A practical synthesis of m-prostaglandin E synthase-1 inhibitor MK-7285

Article abstract of DOI:10.1021/jo901798d

(Chemical Equation Presented) A practical, kilogram-scale chromatography-free synthesis of mPGE synthase I inhibitor MK-7285 is described. The route features a convergent assembly of the core phenanthrene unit via amide-directed ortho-metalation and proxi

Full text of DOI:10.1021/jo901798d

pubs.acs.org/joc
A Practical Synthesis of m-Prostaglandin E Synthase-1 Inhibitor
MK-7285
Francis Gosselin,*,† Stephen Lau,† Christian Nadeau,† Thao Trinh,† Paul D. O’Shea,† and
Ian W. Davies‡
†Department of Process Research, Merck Frosst Centre for Therapeutic Research,
16711 Route Transcanadienne, Kirkland, Queꢀbec, Canada H9H 3L1, and ‡Department of Process Research,
Merck Research Laboratories, Rahway, P.O. Box 2000, New Jersey 07065
francis_gosselin@merck.com
Received August 20, 2009
A practical, kilogram-scale chromatography-free synthesis of mPGE synthase I inhibitor MK-7285 is
described. The route features a convergent assembly of the core phenanthrene unit via amide-directed
ortho-metalation and proximity-induced anionic cyclization, followed by imidazole synthesis and late-
stage cyanation.
Introduction
mPGES-1 inhibitor discovered at Merck Frosst (Fig-
ure 1).4 As part of an ongoing drug development program
in our laboratories, we required multikilogram amounts of
MK-7285 (1). Herein we wish to report a practical synthesis
of MK-7285 suitable for preparation of multikilogram
amounts of the active pharmaceutical ingredient.
The modulation of prostaglandin metabolism is at the center
of current anti-inflammatory therapies.1 Nonsteroidal anti-in-
flammatory drugs (NSAIDs) and cyclooxygenase-2 (COX-2) in-
hibitors block the activity of cyclooxygenases and their ability to
convert arachidonic acid into prostaglandin H2 (PGH2). PGH2
can be subsequently metabolized by terminal prostaglandin
synthases to the corresponding biologically active prostaglan-
dins: prostacyclin (PGI2), thromboxane TxA2, PGD2, PGF2R,
and PGE2. The transformation of PGH2 to PGE2 by prosta-
glandin E synthases (PGES) may represent a key step in the
propagation of inflammatory stimuli. In particular, the micro-
somal enzyme prostaglandin E synthase-1 (mPGES-1) is indu-
cible after exposure to pro-inflammatory stimuli.2 Selective inhi-
bitors of mPGES-1 are currently considered as potential anti-
inflammatory therapeutics for the treatment of diseases such as
osteoarthritis, rheumatoid arthritis, and acute or chronic pain.3
MK-7285, a structurally complex substituted phenan-
throbis(cyano)phenylimidazole, is a potent and selective
FIGURE 1. Structure of mPGES-1 inhibitor MK-7285.
Our synthetic strategy was based on disconnection of the
phenanthroimidazole ring in 1 leading to phenanthrene dione 2
(Scheme 1). The substituted phenanthrene core would then be
^ ꢀ
(4) (a) Giroux, A.; Boulet, L.; Brideau, C.; Chau, A.; Claveau, D.; Cote, B.;
Ethier, D.; Frenette, R.; Gagnon, M.; Guay, J.; Guiral, S.; Mancini, J.; Martins, E.;
(1) Funk, C. D. Science 2001, 294, 1871.
(2) (a) Jakobsson, P. J.; Thoren, S.; Morgenstern, R.; Samuelsson, B.
Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 7220. (b) Murakami, M.; Naraba, H.;
Tanioka, T.; Semmyo, N.; Nakatani, Y.; Kojima, F.; Ikeda, T.; Fueki, M.;
Ueno, A.; Oh, S.; Kudo, I. J. Biol. Chem. 2000, 275, 32783. (c) Murakami,
M.; Kudo, I. Prog. Lipid Res. 2004, 43, 3.
ꢀ
ꢀ
Masse, F.; Methot, N.; Riendeau,D.;Rubin, J.;Xu, D.; Yu,H.; Ducharme, Y.;Friesen,
^ ꢀ
R. W. Bioorg. Med. Chem. Lett. 2009, in press. (b) Cote, B.;Boulet, L.;Brideau, C.;
Claveau, D.; Ethier, D.; Frenette, R.; Gagnon, M.; Giroux, A.; Guay, J.; Guiral, S.;
ꢀ
ꢀ
Mancini, J.; Martins, E.; Masse, F.; Methot, N.; Riendeau, D.; Rubin, J.; Xu, D.; Yu,
(3) Kojima, F.; Kato, S.; Kawai, S. Fund. Clin. Pharmacol. 2005, 19, 255–261.
H.; Ducharme, Y.; Friesen, R. W. Bioorg. Med. Chem. Lett. 2007, 17, 6816.
7790 J. Org. Chem. 2009, 74, 7790–7797
Published on Web 09/24/2009
DOI: 10.1021/jo901798d
r
2009 American Chemical Society

Gosselin et al.
JOCArticle
SCHEME 1. Retrosynthetic Analysis
SCHEME 2. Synthesis of Bromide 3
SCHEME 3. Synthesis of 2-Cyclopropylethanol 9
elaborated through a convergent assembly of bromide 3 and
amide 4 via directed ortho-metalation (DoM)/in situ Suzuki
cross-coupling,5 followed by a proximity-induced anionic cy-
clization (i.e., directed remote metalation, DreM).6
Results and Discussion
The synthesis of bromide 3 began with diazotization of
commercially available 3-bromo-4-methylaniline 5 with iso-
butyl nitrite in the presence of BF3 Et2O to afford arenedia-
Synthesis of Amide Intermediate. The 2-cyclopropyletha-
nol side chain 9 was prepared via cyclopropanation of
3-buten-1-ol 8 with Me3Al and CH2I2 in dichloromethane
(Scheme 3).10 After an exothermic and methane-generating
quench, careful concentration afforded the iodomethane-
free desired alcohol that was fractionally distilled at atmo-
spheric pressure to provide 2-cyclopropylethanol 9 in 64-
83% yield. We found that residual CH2I2 led to competing
alkylation of 2-cyclopropylethanol. Residual CH2I2 was
destroyed by treatment of the crude alcohol with 1,4-
diazabicyclo[2.2.2]octane (DABCO, 200 mol %) prior to
performing the SNAr step. Simple filtration of the resulting
DABCO ammonium salt afforded a solution of 2-cyclopro-
pylethanol 9 in toluene ready for the SNAr step.
3
zonium tetrafluoroborate salt 6 as a bench-stable solid in
>98-99% isolated yield (Scheme 2).7 We recently reported
conditions for metal-free Meerwein arylation of olefins to
provide R-aryl methyl ketones.8 We applied these conditions
toward preparation of ketone 7. The Meerwein arylation was
performed by slow addition of aqueous KOAc to a stirred
slurry of arenediazonium tetrafluoroborate salt 6 in isoprope-
nyl acetate/CH3CN/H2O at rt to afford the desired crude
ketone 7 in 80% HPLC assay yield.9 Slow addition of aqueous
KOAc was required to control the rate of N2 evolution from
the reaction. Formation of the tertiary alcohol initially proved
problematic due to competing ketone enolization and moder-
Treatment of 4-fluorobenzoyl chloride 10 with diethyl-
amine under Schotten-Bauman conditions in the presence of
aqueous K2CO3 as base cleanly afforded the corresponding
amide 11 that was 99.4A% by HPLC (Scheme 4). While the
reaction could be performed in other solvents such as
MTBE, we opted for toluene in order to avoid a solvent-
switch for the subsequent SNAr reaction. The SNAr reaction
was performed in toluene, and potassium tert-pentoxide was
used as base since it led to better conversions than t-BuOK
(33%) and KN(SiMe3)2 (95%).11 Thus, reaction of 11 with
120 mol % of 2-cyclopropylethanol 9 led to complete con-
version after 16 h at 90 °C and afforded amide 4 in 94% yield.
Synthesis of Phenanthrene Dione Intermediate. We envi-
saged two complementary approaches for the formation of
ate conversion to 3. Addition of MeLi LiBr gave tertiary
3
alcohol 3 in 72% assay yield at 85% conversion (15% ketone
remaining). Other conditions (Me3Al, MeMgCl, MeMgBr,
MeLi, etc.) led to decreased conversion to the desired tertiary
alcohol. In smaller scale experiments this reaction reproduci-
bly afforded >95% conversion to the desired alcohol 3. The
difference in conversion was attributed to slightly higher
internal temperature on a kilogram scale.
(5) (a) de Silva, S. O.; Reed, J. N.; Snieckus, V. Tetrahedron Lett. 1978, 19,
5099. (b) Cheng, W.; Snieckus, V. Tetrahedron Lett. 1987, 28, 5097.
(c) Reviewed in: Anctil, E. J.-G.; Snieckus, V. Metal-Catalyzed Cross-Coupling
Reactions, 2nd ed.; Wiley-VCH: New York, 2004; Vol. 2, pp 761-813.
(6) (a) Wang, X.; Snieckus, V. Tetrahedron Lett. 1991, 32, 4879. (b) Wang,
X.; Snieckus, V. Tetrahedron Lett. 1991, 32, 4883. (c) Cai, X.; Brown, S.;
Hodson, P.; Snieckus, V. Can. J. Chem. 2004, 82, 195. (d) A related approach
has been reported recently: Limanto, J.; Dorner, B. T.; Hartner, F. W.; Tan,
L. Org. Proc. Res. Dev. 2008, 12, 1269.
(7) Doyle, M. P.; Bryker, W. J. J. Org. Chem. 1979, 44, 1572.
(8) Molinaro, C.; Mowat, J.; Gosselin, F.; O’Shea, P. D.; Marcoux, J.-F.;
Angelaud, R.; Davies, I. W. J. Org. Chem. 2007, 72, 1856.
(9) Diazonium tetrafluoroborate salt 6 was deemed bench-stable as it
exhibited an exotherm of 35.8 cal/g initiating at 100 °C and a smaller exotherm
of 3.5 cal/g initiating at 175 °C as measured by differential scanning calorimetry.
(10) (a) Maruoka, K.; Fukutani, Y.; Yamamoto, H. J. Org. Chem. 1985,
50, 4412. (b) Russo, J. M.; Price, W. A. J. Org. Chem. 1993, 58, 3589. (c) For a
review, see: Charette, A. B.; Beauchemin, A. Org. React. 2001, 58, 1.
(11) (a) Kim, A.; Powers, J. D.; Toczko, J. F. J. Org. Chem. 2006, 71,
2170. (b) Rodriguez, J. R.; Agejas, J.; Bueno, A. B. Tetrahedron Lett. 2006, 47,
5661. (c) Woiwode, T. F.; Rose, C.; Wandless, T. J. J. Org. Chem. 1998, 63, 9594.
J. Org. Chem. Vol. 74, No. 20, 2009 7791

Gosselin et al.
JOCArticle
SCHEME 4. Synthesis of Amide 4
SCHEME 5. Arylation Approaches to Biaryl 12a/12b
SCHEME 6. Palladium-Catalyzed Direct Arylation Approach
SCHEME 7. In Situ Cross-Coupling
biaryl 12a/12b. In a first approach, we envisioned a direct
metal-catalyzed arylation between amide 11 and bromide 3
(Scheme 5).12 In a second approach, we expected that
appropriate functionalization of amides 4 or 11 could be
accomplished through a directed ortho-metalation (DoM)
and boronation sequence to afford boronic acids 14a/14b.
Subsequent cross-coupling with bromide 3 would lead to the
desired biaryl intermediate 12a/12b.
studies involving the stepwise DoM/boronation sequence on
amide 11 using s-BuLi or t-BuLi as the base led only to the
formation of benzophenone condensation product 16 (eq 1).
Direct arylation of 4-fluoroamide 11 with bromide 3
afforded promising conversion and selectivity of mono vs
diarylation in Pd-catalyzed reactions (Scheme 6).12c,d How-
ever careful H NMR spectroscopic analysis showed that
arylation occurred at the ortho position to the 4-fluoro
substituent to yield undesired biaryl 15.
Boronation could be achieved by performing the DoM in
presenceoftriisopropylborate, leading toa mixtureofboronic
acids 14a/17 with a maximum 5:1 ratio of ortho/meta regio-
isomers using LDA as the base in DME (eq 2). Unfortunately,
after workup the mixture of boronic acids 14a/17 was water-
soluble across the pH range and the isomers could not be
easily separated by nonchromatographic means.
1
In light of these results, we next turned our attention to a
DoM/cross-coupling route to the biaryl intermediate. Initial
(12) (a) Campeau, L.-C.; Stuart, D. R.; Leclerc, J.-P.; Bertrand-Laperle,
M.; Villemure, E.; Sun, H.-Y; Lasserre, S.; Guimond, N.; Lecavallier, M.;
Fagnou, K. J. Am. Chem. Soc. 2009, 131, 3291. (b) Campeau, L.-C.; Parisien,
M.; Jean, A.; Fagnou, K. J. Am. Chem. Soc. 2006, 128, 581. (c) Lafrance, M.;
Rowley, C. N.; Woo, T. K.; Fagnou, K. J. Am. Chem. Soc. 2006, 128, 8754.
(d) Lafrance, M.; Shore, D.; Fagnou, K. Org. Lett. 2006, 8, 5097.
Nevertheless, we found that the corresponding mixture of
lithium boronate esters 18a/18b underwent in situ Suzuki
7792 J. Org. Chem. Vol. 74, No. 20, 2009

Gosselin et al.
JOCArticle
SCHEME 8. Potential Alternative Sequence of Steps to Biaryl 12b
SCHEME 9. Improved DoM/Cross-Coupling to Biaryl 12b
SCHEME 10. Intramolecular Anionic Cyclization to 20
bromide 3 even at >5 mol % catalyst loading. Treatment
of crude 3 with Darco KB carbon (50 wt %) restored the
performance of the cross-coupling, allowing the reaction
to proceed to completion at catalyst loadings between
2.5 and 5 mol %. The DoM/boronation/in situ Suzuki
cross-coupling protocol routinely produced biaryl 12b in
the range of 88-98% yield.
Several issues arose during the optimization of the intra-
molecular anionic cyclization of biaryl 12b to phenanthren-
9-ol 20 (Scheme 10). The outcome of the cyclization reaction
was sensitive to substrate concentration. At concentrations
above 10 mL/g (0.24 M), byproducts arising from intermo-
lecular condensation of 12b were observed as ascertained by
LC-MS analysis of the crude reaction mixture (5-10A%
HPLC). We found that treatment of crude biaryl 12b with
Darco KB carbon prior to cyclization was required in order to
minimize both the charge of lithium diethylamide for cycliza-
tion of 12b and the observed emulsification in the aqueous
workup of phenanthren-9-ol 20. We also found that extractive
workup of crude 20 with aqueous LiOH followed by carbon
treatment of the wet organic stream with 50 wt % Darco KB
carbon were required to reject a number of unidentified
impurities and to provide material of sufficient purity for
the subsequent oxidation step. This protocol afforded
phenanthren-9-ol 20 in yields ranging from 62 to 75% yield.
Oxidation of phenantren-9-ol 20 to diketone 2 was per-
formed through a dibromination/hydrolysis sequence as de-
scribed in Scheme 11.15 Various brominating agents (HBr, Br2,
NBS) and reaction solvents were investigated for this trans-
formation but only low to moderate conversions (40-78%)
and/or decomposition were observed. Addition of N,N-diiso-
propylethylamine only led to minor improvement (83%
conversion). Better conversion was observed with tert-butyla-
mine as base, but competing formation of the tert-butyl
hemiaminal of diketone 2 afforded lower isolated yields
(61%) from chromatographed biaryl 12b.16 The reagent
cross-coupling with bromide 3 using 5 mol % PdCl2 dppf as
3
the catalyst with aqueous K3PO4 in DME to provide the
desired biaryl 12a along with the corresponding meta-regioi-
somer as a ∼5:1 mixture of regioisomers (Scheme 7).
The undesired boronate regioisomer was hypothesized to
arise from the influence of the electron-withdrawing fluorine
atom on the aromatic ring in 11 which renders the protons
ortho to the fluorine more acidic.13 We considered that we
could potentially override the ortho effect by simply chan-
ging the synthetic sequence of reactions and performing the
SNAr displacement first to form 4-(2-cyclopropylethoxy)-
benzamide 4 (Scheme 8). The ether side chain should be a
weaker ortho-director than the amide, thus resulting in
increased regioselectivity of metalation/boronation.14
When we subjected amide 4 to the DoM/boronation
conditions (LDA, B(OiPr)3, DME or THF) only the desired
lithium boronate regioisomer 19 was obtained (Scheme 9).
Once again, the high solubility of the corresponding boronic
acid (not shown) in water precluded its isolation. Gratify-
ingly, in situ Suzuki cross-coupling between bromide 3 and
lithium boronate 19 using PdCl2 dppf as catalyst afforded
3
the desired biaryl 12b in high yield. Optimization showed
that additional base was not required and that simple
addition of H2O (2 mL/g) was sufficient. Furthermore,
catalyst levels could be lowered from 5 to 1 mol % using
purified bromide 3 without affecting conversion or yield.
Unfortunately, the reaction stalled at 30-50% conver-
sion when the cross-coupling was performed using crude
(15) (a) Nagano, T. J. Org. Chem. 1957, 22, 817. (b) Cushman, M.; Patel,
H. H.; Scheuring, J.; Bacher, A. J. Org. Chem. 1992, 57, 5630.
(16) LCMS analysis of the reaction mixtures showed a peak with [M þ H]
438 which corresponds to the hemiaminal. 1H NMR spectroscopic analysis
of crude reaction mixture in acetone-d6 showed multiple signals in the
aromatic region as well as large singlet at 1.45 ppm which account for the
t-Bu group.
(13) (a) Streitweiser, A; Mares, F. J. Am. Chem. Soc. 1968, 90, 644.
(b) Streitweiser, A.; Scannon, P. J.; Niemeyer, H. M. J. Am. Chem. Soc. 1975,
97, 7936. (c) Bridges, A. J.; Lee, A.; Maduakor, E. C.; Schwartz, C. E.
Tetrahedron Lett. 1992, 33, 7495.
(14) (a) Winkle, M. R.; Ronald, R. C. J. Org. Chem. 1982, 47, 2101.
(b) Fraser, R. R.; Bresse, M.;Mansour, T. S. J. Am. Chem. Soc. 1983, 105, 7790.
J. Org. Chem. Vol. 74, No. 20, 2009 7793

Gosselin et al.
JOCArticle
SCHEME 11. Synthesis of Diketone 2
rejection of these impurities proved to be extremely proble-
matic at the API tosylate salt stage. The cyanation reaction
conditions were then investigated further in order to avoid
formation of 26b and 26c. Increasing the amount of CuCN to
400 mol % afforded higher conversion (>99%) but had no
significant effect on the impurity profile as ascertained by
HPLC. Slow addition of reagents was then investigated but
showed no major improvement in terms of impurities gen-
eration. Various cyanide sources [including KCN, NaCN,
Zn(CN)2, and K4Fe(CN)6]18 as well as Pd-catalyzed cyana-
tions19 gave low or no conversion toward the desired bis-
(nitrile)imidazole 26a. Solvent investigation showed that
reduction product 26c was not observed in DMAc at
120 °C. However the reaction suffered from lower conver-
sion (e95%) and significant hydrolysis of both nitriles to
afford bis(amide)imidazole 26f (5-10A%). We found that
addition of imidazole (100 mol %) greatly increased the
reaction rate in DMAc at 120 °C and led to higher conver-
sions presumably due to stabilization of CuCN. A screen of
nitrogen base additives showed that 1-methylimidazole (200
mol %) led to highest increases in both reaction rates and
conversions (>99%) and minimized the formation of by-
product at 90 °C (Scheme 14).20
These conditions also allowed reduction of the CuCN
charge to 250 mol % instead of the 400 mol % usually
required to reach g99% conversion. The reaction showed
1.3A% residual 26b after 18 h but then required an additional
6 h to reduce the level down to 0.3-0.8A%. Careful LC-MS
analysis showed that the residual 26b was in fact the chloro
analogue 26d that came from aldehyde 23.21 The level of 26d
was then conveniently controlled with extended reaction time
without adverse effect on reaction yield or purity profile.
Workup of the reaction required the use of 2-methyl-THF in
order to avoid problematic tar formation.22 The desired MK-
7285 free base was obtained as a THF solution (3.02 kg, 99%
SCHEME 12. Synthesis of Aldehyde 23
of choice for this reaction proved to be 1,3-dibromo-5,5-
dimethylhydantoin because it did not require the addition of
base and led to similar yields (60%) of diketone 2 without
purification of biaryl 12b and phenanthren-9-ol 20.17 The
reaction could be performed in a variety of solvents (DME,
DMF, iPAc, MeCN) but showed some sensitivity to oxygen
since degassing the solution of 20 prior to the reaction in-
creased the yield by 5%. Hydrolysis of dibromoketone 21 to
the diketone 2 required treatment with 1 M aqueous NaOH
and heating at 50 °C for 16 h to reach complete conversion.
Milder bases and lower temperatures led to lower conversion.
The yield of the oxidation of 20 to diketone 2 increased to 82%
after implementing the extractive workup with LiOH and
carbon treatment of 20. The crude product (87.8A% HPLC)
was recrystallized in iPAc/heptane to afford phenanthrene
dione 2 as a brown solid in 84% yield.
Synthesis of Aldehyde Intermediate. 2,6-Dibromo-4-fluoro-
benzaldehyde 23 was obtained by halogen-metal exchange
of 1,3-dibromo-5-fluoro-2-iodobenzene 22 with isopropyl-
magnesium chloride, followed by formylation of the resulting
arylmagnesium chloride with DMF. In THF as solvent,
competing reduction of 22 by i-PrMgCl afforded 23/24 ratios
of 85:15. In comparison, switching to toluene as solvent
improved the ratio of 23/24 to 95:5 (Scheme 12) and afforded
aldehyde 23 in 72% yield after recrystallization from 2-propa-
nol/water.
End-Game Chemistry. Condensation of phenanthrene di-
ketone 2 with 2,6-dibromo-4-fluorobenzaldehyde 23 in the
presence of ammonium acetate in AcOH for 1 h at 90 °C
afforded dibromoimidazole 25 in 85% yield (Scheme 13).
When crude aldehyde 23 was used, we observed late-running
impurities (unidentified) in the HPLC chromatogram of
crude imidazole 25. Fortunately, we found that simple
recrystallization of aldehyde 23 from 2-propanol/water
eliminated these impurities.
Preliminary attempts at the cyanation of 25 using
250 mol % of CuCN in DMF at 90 °C for 16 h afforded
>95% conversion to MK-7285 free base 26a along with
significant monocyanation and reduction products 26b/26c
(4-6A% and 2-6A%, respectively). Not surprisingly,
(18) (a) For a review, see: Ellis, G. P.; Romney-Alexander, T. Chem. Rev.
1987, 87, 779–794. (b) Merz, V.; Weith, W. Ber. 1877, 10, 746.
(19) For a review, see: (a) Sundermeier, M.; Zapf, A.; Beller, M. Eur. J.
Inorg. Chem. 2003, 3513. For several examples, see: (b) Anderson, B. A.;
Bell, E. C.; Ginah, F. O.; Harn, N. K.; Pagh, L. M.; Wepsiec, J. P. J. Org.
Chem. 1998, 63, 8224. (c) Tschaen, D. M.; Desmond, R.; King, A. O.; Fortin,
M. C.; Pipik, B.; King, S.; Verhoeven, T. R. Synth. Commun. 1994, 24, 887.
(d) Weissman, S. A.; Zewge, D.; Chen, C. J. Org. Chem. 2005, 70, 1508.
€
€
(20) (a) Schareina, T.; Zapf, A.; Magerlein, W.; Muller, N.; Beller, M.
€
€
Synlett 2007, 555. (b) Schareina, T.; Zapf, A.; Magerlein, W.; Muller, N.;
Beller, M. Chem.;Eur. J. 2007, 13, 6249.
(21) We presume that the formation of the chloroaldehyde analogue
occurs during the halogen-metal exchange reaction via benzyne formation.
Careful GC-MS analysis of starting material 22 did not show any chloro-22.
(a) Villieras, J. Bull. Soc. Chim. Fr. 1967, 5, 1520. (b) Villieras, J.; Kirschleger,
B.; Tarhouni, R.; Rambaud, M. Bull. Soc. Chim. Fr. 1986, 470. (c) Reviewed
in: Knochel, P.; Dohle, W.; Gommermann, N.; Kneisel, F. F.; Kopp, F.;
Korn, T.; Sapountzis, I.; Vu, V. A. Angew. Chem., Int. Ed. 2003, 42, 4302-
4320. (d) Wang, X.-J.; Sun, X.; Zhang, L.; Xu, Y.; Krishnamurthy, D.;
Senanayake, C. H. Org. Lett. 2006, 8, 305–307.
(17) (a) Alam, A. Synlett 2005, 2403. (b) Chassaing, C.; Haudrechy, A.;
Langlois, Y. Tetrahedron Lett. 1997, 38, 4415.
(22) Aycock, D. F. Org. Proc. Res. Dev. 2007, 11, 156.
7794 J. Org. Chem. Vol. 74, No. 20, 2009

Gosselin et al.
JOCArticle
SCHEME 13. Synthesis of Imidazole 25
SCHEME 14. Cyanation of Imidazole 25
SCHEME 15. Formation of MK-7285 Tosylate
yield, HPLC: 95.6A%, 0.3A% 26d, 0.8A% 26f, ion chroma-
tography: <1.3 ppm residual cyanide).
MK-7285 tosylate was identified as a stable crystalline and
bioavailable form. Tosylate salt formation was performed by
heating a solution of MK-7285 crude free base 26a in THF at
sually clean 50 L round-bottom flask equipped with an N2 inlet,
thermocouple, addition funnel, and mechanical stirrer was
charged with diisopropylamine (3.64 L, 25.6 mol) and THF
(13 L) and cooled to an internal temperature of -5 °C.
n-Butyllithium (10.23 L, 25.6 mol, 2.5 M in hexane) was added
over 1-2 h, keeping the internal temperature below 0 °C. The
mixture was then stirred for 30 min at -5 °C. A visually clean
100 L round-bottom flask equipped with an N2 inlet, thermo-
couple, addition funnel, and mechanical stirrer was charged
with N,N-diethylbenzamide 4 (3.75 kg, 89.07 wt %, 12.78 mol),
THF (25 L), and triisopropyl borate (5.94 L, 25.6 mol). The
mixture was cooled to an internal temperature of -25 °C. The
LDA solution formed in the 50 L reactor was added to the amide
solution over 2 h, keeping the internal temperature below
-20 °C. The batch was stirred for 2 h for complete consumption
of amide. Water (8 L) was then added rapidly, and the reaction
mixture was warmed to 20 °C. The mixture was degassed with
N2 for 10 min. A solution of aryl bromide 3 (7.4 kg, 30 wt % in
toluene, 9.13 mol) was added, and the mixture was continued to
60 °C and adding a solution of p-TsOH H2O in THF over 30
3
min (Scheme 15). Residual THF levels in the API proved
problematic (1.9 wt %). Heating the API (1) under high
vacuum/dry N2 sweep at temperatures up to 60 °C failed to
reduce substantially the levels of residual THF (1.6 wt %). A
reslurry of MK-7285 tosylate in heptane eliminated com-
pletely THF and left 0.07 wt % heptane. However, we found
that simply drying the batch on the frit under low vacuum
under atmospheric moisture reduced THF levels to 0.08 wt %
in only 2-3 h at rt. The desired MK-7285 tosylate salt (1) was
obtained as a bright yellow solid in 80% yield, HPLC: 98.6A%
1, 0.34A%, 1d, 0.53A% “mono-amide” 1e, 2.2 wt % water,
ICPMS analysis: residual Fe, Pd, Cu <10 ppm.
be degassed for an additional 10 min. Solid PdCl2 dppf (250 g,
3
Conclusion. In conclusion, we have reported a robust
and practical chromatography-free synthesis of mPGES-1
inhibitor MK-7285 tosylate that is suitable for preparation
of multikilogram amounts of API.
0.342 mol) was added while the solution continued to be
degassed for an additional 15 min. A reflux condenser was
fitted, and the mixture was very slowly brought to an internal
temperature of 65 °C over 1-2 h in order to control the rate of
reflux, refluxing for a total of 12 h under N2. The reaction
mixture was then cooled to 30 °C and transferred to a visually
clean 160 L extractor, and the reaction vessel was rinsed with 2 L
of toluene. The remaining residue in the flask was rinsed with
water (24 L) and added to the extractor. The layers were cut, and
Experimental Section
5-(2-Cyclopropylethoxy)-50-(2-hydroxy-2-methylpropyl)-20-
methylbiphenyl-2-carboxylic Acid Diethyl Amide (12b). A vi-
J. Org. Chem. Vol. 74, No. 20, 2009 7795

Gosselin et al.
JOCArticle
the organic layer was washed for with 1 N NaOH (17 L). MTBE
(17 L) was then added to the organic layer and washed with 1 N
aqueous HCl (17 L), followed by H2O (2 ꢀ 17 L). The resulting
organic layer was then combined with a second batch of the
same size. The combined organic layers were then treated with
Darco KB (4 kg, 50 wt % based on combined total mass of
product) for 1 h at rt. The batch was filtered on Solka-floc, and
the filter cake was washed with MTBE (2 ꢀ 17 L) and toluene
(2 ꢀ 17 L). The filtrate was line-filtered, concentrated, and flushed
with toluene (2 ꢀ 17 L) and THF (2 ꢀ 17 L) to afford 7.69 kg of
biaryl amide 12b (7.69 kg, 40 wt % in THF, 98.7% assay yield):
1H NMR (400 MHz, acetone-d6) δ 7.23 (d, 1H, J = 8.4 Hz),
7.16-7.04 (m, 3H), 6.97 (dd, 1H, J = 8.4, 2.6 Hz), 6.78 (s, 1H),
4.11 (t, 2H, J = 6.6 Hz), 3.24 (s, 1H), 3.02 (bs, 4H), 2.68 (s, 2H),
2.17 (s, 3H), 1.68 (q, 2H, J = 6.6 Hz), 1.13 (s, 6H), 0.92-0.86 (m,
4H), 0.71 (t, 3H, J = 7.1 Hz), 0.49-0.39 (m, 2 H), 0.16-0.10 (m,
2H); 13C NMR (101 MHz, acetone-d6) δ 169.3, 158.7, 140.3,
139.0, 135.5, 131.7, 130.2, 129.7, 129.2, 127.6, 115.7, 113.0,
111.5, 69.9, 67.9, 49.3, 42.3, 37.6, 34.2, 19.3, 13.3, 11.7, 7.6,
3.8; HRMS calcd for C27H37NO3 [M þ H] 424.2846, found
424.2861. HPLC analysis: column: Zorbax XDB C18, 4.6 ꢀ 30
mm, 1.8 μm; eluent: 0.1% aqueous H3PO4/0.1% H3PO4 in
CH3CN (85:15 to 10:90 over 7 min, hold 1 min); flow: 1.5 mL/
min; det: 220 nm; temp: 20 °C; inj: 5 μL, post-time: 2.00 min.
Diethyl benzamide 4: 4.66 min; aryl bromide 3: 4.30 min; biaryl
amide 12b: 5.63 min.
for 1 h at rt. Solutions from both extractors were filtered over a
single bed of Solka-floc, washing with 2 ꢀ 20 L of toluene. The
filtrate was concentrated and flushed with toluene (2 ꢀ 20 L) to
afford 4.03 kg of phenanthren-9-ol 20 (4.03 kg, 63.3% assay
yield, 30 wt % in toluene). An analytical sample was isolated by
chromatography on silica gel using hexanes/EtOAc (10:1) as
eluent. Phenanthren-9-ol was isolated as an unstable 2:1 mixture
of enol/keto form. Enol form: 1H NMR (400 MHz, acetone-d6)
δ 9.08 (bs, 1H), 8.51 (s, 1H), 8.36-8.28 (m, 1H), 8.19 (d, 1H, J =
2.4), 7.57 (d, 1H, J = 8.1), 7.43 (dd, 1H, J = 8.1, 1.52), 7.28 (dd,
1H, J = 9.0, 2.4), 7.00 (s, 1H), 4.25 (t, 2H, J = 6.5), 2.99 (s, 2H),
1.77-1.69 (m, 2H), 1.26 (s, 6H), 0.98-0.90 (m, 1H), 0.56-0.42
(m, 3H), 0.21-0.12 (m, 3H); 13C NMR (101 MHz, acetone-d6) δ
4.7, 8.5, 29.7, 35.1, 50.7, 68.7, 71.0, 103.7, 105.8, 107.5, 116.7,
121.6, 125.0, 125.1, 126.6, 130.7, 133.2, 134.0, 134.6, 151.5,
1
159.2. Keto form: H NMR (400 MHz, acetone-d6) δ 8.65 (s,
1H), 8.42 (d, 1H, J = 9.0), 8.09 (s, 1H), 7.35 (dd, 1H, J = 9.0,
2.3), 7.25 (d, 1H, J = 8.3), 7.12 (d, 1H, J = 8.3), 4.32 (t, 2H, J =
6.5), 3.50 (s, 2H), 2.97 (s, 2H), 1.81-1.75 (m, 2H), 1.25 (s, 3H),
0.98-0.90 (m, 1H), 0.56-0.42 (m, 2H), 0.21-0.12 (m, 2H); 13
C
NMR (101 MHz, acetone-d6) δ 4.7, 8.5, 29.7, 35.1, 50.5, 60.5,
68.9, 71.0, 105.9, 116.8, 121.7, 125.2, 125.3, 126.2, 126.8, 130.9,
133.1, 134.3, 134.8, 159.6, 170.9; HRMS calcd for C23H27O3
[M þ H] 351.1955, found 351.1952. HPLC analysis: Column:
Zorbax XDB C18, 4.6 ꢀ 30 mm, 1.8 μm; eluent: 0.1% aqueous
H3PO4/0.1% H3PO4 in CH3CN (85:15 to 10:90 over 7 min, hold
1 min); flow: 1.5 mL/min; det: 220 nm; temp: 20 °C; inj: 5 μL,
post-time: 2.00 min. Biaryl amide 12b: 5.63 min; phenanthren-9-
ol 20: 5.14 min.
Note on Biaryl 12b NMR Spectroscopic Analysis. NEt2 signals
are unusual. In CDCl3, the ethyl groups are not equivalent and
difficult to assign. In acetone-d6, one CH3 (0.71) is well resolved
while the other CH3 (0.92-0.86) is not, and the methylenes
(3.02) are similarly unresolved. This complicates the already
complex 13C NMR analysis, where some tertiary and quatern-
ary carbons in the aromatic system either do not show up or are
low, broad signals.
3-(2-Cyclopropylethoxy)-6-(2-hydroxy-2-methylpropyl)phe-
nanthrene-9,10-dione (2). Crude phenanthren-9-ol 20 (27.73 kg
at 13.5 wt %; 3.74 kg, 10.67 mol) that already contained 27.8 L
of toluene was diluted with toluene (9.6 L). The solution was
degassed for 30 min with subsurface nitrogen sparge, cooled to
-5 °C, and protected from light. Solid 1,3-dibromo-5,5-di-
methylhydantoin (4.58 kg, 16.01 mol) was added portionwise
over 1 h, and the internal reaction temperature was maintained
under 9 °C. The reaction mixture was warmed to rt over 30 min
and kept at this temperature for 2.5 h. Aqueous NaOH (56 L,
1 M) was added, and the mixture was stirred vigorously at 50 °C
for 16 h. The reaction mixture was cooled to rt, and the layers
were cut. The aqueous layer was back-extracted with toluene
(18.7 L), and the combined organic layers were washed with 1 M
aqueous NaOH (2 ꢀ 37.4 L) and 1% aqueous NaCl (2 ꢀ 37.4 L).
HPLC assay of the toluene solution indicated 53.41 kg, 5.95 wt %
in toluene for a total mass 3.18 kg of diketone 2 (81.7% yield). The
organic layer was concentrated and flushed with toluene
(20 L) and then with iPAc (2 ꢀ 40 L). The slurry was warmed
to 80 °C until a homogeneous solution was obtained. 1H NMR
analysis of the solution showed a 1:25.1:0.15 mol ratio of diketone/
iPAc/toluene (87 wt % iPAc; 0.5 wt % toluene). The diketone
solution already contained 22.3 kg (25.6 L) of iPAc. Fresh iPAc
(4.6 L) was added, and the solution was heated to
80 °C. The solution was cooled to rt over 2 h, and heptane
(90.6 L) was added over 1 h. The resulting slurry was stirred at
rt for 16 h and then filtered. The filter cake was washed with iPAc/
heptane (1:3, 2 ꢀ 31.8 L, 1 ꢀ 10 L). The batch was dried under low
vacuum on the frit for 64 h. Diketone 2 was obtained as a brown
solid (3.5 kg, 76.7 wt % for a total mass 2.68 kg, 84% yield, 69%
6-(2-Cyclopropylethoxy)-3,3-(2-hydroxy-2-methylpropyl)phe-
nanthren-9-ol (20). A visually clean 50 L round-bottom flask
equipped with an N2 inlet, thermocouple, 5 L addition funnel,
and mechanical stirrer was charged with diethylamine (6.64 L,
63.5 mol) and THF (24 L) and cooled to an internal temperature
of -5 °C. n-Butyllithium (25.4 L, 63.5 mol, 2.5 M in hexane) was
added over 1-2 h keeping internal temperature below 0 °C. The
mixture was stirred for 30 min at -5 °C. A visually clean 160 L
extractor equipped with an N2 inlet, thermocouple, addition
funnel, and glycol cooling lines was charged with biaryl amide
12b (7.69 kg, 40 wt % in THF, 18.15 mol, contained 13 L of
THF) and diluted with THF (67 L) to give a total THF volume
of 80 L. The lithium diethylamide solution formed in the 50 L
reactor was added to the biaryl amide 12b solution over 2 h,
keeping the internal temperature below 0 °C. Half of the
reaction mixture was then transferred into another visually
clean 160 L extractor equipped with an N2 inlet, thermocouple,
and glycol cooling lines in order to ease stirring due to the large
volume. The batch was very thick and became somewhat
difficult to stir for 30-60 min. Both extractors were kept at
-5 °C and stirred for 14 h for complete conversion of the biaryl
amide. For each extractor: The reaction was quenched with 40 L
of H2O and diluted with 40 L of hexanes with mixing and
warming to 20 °C. The layers were cut and the aqueous layers
kept. The organic layer was then extracted with 1 N aqueous
LiOH (5 ꢀ 20 L) for 15-20 min each. The combined aqueous
layer was washed with hexanes (20 L) and then divided into two
batches due to the large volume. For each batch: Toluene (10 L)
was added to the aqueous layer with stirring, and the aqueous
layer was acidified to pH 1-2 with 12 N HCl (∼8 L). The layers
were cut, and the aqueous layer was extracted with toluene (2 ꢀ
10 L). The combined toluene layers from both batches were
washed with water (2 ꢀ 20 L). Darco KB (1 kg, 50 wt % based on
mass of product) was added to the toluene solution and stirred
1
yield over oxidation/recrystallization): mp = 127 °C; H NMR
(500 MHz, CDCl3) δ 8.14 (1 H, d, J = 8.7 Hz), 8.09 (1 H, d, J =
7.9 Hz), 7.80 (1 H, s), 7.43 (1 H, d, J = 2.3 Hz), 7.32 (1 H, d, J =
8.0 Hz), 6.91 (1 H, dd, J = 8.7, 2.3 Hz), 4.19 (2 H, t,
J = 6.6 Hz), 2.90 (2 H, s), 1.75 (2 H, q, J = 6.7 Hz), 1.32 (6 H,
s), 0.94-0.85 (1 H, m), 0.56-0.51 (2 H, m), 0.19-0.14 (2 H, m);
13C NMR (125 MHz, CDCl3) δ 180.4, 178.6, 165.5, 146.9, 138.0,
135.1, 133.6, 131.9, 130.0, 129.8, 125.9, 124.7, 114.5, 110.5, 70.9,
68.8, 50.0, 34.2, 29.6, 7.7, 4.4; IR (cm-1, NaCl thin film): 3583,
7796 J. Org. Chem. Vol. 74, No. 20, 2009

Gosselin et al.
JOCArticle
3475, 3077, 3001, 2970, 2875, 1672, 1659, 1592, 1311, 1240, 1223;
HRMS calcd for C23H25O4 [M þ H] 365.1753; found 365.1750.
Conversion and wt % were determined by HPLC with a 4.6 mm ꢀ
25 cm ACE C18 column (0.1% aq H3PO4/CH3CN 70:30 to 40:60
over 10 min, to 35:65 over 5 min, to 5:95 over 10 min, hold 5 min,
1.5 mL/min, 220 nm, 35 °C); diketone 2, tR = 14.17 min.
MK-7285 tosylate in iPAc and saturated aqueous NaHCO3:mp=
178-180 °C; 1H NMR (500 MHz, DMSO-d6) δ 14.21-13.84 (1
H, br), 8.65 (1 H, s), 8.56 (1 H, d, J = 8.2 Hz), 8.56 (1H, d, J = 8.2
Hz), 8.44-8.27 (2 H, br), 8.25 (1 H, s), 7.63 (1 H, d, J = 8.2 Hz),
7.42 (1 H, d, J = 8.7 Hz), 4.29 (2 H, t, J = 6.5 Hz), 2.97 (2 H, s),
1.74 (2 H, q, J = 6.6 Hz), 1.15 (6 H, s), 0.97-0.90 (1 H, m),
0.52-0.46 (2 H, m), 0.21-0.17 (2 H, m); 13C NMR (125 MHz,
DMSO-d6) δ 161.1 (d, J = 255.1 Hz), 157.1, 140.8, 136.6 (br),
134.3, 134.3, 130.3, 129.7 (br), 127.2, 126.0, 125.8, 125.7, 123.5,
121.0, 116.6, 115.6, 115.5, 115.4, 115.3, 107.2, 69.6, 68.0, 49.5,
33.8, 29.4, 7.8, 4.3; 19F NMR (375 MHz, CDCl3) δ - 108.9; IR
(cm-1, NaCl thin film) 3498, 3431, 3075, 2964, 2865, 2222, 1629,
1592, 1506, 1453, 1300, 1216; HRMS calcd for C32H27FN4O2 [M]
518.2196, found 518.2193. Conversion and yield were determined
by HPLC with a 4.6 mm ꢀ 15 cm Waters Symmetry C18 column
(pH 6.8 phosphate buffer/CH3CN 47:53 hold 18 min, to 35:65
over 2 min, to 10:90 over 0.1 min, hold 2 min, to 47:53 over 0.1
min, hold 4 min, 1 mL/min, 220 nm, 35 °C); MK-7285 tR = 10.21
min.
1-[9-(2-Cyclopropylethoxy)-2-(2,6-dibromo-4-fluorophenyl)-
1H-phenanthro[9,10-d]imidazol-9-yl]-2-methylpropan-2-ol (25).
A visually clean 160 L extractor equipped with a dropping
funnel and a N2 inlet was charged with glacial acetic acid (13.2
L). Solid diketone 2 (2.63 kg, 7.22 mol), solid aldehyde 23 (2.02
kg, 7.17 mol), and NH4OAc (5.56 kg, 72.2 mol, 1000 mol %)
were added. The resulting suspension was heated to an internal
temperature 93.6-96.0 °C for 1 h. HPLC showed 98.0% con-
version to imidazole 25. The reaction mixture was cooled to
40 °C, diluted with iPAc (26.4 L), and washed with H2O (26.4 L);
the aqueous layer was pH 4. The extractor was charged with
H2O (26.4 L) and 1 N aqueous NaOH (20.0 L), and the layers
were cut. The organic layer was washed with 1 N aqueous
NaOH (20 L); the aqueous layer was pH 9.5. The layers were
cut, and the organic layer was treated with Darco G-60 (5.26 kg,
200 wt %). The suspension was stirred at 23.4 °C for 1 h, and the
solids were filtered through Solka-floc and rinsed with iPAc (2 ꢀ
13.2 L). The solution was concentrated, and the solids were
dried under vacuum at 30 °C overnight. Imidazole 25 was
obtained as light brown solid: 3.85 kg, 85.3% assay yield, 94.0
A% HPLC; an analytical sample was prepared by chromato-
graphy on silica gel using EtOAc/hexane (1:1) as eluent: mp =
2-[9-(2-Cyclopropylethoxy)-6-(2-hydroxy-2-methylpropyl)-1H-
phenanthro[9,10-d]imidazo-2-yl]-5-fluoroisophthalonitrile 4-Me-
thylbenzenesulfonate (MK-7285 Tosylate, 1). A visually clean
50 L round-bottom flask equipped with a mechanical stirrer, a
dropping funnel, a reflux condenser, and a N2 inlet was charged
with line-filtered MK-7285 crude free base 26a (3.0 kg, 5.785 mol,
95.6A% HPLC, 21.5 wt % in THF, KF 364 ppm, 9.7 wt %
residual iPAc, 0.3A% HPLC 26d, 0.8A% HPLC tR = 4.8 min
26f). The flask was rinsed with line-filtered THF (2 ꢀ 625 mL).
The dark brown solution was heated to 60 °C. A solution of line-
1
225.5-226.0 °C; H NMR (400 MHz, CDCl3) δ 11.85 (br s, 1
H), 8.64 (br s, 1H), 8.46 (s, 1 H), 8.16 (s, 1 H), 8.04 (br s, 1 H),
7.42-7.29 (br m, 3 H), 7.23-7.02 (br m, 1 H), 4.23 (t, 2H, J =
6.67 Hz), 3.00 (s, 1 H), 1.78 (q, 2H, J = 6.78 Hz), 1.29 (s, 6 H),
1.01-0.88 (m, 1 H), 0.58-0.52 (m, 2 H), 0.19 (q, 2H, J = 5.0
Hz); 13C NMR (126 MHz, CDCl3) δ 162.1 (d, J = 257.6 Hz),
157.1, 145.9, 134.4, 130.4 (d, J = 14.3 Hz), 129.9 (br), 129.4 (br),
128.0, 125.8 (d, J = 10.3 Hz), 125.0, 124.1, 122.5, 119.6 (d, J =
filtered p-TsOH H2O (1.10 kg, 5.785 mol) in THF (7 L) was
3
added over 30 min, and the batch was aged at 60 °C for 30 min.
Crystallization began at 10-15 min to afford a light slurry. The
batch was allowed to cool to rt and stirred at rt for 2-3 h. The
slurry was filtered, and the filter cake was washed with line-
filtered THF (2 ꢀ 10 L). The solids were dried on the filter pot
overnightatrtand thenunder vacuum/N2 sweepat 30 °C over the
weekend and then dried on the frit for 5 h. MK-7285 tosylate 1
was obtained as a bright yellow solid: 3.2 kg, 80% yield, 98.6A%
HPLC, 0.34A% HPLC 1d, 0.53A% HPLC 1e; HPLC analysis 1e
14.8 Hz), 116.1, 107.2, 71.0, 68.4, 50.0, 34.5, 29.1, 7.8, 4.4; 19
NMR (377 MHz, CDCl3) δ -107.0 (reference -62.7); IR (cm-1
F
,
CHCl3) 2969, 1592, 1561, 1451, 1423, 1230, 908, 732; HRMS
calcd for C30H28Br2FN2O2 [M þ H] 625.0502, found 625.0504.
Conversion and yield were determined by HPLC with a 4.6 ꢀ
250 mm Zorbax RX-C8 column (0.1% aqueous H3PO4/CH3CN
70:30 to 5:95 over 25 min, hold 5 min, 2 mL/min, 220 nm, 35 °C);
t
R = 5.26; MK-7285 tR = 12.27; 1d tR = 16.02; mp 174-176 °C
dec; 1H NMR (DMSO-d6) δ 8.70 (s, 1H), 8.59 (d, 2H, J = 8.5),
8.39 (d, 1H, J = 26.0), 8.38 (d, 1H, J = 25.0), 8.29 (d, 1H, J =
2.0), 7.68(d, 1H, J = 8.0), 7.50 (d, 2H, J = 8.0), 7.47 (dd, 1H, J =
2.0, 9.0), 7.14 (d, 2H, J = 8.0), 4.31 (dd, 2H, J = 6.0, 6.5), 3.00 (s,
2H), 2.30 (s, 3H), 1.75 (dd, 2H, J = 6.5, 13.0), 1.17 (s, 6H), 0.96
(bm, 1H), 0.49 (m, 2H), 0.20 (dd, 2H, J = 5.0, 9.5); 13C NMR
(DMSO-d6) δ 162.7, 160.7, 157.7, 144.8, 140.0, 138.3, 137.5,
132.3, 131.1, 130.7, 130.3, 130.2, 128.3, 127.5, 126.0 (J = 25.0),
125.9, 125.5, 123.7, 121.8, 121.2, 117.3, 116.9, 115.9 (J = 10.6),
115.3, 107.4, 69.6, 68.1, 49.5, 33.9, 29.4, 20.8, 7.9, 4.3; 19F NMR
(DMSO-d6) δ -106.9; HRMS calcd for C32H28FN4O2 [M þ H]
519.2191, found 519.2196.
t
R = 12.2 min.
2-[9-(2-Cyclopropylethoxy)-6-(2-hydroxy-2-methylpropyl)-1H-
phenanthro[9,10-d]imidazo-2-yl]-5-fluoroisophthalonitrile (26a).
Crude dibromoimidazole 25 (3.673 kg, 5.86 mol) was flushed
with DMAc (5.5 L) to remove any residual iPAc and was diluted
in DMAc (29.2 L). 1-Methylimidazole (934.8 mL, 11.73 mol, 200
mol %) and CuCN (1.313 kg, 14.66 mol, 250 mol %) were added.
DMAc (2 L) was added, and the reaction mixture was degassed
using a subsurface nitrogen sparge for 20 min and then heated to
90 °C for 24 h. The batch was cooled to rt and transferred into
an cylindrical extractor containing 2-methyltetrahydrofuran
(36.7 L). Aqueous NH4Cl/NH4OH buffer [saturated aqueous
NH4Cl/30% aqueous NH4OH/water (4:1:3), 36.7 L] was then
added, the mixture was stirred vigorously for 10 min, the layers
were cut, and the aqueous layer was back-extracted with iPAc
(36.7 L). The combined organic layers were washed with NH4Cl/
NH4OH buffer (3 ꢀ 36.7 L) with 10 min of vigorous stirring for
each wash. The organic layer was then washed with water (36.7
L), concentrated, and flushed with iPAc (22 L) and then with
THF (2 ꢀ 36.7 L). HPLC assay of the concentrated solution
indicated 14.05 kg, 21.5 wt % in THF for a total mass 3.02 kg of
MK-7285 free base 26a (99% yield). 1H NMR analysis showed a
1:20.75:0.545 mol ratio of MK-7285/THF/iPAc (9.7 wt % iPAc
vs MK-7285). The MK-7285 solution already contained 9.815 L
of THF. An analytical sample was obtained by salt-break of
Acknowledgment. We thank Dr. Wayne Mullett, Mr.
Claude Briand, and Mr. Ravi Sharma for analytical research
support, Dr. Thomas Novak for HRMS data, and Robert
Reamer and Dr. Dan Sorensen for help with NMR analyses
of compounds 12b, 15, and 20. We also thank Professors
Barry M. Trost (Stanford University) and Paul Knochel
€
(Ludwig Maximilian Universitat Munchen) for helpful dis-
€
cussions.
Supporting Information Available: Experimental proce-
dures for compounds 3, 4, 6, 7, 11, and 23. Copies of 1H, 13C,
and 19F NMR spectra of compounds 1-4, 6, 7, 11, 12b, 20, 23,
25, and 26a. This material is available free of charge via the
Internet at http://pubs.acs.org.
J. Org. Chem. Vol. 74, No. 20, 2009 7797