Nine-Step Stereoselective Synthesis of Islatravir from Deoxyribose

Article abstract of DOI:10.1021/acs.orglett.0c00239

A stereoselective nine-step synthesis of the potent HIV nucleoside reverse transcriptase translocation inhibitor (NRTTI) islatravir (EfdA, MK-8591) from 2-deoxyribose is described. Key findings include a diastereodivergent addition of an acetylide nucleophile to an enolizable ketone, a chemoselective ozonolysis of a terminal olefin and a biocatalytic glycosylation cascade that uses a unique strategy of byproduct precipitation to drive an otherwise-reversible transformation forward.

Full text of DOI:10.1021/acs.orglett.0c00239

pubs.acs.org/OrgLett
Letter
Nine-Step Stereoselective Synthesis of Islatravir from Deoxyribose
Christopher C. Nawrat,* Aaron M. Whittaker,* Mark A. Huffman, Mark McLaughlin, Ryan D. Cohen,
Teresa Andreani, Bangwei Ding, Hongming Li, Mark Weisel, and David M. Tschaen
Cite This: https://dx.doi.org/10.1021/acs.orglett.0c00239
Read Online
sı
*
Metrics & More
Article Recommendations
Supporting Information
ACCESS
ABSTRACT: A stereoselective nine-step synthesis of the potent HIV nucleoside reverse transcriptase translocation inhibitor
(NRTTI) islatravir (EfdA, MK-8591) from 2-deoxyribose is described. Key findings include a diastereodivergent addition of an
acetylide nucleophile to an enolizable ketone, a chemoselective ozonolysis of a terminal olefin and a biocatalytic glycosylation
cascade that uses a unique strategy of byproduct precipitation to drive an otherwise-reversible transformation forward.
he World Health Organisation (WHO) estimates that
Tover 37 million people worldwide are currently living
with HIV.1 Although modern antiretroviral drugs are highly
effective and the death rate from HIV/AIDS is at its lowest
since the 1980s, the number of infected individuals continues
to rise by almost 2 million in 2016 alone. Today, nucleoside
reverse transcriptase inhibitors (NRTIs) are a mainstay of
modern antiretroviral therapy, and five drugs in this class are
currently approved HIV treatment by the FDA.2a However,
with no cure in sight, patients often take such drugs for
decades; therefore new medicines that address the issues of
side effects, efficacy, resistance, and dosing with existing
treatments are still highly sought after.2b,c
Islatravir (EfdA, MK-8591, 1) is an adenosine-based
inhibitor of reverse transcriptase translocation (i.e., an
NRTTI)2d discovered by researchers at the Yamasa Corpo-
ration and their academic collaborators in 2004 (see Scheme
1).3,4 It is the first NRTTI reported to date and has shown
great promise in preclinical studies, including remarkable
potency and a long intracellular half-life. As befits its promise
as a drug candidate, several total syntheses of 1 have been
reported in the literature thus far.5
In 2017, researchers at Merck & Co., Inc. disclosed a
scalable route amenable to the production of multikilogram
quantities of islatravir from inexpensive dihydroxyacetone.5d
Although this route is highly efficient (17% overall yield), it is
relatively long (14 steps) and suffers from a low yielding and
poorly diastereoselective penultimate glycosylation step. More
recently, the development of an efficient and highly stereo-
selective biocatalytic glycosylation solved this latter problem,
allowing 1 to be prepared from ribose 5′-phosphate 2 with
near-perfect stereoselectivity (Scheme 1).6 Although ultimately
it was found that 2 could also be prepared enzymatically,
before this biocatalytic route became available, we sought a
chemical synthesis of 2 that would allow us to leverage this
new enzyme-mediated glycosylation method. Unfortunately,
attempts to adapt published syntheses of 1 or similar
nucleosides for the preparation of phosphate 2 resulted in
lengthy, inefficient synthetic sequences, and highlighted the
need for a completely new approach to this key intermediate.
With this goal in mind, we envisioned a chiral pool
approach, predicated on the use of the known 2-deoxy-D-
ribose-derived ketone 5 as starting material (see Scheme 1).
Achieving an efficient and diastereoselective acetylide addition
to 5 to correctly set the stereochemistry at the fully substituted
stereogenic center in masked triol 4a would be a key challenge
of this approach, because of the well-known difficulty of
achieving axial addition to α-substituted cyclohexanones.7 If
this could be accomplished, then acetal 4a contains all the
necessary carbon atoms and stereogenic centers found in
ribose 2. From here, acetal deprotection, phosphorylation of
the primary alcohol, selective ozonolysis of the terminal olefin
and deprotection were expected to give the target ribose-5-
phosphate 2, whose biocatalytic conversion to 1 has been
demonstrated.6
Received: January 16, 2020
https://dx.doi.org/10.1021/acs.orglett.0c00239
© XXXX American Chemical Society
Org. Lett. XXXX, XXX, XXX−XXX
A

Organic Letters
pubs.acs.org/OrgLett
Letter
a
Scheme 1. Retrosynthetic Analysis of 1
Table 1. Addition of Metalated Ethynyltrimethylsilanes
a
1
Diastereomeric ratio determined by H NMR.
Scheme 2. Axial Approach of Alkyne That Is Required To
a
Give the Desired Diastereomer 6a
a
PNP = purine nucleoside phosphorylase and PPM = phosphopento-
mutase.
a
Although ketone 5 has been used as a synthetic intermediate
in the literature,8 it was impractical to obtain the quantities
necessary to develop our route using the reported procedures,
which required three chromatographic purifications. Thus,
significant development of the literature chemistry was
performed, in order to enable preparation of multihundred-
gram quantities of 5 (see the Supporting Information for
details). Crucially, although ketone 5 is a low-melting and
difficult-to-handle wax, we found that it readily forms a stable
crystalline hydrate. This enabled us to purify this material by
crystallization from aqueous acetonitrile, which, along with
other changes, allowed us to eliminate two chromatographic
purification steps from the reported sequence.9 With ketone 5
now readily available, we began to study the key acetylide
addition reaction, the stereochemical outcome of which was
uncertain at the outset of these studies. First, we examined the
use of TMS-protected acetylide nucleophiles, as these can be
conveniently prepared from inexpensive trimethylsilylacetylene
without the need for handling or generating acetylene, which is
extremely hazardous on scale,10 and allowed us to screen a
range of alkynylmetals (see Table 1). Disappointingly,
although a range of protected acetylide nucleophiles was
found to react with 5 in good yield, none of these favored the
required diastereomer. This is perhaps unsurprising, since
formation of the desired diastereomer would require the axial
approach of the acetylide, which is a trajectory that is generally
disfavored for larger nucleophiles on steric grounds (see
Scheme 2).7,11
The stereochemical outcome was confirmed by NoE analysis.
to the desired stereochemistry in our case.12 Unfortunately, we
were not able to obtain any of the desired product from the
reaction of sodium acetylide with ketone 5,13 and lithium
acetylide ethylenediamine complex was similarly unsuccessful
(see Table 2, entries 1 and 2).14 Although we found that
ethynylmagnesium bromide underwent addition in reasonable
yield, this again favored the undesired equatorial addition
product (Table 2, entry 3). Given the promising literature
precedent for the use of sodium acetylide, we again returned to
this reagent. Since sodium acetylide is highly insoluble and
seemingly too basic to be an effective nucleophile in our
system, a way to modify its reactivity was sought. After some
experimentation, we were delighted to find that sodium
trimethylethynylaluminateformed by premixing equimolar
amounts of sodium acetylide and trimethylaluminum in a
mixture of THF and toluenewas a singularly effective
acetylide source, giving the desired product in high yield and
with better than 19:1 diastereoselectivity (Table 2, entry 4).15
Upon further exploration, this aluminate addition reaction
exhibits many surprising features. Interestingly, the diaster-
eoselectivity and reaction profile were found to be highly
sensitive toward the ratio of trimethylaluminum to sodium
acetylide, with the maximum yield and dr being obtained with
precisely equimolar quantities of both (Table 2, entry 4). Even
a small excess of sodium acetylide was found to result in lower
yields, accompanied by the formation of numerous unidenti-
fied byproducts, underscoring the incompatibility of 5 with this
reagent. Conversely, the use of excess trimethylaluminum was
found to greatly decrease the diastereoselectivity of the
In view of these findings, we decided to investigate the use of
smaller acetylide nucleophiles. Encouragingly, the addition of
sodium acetylide to 4-tert-butylcyclohexanone has long been
known to favor the product of axial additioncorresponding
B
https://dx.doi.org/10.1021/acs.orglett.0c00239
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
pubs.acs.org/OrgLett
Letter
trimethylaluminum are inexpensive and widely available in
bulk, because of their industrial importance.
Table 2. Addition of Unprotected Acetylides to 5
Returning to silylated alkynes with this result in mind, we
were curious about how the aluminate generated from
ethynyltrimethylsilane would behave in the addition reaction
and how its diastereoselectivity would compare to the reagents
already screened (see Scheme 3). Although this aluminate
Scheme 3. Unexpected Stereodivergence in the Addition of
Alkyne Nucleophiles to Ketone 5
reagent has not been previously reported, we found that it
could be formed as expected by deprotonation with NaHMDS,
followed by treatment of the sodiated alkyne with
trimethylaluminum. To our surprise, we found that addition
of this aluminate nucleophile strongly favored the desired
diastereomer 6a, in contrast to all other metalated alkynes of
this type, which all predominantly produce unwanted alcohol
6b. This unexpected stereodivergence in the addition of
acetylides therefore allows preparation of both epimers of 6
with good diastereoselectivity, a feature that may be useful in
the preparation of other 4′-substituted nucleosides. The origin
of the unique axial selectivity of aluminate reagents is currently
unknown, and it is not clear if these reagents represent a
general solution to this problem. Regardless, we suspect that
the general difficulty in accesing organosodium species may
limit this tactic to alkyne nucleophiles, unless aluminates
generated from other more readily prepared organometallic
nucleophiles display similar reactivity.
With a scalable and highly diastereoselective acetylide
addition in hand, we now turned our attention to advancing
alcohol 4a toward the target (see Scheme 4). First,
deprotection of the benzylidene acetal was accomplished
using hydrochloric acid in methanol.17 Next, we studied the
regioselective phosphorylation of the primary alcohol in triol 7
using diethyl chlorophosphate. Initially, it was found that the
reaction suffered from poor selectivity and produced cyclic
phosphate byproducts when aged in a pyridine solvent or run
with high loading of nucleophilic catalysts (e.g., 20 mol %
DMAP).18 A favorable balance of reaction rate and selectivity
was eventually obtained with 1 mol % DMAP in dichloro-
methane, delivering phosphate 3 in good yield.
a
1
b
Diastereomeric ratio determined by H NMR. Using AlMe3 (2.0
equiv) and NaCCH (2.0 equiv). cUsing AlMe3 (2.5 equiv) and
d
NaCCH (2.0 equiv). Using AlMe3 (3.0 equiv) and NaCCH (2.0
equiv). eIsolated yield on 822 g scale (see the Supporting Information
f
for details). Using AlMe3 (1.92 equiv) and NaCCH (1.55 equiv).
reaction, eventually even causing the unwanted epimer 4b to
be favored (Table 2, entries 5 and 6). The origin of this
reversal is not immediately clear, but it may be that the excess
trimethylaluminum coordinates to the carbonyl oxygen in a
fashion that disfavors the axial approach of the nucleophile,
encouraging unwanted equatorial attackessentially perform-
ing the opposite function to the much bulkier Yamamoto
aluminum Lewis acids that have been employed to block
equatorial attack upon cyclohexanones.16 The aluminate is a
far weaker Lewis acid than trimethylaluminum and likely
cannot perform this role alone. Unfortunately, attempts to
study the reaction by NMR spectroscopy were complicated by
higher-order aggregates and further investigation of this
unusual effect is ongoing.
Balancing these competing scenarios, it was found that,
because of difficulties in accurately charging the heterogeneous
slurry of sodium acetylide that is available commercially, it was
best to use a 5% excess of trimethylaluminum in toluene to
ensure reproducibly high yields with acceptable diastereose-
lectivity. Under these optimized conditions, the reaction was
run on 822 g scale, giving 72% yield and 8.4:1 dr. Aside from
its unique diastereoselectivity in this reaction, note that,
compared to more common acetylide anion sources, such as
ethynylmagnesium bromide, sodium trimethylethynylalumi-
nate possesses additional advantages in cost and large-scale
availability. Indeed, we could only identify one U.S.-based
supplier able to provide ton-scale quantities of the former (and
only as a 0.5 M solution), whereas sodium acetylide and
Ozonolysis of substrate 3 invoked concerns surrounding
chemoselectivity between the alkyne and olefin, tolerance of
sensitive functional groups such as a tertiary propargylic
alcohol and primary phosphate, and the isolation of a
potentially unstable product.19 Initial experiments generated
a mixture of products that were characterized as the lactol and
the related methyl glycoside 8. We were pleased to discover
that either product could be selectively formed by varying the
pH and solvent, allowing us to isolate the more stable methyl
glycoside from an acidic mixture of 1:3 MeOH:CHCl3. Under
C
https://dx.doi.org/10.1021/acs.orglett.0c00239
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
pubs.acs.org/OrgLett
Letter
glucose 1-phosphate).20,21 Without the driving force provided
by this phosphate-scavenging system, the otherwise-reversible
glycosylation cascade from 2 to 1 does not proceed to high
conversion. However, when 9 is used as starting material, the
phosphate produced in the reaction is precipitated as a highly
insoluble calcium salt, driving the reaction to completion even
in the absence of sucrose/sucrose phosphorylase.22 The crude
API obtained from this reaction could be dissolved in
dimethylformamide (DMF), filtered away from the inorganic
calcium phosphate byproducts and recrystallized upon the
addition of water to deliver the product in >99% purity and
82% isolated yield.
Scheme 4. Completion of Islatravir
We have reported a highly stereoselective synthesis of
islatravir in six steps (36% overall yield) from known ketone 5
(nine steps overall from inexpensive 2-deoxyribose). Of
particular interest is the highly diastereoselective addition of
sodium trimethylethynylaluminate, which is an inexpensive and
underutilized acetylide equivalent whose use was demonstrated
on a >800 g scale, as well as a selective ozonolysis of the
terminal olefin, and an enzymatic glycosylation cascade driven
forward by the precipitation of calcium phosphate. This
sequence represents an expedient approach to 4′-substituted
ribose derivatives that may be generalizable for other non-
natural nucleosides.
ASSOCIATED CONTENT
* Supporting Information
■
sı
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/acs.orglett.0c00239.
1
Experimental procedures, characterization data, and H
optimized conditions, we successfully performed the ozonol-
ysis on 25 g scale to deliver the desired methyl glycoside 8 in
83% yield.
and 13C NMR spectra of all new compounds (PDF)
AUTHOR INFORMATION
Corresponding Authors
■
Methyl glycoside 8 could subsequently be fully dealkylated
using bromotrimethylsilane, and after quenching with water,
the aqueous solution of lactol could be used directly in the
enzymatic glycosylation step. Unfortunately, this protocol
generated product streams containing high levels of Br ions
that adversely affected the efficiency of the enzymatic
glycosylationparticularly the purine nucleoside phosphor-
ylaserequiring high dilution and high enzyme loading in
order to overcome the negative impact of halide ions on this
biocatalyst.6 This result, taken in combination with a notable
lack of solid intermediates throughout the described synthesis,
caused us to target an isolation at the lactol stage utilizing the
phosphate group as a handle. Initial results indicated that salts
with Ca2+, Mn2+, Mg2+, or NH4+ were isolable, but purity,
stability, and ease of isolation guided us toward the Ca2+ salt.
After some optimization, it was found that we could isolate the
stable calcium phosphate lactol 9 in high yield by addition of
Ca(OAc)2 to the crude aqueous lactol solution, followed by
trituration and filtration of the highly insoluble solid with
acetonitrile.
To our delight, isolated calcium lactol phosphate 9 far
outperformed the aqueous stream of 2 from the deprotection
that had originally been used in the final enzymatic
glycosylation cascade. Furthermore, the use of 9 as an
intermediate provided an additional advantage, because it
allowed us to simplify the glycosylation protocol. As originally
published,6 using the parent phosphoric acid 2 (instead of its
calcium salt 9), the biocatalytic glycosylation requires the
addition of sucrose phosphorylase and sucrose in order to
sequester the phosphate ions produced in the reaction (as
Christopher C. Nawrat − Department of Process Research and
Development, Merck & Co., Inc., Rahway, New Jersey 07065,
United States; orcid.org/0000-0003-4550-9954;
Email: christopher.nawrat@merck.com
Aaron M. Whittaker − Department of Process Research and
Development, Merck & Co., Inc., Rahway, New Jersey 07065,
United States; orcid.org/0000-0002-8348-5490;
Email: aaron.whittaker@merck.com
Authors
Mark A. Huffman − Department of Process Research and
Development, Merck & Co., Inc., Rahway, New Jersey 07065,
United States
Mark McLaughlin − Department of Process Research and
Development, Merck & Co., Inc., Rahway, New Jersey 07065,
United States; orcid.org/0000-0003-0595-4754
Ryan D. Cohen − Department of Process Research and
Development, Merck & Co., Inc., Rahway, New Jersey 07065,
United States
Teresa Andreani − Department of Process Research and
Development, Merck & Co., Inc., Rahway, New Jersey 07065,
United States
Bangwei Ding − Department of Process Research and
Development, Merck & Co., Inc., Rahway, New Jersey 07065,
United States
Hongming Li − Department of Process Research and
Development, Merck & Co., Inc., Rahway, New Jersey 07065,
United States
D
https://dx.doi.org/10.1021/acs.orglett.0c00239
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
pubs.acs.org/OrgLett
Letter
low toxicity and prevent the emergence of resistant HIV mutants.
Proc. Jpn. Acad., Ser. B 2011, 87, 53−58.
Mark Weisel − Department of Process Research and
Development, Merck & Co., Inc., Rahway, New Jersey 07065,
United States
David M. Tschaen − Department of Process Research and
Development, Merck & Co., Inc., Rahway, New Jersey 07065,
United States
(4) In 2012, Merck & Co., Inc. (Kenilworth, NJ, USA) licensed the
compound from Yamasa for further development.
(5) Previous syntheses of 1: (a) Kageyama, M.; Nagasawa, T.;
Yoshida, M.; Ohrui, H.; Kuwahara, S. Enantioselective total synthesis
of the potent anti-HIV nucleoside EfdA. Org. Lett. 2011, 13, 5264−
5266. (b) Kageyama, M.; Miyagi, T.; Yoshida, M.; Nagasawa, T.;
Ohrui, H.; Kuwahara, S. Concise synthesis of the anti-HIV nucleoside
EfdA. Biosci., Biotechnol., Biochem. 2012, 76, 1219−1225. (c) Fukuya-
ma, K.; Ohrui, H.; Kuwahara, S. Synthesis of EFdA via a
diastereoselective aldol reaction of a protected 3-keto furanose. Org.
Lett. 2015, 17, 828−831. (d) McLaughlin, M.; Kong, J.; Belyk, K. M.;
Chen, B.; Gibson, A. W.; Keen, S. P.; Lieberman, D. R.; Milczek, E.
M.; Moore, J. C.; Murray, D.; Peng, F.; Qi, J.; Reamer, R. A.; Song, Z.
J.; Tan, L.; Wang, L.; Williams, M. J. Enantioselective synthesis of 4′-
ethynyl-2-fluoro-2′-deoxyadenosine (EFdA) via enzymatic desymmet-
rization. Org. Lett. 2017, 19, 926−929. (e) Kamata, M.; Takeuchi, T.;
Hayashi, E.; Nishioka, K.; Oshima, M.; Iwamoto, M.; Nishiuchi, K.;
Kamo, S.; Tomoshige, S.; Watashi, K.; Kamisuki, S.; Ohrui, H.;
Sugawara, F.; Kuramochi, K. Synthesis of nucleotide analogues, EFdA,
EdA and EdAP, and the effect of EdAP on hepatitis B virus
replication. Biosci., Biotechnol., Biochem., 84, 217−227.
(6) 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. Design of an in vitro biocatalytic cascade for the
manufacture of islatravir. Science 2019, 366, 1255−1259.
(7) The reaction of dioxanones with methylmagnesium chloride and
lithium aluminum hydride has long been known to be highly selective
for the desired axial addition products, as a result of the greater
torsional strain that develops in the transition state for equatorial
addition (in fact, addition to dioxanones is generally more axial-
selective than for the parent cyclohexanones, because of the shorter
C−O bonds). See: (a) Urbansky, M.; Davis, C. E.; Surjan, J. D.;
Coates, R. M. Synthesis of enantiopure 2-C-methyl-d-erythritol 4-
phosphate and 2,4-cyclodiphosphate from d-arabitol. Org. Lett. 2004,
6, 135−138. (b) Wu, Y. D.; Houk, K. N. Electronic and
conformational effects on pi-facial stereoselectivity in nucleophilic
additions to carbonyl compounds. J. Am. Chem. Soc. 1987, 109, 908−
910. However, addition of larger nucleophiles such as ethynylmagne-
sium bromide has been reported to be relatively unselective in a
similar system, highlighting the challenge of achieving selective axial
attack of larger carbon nucleophiles under substrate control:
(c) Carda, M.; Casabo, P.; Gonzalez, F.; Rodriguez, S.; Domingo,
L. R.; Marco, J. A. Diastereoselectivity of the reactions of
organometallic reagents with protected d-and l-erythrulose 1, 3-O-
ethylidene acetals. Tetrahedron: Asymmetry 1997, 8, 559−577.
(8) Urosa, A.; Marcos, I. S.; Diez, D.; Padron, J. M.; Basabe, P.
Synthesis of luffarin L and 16-epi-luffarin L using a temporary silicon-
tethered ring-closing metathesis reaction. J. Org. Chem. 2015, 80,
6447−6455.
Complete contact information is available at:
https://pubs.acs.org/10.1021/acs.orglett.0c00239
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The authors wish to thank Francois Levesque for assistance
with ozonolysis equipment, Kevin Maloney and Guy
Humphrey for helpful chemistry discussions, and Louis-
Charles Campeau for suggestions regarding this manuscript,
and Wilfredo Pinto is gratefully recognized for his assistance
with the collection of HRMS data (all at Merck & Co., Inc.,
Kenilworth, NJ, USA). Ajay Ryerson, Greg Wells, Mark
Pratton and Andrew Carpenter (AMPAC Fine Chemicals,
LLC) are acknowledged for studies into the scaleup of the
alkyne addition. Shiwei Wang and Jinlong Zhao (WuXi
AppTec Co., Ltd.) are thanked for their help with the scaleup
of ketone 5 and the optimization of the literature procedures
required to achieve this.
REFERENCES
■
(1) HIV/AIDS Data and Statistics. Available via the Internet at:
http://www.who.int/hiv/data/en/ (accessed Nov 15, 2018).
(2) (a) FDA-Approved HIV Medicines. Available via the Internet at:
https://aidsinfo.nih.gov/understanding-hiv-aids/fact-sheets/21/58/
fda-approved-hiv-medicines (accessed Jan 16, 2020). For information
on NRTIs, see: (b) De Clercq, E. The Nucleoside reverse
transcriptase inhibitors, nonnucleoside reverse transcriptase tnhibi-
tors, and protease inhibitors in the treatment of HIV infections
(AIDS). Adv. Pharmacol. 2013, 67, 317−358. For a broader view of
anti-viral drugs, see: (c) De Clercq, E. Antiretroviral drugs. Curr. Opin.
Pharmacol. 2010, 10, 507−515. Conversely, for a discussion of the
NRTTI mechanism of action, see: (d) Michailidis, E.; Marchand, B.;
Kodama, E. N.; Singh, K.; Matsuoka, M.; Kirby, K. A.; Ryan, E. M.;
Sawani, A. M.; Nagy, E.; Ashida, N.; Mitsuya, H.; Parniak, M. A.;
Sarafianos, S. G. DNA: Replication, repair, recombination, and
chromosome dynamics. J. Biol. Chem. 2009, 284, 35681−35691.
(3) Discovery: (a) Kohgo, S.; Ohrui, H.; Kodama, E.; Matsuoka, M.;
Mitsuya, H. 4’-C-Substituted-2-halo-adenosine derivative. Can. Patent
No. CA 2502109, March 22, 2005. (b) Kohgo, S.; Yamada, K.;
Kitano, K.; Iwai, Y.; Sakata, S.; Ashida, N.; Hayakawa, H.; Nameki, D.;
Kodama, E.; Matsuoka, M.; Mitsuya, H.; Ohrui, H. Design, efficient
synthesis, and anti-HIV activity of 4′-C-cyano- and 4′-C-ethynyl-2′-
deoxy purine nucleosides. Nucleosides, Nucleotides Nucleic Acids 2004,
23, 671−690. (c) Ohrui, H. 2′-deoxy′4′-C-ethynyl-2-fluoroadenosine,
a nucleoside reverse transcriptase inhibitor, is highly potent against all
human immunodeficiency viruses type 1 and has low toxicity. Chem.
Rec. 2006, 6, 133−143. (d) Ohrui, H.; Hayakawa, H.; Kohgo, S.;
Matsuoka, M.; Kodama, E.; Mitsuya, H. Creation of low toxic reverse-
transcriptase inhibitory nucleosides that prevent the emergence of
drug-resistant HIV variants. Yuki Gosei Kagaku Kyokaishi 2006, 64,
716−721. (e) Ohrui, H.; Kohgo, S.; Hayakawa, H.; Kodama, E.;
Matsuoka, M.; Nakata, T.; Mitsuya, H. 2′-Deoxy-4′-C-ethynyl-2-
fluoroadenosine: A nucleoside reverse transcriptase inhibitor with
highly potent activity against wide spectrum of HIV-1 strains,
favorable toxic profiles, and stability in plasma. Nucleosides, Nucleotides
Nucleic Acids 2007, 26, 1543−1546. (f) Ohrui, H. Development of
modified nucleosides that have supremely high anti-HIV activity and
(9) The parent 2-phenyldioxanone has been reported to form a
similar hydrate: Carlsen, P. H. J.; Sørbye, K.; Ulven, T.; Aasbø, K.;
Klinga, M.; Romerosa, A. Synthesis of benzylidene-protected
dihydroxyacetone. Acta Chem. Scand. 1996, 50, 185−187. [Note
that, prior to the alkyne additions, the ketone hydrate was dehydrated
by azeotropic concentration from toluene to give a dry solution of the
ketone (see the Supporting Information for details).]
(10) Miller, S. A.; Penny, E. Hazards in Handling Acetylene in
Chemical Processes Particularly under Pressure. In Symposium on
Chemical Process Hazards; IChemE Symposium Series No. 7, 1960; pp
87−94.
E
https://dx.doi.org/10.1021/acs.orglett.0c00239
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
pubs.acs.org/OrgLett
Letter
(11) Mehta, G.; Chandrasekhar, J. Electronic control of facial
selection in additions to sterically unbiased ketones and olefins. Chem.
Rev. 1999, 99, 1437−1467.
(21) Hammer-Jespersen, K.; Munch-Petersem, A. Phosphodeoxyr-
ibomutase from escherichia coli purification and some properties. Eur.
J. Biochem. 1970, 17, 397−407.
(22) The related tactic of driving a phosphate-producing reaction to
completion by precipitating insoluble magnesium phosphate has been
reported, for example, for cellodextrin phosphorylase: Zhong, C.;
Luley-Goedl, C.; Nidetzky, B. Product solubility control in
cellooligosaccharide production by coupled cellobiose and cellodex-
trin phosphorylase. Biotechnol. Bioeng. 2019, 116, 2146−2155. In this
case, we found magenisum additives to be less effective than either the
use of a calcium salt as described here or the sucrose/sucrose
phosphorylase system described in ref 6.
(12) Hennion, G. F.; O’Shea, F. X. Ethynylation of 4-t-
Butylcyclohexanone and kinetics of saponification of the ethynylcar-
binol esters. J. Am. Chem. Soc. 1958, 80, 614−617.
(13) Some decomposition of the starting material was observed. We
speculate that the greater acidity of the protons adjacent to the ketone
5 in our system may lead to enolate formation and self-condensation.
Generally, more basic nucleophiles were found to produce greater
levels of side-products with this system.
(14) The parent lithium acetylide was avoided, because of its well-
known propensity to disproportionate to lithium carbide and
acetylene, which makes its use complicated and hazardous on scale:
(a) Mortier, J.; Vaultier, M.; Carreaux, F.; Douin, J.-M. Disproportio-
nation of monolithium acetylide into dilithium carbide and acetylene
is a reversible reaction in tetrahydrofuran at 0 °C. J. Org. Chem. 1998,
63, 3515−3516. (b) Midland, M. M.; Gallou, F. Lithium Acetylide in
Encyclopedia of Reagents for Organic Synthesis [Online] 2006 (
DOI: 10.1002/047084289X.rl035.pub2).
(15) Joung, M. J.; Ahn, J. H.; Yoon, N. M. Sodium
trimethylethynylaluminate, a new chemoselective ethynylating agent.
J. Org. Chem. 1996, 61, 4472−4475. Although this reagent was
described some time ago, to the best of our knowledge, it has not
been applied in a multistep organic synthesis before this report.
(16) A more classical solution to achieve equatorial nucleophilic
attack on cyclohexanones involves the use of the bulky MAD
aluminum Lewis acids. However, these are generally less selective for
α-substituted ketones, presumably because of the steric clash of
neighboring groups with the Lewis acid. See: Maruoka, K.; Itoh, T.;
Yamamoto, H. Methylaluminum Bis(2,6-di-tert-butyl-4-alkylphenox-
ide). A New Reagent for Obtaining Unusual Equatorial and Anti-
Cram Selectivity in Carbonyl Alkylation. J. Am. Chem. Soc. 1985, 107,
4573−4576.
(17) This seemingly simple transformation was complicated by the
extremely high water solubility of triol 7, which precluded a traditional
aqueous workup. An efficient protocol was developed wherein, upon
completion, water was added to the methanolic reaction mixture,
allowing the benzaldehyde generated to be removed by extraction
with heptanes, which is immiscible with the methanol−water phase
containing the product. Saturating the aqueous mixture with K2CO3
allowed (i) phase separation of the MeOH (where the triol resides)
from the water, and (ii) stabilization of the triol intermediate, which
was found to decompose if concentrated from low pH solutions.
Although unnecessary for downstream chemistry, residual salts could
be removed by slurrying the crude product in TBME, followed by
filtration. This procedure provided 44 g of triol 7 in 96% isolated yield
and 94 wt% purity. For further discussion of the phase separation of
miscible liquids, see: Hyde, A. M.; Zultanski, S. L.; Waldman, J. H.;
Zhong, Y.-L.; Shevlin, M.; Peng, F. General principles and strategies
for salting-out informed by the Hofmeister series. Org. Process Res.
Dev. 2017, 21, 1355−1370.
(18) Phosphate ester migration and cyclization of similar
compounds under basic conditions has been reported. See: Brown,
D. M.; Todd, A. R. Nucleotides. Part X. Some observations on the
structure and chemical behaviour of the nucleic acids. J. Chem. Soc.
1952, 52−58.
(19) For examples of industrial-scale ozonolysis, see: (a) Van
Ornum, S. G.; Champeau, R. M.; Pariza, R. Ozonolysis applications in
drug synthesis. Chem. Rev. 2006, 106, 2990−3001. (b) Fisher, T. J.;
Dussault, P. H. Alkene ozonolysis. Tetrahedron 2017, 73, 4233−4258.
For chemoselective ozonolysis of alkenes over alkynes, see ref 5a and:
McCurry, P. M., Jr.; Abe, K. Selective ozonolysis of an enyne
derivative. The synthesis of cis-jasmone. Tetrahedron Lett. 1974, 15,
1387−1390.
(20) Tischler, W.; Hans-Georg, I.; Barzu, O.; Sakamoto, H.;
Pistotnik, E.; Marliere, P.; Pochet, S. Enzymatic Synthesis of
Deoxyribonucleosides. U.S. Patent No. US07229797B1, 2007.
F
https://dx.doi.org/10.1021/acs.orglett.0c00239
Org. Lett. XXXX, XXX, XXX−XXX