Development of a Direct Photocatalytic C-H Fluorination for the Preparative Synthesis of Odanacatib

Article abstract of DOI:10.1021/acs.orglett.5b02532

Late-stage C-H fluorination is an appealing reaction for medicinal chemistry. However, the application of this strategy to process research appears less attractive due to the formation and necessary purification of mixtures of organofluorines. Here we dem

Full text of DOI:10.1021/acs.orglett.5b02532

Letter
pubs.acs.org/OrgLett
Development of a Direct Photocatalytic C−H Fluorination for the
Preparative Synthesis of Odanacatib
Shira D. Halperin,† Daniel Kwon,† Michael Holmes,† Erik L. Regalado,‡ Louis-Charles Campeau,‡
Daniel A. DiRocco,*,‡ and Robert Britton*,†
†Department of Chemistry, Simon Fraser University, Burnaby, British Columbia, Canada
‡Department of Process Chemistry, Merck Research Laboratories, Rahway, New Jersey, United States
S
* Supporting Information
ABSTRACT: Late-stage C−H fluorination is an appealing reaction for
medicinal chemistry. However, the application of this strategy to process
research appears less attractive due to the formation and necessary
purification of mixtures of organofluorines. Here we demonstrate that γ-
fluoroleucine methyl ester, an intermediate critical to the large-scale
synthesis of odanacatib, can be accessed directly from leucine methyl
ester using a combination of the decatungstate photocatalyst and N-
fluorobenzenesulfonimide in flow. This efficient C−H fluorination reaction compares favorably with several generations of
classical γ-fluoroleucine process syntheses.
he importance of fluorine to drug discovery cannot be
fluoroleucine is an essential structural component of
Toverstated. Fluorination of drug leads is known to have odanacatib, where fluorination prevents hydroxylation of the
leucine subunit by CYP3A, significantly extending drug half-life
when compared to the parent lead candidate 1.11 Considering
the importance of (2S)-γ-fluoroleucine to the development of
odanacatib, its synthesis has been studied extensively. In fact,
several scalable and robust synthetic routes to this important
amino acid have already been published,12,13 five of which were
developed at Merck and are highlighted in Figure 1. In general,
the major challenges faced in each synthesis of γ-fluoroleucine
are the controlled introduction and preservation of the α-chiral
amine and tertiary alkyl fluoride functions. Consequently,
syntheses of γ-fluoroleucine have relied on chiral catalysis,13b,c
biocatalysis,13d the chiral pool,13a and chiral resolution13e to
introduce the α-chiral amine function. However, incorporating
the tertiary alkyl fluoride has only been possible using classical
alkene or epoxide hydrofluorination reactions. A selective C−H
fluorination of leucine that directly affords γ-fluoroleucine
would clearly represent the most efficient approach to this key
building block.
One of our laboratories has recently reported a photo-
catalytic fluorination of unactivated C−H bonds.10 This
reaction involves photoexcitation of a decatungstate catalyst
to produce an excited state intermediate capable of hydrogen
atom abstraction from various aliphatic substrates.14 The
resulting carbon-centered radical can then engage in fluorine
atom transfer reactions with N-fluorobenzenesulfonimide
(NSFI) in accordance with processes identified by Sammis
and Paquin.15 This reaction provides direct access to a range of
fluorinated organics in modest to good yield and is unique
among late-stage C−H fluorination reactions in the use of
profound effects on binding affinity, lipophilicity, and
absorption, distribution, metabolism, excretion, and toxicity
(ADME-tox).1 While the vast majority of fluorinated
pharmaceuticals are derived from commodity or readily
prepared trifluoromethyl or fluoroaryl building blocks, the
development of selective fluorination reagents and reactions
have provided new and enabling tools for medicinal chemistry.2
For example, C−H fluorination reactions reported by Sanford,3
Groves,4 Lectka,5 Hartwig,6 and others2c,7 provide a means to
directly fluorinate advanced drug candidates and attenuate
metabolism. However, the increasing sophistication of these
late-stage C−H fluorination processes can also present
challenges that complicate or protract large-scale efforts
required to support clinical studies and commercialization.8
In fact, it has recently been perpended whether advances in late
stage C−H fluorination are indeed useful or simply a “fancy
novelty”.8 Thus, despite obvious promise, the practical
application of late-stage C−H fluorination can be encumbered
by poor conversion or selectivity and consequently complex
and cost-prohibitive purifications. Here we report the direct
synthesis of γ-fluoroleucine, a crucial intermediate in the
synthesis of odanacatib.9 Importantly, by pairing our photo-
catalytic late-stage C−H fluorination strategy10 with high-
throughput reaction optimization and flow photochemistry, γ-
fluoroleucine was made readily available with dramatic
reductions in cost, time, reactor space, and waste produced.
While singular, the demonstration of this process on large-scale
highlights a bright future for late-stage C−H fluorination in
drug discovery and development.
Odanacatib (2) is a potent and selective cathepsin K
inhibitor currently in clinical trials for the treatment of
osteoporosis.9 The unusual fluorinated amino acid (2S)-γ-
Received: September 2, 2015
© XXXX American Chemical Society
A
DOI: 10.1021/acs.orglett.5b02532
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
over the course of 16 h. Repetition of this reaction in a mixture
of CH3CN and H2O resulted in improved conversion, albeit to
a mixture of the ester 10 and corresponding hydrolysis product
(not shown). Unfortunately, reaction with (2S)-leucine methyl
ester (8) failed to provide the desired γ-fluoro product 11 and
instead delivered an N-benzenesulfonamide derivative (entry
3). In an effort to avoid side reactions between the free amine
and NFSI, the fluorination was repeated with the HCl salt 9,
and we were delighted to find that small amounts of the γ-
fluoroleucine·HCl salt 12 were produced along with the
corresponding chloroleucine 13 (entries 4 and 5). The
chlorination of 9 was surprising, and several experiments
were carried out to gain further insight into this reaction. For
example, when the reaction described in entry 4 was repeated
without NFSI, no chloroleucine was produced. Replacing NFSI
with Na2S2O8 resulted in chlorination of leucine methyl ester in
a similar yield (∼30%). Notably, this latter reaction only
returned starting material when repeated without TBADT and
irradiation. Considering these facts, and the reduction potential
of NFSI (E° = −0.78 V vs SCE)2b and TBADT (E° = 2.85 V vs
SHE),18 the formation of chloroleucine 13 may involve
oxidation of chloride to molecular chlorine by photoexcited
TBADT18 and subsequent reaction of chlorine with the
generated carbon radical. Alternatively, single electron transfer
from the carbon radical to NFSI15a or oxidation of the carbon
radical by TBADT19 would lead to a carbocation intermediate
that could react with associated chloride. In an effort to
dissociate the chloride and avoid the formation of the
chloroleucine analogue 13, the fluorination reaction was
repeated in a mixture of CH3CN−H2O (entry 6). Gratifyingly,
under these conditions, the γ-fluoroleucine derivative 12 was
produced as the major product in excellent yield. The reaction
described in entry 6 was also repeated in a standard 5 mm
Figure 1. Five generations of large-scale γ-fluoroleucine syntheses
developed at Merck.
NFSI.3−7 Importantly, the decreased reactivity of NFSI relative
to other fluorine transfer agents (e.g., Selectfluor)16 also
renders this reaction compatible with a wide range of functional
groups.10 Considering the importance of γ-fluoroleucine to the
clinical development of odanacatib, we recognized that a direct
fluorination of leucine would have a significant impact on the
cost and ease of preparation of γ-fluoroleucine as well as the
waste produced by this process.17
Table 1 highlights our initial attempts to effect the
fluorination of (2S)-leucine. For simplicity, we first explored
this reaction on the corresponding desamino compound 7
using our optimized reaction conditions.10 As indicated in entry
1, we observed slow conversion to the 4-fluoro derivative 10
1
NMR tube and monitored by H NMR spectroscopy (Figure
1
2). As indicated in the stacked H NMR spectra, after only 1.5
Table 1. Fluorination of Analogues and Derivatives of
Leucine
1
Figure 2. H NMR spectra (600 MHz, CD3CN/D2O) of the C−H
fluorination of leucine methyl ester·HCl salt (9) (Table 1, entry 6)
taken at various time intervals. Resonances corresponding to the
methyl ester of γ-fluoroleucine·HCl salt (12) are highlighted with gray
boxes. The resonances that appear in the downfield region (δ > 7.0
ppm) indicate the conversion of (a) NFSI into (d) NHSI.
a
entry reactant
solvent
concn
product (yield)
b
1
2
3
4
5
6
7
7
8
9
9
9
CH3CN
1.0 M
1.0 M
0.3 M
0.2 M
2 M
10 (28%)
c
f
d
CH3CN−H2O
CH3CN
10 (40%)
h the conversion reached ∼25%, which doubled after 3 h.
Finally, after 16 h, less than 5% of the starting material
remained. Considering the efficiency of this process, production
of minimal byproducts, and reliance on commodity chemicals,
we were keen to evaluate whether this late-stage fluorination
was indeed amenable to large-scale and could supplant existing
process routes to γ-fluoroleucine (Figure 1).
Further optimization of the C−H fluorination reaction
described in Table 1 (entry 6) involved high-throughput
experimentation (HTE) using a recently developed photo-
chemical screening platform20 and MISER-LCMS analysis.21 As
e
11 (0%)
CH3CN
12 (35%), 13 (43%)
12 (18%), 13 (37%)
12 (83%), 13 (<5%)
CH3CN
CH3CN−H2O
0.2 M
a
Yield based on analysis of 1H NMR spectra recorded on crude
reaction mixtures using an internal standard. 0.1 equiv of NaHCO3
was added. Ratio of CH3CN/H2O = 2:1. Combined yield of methyl
ester and corresponding carboxylic acid. The major isolated product
b
c
d
e
was the N-benzenesulfonamide derivative of leucine methyl ester.
f
Ratio of CH3CN/H2O = 2:1.
B
DOI: 10.1021/acs.orglett.5b02532
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
reproduced on a 20 mg scale and provided a 97% assay yield
after 18 h of irradiation.
In order to evaluate the utility of this direct fluorination for
the large-scale production of odanacatib, a 365 nm flow
photoreactor was built following the basic design principles
described by Booker-Milburn.23 The reactor was calculated to
have a total light output of 32 W (λmax = 365 nm) and consists
̋
of 100 ft of 1/16 I.D. PFA tubing (total reactor volume = 60
mL).22 Using the optimized reaction conditions (vide supra)
fluorination was performed on 190 mmol scale (46.5 g), at 0.2
M (9:1 MeCN/water), and pumped through the reactor at 0.5
mL/min (2 h residence time). After the total volume was
collected, the mixture was concentrated under reduced pressure
followed by azeotropic distillation with 2-MeTHF. Following
this process, (S)-γ-fluoroleucine methyl ester bisulfate could be
isolated by direct precipitation from 2-MeTHF (44.7 g, 90%)
(Figure 4).
Figure 3. High-throughput optimization and MISER analysis of
reaction conditions. Experiments performed on 10 μmol scale and
irradiated with 365 nm LEDs for 16 h. Data are depicted as a heat map
showing relative product concentrations based on integration of
LCMS mass response. The standardized peaks displayed in the heat
map are reproduced from the LCMS and correspond to the peak for
the desired γ-fluoroleucine derivative.
depicted in Figure 3, we evaluated several reaction parameters
in a 96 vial, 365 nm photoreactor on 10 μmol scale, including
the amino acid salt, catalyst loading, concentration, and solvent
composition. Additionally, we explored the effect of the catalyst
counterion in hopes of replacing tetrabutylammonium with a
less expensive sodium counterion.22
Figure 4. Scalable flow synthesis of γ-fluoroleucine and reactor design.
In conclusion, a robust, one-step synthesis of (S)-γ-
fluoroleucine methyl ester has been developed that relies on
a direct photocatalytic C−H fluorination reaction and compares
well to previous, large-scale syntheses of (S)-γ-fluoroleucine
esters (Figures 1 and 5). This process has been executed on
multigram scale with little deviation in conversion or yield from
exploratory small-scale reactions (<100 mg). This investigation
As highlighted in the heat map (Figure 3) we were pleased to
find that the sodium salt of decatungstate (rows 1−6) is
superior to TBADT in catalyzing the fluorination of all salts of
leucine methyl ester. Additionally, the HCl salt 9 is least
compatible with the desired process due to competitive
chlorination (see separate MISERgram for production of
chloroleucine, bottom of Figure 3). While the addition of
water improved the outcome of reactions involving the HCl salt
9 (e.g., A3 and A6 in heat map), owing to the limited solubility
of NFSI in aqueous acetonitrile, a highly solvent intensive
process would be required to both maintain a homogeneous
reaction and sufficiently limit the production of chloroleucine.
Alternatively, chloroleucine byproducts were avoided through
the use of either bisulfate or mesylate (MSA) salts, which also
proved to be relatively insensitive to solvent composition or
concentration (rows C−F in heat map). As highlighted in rows
G and H, reactions involving the napthalene disulfonic acid
(NDA) salt proved problematic due to decreased solubility and
required both a higher catalyst loading and water content to
reach reasonable levels of conversion. Considering these results
as well as ease of isolation and downstream processing, the
bisulfate salt was ultimately selected as the preferred salt. The
optimal reaction conditions (row C, column 2) were
Figure 5. Comparison of overall yield and number of steps for the
previous (Figure 1) and present syntheses of (S)-γ-fluoroleucine
esters. Numbers on horizontal axis represent individual large-scale
synthesis campaigns for (S)-γ-fluoroleucine esters (see Figure 1) with
the present synthesis represented by a green bar (6).
C
DOI: 10.1021/acs.orglett.5b02532
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
McKay, D. J.; Percival, M. D.; Riendeau, D.; Rodan, S. B.; Therien, M.;
Truong, V.-L.; Wesolowski, G.; Zamboni, R.; Black, W. C. Bioorg. Med.
Chem. Lett. 2006, 16, 1985. (b) Gauthier, J. Y.; Chauret, N.; Cromlish,
W.; Desmarais, S.; Duong, L. T.; Falgueyret, J.-P.; Kimmel, D. B.;
highlights the remarkable efficiency and selectivity of the
decatungstate-catalyzed fluorination reaction,10 where the final
purification of (S)-γ-fluoroleucine methyl ester could be
accomplished by straightforward precipitation and filtration.
Importantly, this first demonstration of late-stage fluorination
of an unactivated C−H bond for process research purposes
further emphasizes the important role new C−H fluorination
strategies can play in drug discovery and development.8
́
Lamontagne, S.; Leger, S.; LeRiche, T.; Li, C. S.; Masse, F.; McKay, D.
J.; Nicoll-Griffith, D. A.; Oballa, R. M.; Palmer, J. T.; Percival, M. D.;
Riendeau, D.; Robichaud, J.; Rodan, G. A.; Rodan, S. B.; Seto, C.;
Therien, M.; Truong, V.-L.; Venuti, M. C.; Wesolowski, G.; Young, R.
N.; Zamboni, R.; Black, W. C. Bioorg. Med. Chem. Lett. 2008, 18, 923.
(12) (a) Papageorgiou, C.; Borer, X.; French, R. R. Bioorg. Med.
Chem. Lett. 1994, 4, 267. (b) Padmakshan, D.; Bennett, S. A.; Otting,
G.; Easton, C. J. Synlett 2007, 2007, 1083.
ASSOCIATED CONTENT
* Supporting Information
■
S
(13) (a) Truong, V. L.; Gauthier, J. Y.; Boyd, M.; Roy, B.; Scheigetz,
J. Synlett 2005, 8, 1279. (b) Nadeau, C.; Gosselin, F.; O’Shea, P. D.;
Davies, I. W.; Volante, R. P. Synlett 2006, 2, 291. (c) Humphrey, G.;
Chung, C. K.; Rivera, N. R.; Belyk, K. PCT Patent Application WO
2013/148550 A1. (d) Limanto, J.; Shafiee, A.; Devine, P. N.;
Upadhyay, V.; Desmond, R. A.; Foster, B. R.; Gauthier, D. R., Jr.;
Reamer, R. A.; Volante, R. P. J. Org. Chem. 2005, 70, 2372. (e) Truppo,
M. D.; Hughes, G. Org. Process Res. Dev. 2011, 15, 1033. (f) Rivera, N.;
Pendri, Y. R.; Mende, S.; Reddy, B. N. Patent Application WO2013/
148554 A1.
(14) (a) Tzirakis, M. D.; Lykakis, I. N.; Orfanopoulos, M. Chem. Soc.
Rev. 2009, 38, 2609. (b) Protti, S.; Ravelli, D.; Fagnoni, M.; Albini, A.
Chem. Commun. 2009, 7351. (c) Combs-Walker, L. A.; Hill, C. L. J.
Am. Chem. Soc. 1992, 114, 938. (d) Renneke, R. F.; Pasquali, M.; Hill,
C. L. J. Am. Chem. Soc. 1990, 112, 6585.
(15) (a) Rueda-Becerril, M.; Sazepin, C. C.; Leung, J. C. T.;
Okbinoglu, T.; Kennepohl, P.; Paquin, J.-F.; Sammis, G. M. J. Am.
Chem. Soc. 2012, 134, 4026. (b) Leung, J. C. T.; Chatalova-Sazepin,
C.; West, J. G.; Rueda-Becerril, M.; Paquin, J.-F.; Sammis, G. M.
Angew. Chem., Int. Ed. 2012, 51, 10804.
(16) (a) Toullec, P. Y.; Bonaccorsi, C.; Mezzetti, A.; Togni, A. Proc.
Natl. Acad. Sci. U. S. A. 2004, 101, 5810. (b) Gilicinski, A. G.; Pez, G.
P.; Syvret, R. G.; Lal, G. S. J. Fluorine Chem. 1992, 59, 157.
(17) Wender, P. A. Nat. Prod. Rep. 2014, 31, 433.
(18) Texier, I.; Delaire, J. A.; Giannotti, C. Phys. Chem. Chem. Phys.
2000, 2, 1205.
(19) Jen, S.-F.; Anderson, A. B.; Hill, C. L. J. Phys. Chem. 1992, 96,
5658.
(20) DiRocco, D. A.; Dykstra, K.; Krska, S.; Vachal, P.; Conway, D.
V.; Tudge, M. Angew. Chem., Int. Ed. 2014, 53, 4802.
(21) Welch, C. J.; Gong, X.; Schafer, W.; Pratt, E. C.; Brkovic, T.;
Pirzada, Z.; Cuff, J. F.; Kosjek, B. Tetrahedron: Asymmetry 2010, 21,
1674.
(22) See Supporting Information for full experimental details.
(23) Hook, B. D. A.; Dohle, W.; Hirst, P. R.; Pickworth, M.; Berry,
M. B.; Booker-Milburn, K. I. J. Org. Chem. 2005, 70, 7558.
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.or-
glett.5b02532.
Experimental procedures, high throughput experimenta-
tion, and photochemical flow reactor design (PDF)
AUTHOR INFORMATION
Corresponding Authors
■
*E-mail: rbritton@sfu.ca.
*E-mail: daniel.dirocco@merck.com.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported by an NSERC Discovery Grant to
R.B., an MSFHR Career Investigator Award to R.B., an NSERC
Post-Graduate Scholarship for S.D.H., and SFU Graduate
Fellowships for S.D.H., D.K., and M.H.
REFERENCES
■
(1) (a) Hagmann, W. K. J. Med. Chem. 2008, 51, 4359. (b) O’Hagan,
D. J. Fluorine Chem. 2010, 131, 1071. (c) Bohm, H.-J.; Banner, D.;
Bendels, S.; Kansy, M.; Kuhn, B.; Muller, K.; Obst-Sander, U.; Stahl,
M. ChemBioChem 2004, 5, 637. (d) Kirk, K. L. J. Fluorine Chem. 2006,
̈
̈
127, 1013.
(2) For selected reviews, see: (a) Kirk, K. L. Org. Process Res. Dev.
2008, 12, 305. (b) Liang, T.; Neumann, C. N.; Ritter, T. Angew. Chem.,
Int. Ed. 2013, 52, 8214. (c) Champagne, P. A.; Desroches, J.; Hamel,
J.-D.; Vandamme, M.; Paquin, J.-F. Chem. Rev. 2015, 115, 9073.
(d) Furuya, T.; Kamlet, A. S.; Ritter, T. Nature 2011, 473, 470.
(3) Hull, K. L.; Anani, W. Q.; Sanford, M. S. J. Am. Chem. Soc. 2006,
128, 7134.
(4) Liu, W.; Huang, X.; Cheng, M.-J.; Nielsen, R. J.; Goddard, W. A.;
Groves, J. T. Science 2012, 337, 1322.
(5) Bloom, S.; Pitts, C. R.; Miller, D. C.; Haselton, N.; Holl, M. G.;
Urheim, E.; Lectka, T. Angew. Chem., Int. Ed. 2012, 51, 10580.
(6) Fier, P. S.; Hartwig, J. F. Science 2013, 342, 956.
(7) For recent review, see: Ma, J.-A.; Li, S. Org. Chem. Front. 2014, 1,
712.
(8) Neumann, C. N.; Ritter, T. Angew. Chem., Int. Ed. 2015, 54, 3216.
(9) (a) Langdahl, B.; Binkley, N.; Bone, H.; Gilchrist, N.; Resch, H.;
Rodriguez Portales, J.; Denker, A.; Lombardi, A.; Le Bailly De
Tilleghem, C.; DaSilva, C.; Rosenberg, E.; Leung, A. J. Bone Miner. Res.
2012, 27, 2251. (b) Brixen, K.; Chapurlat, R.; Cheung, A. M.; Keaveny,
T. M.; Fuerst, T.; Engelke, K.; Recker, R.; Dardzinski, B.; Verbruggen,
N.; Ather, S.; Rosenberg, E.; de Papp, A. E. J. Clin. Endocrinol. Metab.
2013, 98, 571.
(10) (a) Halperin, S. D.; Fan, H.; Chang, S.; Martin, R. E.; Britton, R.
Angew. Chem., Int. Ed. 2014, 53, 4690. (b) Nodwell, M. B.; Bagai, A.;
Halperin, S. D.; Martin, R. E.; Knust, H.; Britton, R. Chem. Commun.
2015, 51, 11783.
(11) (a) Li, C. S.; Deschenes, D.; Desmarais, S.; Falgueyret, J.-P.;
Gauthier, J. Y.; Kimmel, D. B.; Leger, S.; Masse, F.; McGrath, M. E.;
́ ́
D
DOI: 10.1021/acs.orglett.5b02532
Org. Lett. XXXX, XXX, XXX−XXX