Selective removal of a benzyl protecting group in the presence of an aryl chloride under gaseous and transfer hydrogenolysis conditions

Article abstract of DOI:10.1016/S0040-4039(03)00846-3

Selective removal of a benzyl protecting group in the presence of an aryl chloride using Pd/C under gaseous and transfer hydrogenolysis conditions is described. The addition of chloride salts to the debenzylation reaction provides excellent selectivity.

Full text of DOI:10.1016/S0040-4039(03)00846-3

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 44 (2003) 4041–4043
Selective removal of a benzyl protecting group in the presence
of an aryl chloride under gaseous and transfer
hydrogenolysis conditions
Jun Li,* Steve Wang, Gerard A. Crispino, Karen Tenhuisen, Ambarish Singh and John A. Grosso
Process Research & Development, Pharmaceutical Research Institute, Bristol-Myers Squibb Co., PO Box 191, New Brunswick,
NJ 08903-0191, USA
Received 20 March 2003; accepted 1 April 2003
Abstract—Selective removal of a benzyl protecting group in the presence of an aryl chloride using Pd/C under gaseous and
transfer hydrogenolysis conditions is described. The addition of chloride salts to the debenzylation reaction provides excellent
selectivity. © 2003 Elsevier Science Ltd. All rights reserved.
4-Aryl-quinolin-2(1H)-ones are modulators of the
large-conductance, calcium-activated potassium (Maxi-
K and BK) channels and are potentially useful in the
treatment of diseases which arise from dysfunction of
cellular membrane polarization and conductance.1 Dur-
ing process development of 4-(5-chloro-2-hydroxy-
phenyl)-3-(2-hydroxyethyl)-6-(trifluoromethyl)-2(1H)-
quinolinone (1), debenzylation of the aryl chloride 2 via
hydrogenolysis with palladium on carbon (Pd/C)
proved troublesome (Scheme 1). Competitive dechlori-
nation afforded the deschloro analog 3, which was
difficult to purge by selective crystallization of 1. Thus,
we needed to develop conditions for a more chemo-
selective hydrogenolysis which minimized formation of
the undesired byproduct.
vided the desired selectivity, but complete removal of
the benzyl group in 2 was not achieved. We thus
returned to palladium catalysts and focused our atten-
tion on the effect of hydrogenolysis conditions on the
course of the reaction. Our initial finding was that the
ratio of 1:3 correlated directly with the dielectric con-
stant of the solvent employed, in line with the observa-
tions of others.2 Thus, the deschloro analog 3 was
observed to form in up to 35% yield in solvents such as
methanol with high dielectric constant (32.7 at 25°C),
while solvents with low dielectric constant such as ethyl
acetate (6.02 at 25°C) or methylene chloride (9.08 at
25°C) provided product with deschloro levels of only
1–2%. Although this initial screening demonstrated a
high selectivity for debenzylation using ethyl acetate,
the product 1 was poorly soluble in this solvent and
tended to precipitate out on the catalyst, dramatically
inhibiting the reaction rate.
In early screening, we found that the use of platinum
oxide (Adams’ catalyst) or platinum on carbon pro-
Scheme 1.
* Corresponding author. Fax +1-732-519-3240; e-mail: jun.li1@bms.com
0040-4039/03/$ - see front matter © 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0040-4039(03)00846-3

4042
J. Li et al. / Tetrahedron Letters 44 (2003) 4041–4043
Selective hydrogenolysis of benzyl groups in the pres-
ence of an aryl chloride in acidic media have been
reported.2–4 When the reaction was carried out over
Pd/C in 5 mol% hydrochloric acid/ethanol, the ratio of
1:3 increased to 40/1. Larger amounts of acid further
increased the selectivity, unfortunately, however, at the
cost of a dramatically slower reaction rate.2
The benefits of chloride salts in the selective transfer
hydrogenolysis of 2 were much more pronounced than
in gaseous hydrogenolysis. In hopes of taking advan-
tage of the dual functions of formic acid as hydrogen
donor and acidic medium, selective hydrogenolysis of 2
with formic acid in methanol was investigated (Table
2). Unfortunately, no reaction occurred under these
conditions. Although ammonium formate has been
reported to be superior to formic acid as a hydrogen
donor due to the high hydrogen-donating ability of
formate ion, it is also highly effective in dehalogenation
of aryl chlorides.6 In line with this report, we found
that transfer hydrogenation of 2 in methanol gave
exclusively deschloro analog 3. On the other hand, in
ethyl acetate an 18:1 ratio of 1:3 was observed. The
chloride salt deactivating effect on selectivity was strik-
ing when ammonium chloride was added to the reac-
tion in MeOH. Although the debenzylation reaction
was slow and stalled around 93% conversion, the
product/deschloro ratio was 172:1.
In a recent study on the active Pd/C species in dechlori-
nations,5 cationic Pd was inferred to be responsible for
chemisorption of substates prior to reduction. We rea-
soned that chloride salts might selectively deactivate the
Pd/C catalyst by chelating to the cationic Pd sites, while
not deactivating the sites for the debenzylation. We
describe herein the results of a study which demonstrate
a novel approach to achieving both high chemoselectiv-
ity and reaction rate by the use chloride salts under
gaseous or transfer hydrogenolysis conditions.
Tetrabutylammonium chloride (TBACl) was examined
as chloride source for hydrogenations over Pd/C in
methanol and ethyl acetate (Table 1). In methanol, the
ratio of product to deschloro (1:3) increased dramati-
cally from 3.3:1 to 57:1 when 1.1 equiv. of TBACl was
used. The level of deschloro 3 could be further
decreased (but not eliminated) by use of higher levels of
TBACl. When a large excess of TBACl (5 equiv.) was
used, a new impurity was observed and the debenzyla-
tion did not go to completion, possibly indicating ‘over-
toxification’ of the catalyst. TBACl also detered
debenzylation in EtOAc. However, in this solvent the
reaction stalled at 25% conversion when only 0.1 equiv.
of TBACl was used.
Sodium hypophosphite has been recommended as a
mild hydrogen donor in the hydrogenolysis of benzyl
ethers.7 The transfer hydrogenation has been proposed
to occur via isomerization of the phosphoryl form of
(NaO)H2PO to the hydroxide (NaO)(HO)PH form fol-
lowed by dissociation of the PꢀH bond.8 In this study,
we found that use of this reagent in MeOH/water
afforded a 17:1 ratio of 1:3. A possible explanation of
this effect is the difference in pH of the reaction mix-
tures: the HCOONH4 reaction mixture is basic (pH
ꢀ9), while that of sodium hypophosphite is moderately
Table 1. TBACl effect on the ratio of 1/3a
TBACl (equiv.)
Solvent
3 (APb)
1 (AP)
2 (AP)
1/3 ratio
0
MeOH
MeOH
MeOH
MeOH
EtOAc
EtOAc
22.8
2.4
1.7
0.9
2.5
–
75.7
96.2
96.6
91.3
94.6
24.6
0.2
–
0.1
1.0
–
3.3:1
40:1
57:1
101:1
39:1
0.5
1.1
5.0
0
0.1
75.4
a Hydrogenations were carried out at ambient temperature and 40 psi over 18 h. The catalyst used was 5% Pd/C from Johnson Matthey. The
loading was 10% w/w relative to 2. The results were determined by HPLC monitored at 235 nm.
b AP stands for HPLC area percent.
Table 2. Ratios of 1/3 and conversions for different hydrogen donors, chloride salts and solvents
Hydrogen donor (equiv.)
Solvent
Chloride
Time (h)
Ratio (1/3)
Conversionc (%)
HCOOH (6 equiv.)
MeOH
MeOH
MeOH
EtOAc
MeOH/water (2/1)
MeOH/water (2/1)
MeOH/water (2/1)
MeOH/water (2/1)
MeOH/water (2/1)
16
16
21
16
16
23
2.5
23
23
–
0
100
93
aHCOONH4 (2 equiv.)
aHCOONH4 (2 equiv.)
aHCOONH4 (2 equiv.)
bNaH2PO2 (2 equiv.)
bNaH2PO2 (2 equiv.)
bNaH2PO2 (2 equiv.)
bNaH2PO2 (2 equiv.)
bNaH2PO2 (2 equiv.)
0/100
172/1
18/1
17/1
5/1
120/1
33/1
82/1
NH4Cl (8.8 equiv.)
76
100
100
98
100
86
NaCl (15 wt%)
NaCl (15 wt%)
NaCl (20 wt%)
a Reaction conditions: 2 (50 mg, 0.11 M), catalyst (10% Pd/C from Degussa, 10% w/w), 65°C.
b Reaction condition: 2 (0.08 M), catalyst (10% Pd/C from Johnson Matthey), 65°C.
c (Product+deschloro)/(product+deschloro+starting material).

J. Li et al. / Tetrahedron Letters 44 (2003) 4041–4043
4043
ratio of the product/deschloro was between 88:1 to
120:1 with 98% conversion after 2.5 h.
Based on the above results, the transfer hydrogenolysis
reaction was demonstrated on gram-scale using sodium
hypophosphite in the presence of 12 wt% NaCl solution
to afford product 1 in 93% isolated yield with product/
deschloro ratio as high as 120:1.
In conclusion, we have demonstrated a highly selective
gaseous and transfer hydrogenolysis of a benzyl pro-
tecting group in the presence of an aryl chloride in the
synthesis of quinolinone (1) using chloride salts to
suppress an undesired dehalogenation reaction.
Figure 1. Reaction conditions: 2 (50 mg, 0.08 M), NaH2PO2 (2
equiv.), catalyst (10% Pd/C, 10% w/w), MeOH/aq. NaCl, 2:1
(v/v), temperature (65°C).
Acknowledgements
acidic (pH ꢀ4). As noted earlier, dehalogenation is
faster under basic and neutral conditions but slower in
the presence of acid.2
The authors thank Drs. John Scott, Richard Mueller,
Edward Delaney, and Wendel Doubleday for critical
review of the manuscript. Drs. Jianji Wang, Y. Lear,
W.-S. Li, R. Polniaszek, P. Vemishetti, B. Santiago,
and Wansheng Liu for helpful discussions.
Employing sodium hypophosphite, it was found that
sodium chloride effectively suppresses the level of
deschloro byproduct, even under the prolonged heating
conditions (23 h) required to achieve complete debenzyl-
ation. We also found that the length of reaction time as
well as the amount of NaCl affected the ratio of 1:3
(Table 2). In order to investigate the effect resulted
from the concentration of NaCl and to achieve com-
plete debenzylation across the range of concentrations,
the results of the following reactions were compared
after 23 h (Fig. 1). Hydrogenolysis reactions were con-
ducted with various mixtures of MeOH and aq. NaCl
in which the MeOH/aq. NaCl ratio was maintained at
2:1 v/v, but the amount of NaCl in the aqueous compo-
nent was varied from 0–20 wt%. As shown in Figure 1,
the ratio of product/deschloro increased from 5:1 to
33:1 as the NaCl concentration in the aqueous compo-
nent increased from 0 to 18 wt%. However, the ratio
reached a plateau around 15 wt% NaCl. When 20 wt%
NaCl was used, the debenzylation reaction stalled at
86% conversion, although the ratio of product/
deschloro was 82:1. No further dechlorination was
observed, indicating ‘over-toxification’ of dechlorina-
tion site on the catalyst as well. To avoid the potential
catalyst poisoning due to high salt concentration and to
maximize the ratio of the product/deschloro without
compromising the yield, the optimal NaCl concentra-
tion was determined to be 15 3 wt% and the average
References
1. (a) Sit, S.-Y.; Meanwell, N. A. US 5,892,045, April 6,
1999; (b) Hewawasam, P.; Starrett, J. E., Jr.; Swartz, S. G.
US 5,972,961, October 26, 1999; (c) Hewawasam, P.;
Starrett, J. E., Jr. US 6,184,231, February 6, 2001.
2. Studer, M.; Blaser, H.-U. J. Mol. Catal. A: Chem. 1996,
112, 437–445.
3. Freifelder, M. Practical Catalytic Hydrogenation; Wiley-
Interscience: New York, 1971.
4. (a) Fitzpatrick, L.; Sala, T.; Sargent, M. V. J. Chem. Soc.,
Perkin Trans. 1 1980, 85–89; (b) Jendralla, H.; Granzer,
E.; Kerekjarto, B. V.; Krause, R.; Schacht, U.; Baader, E.;
Bartmann, W.; Beck, G.; Bergmann, A.; Kesseler, K.;
Wess, G.; Chen, L.-J.; Granata, S.; Herchen, J.; Kleine,
H.; Schussler, H.; Wagner, K. J. Med. Chem. 1991, 34,
2962–2983.
5. Gomez-Sainero, L. M.; Seoane, X. L.; Fierro, J. L. G.;
Arcoya, A. J. Catal. 2002, 209, 279–288.
6. Rajagopal, S.; Spatola, A. F. J. Org. Chem. 1995, 60,
1347–1355.
7. Sala, R.; Doria, G.; Passarotti, C. Tetrahedron Lett. 1984,
25, 4565–4568.
8. Dorfman, Y. A.; Aleshkova, M. M. Kinet. Catal. 1998, 39,
852–858.