A Selective Synthesis of 2,2-Difluorobicyclo[1.1.1]pentane Analogues: "BCP-F2"

Article abstract of DOI:10.1021/acs.orglett.9b02026

The bicyclo[1.1.1]pentane (BCP) motif has been utilized as bioisosteres in drug candidates to replace phenyl, tert-butyl, and alkynyl fragments in order to improve physicochemical properties. However, bceause of the difficulty of synthesis, most BCP analo

Full text of DOI:10.1021/acs.orglett.9b02026

Letter
pubs.acs.org/OrgLett
Cite This: Org. Lett. XXXX, XXX, XXX−XXX
A Selective Synthesis of 2,2-Difluorobicyclo[1.1.1]pentane
Analogues: “BCP‑F2”
Xiaoshen Ma,*,† David L. Sloman,† Yongxin Han,† and David J. Bennett†
†Department of Discovery Chemistry, Merck & Co., Inc., 33 Avenue Louis Pasteur, Boston, Massachusetts 02115, United States
S
* Supporting Information
ABSTRACT: The bicyclo[1.1.1]pentane (BCP) motif has been utilized as
bioisosteres in drug candidates to replace phenyl, tert-butyl, and alkynyl fragments
in order to improve physicochemical properties. However, bceause of the difficulty of
synthesis, most BCP analogues prepared only bear 1,3-“para”-substituents. We report
the first selective synthesis of 2,2-difluorobicyclo[1.1.1]pentanes via difluorocarbene
insertion into bicyclo[1.1.0]butanes. Moreover, this methodology should inspire future
studies on synthesis of other “ortho/meta-substituted” BCPs via similar mechanisms.
ioisostere replacement is a fruitful strategy in fine-tuning
The synthesis of functionalized BCP building blocks has
been challenging, because of the high ring strain of the system.2
A successful strategy for the synthesis of functionalized BCPs
employs [1.1.1]propellane (4) as the starting material (Scheme
2A).2 Early works in this field dated back to the 1970s.7
Recently, de Meijere and co-workers achieved the synthesis of
iodoalkylation of 4 with ultraviolet light irradiation to provide
5.8 Several groups also developed coupling reactions involving
5 for further funtionalization.8,9 In 2011, Pfizer reported the
synthesis of BCP hydrazine derivative 6 using hydrogen atom
transfer chemistry.10 The BCP hydrazine derivative 6 could be
further converted to 1-bicyclo[1.1.1]pentylamine.10 Baran and
co-workers reported a facile protocol achieving BCP analogues
(7) with tertiary amine functionalities, using a strain release
amination strategy.11 Built on the radical and anionic opening
of [1.1.1]propellane (4), the Uchiyama12 and Knochel13
groups developed strategies to obtain difunctionalized BCP
analogues 8 and 9, respectively. Shelp and Walsh reported the
synthesis of BCP benzylamines 10 by opening 4 with 2-azaallyl
anions.14 Although these methodologies allow the access of
various substituted BCP analogues, the usage of 4 as the
starting material limits the functionalities on the methylene
positions of the BCP moiety. In the meantime, a less-
developed strategy toward the synthesis of BCP analogues
involves a dichlorocarbene insertion into bicyclo[1.1.0]butanes
12 (Scheme 2B, the dichloromethylene functionality “CCl2” is
usually reduced to “CH2” in subsequent steps).15 The major
issue of this route was the low yields in both the
dichlorocarbene insertion steps (12 → 13) and the reduction
steps (13 → 14).15 However, this strategy inspired us to
explore the potential of achieving BCP analogue synthesis by
incorporating other carbene insertion processes. Since the
incorporation of fluorine atoms can significantly modify the
physicochemical properties of drug candidates,16 we wish to
prepare 2,2-difluorobicyclo[1.1.1]pentanes 16. The additional
B
the physicochemical properties of drug candidates.1 Since
the 1,3-disubstituted bicyclo[1.1.1]pentane (BCP) motif
shares a similar geometry with 1,4-disubstituted benzene,
BCP has been used as bioisosteres for phenyl groups.2 The first
example of such replacement was reported by Pellicciari and
co-workers in 1996, using 1 (Scheme 1) as a novel mGluR1
Scheme 1. Bicyclo[1.1.1]pentane Motif Used As
Bioisosteres of para-Substituted Phenyl (1), tert-Butyl (2),
and Alkynyl Moieties (3)
antagonist.3 Similar strategies have been applied to a γ-
secretase inhibitor,4 an LpPLA2 inhibitor,5 and several other
central nervous system drug candidates.6 It has been
demonstrated that using BCP as the phenyl bioisostere
replacement can significantly improve aqueous solubility,
membrane permeability, and in vitro metabolic stability.1a
BCP fragments also have been applied as tert-butyl (2, Scheme
1) and alkynyl isosteres (3, Scheme 1).
Received: June 12, 2019
© XXXX American Chemical Society
A
DOI: 10.1021/acs.orglett.9b02026
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
employed as the dichlorocarbene precursor in reported
syntheses of 13,15 we first evaluated the analogous difluor-
ocarbene precursors (Table 1, entries 1 and 2). Sodium
Scheme 2. (A) Recent Advances in the Syntheses of
Functionalized BCP from [1.1.1]Propellane (4); (B) A Less
Developed Synthesis of BCP from Bicyclo[1.1.0]butane
(12); (C) This Work
Table 1. Condition Optimization for BCP-F2 Synthesis
a
b
entry
conditions
yield (%)
1
2
3
4
5
6
7
8
ClF2CO2Na, diglyme, 177 °C
BrF2CO2Na, diglyme, 150 °C
(34)
13
3
40
56
45
61
66
FSO2CF2CO2TMS, NaF, toluene, 150 °C
FSO2CF2CO2TMS, NaF, MeOBz, 105 °C
FSO2CF2CO2TMS, NaF, tBuOAc, 90 °C
FSO2CF2CO2TMS, NaF, EtOAc, 75 °C
FSO2CF2CO2TMS, NaF, CH3CN, 90 °C
FSO2CF2CO2TMS, NaF, dioxane, 90 °C
a
For each reaction, 15a (20 mg) was employed. For detailed reaction
b
conditions, see the Supporting Information. Yields determined by
19F NMR using hexafluorobenzene as the internal standard. Isolated
yields are shown in parentheses.
chlorodifluoroacetate afforded the desired product 16a in 34%
isolated yield (entry 1), which is superior than sodium
bromodifluoroacetate (13% NMR yield, Table 1, entry 2). The
usage of diglyme as the reaction solvent and the difficulties in
reaction operations (addition of slurries of the sodium salts
into the starting material at high temperatures) encouraged us
to explore the possibility of utilizing other difluorocarbene
precursors.19 Although only 3% of the desired product was
observed by 19F NMR when using trimethylsilyl 2-
fluorosulfonyl-2,2-difluoroacetate (TFDA) as the difluorocar-
bene precursor in toluene at 150 °C (Table 1, entry 3), a
systematic decrease of reaction temperature by changing the
reaction solvent showed that 90 °C was the optimal reaction
temperature (Table 1, entries 3−8). Finally, the optimized
conditions, using TFDA as the difluorocarbene precursor,
activated by a substoichiometric amount of sodium fluoride in
dioxane at 90 °C, afforded the desired product 16a in 66%
yield by 19F NMR (Table 1, entry 8).
We explored the preliminary substrate scope of BCP-F2
synthesis on preparative scales (Scheme 3, 100 mg scales). The
standard product 16a was obtained in 56% isolated yield under
optimized conditions (Table 1, entry 8). The lower yield of the
unsubstituted product 16b (34%), compared to 16a was
attributed to the lower stability of the corresponding
bicyclo[1.1.0]butane starting material at the reaction temper-
ature. A variety of substituents at the aromatic ring have been
examined. Generally, substrates with electron-withdrawing
groups on the aromatic ring exhibited superior reactivities,
compared to those with electron-donating groups. As shown,
substrates with para-trifluoromethyl (16a, 56%), para-bromo
(16c, 41%), para-fluoro (16d, 53%), and para-methoxycar-
bonyl (16e, 63%) functionalities were obtained with higher
yields than the unsubstituted analogue 16b (35%). For para-
methoxy analogue 16f, complicated decompositions of the
starting material were observed under reaction conditions. This
difluoromethylene functionality in 16 can both finely tune the
physicochemical properties of the BCP fragment and introduce
additional interactions (e.g., fluorine−dipole interaction and
hydrogen bonding interaction)16 with the target protein that is
not available with traditional BCP moieties. To our knowledge,
there has been only two reports of fluorinated BCP analogues
prior to our work.17 The reported BCP fluorination conditions
were harsh and mixtures of multifluorinated BCP analogues
were synthesized.17 Michl and co-workers also reported the
synthesis of chlorinated BCP analogues in 2019.18 Numerous
attempts for selective synthesis of fluorinated BCP analogues
have been ongoing for decades in medicinal chemistry
research; however, success was scarce. Herein, we report the
first selective synthesis of 2,2-difluorobicyclo[1.1.1]pentane
analogues, which we named BCP-F2, as an exotic functional
group for potential applications in medicinal chemistry.
Bicyclo[1.1.0]butane 15a was prepared as the standard
substrate for condition optimizations, because of the ease of
analysis by 19F NMR. Since sodium trichloroacetate was
B
DOI: 10.1021/acs.orglett.9b02026
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
a
These results led to the mechanistic proposal that is
described in Scheme 4A. The bicyclo[1.1.0]butane starting
Scheme 3. Preliminary Substrate Scope
Scheme 4. (A) Proposed Mechanism for BCP-F2 Synthesis;
(B) Unproductive Pathways of I under Heat
material I is in resonance to the diradical intermediate II.
Meanwhile, fluoride-catalyzed decomposition of TFDA
provided difluorocarbene III, releasing trimethylsilyl fluoride
(TMSF), carbon dioxide (CO2), and sulfur dioxide (SO2). The
desired product IV could be afforded either via direct insertion
of the difluorocarbene III into the C−C bond in I or by
stepwise radical addition with II. The overall moderate yields
of the BCP-F2 synthesis are attributed to the potential
unproductive decomposition pathways of I (Scheme 3B; see
Scheme S1 in the Supporting Information).2
To study the chemostability of the BCP-F2 functionality, we
prepared bicyclo[1.1.0]butane analogue 15n (Scheme 5). The
BCP-F2 analogue 16n could be obtained in 47% isolated yield;
however, for the ease of operation, we subjected the crude
reaction mixture after BCP-F2 synthesis directly to further
functionalization without purification. We first tested the
stability of BCP-F2 fragment under extreme conditions, i.e.,
strongly acidic environment to achieve 17a and strongly
reductive conditions to achieve 17b. Treating the crude
reaction mixture with anhydrous hydrochloric acid provided
the corresponding BCP-F2 carboxylic acid 17a in 40% isolated
yield. Reducing BCP-F2 intermediate 16n by lithium
aluminum hydride (LiAlH4) in refluxing tetrahydrofuran
(THF) provided the corresponding alcohol 17b in 32%
isolated yield. We then tested the compatibility of BCP-F2
functionality in three of the most commonly used trans-
formations in medicinal chemistry (i.e., palladium-catalyzed
C−N coupling, Suzuki coupling, and Stille coupling reactions).
Tandem BCP-F2 formation and C−N cross coupling provided
the azetidine derivative 17c in 39% yield. Telescoping BCP-F2
formation and Suzuki coupling with 4-pyridylboronic acid
provided 17d in 31% yield over two steps. The Stille coupling
product 17e was obtained in 29% isolated yield when coupled
with 2-(tributylstannyl)oxazole in the presence of tetrakis-
(triphenylphosphine)palladium(0) [Pd(PPh3)4], copper(I)
iodide (CuI), and cesium fluoride (CsF) in refluxing N,N-
a
All yields are isolated yields. Bicyclo[1.1.0]butane (100 mg), NaF
(50 mol %), TFDA (3.00 equiv), dioxane (0.33 M), 90 °C, unless
otherwise noticed. bReaction conducted on a 700 mg scale. cComplex
decomposition of the corresponding bicyclo[1.1.0]butane starting
material was observed.
is because the bicyclo[1.1.0]butane substrate 15f (not shown,
see the Supporting Information) was highly unstable, even at
room temperature in solutions. This electronic trend was also
observed with meta-substituted analogues, as the yields
decrease with the increasing electron density of the aromatic
substituent on the bicyclo[1.1.0]butane starting material
[meta-fluoro analogue 16g (47%) > meta-methyl analogue
16h (32%) > meta-methoxy analogue 16i (7%)]. The 4-chloro-
3-fluoro analogue 16j was obtained in 21% yield. We also
examined the nature of the ester group and both benzyl (16k,
44%) and tert-butyl esters (16l, 54%) were obtained in
synthetically useful yields. To demonstrate the necessity of the
aromatic group, we prepared the monosubstituted
bicyclo[1.1.0]butane substrate 15m (not shown; see the
Supporting Information). Under optimized conditions, the
desired product 16m was not observed.
C
DOI: 10.1021/acs.orglett.9b02026
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
al. Representative examples of bicyclo[1.1.0]butanes that failed
in preparation are summarized in Scheme 7. The ortho-
Scheme 5. Further Functionalization of BCP-F2
Analogues:
a b
,
Scheme 7. Substrates That Failed in Preparation
substituted analogues (e.g., 18a and 18b) were not suitable for
the synthetic route. Electron-poor aryl analogues (e.g., 18c)
also failed because of low nucleophilicity of the corresponding
organometallic reagents in the addition reaction into the
ketone functionality of 11. Another major limitation for the
synthesis of bicyclo[1.1.0]butanes is the incompatibility with
heteroaryl substituents (e.g., 18d and 18e).
Note that the yield for this transformation (formation of
16a; see Scheme 8) is similar to the analogue obtained with
dichlorocarbene analogue 19.5
Scheme 8. Comparison of Dichlorocarbene and
Difluorocarbene Insertion Results Toward the Synthesis of
Darapladib Analogues (A) 20 and (B) 20-F2
a
b
Yields over two steps. Mechanism routes: (a) HCl (10.0 equiv),
dioxane, 24 °C, 12 h, 40%; (b) LiAlH4 (1.00 equiv), THF, reflux, 15
min, 32%. (c) XantPhos Pd G3 (10 mol %), 3-(difluoromethyl)-
azetidine hydrochloride (3.00 equiv), cesium carbonate (5.00 equiv),
dioxane, 120 °C, 4 h, 39%; (d) XPhos Pd G2 (10 mol %), 4-
pyridylboronic acid (3.00 equiv), sodium carbonate (3.00 equiv),
dimethoxyethane, 90 °C, 1.5 h, 31%; and (e) Pd(PPh3)4 (10 mol %),
CuI (10 mol %), CsF (2.00 equiv), 2-(tributylstannyl)oxazole (2.00
equiv), DMF, 130 °C, 1 h, 29%.
dimethyl formamide (DMF). These apparently harsh reaction
conditions were intentionally chosen to establish the stability
of the BCP-F2 functionality under forced conditions.
Finally, to demonstrate the scalability of this methodology,
we conducted the synthesis of 16o with 1.00 g of 15o under
the optimized conditions (Scheme 6). BCP-F2 analogue 16o
was obtained in 38% isolated yield (463.7 mg), which is
comparable to the 37% isolated yield of 16o obtained on a
100-mg-scale reaction.
Scheme 6. Gram-Scale Reaction
We also conducted some preliminary in silico profiling of
BCP-F2 analogue 21, compared to its analogue 22. The
selective data that are summarized in Table 2 are predicted
properties that have been calculated using proprietary
models.20 This table shows that, although the log D and
polar surface area (PSA) values are similar between BCP-F2
analogue 21 (log D = 0.63, PSA = 44.6) and BCP analogue 22
(log D = 0.50, PSA = 44.9), the c log P value of 21 (c log P =
1.59) is more than 0.5 log units lower than 22 (c log P = 2.11).
The difluorosubstituents on 21 also rendered a lower pKa of
the carboxylic acid functionality (pKa = 3.17), compared to 22
(pKa = 4.81). Further in vitro and in vivo profilings of BCP-F2
analogues are underway in our laboratory.
Currently, we are investigating further scale-up procedures
for 16o. We are also in the process of developing optimal
conditions to oxidatively convert the phenyl group in 16o to
the carboxylic acid functionality to achieve derivatives with
broader application for medicinal chemistry.
A major limitation of this transformation is the potential
difficulty in accessing the bicyclo[1.1.0]butane starting materi-
To conclude, we report the first selective synthesis of 2,2-
difluorobicyclo[1.1.1]pentane analogues, which we have
D
DOI: 10.1021/acs.orglett.9b02026
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
̈
(2) Dilmac,̧ A. M.; Spuling, E.; de Meijere, A.; Brase, S. Angew.
Chem., Int. Ed. 2017, 56, 5684.
Table 2. Predicted Properties
(3) (a) Pellicciari, R.; Raimondo, M.; Marinozzi, M.; Natalini, B.;
Costantino, G.; Thomsen, C. J. Med. Chem. 1996, 39, 2874.
(b) Costantino, G.; Maltoni, K.; Marinozzi, M.; Camaioni, E.;
Prezeau, L.; Pin, J.-P.; Pellicciari, R. Bioorg. Med. Chem. 2001, 9, 221.
(4) Stepan, A. F.; Subramanyam, C.; Efremov, I. V.; Dutra, J. K.;
O’Sullivan, T. J.; DiRico, K. J.; McDonald, W. S.; Won, A.; Dorff, P.
H.; Nolan, C. E.; Becker, S. L.; Pustilnik, L. R.; Riddell, D. R.;
Kauffman, G. W.; Kormos, B. L.; Zhang, L.; Lu, Y.; Capetta, S. H.;
Green, M. E.; Karki, K.; Sibley, E.; Atchison, K. P.; Hallgren, A. J.;
Oborski, C. E.; Robshaw, A. E.; Sneed, B.; O’Donnell, C. J. J. Med.
Chem. 2012, 55, 3414.
(5) Measom, N. D.; Down, K. D.; Hirst, D. J.; Jamieson, C.; Manas,
E. S.; Patel, V. K.; Somers, D. O. ACS Med. Chem. Lett. 2017, 8, 43.
(6) Lea IV, P. M.; Faden, A. I. CNS Drug Rev. 2006, 12, 149.
(7) (a) Hall, H. K.; Smith, C. D.; Blanchard, E. P.; Cherkofsky, S. C.;
Sieja, J. B. J. Am. Chem. Soc. 1971, 93, 121. (b) Wiberg, K. B.;
Waddell, S. T. J. Am. Chem. Soc. 1990, 112, 2194. (c) Kaszynski, P.;
McMurdie, N. D.; Michl, J. J. Org. Chem. 1991, 56, 307. (d) Kaszynski,
P.; Friedli, A. C.; Michl, J. J. Am. Chem. Soc. 1992, 114, 601.
(8) Messner, M.; Kozhushkov, S. I.; de Meijere, A. Eur. J. Org. Chem.
2000, 2000, 1137.
named BCP-F2. The reaction proceeds rapidly under mild
conditions with commercially available reagents. We proposed
a diradical−carbene combination mechanism. The desired
products were obtained with synthetically useful yields on 100-
mg scales. We further demonstrated that the BCP-F2
functionality is stable under harsh reaction conditions. We
believe this BCP-F2 work should find application as an exotic
functional group in medicinal chemists’ journey to escape the
“flatland.”1a Currently, further studies to broaden the substrate
scope as well as using other carbene precursors to provide 2-
substituted and/or 2,2-disubstituted bicyclo[1.1.1]pentane
analogues are underway in our laboratory.
(9) (a) Daniel Rehm, J. D.; Ziemer, B.; Szeimies, G. Eur. J. Org.
̈
Chem. 1999, 1999, 2079. (b) Patzel, M.; Sanktjohanser, M.; Doss, A.;
Henklein, P.; Szeimies, G. Eur. J. Org. Chem. 2004, 2004, 493. (c) de
Meijere, A.; Zhao, L.; Belov, V. N.; Bossi, M.; Noltemeyer, M.; Hell,
S. W. Chem. - Eur. J. 2007, 13, 2503. (d) Caputo, D. F. J.; Arroniz, C.;
Durr, A. B.; Mousseau, J. J.; Stepan, A. F.; Mansfield, S. J.; Anderson,
̈
E. A. Chem. Sci. 2018, 9, 5295.
(10) Bunker, K. D.; Sach, N. W.; Huang, Q.; Richardson, P. F. Org.
Lett. 2011, 13, 4746.
(11) Gianatassio, R.; Lopchuk, J. M.; Wang, J.; Pan, C.-M.; Malins,
L. R.; Prieto, L.; Brandt, T. A.; Collins, M. R.; Gallego, G. M.; Sach,
N. W.; Spangler, J. E.; Zhu, H.; Zhu, J.; Baran, P. S. Science 2016, 351,
241.
ASSOCIATED CONTENT
■
S
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.or-
glett.9b02026.
(12) Kanazawa, J.; Maeda, K.; Uchiyama, M. J. Am. Chem. Soc. 2017,
139, 17791.
(13) Makarov, I. S.; Brocklehurst, C. E.; Karaghiosoff, K.; Koch, G.;
Knochel, P. Angew. Chem., Int. Ed. 2017, 56, 12774.
(14) Shelp, R. A.; Walsh, P. J. Angew. Chem., Int. Ed. 2018, 57,
15857.
Experimental procedures and spectroscopic data for all
new compounds (PDF)
(15) Applequist, D. E.; Renken, T. L.; Wheeler, J. W. J. Org. Chem.
1982, 47, 4985.
AUTHOR INFORMATION
■
(16) Gillis, E. P.; Eastman, K. J.; Hill, M. D.; Donnelly, D. J.;
Meanwell, N. A. J. Med. Chem. 2015, 58, 8315.
(17) Levin, M. D.; Hamrock, S. J.; Kaszynski, P.; Shtarev, A. B.;
Levina, G. A.; Noll, B. C.; Ashley, M. E.; Newmark, R.; Moore, G. G.
I.; Michl, J. J. Am. Chem. Soc. 1997, 119, 12750.
Corresponding Author
*Email: xiaoshen.ma@merck.com.
ORCID
Xiaoshen Ma: 0000-0002-0713-2896
Notes
̌
̌
́
̌
́
̌
́
́
(18) Kaleta, J.; Roncevic, I.; Císarova, I.; Dracínsky, M.; Solínova, V.;
̌
̌
Kasicka, V.; Michl, J. J. Org. Chem. 2019, 84, 2448.
(19) Ni, C.; Hu, J. Synthesis 2014, 46, 842.
The authors declare no competing financial interest.
(20) Sanders, J. M.; Beshore, D. C.; Culberson, J. C.; Fells, J. I.;
Imbriglio, J. E.; Gunaydin, H.; Haidle, A. M.; Labroli, M.; Mattioni, B.
E.; Sciammetta, N.; Shipe, W. D.; Sheridan, R. P.; Suen, L. M.; Verras,
A.; Walji, A.; Joshi, E. M.; Bueters, T. J. Med. Chem. 2017, 60, 6771.
ACKNOWLEDGMENTS
■
We gratefully acknowledge colleagues at Merck & Co., Inc., Dr.
Andrew M. Haidle, Dr. Matthew J. Mitcheltree, Dr. Min Lu,
Dr. Charles S. Yeung, Dr. Anthony J. Roecker, and Craig R.
Gibeau for constructive discussion and valuable criticisms for
the manuscript. We thank colleagues at Merck & Co., Inc., Dr.
Josep Sauri, Dr. Donovon Adpressa, and Dr. Adam Beard for
analytical characterization of the final compounds.
REFERENCES
■
(1) (a) Lovering, F.; Bikker, J.; Humblet, C. J. Med. Chem. 2009, 52,
6752. (b) Brown, N.; Mannhold, R. Bioisosteres in Medicinal
Chemistry; Wiley−VCH: Weinheim, Germany, 2012. (c) Mykhailiuk,
P. K. Org. Biomol. Chem. 2019, 17, 2839. (d) Kanazawa, J.; Uchiyama,
M. Synlett 2019, 30, 1.
E
DOI: 10.1021/acs.orglett.9b02026
Org. Lett. XXXX, XXX, XXX−XXX