Pd-catalyzed deoxygenation of mandelate esters

Article abstract of DOI:10.1021/jo8021499

(Chemical Equation Presented) A new approach to the synthesis of phenylacetic acids and esters has been developed via the palladium-catalyzed deoxygenation of mandelate esters.

Full text of DOI:10.1021/jo8021499

SCHEME 1. Proposed Synthesis of 4
Pd-Catalyzed Deoxygenation of Mandelate Esters
Jacqueline E. Milne,* Jerry A. Murry, Anthony King, and
Robert D. Larsen
Chemical Process R&D, Amgen Inc., One Amgen Center
DriVe, Thousand Oaks, California 91320; milne@amgen.com
milne@amgen.com
ReceiVed September 25, 2008
SCHEME 2. Preparation of the Intermediate Phosphate 6
A new approach to the synthesis of phenylacetic acids and
esters has been developed via the palladium-catalyzed
deoxygenation of mandelate esters.
number of them are commercially available. We envisioned that
4-hydroxy-3-methoxymandelic acid 2 (commonly used for the
synthesis of homovanillic acid) could be elaborated to intermediate
3 and reduced to afford 4. Attempts at a global reduction by metal-
catalyzed hydrogenation resulted in exclusive nitro reduction
without any evidence of C-O bond cleavage.6 As a result, we
investigated several alternative reduction protocols to effect this
transformation. Herein, we report the development and application
of a new approach to the synthesis of phenylacetic acids and esters
from mandelate esters which involves activation of the R-hydroxy
group as the phosphate ester and subsequent palladium-catalyzed
deoxygenation.
Aryl acetic acids and derivatives are an important class of
organic molecules that are prevalent in both pharmaceuticals
and natural products. They show a broad range of biological
activities1 including antibacterial, analgesic, and antiviral and
are one of the main classes of nonsteroidal anti-inflammatory
drugs. During a recent development program, we required an
efficient route to the advanced intermediate 4 (Scheme 1).2
After reviewing the relevant literature methods3,4 we became
interested in the reduction of mandelic acids as a convenient route
to this moiety. Mandelic acids are easily synthesized5 and a large
Our initial investigation utilized methyl mandelate 5 as a
model substrate for the deoxygenation reaction. The phosphate
6 was prepared in situ (Scheme 2) and the crude reaction mixture
was directly subjected to a screen of metal-catalyzed reduction
conditions.7 A variety of metal catalysts (Pd, Ni, Cu, and Fe)
were examined with NaBH4 as the reducing agent and only Pd
catalysts afforded significant amounts of product (Table 1,
entries 2-6). Various Pd/ligand combinations were examined
(entries 10-14). Pd(ACN)2Cl2/BINAP was identified as a
particularly active catalyst system (entry 8) and was shown to
be effective at loadings of 0.5% Pd (Table 1, entry 9).
The effects of solvent, temperature, and reducing agent were
subsequently examined. It was shown that although the reaction
could proceed in THF (Table 2, entry 3), slightly higher yields
were obtained in the presence (ca. 35-40%) of a cosolvent
(Table 2, entry 1-2). Use of other solvents such as MTBE,
PhMe, and DMF for both the phosphate formation and deoxy-
genation gave lower yields of the desired phenylacetate (Table
2, entries 4-6). At room temperature, the reaction provided 75%
(1) (a) Tuncel, G.; Nergiz, C. Lett. Appl. Microbiol. 1993, 17, 300–302. (b)
Naylor, A.; Judd, D. B.; Brown, D. S. EP 343,900, 1991. (c) Singh, R.; Mall,
T. P. Sci. Cult. 1979, 45, 66–68. (d) Rieu, J.-P.; Boucherle, A.; Cousse, H.;
Mouzin, G. Tetrahedron 1986, 42, 4095–4131.
(2) Medina, J. C.; Liu, J.; Fu, Z.; Schmitt, M.; Wang, Y.; Derek, M. C.;
Tang, H. L.; Sullivan, T.; Tonn, G.; Collins, T. Discovery of AMG 009: A
CRTH2 and DP dual antagonist. Abstracts of Papers; 234th National Meeting
of the ACS, Boston, MA, Aug 19-23, 2007; American Chemical Society:
Washington, DC, 2007; No. 883609.
(3) Recent syntheses of phenylacetic acids and esters: Pd-catalyzed
R-arylation: Moradi, W. A.; Buchwald, S. L. J. Am. Chem. Soc. 2001, 123,
7996–8002. Hama, T.; Hartwig, J. F. Org. Lett. 2008, 10, 1549–1552.
Rh-catalyzed carbonylation: Giroux, A.; Nadeau, C.; Han, Y. Tetrahedron Lett.
2000, 41, 7601–7604. One carbon elongation of aldehydes: Huh, D. H.; Jeong,
J. S.; Lee, H. B.; Ryu, H.; Kim, Y. G. Tetrahedron 2002, 58, 9925–9932. Zhou,
G.-B.; Zhang, P.-F.; Pan, Y.-J. Tetrahedron, 2005, 61, 5671–5677. Pd-catalyzed
carbonylation: Kohlpaintner, C. W.; Beller, M. J. Mol. Catal. A: Chem. 1997,
116, 259–267. Willerodt-Kindler: Alam, M. M.; Adapa, S. R. Synth. Commun.
2003, 33, 59–63. Reduction of mesylates: Inokuchi, T.; Sugimoto, T.; Kusumoto,
M.; Torii, S. Bull. Chem. Soc. Jpn. 1992, 65, 3200–3202. Nakamura, K.; Yasui,
S.; Ohno, A.; Oka, S. Tetrahedron Lett. 1983, 24, 2001–2004.
(4) Reduction of Mandelic acids: Mitchell, A.; Bailey, T. Br. Patent 2 078
718, 1982. Spielmann, W.; Schaeffer, G. DE 3 001 511, 1980. . Vallejos, J.-C.;
Christidis, Y. EP 0 221 815, 1986. PhilipN. E. Br. Patent 1 576 333, 1976.
Vallejos, J.-C.; Legrand, O.; Christidis, Y. Bull. Soc. Chim. Fr. 1997, 134, 101–
104. Ronchin, L.; Toniolo, L.; Cavinato, G. Appl. Catal., A 1997, 165, 133–
145.
(6) The aromatic amino group may suppress C-O bond cleavage: Sajiki, H.
Tetrahedron Lett. 1995, 36, 3465–3468.
(5) Synthesis of mandelic acids: (a) Corson, B. B.; Dodge, R. A.; Harris,
S. A.; Yeaw, J. S. Org. Synth. 1926, 6, 58–62. (b) Bjorsvik, H.-R.; Liguori, L.;
Minisci, F. Org. Process Res. DeV. 2000, 4, 534–543. (c) Weist, S.; Kittel, C.;
Bischoff, D.; Bister, B.; Pfeifer, V.; Nicholson, G. J.; Wohlleben, W.; Sussmuth,
R. D. J. Am. Chem. Soc. 2004, 126, 5942–5943. (d) Gathergood, N.; Zhuang,
W.; Jorgensen, K. A. J. Am. Chem. Soc. 2000, 122, 12517–12522.
(7) The use of triflates, sulfonates, and carbonates was also investigated.
Formation of the triflate of methyl mandelate was unsuccessful possibly due to
instablity at RT: Feenstra, R. W.; Stokkingreef, E. H. M.; Nivard, R. J. F.;
Ottenheijm, H. C. J. Tetrahedron 1988, 44, 5583–5595. The mesylate, tosylate,
and ethyl carbonate did not give high yields of desired product 7, when subjected
to the Pd-catalyzed deoxygenation.
10.1021/jo8021499 CCC: $40.75
Published on Web 11/18/2008
2009 American Chemical Society
J. Org. Chem. 2009, 74, 445–447 445

TABLE 1. Deoxygenation of Methyl Mandelate Catalyst
Optimization
TABLE 3. Substrate Scope
catalyst
yield (%)b
1
2
3
4
none
0
16
14
8
20 mol % Ni(acac)2
20 mol % Ni(dppf)Cl2
10 mol % CuI/1,10-phenanthroline
20 mol % Fe(acac)3
5
0
6
7
8
9
10
11
12
13
14
2 mol % Pd(dippf)Cl2
95
27
100
94
87
88
49
29
28
2 mol % Pd(ACN)2Cl2 /no ligand
2 mol % Pd(ACN)2Cl2/BINAP
0.5 mol % Pd(ACN)2Cl2/BINAP
2 mol % Pd(OAc)2/BINAP
2 mol % Pd(OAc)2/Cy-JohnPhos
2 mol % Pd(OAc)2 /XantPhos
2 mol % Pd(OAc)2/t-Bu3P
2 mol % Pd(OAc)2/n-Bu3P
a All phosphates prepared in THF. b hplc assay yield. c dippf )
1,1′-bis(di-isopropylphosphino)ferrocene; Cy-JohnPhos ) 2-(dicyclohexyl-
phosphino)biphenyl.
TABLE 2. Deoxygenation of Methyl Mandelate Reaction
Optimizationa
a Yields on 5 mmol scale. b Isolated after hydrolysis to the acid.
c Phosphate prepared at 45 °C, using 1 equiv of DMAP.
reductant
NaBH4
NaBH4
NaBH4
NaBH4
NaBH4
NaBH4
NaBH4
solvent
T (°C)
yield (%)b
2-hydroxy-2-phenylacetate was subjected to our deoxygenation
protocol despite the lability of these groups under other Pd-
catalyzed reduction conditions8 (entry 5). Furthermore, this method
tolerates steric hindrance at the R-position and can be applied to
the synthesis of phenyl propanoic acids (entry 7). Despite the
additional steric bulk, C-O cleavage to the desired product was
achieved in 85% yield. The success of this tertiary alcohol example
is noteworthy since the olefin side product from the possible
competing ꢀ-hydride elimination pathway was not observed.
Employing similar benzylic phosphates containing ꢀ-hydrogens in
Pd-catalyzed cross-coupling reactions has shown ꢀ-hydride elimi-
nation to dominate.9,10
1
2
3
4
5
6
7
8
9
THF/diglyme
THF/DME
THF
PhMe
MTBE
75
75
75
75
75
75
40
rt
100
95
88
55
76
20
97
75
95
95
DMF
THF/DME
THF/DME
THF/DME
THF/DME
NaBH4
Et3N/HCO2H
HCO2NH4
75
75
10
a All phosphates prepared in THF except entries 4, 5, and 6. In these
cases phosphate was prepared in solvent stated in table b hplc assay
yield.
To gain further insight, we examined the deoxygenation of
(S)-methyl 2-hydroxy-2-phenylpropanoate (100% ee). The
phosphate was prepared without loss of chiral purity. However,
since the deoxygenation step presumably proceeds via a Pd
enolate, partial racemization was observed, regardless of whether
rac- or (S)-BINAP was employed (Scheme 3).
Encouraged by the scope of our method we returned our focus
to the synthesis of 4. Fortunately, subjection of phosphate 10
to our Pd-catalyzed reduction conditions concomitantly reduced
both the C-O and nitro groups11 to directly provide the target
yield of the desired product 7. However, an increase in the
reaction temperature to just 40 °C improved the conversion to
afford 7 in 97% yield (Table 2, entries 7 and 8). Interestingly,
the reaction is not limited to the use of NaBH4, which can be
replaced by either Et3N/formic acid or ammonium formate as
the reducing agent (Table 2, entries 9 and 10). In addition, it is
worth noting that while all reaction optimization was performed
under an inert atmosphere, 94% of 7 could be obtained by using
2% Pd without the exclusion of air.
With the optimal conditions in hand, we next examined the scope
of the method. A range of mandelate esters were subjected to the
deoxygenation procedure and the phenylacetic acid/ester products
were obtained in yields ranging from 73% to 99%. As shown in
Table 3, examples include substrates with both electron-rich and
-poor substituents on the aromatic ring. It is noteworthy that the
deoxygenation of methyl 2-(4-chlorophenyl)-2-hydroxyacetate
(entry 4) proceeds without concomitant cleavage of the C-Cl bond.
In addition, the benzyl group remained intact when benzyl
(8) Benzyl removal: Felix, A. M.; Heimer, E. P.; Lambros, T. J.; Tzougraki,
C.; Meienhofer, J. J. Org. Chem. 1978, 43, 4194–4196.
(9) Coupling of benzylic phosphates: McLaughlin, M. Org. Lett. 2005, 7,
4875–4878.
(10) The product of ꢀ-elimination methyl 2-phenylacrylate is not observed
in the reaction mixture. It is unlikely that this is a major intermediate in the
reaction pathway. When subjected to the Pd-catalyzed deoxygenation conditions
described in this paper, a number of products are formed.
(11) Pd-catalyzed nitro reductions: Petrini, M.; Ballini, R.; Rosini, G.
Synthesis 1987, 71, 3–714. Rahaim, R. J., Jr.; Maleczka, R. E., Jr. Synthesis
2006, 331, 6–3340.
446 J. Org. Chem. Vol. 74, No. 1, 2009

SCHEME 3
obtained in 70% yield. However, 11 was not readily formed
under the reaction conditions.12 Therefore, we suggest that the
most likely rationale is described by path B. In this case, the
role of the chloride is to replace the phosphate as a ligand on
Pd, forming the Pd(II)Cl complex, which is capable of undergo-
ing transmetalation/reduction. There is much precedent that the
presence of chloride ion has a positive effect on the Pd-catalyzed
couplings of aryl and vinyl triflates.13 Similarly, LiCl has been
employed to facilitate the cross-coupling reactions of
phosphates.13d
SCHEME 4
In summary, we have demonstrated a new procedure for the
synthesis of phenylacetic acids and esters from methyl mande-
lates. This approach required activation of the R-hydroxyl by
the preparation of an intermediate phosphate, which was reduced
with sodium borohydride by using a Pd catalyst. The ready
availability of mandelate esters should make this an attractive
method. The generality and functional group compatibility of
this deoxygenation was demonstrated by using a variety of
substrates.
Experimental Section
Methyl 2-(4-(2-Amino-4-(butylcarbamoyl)phenoxy)-3-meth-
oxyphenyl)acetate (4). Methyl 2-(4-(4-(butylcarbamoyl)-2-nitro-
phenoxy)-3-methoxyphenyl)-2-hydroxyacetate (1 g, 2.3 mmol, 1
equiv) and DMAP (11 mg, 0.09 mmol, 4%) were dissolved in THF
(10 mL) in a vial sealed with a rubber septum. Triethylamine (0.64
mL, 6.4 mmol, 2 equiv) was added dropwise followed by diphenyl
phosphoryl chloride (0.60 mL, 2.88 mmol, 1.25 equiv). The reaction
was stirred at rt until 100% complete (checked by hplc).
SCHEME 5
Pd(ACN)2Cl2 (30 mg, 5 mol %) and BINAP (72 mg, 5 mol %)
were weighed into a glass vial and sealed with a rubber septum.
The vial was evacuated and filled with N2 three times. The
phosphate reaction mixture was transferred by syringe with THF
(2 mL) wash. NaBH4 (9.2 mL, 0.5 M solution in diglyme, 2 equiv)
was added dropwise with a syringe. The reaction was heated to 75
°C until complete (checked by hplc). The resulting crude reaction
was filtered, concentrated, and purified by column chromatography
(silica, eluent heptane:EtOAc 20:1 to 1:1) to yield solid methyl
2-(4-(2-amino-4-(butylcarbamoyl)phenoxy)-3-methoxyphenyl)ac-
1
etate (0.546 g, 1.4 mmol, 61%); mp 108.3-109.5 °C; H NMR
(400 MHz, DMSO-d6) δ ppm 8.13 (t, J ) 5.5 Hz, 1H), 7.24 (s,
1H), 7.07 (s, 1H), 6.93 (dd, J ) 8.1, 2.0 Hz, 1H), 6.85 (m, 2H),
6.44 (d, J ) 8.5 Hz, 1H), 5.04 (s, 2H), 3.76 (s, 3H), 3.68 (s, 2H),
3.63 (s, 3H), 3.20 (q, J ) 6.5 Hz, 2H), 1.47 (m, 2H), 1.30 (m, 2H),
0.89 (t, J ) 7.5 Hz, 3H); 13C NMR (101 MHz, DMSO-d6) δ ppm
171.6, 166.3, 150.5, 146.0, 143.1, 138.4, 131.0, 130.0, 121.8, 120.0,
115.1, 114.5, 114.4, 55.7, 51.7, 38.7, 31.3, 19.6, 13.7; IR (neat)
3443, 3364, 3293, 1737, 1631; HRMS calcd for C21H27N2O5 [M
+ H]+ 387.1914, found 387.1904
4 (Scheme 4). Interestingly, in this case, diglyme cosolvent was
necessary to obtain a moderate yield (61%). Reactions conducted
solely in THF gave 66% conversion of phosphate 10, affording
only 2% of the desired product 4. The main product in this
case was the nitro methyl ester, a result of selective C-O
cleavage of the phosphate. At lower temperatures and catalyst
loadings inferior yields were obtained.
Several mechanistic scenarios can be envisioned for the
described transformation. Our optimized protocol involved using
solutions of phosphate that were prepared in situ and directly
subjected to the reduction conditions. To our surprise, when
the phosphate 6 was isolated by aqueous workup and subjected
to the reaction conditions only 15% of the desired product was
observed. Interestingly, when repeated with the addition of 2
equiv of LiCl or Et3NHCl (a byproduct from the phosphate
formation), 68% and 99% yield of the phenylacetate 7 was
obtained, respectively. To explain this observation we envisioned
that the reaction could proceed via two plausible pathways as
depicted in Scheme 5. Path A involves intermediacy of the
chloro compound 11, which could form by substitution of the
phosphate group with a chloride. In support of this theory, it
was shown that when ethyl R-chlorophenylacetate was treated
under the deoxygenation conditions, ethyl phenylacetate was
Acknowledgment. The authors wish to thank Jenny Chen
for her help in chiral hplc analysis. Mr. Jim Yang is acknowl-
edged for early experiments.
Supporting Information Available: Experimental proce-
dures, compound characterization data, and copies of 1H NMR
and 13C NMR spectra. This material is available free of charge
via the Internet at http://pubs.acs.org.
JO8021499
(12) When the crude solution of phosphate was heated to 75 °C in either the
presence or absence of catalyst conversions to the chloro compound were 39%
and 9%, respectively
(13) Role of LiCl: (a) Echavarren, A. M.; Stille, J. K. J. Am. Chem. Soc.
1987, 109, 5478–5486. (b) Scott, W. J.; Crisp, G. T.; Stille, J. K. J. Am. Chem.
Soc. 1984, 106, 4630–4632. (c) Scott, W. J.; Stille, J. K. J. Am. Chem. Soc.
1986, 108, 3033–3040. (d) Nicolaou, K. C.; Shi, G.-Q.; Gunzner, J. L.; Gartner,
P.; Yang, Z. J. Am. Chem. Soc. 1997, 119, 5467–5468.
J. Org. Chem. Vol. 74, No. 1, 2009 447