P. N.; Duan, Da.; Forstater, J.
H.; Grosser, S. T.; Halsey, H. M.; Hughes, G. J.; Jo, J.; Joyce, L. A.;
Kolev, J. N.; Liang, J.; Maloney, K. M.; Mann, B. F.; Marshall, N. M.;
McLaughlin, M.; Moore, J. C.; Murphy, G. S.; Nawrat, C. C.; Nazor, J.;
Novick, S.; Patel, N. R.; Rodriguez-Granillo, A.; Robaire, S. A.; Sherer,
E.; Truppo, M. D.; Whittaker, A. M.; Verma, D.; Xiao, L.; Xu, Y.; Yang, H.
Science 2019, 366, 1255–1259.
Chem. 2008, 73, 4688-4690; c) G. A. Molander, D. J. Cooper, J. Org.
Chem. 2007, 72, 3558-3560.
[28] a) D. Gauss, B. Schoenenberger, R. Wohlgemuth, Carbohydr. Res.
2014, 389, 18-24; b) M. Binczarski, J. Berlowska, A. Stanishevsky, I.
Witonska, RSC Adv. 2016, 6, 92420-92427.
[29] a) K. Hammer-Jespersen, A. Munch-Petersen, Eur. J. Biochem. 1970,
17, 397-407; b) T. D. Panosian, D. P. Nannemann, G. R. Watkins, V. V.
Phelan, W. H. McDonald, B. E. Wadzinski, B. O. Bachmann, T. M.
Iverson, J. Biol. Chem. 2011, 286, 8043-8054.
[9]
a) N. R. Patel, C. C. Nawrat, M. McLaughlin, Y. Xu, M. A. Huffman, H.
Yang, H. Li, A. M. Whittaker, T. Andreani, F. Lévesque, A. Fryszkowska,
A. Brunskill, D. M. Tschaen, K. M. Maloney, Org. Lett. 2020, 22, 4659-
[30] See SI for details.
[31] In our previous report (reference 8) we obtained high ee of the product
from high ee 2. This is the first demonstration of a resolution using the
same glycosylation cascade.
4
664; b) C. C. Nawrat, A. M. Whittaker, M. A. Huffman, M. McLaughlin,
R. D. Cohen, T. Andreani, B. Ding, H. Li, M. Weisel, D. M. Tschaen,
Org. Lett. 2020, 22, 2167-2172.
[
[
[
10] In addition to the higher yield and shorter step count, reference 8-9
begin from custom materials. The synthesis described herein starts
from commodity chemicals available in multi-ton quantities.
11] a) Y. Yang, C. Qu, X. Chen, K. Sun, L. Qu, W. Bi, H. Hu, R. Li, C. Jing,
D. Wei, S. Wei, Y. Sun, H. Liu, Y. Zhao, Org. Lett. 2017, 19, 5864-
5
867; b) C. Sun, R. Bittman, J. Org. Chem. 2004, 69, 7694-7699.
12] P. B. Huleatt, M. L. Khoo, Y. Y. Chua, T. W. Tan, R. S. Liew, B. Balogh,
R. Deme, F. Goloncser, K. Magyar, D. P. Sheela, H. K. Ho, B. Sperlagh,
P. Matyus, C. L. Chai, J. Med. Chem. 2015, 58, 1400-1419.
[
[
[
[
13] K. S. Babu, X. C. Li, M. R. Jacob, Q. Zhang, S. I. Khan, D. Ferreira, A.
M. Clark, J. Med. Chem. 2006, 49, 7877-7886.
14] B. Zheng, R. J. Fox, M. Sugiyama, A. Fritz, M. D. Eastgate, Org.
Process Res. Dev. 2014, 18, 636-642.
15] V. R. Sabbasani, H. Lee, Y. Xia, D. Lee, Angew. Chem. Int. Ed. Engl.
2
016, 55, 1151-1155.
16] a) B. M. Trost, A. H. Weiss, Adv. Synth. Catal. 2009, 351, 963-983; b) T.
Bauer, Coord. Chem. Rev. 2015, 299, 83-150; c) P. G. Cozzi, Angew.
Chem. 2003, 115, 3001-3004; d) Y. Zheng, Y. Tan, K. Harms, M.
Marsch, R. Riedel, L. Zhang, E. Meggers, J. Am. Chem. Soc. 2017, 139,
4
322-4325; e) B. Bieszczad, D. G. Gilheany, Angew. Chem. Int. Ed.
Engl. 2017, 56, 4272-4276.
[17] D. Park, C. I. Jette, J. Kim, W. O. Jung, Y. Lee, J. Park, S. Kang, M. S.
Han, B. M. Stoltz, S. Hong, Angew. Chem. Int. Ed. Engl. 2020, 59, 775-
7
79.
[
18] a) H. Du, A. H. Velders, P. J. Dijkstra, J. Sun, Z. Zhong, X. Chen, J.
Feijen, Chem. Eur. J. 2009, 15, 9836-9845; b) A. Yeori, I. Goldberg, M.
Shuster, M. Kol, J. Am. Chem. Soc. 2006, 128, 13062-13063; c) J. Sun,
Z. Dai, M. Yang, X. Pan, C. Zhu, Synthesis 2008, 2008, 2100-2104; d)
J. Sun, C. Zhu, Z. Dai, M. Yang, Y. Pan, H. Hu, J. Org. Chem. 2004, 69,
8
5
500-8503; e) H. Egami, T. Katsuki, Angew. Chem. Int. Ed. 2008, 47,
171-5174; f) Z. Dai, M. Shao, X. Hou, C. Zhu, Y. Zhu, Y. Pan, Appl.
Organomet. Chem. 2005, 19, 898-902; g) Z. Dai, C. Zhu, M. Yang, Y.
Zheng, Y. Pan, Tetrahedron: Asymmetry 2005, 16, 605-608; h) A. Yeori,
S. Groysman, I. Goldberg, M. Kol, Inorg. Chem. 2005, 44, 4466-4468.
19] E. N. Jacobsen, W. Zhang, M. L. Guler, J. Am. Chem. Soc. 1991, 113,
[
[
6
703-6704.
20] 7 is available as a 0.5 M THF solution at $154 / 800 mL, while EtMgCl
is available as THF solution at $226 800 mL
www.sigmaaldrich.com on Nov 8th 2019). Therefore, catalyst
a
2
M
/
(
formation with the more concentrated EtMgCl represents cost and
volume reductions, which are important considerations on scale.
21] D. Guillaneux, S.-H. Zhao, O. Samuel, D. Rainford, H. B. Kagan, J. Am.
Chem. Soc. 1994, 116, 9430-9439.
[
[22] S. Shaw, J. D. White, Chem. Rev. 2019, 119, 9381-9426.
2
[23] The crystals of 8 were formed amongst oils and LiOH/Mg(OH) , so
were not clean enough to be used directly as a catalyst.
[24] H. Elias, F. Stock, C. Rohr, Acta Cryst. C 1997, 53, 862-864.
[
25] D. A. Atwood, J. A. Jegier, K. J. Martin, D. Rutherford, Organometallics
995, 14, 1453-1460.
1
[
26] a) S. Goswami, A. K. Adak, Tetrahedron Lett. 2002, 43, 503-505; b) R.
Persky, A. Albeck, J. Org. Chem. 2000, 65, 5632-5638; c) D. R.
Edwards, C. T. Liu, G. E. Garrett, A. A. Neverov, R. S. Brown, J. Am.
Chem. Soc. 2009, 131, 13738-13748.
[27] a) R. Willand-Charnley, T. J. Fisher, B. M. Johnson, P. H. Dussault, Org.
Lett. 2012, 14, 2242-2245; b) C. E. Schiaffo, P. H. Dussault, J. Org.
6
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.202003091
RESEARCH ARTICLE
Entry for the Table of Contents
A 5-step enantioselective synthesis of the potent anti-HIV nucleoside islatravir is reported. The highly efficient route was enabled by a
novel enantioselective alkynylation of an α,β-unsaturated ketone, a unique ozonolysis-dealkylation cascade in water, and an
enzymatic aldol-glycosylation cascade.
7
This article is protected by copyright. All rights reserved.