Utilization of sequential palladium-catalyzed cross-coupling reactions in the stereospecific synthesis of trisubstituted olefins

Article abstract of DOI:10.1021/op900015g

A stereospecific synthesis of the drug-candidate 1 is described. The synthetic sequence, aimed at accomplishing modularity and cost savings, features a series of organometallic steps to afford stereospecifically the desired trisubstituded olefin active pharmaceutical ingredient. Key developments consist of a mild Sonogashira reaction of aryl bromide 7a with the polymerization prone propargyl alcohol and a stereospecific hydroalumination, Zn/Al exchange, and Pd-catalyzed cross-coupling sequence facilitated by the commercially available PEPPSI catalyst.

Full text of DOI:10.1021/op900015g

Organic Process Research & Development 2009, 13, 598–606
Utilization of Sequential Palladium-Catalyzed Cross-Coupling Reactions in the
Stereospecific Synthesis of Trisubstituted Olefins
Ioannis N. Houpis,*,† Didier Shilds,† Ulrike Nettekoven,‡ Anita Schnyder,‡ Erhart Bappert,‡ Koen Weerts,† Martine Canters,†
and Wim Vermuelen†
Johnson and Johnson PRD, API DeVelopment, Turnhoutseweg 30, 2340 Beerse, Belgium, and SolVias A.G., Business Unit
Synthesis and Catalysis, Mattenstrasse 22, 4002 Basel, Switzerland
Scheme 1. Retrosynthetic analysis
Abstract:
A stereospecific synthesis of the drug-candidate 1 is described. The
synthetic sequence, aimed at accomplishing modularity and cost
savings, features a series of organometallic steps to afford ste-
reospecifically the desired trisubstituded olefin active pharmaceuti-
cal ingredient. Key developments consist of a mild Sonogashira
reaction of aryl bromide 7a with the polymerization prone
propargyl alcohol and a stereospecific hydroalumination, Zn/Al
exchange, and Pd-catalyzed cross-coupling sequence facilitated by
the commercially available PEPPSI catalyst.
Introduction
In the course of the last several decades, there has been an
explosion of transition metal-catalyzed cross-coupling reactions
of aromatic and aliphatic electrophiles starting with, literally,
the entire gamut of organometallic reagents (B, Mg, Sn, In,
Cu, Mn, etc.).1
These reactions have found particular use in heterocyclic
chemistry and are applied vigorously in the pharmaceutical
industry.2
One area where these reactions can be most useful is the
stereospecific construction of trisubstituted olefins by combining
hydro- or carbometalation with cross-coupling or alkylation
protocols. The application of these methods still presents a
challenge despite significant advances in the field.3
In this contribution we describe the development and
optimization of an “all organometallic” synthetic sequence for
the synthesis of 1 (Scheme 1) where several challenging
catalytic steps had to be developed in order to accommodate
the specific reactivity of our substrate. Although 1 does not
possess significant structural complexity, this actually places
an additional demand to develop a most efficient and economi-
cally advantageous synthetic sequence.
Results and Discussion
Retrosynthetic Analysis. Our strategy focused on a se-
quence of cross-coupling reactions in the early steps of the
synthesis leaving for the end the allylic amination reaction in
order to avoid the use of transition metals in the final step, thus
reducing the chances of transition metal contamination in the
final product.
We envisioned that 1 could be derived by alkylation of an
activated derivative of 3 with sarcosine hydrochloride 2. In turn,
the allylic alcohol 3 could be prepared via a stereospecific cross-
coupling reaction of a nucleophilic (4a)4 or electrophilic (4b)
thiophene derivative with an appropriately functionalized
* Author for correspondence. E-mail: yhoupis@its.jnj.com.
† Johnson and Johnson PRD.
‡ Solvias A.G.
(1) (a) Anctil, E. J.-G.; Snieckus, V., Metal-Catalyzed Cross Couping
Reactions, 2nd ed.; Wiley: New York, 2004; pp 761-814. (b) Whisler,
M. C.; MacNeil, S.; Snieckus, V.; Beak, P. Angew. Chem., Int. Ed.
2004, 43, 2206. (c) Suzuki, A. Metal-Catalyzed Cross Couping
Reactions; Wiley-VCH: New York, 1998; p 49.
(2) King, A. O.; Yasuda, N. Organometallics in Process Chemistry;
Springer-Verlag: Berlin, 2004; pp 205-245.
(3) (a) Huang, Z.; Negishi, E. J. Am. Chem. Soc. 2007, 129, 14788. (b)
Langille, N. F.; Jamison, T. F. Org. Lett. 2006, 8, 3761. Yin, N.; Wang,
G.; Qian, M.; Negishi, E. Angew. Chem., Int. Ed. 2006, 45, 2916.
Reiser, O. Angew. Chem., Int. Ed. 2006, 45, 2838. Hoveyda, A. H.;
Evans, D. A.; Fu, G. C. Chem. ReV. 1993, 93, 1307.
(4) The reaction of the commercially available boronic acid 4a with the
vinyl iodide 9 has been carried out and is successful. It is described
in the Experimental Section but will not be discussed in detail here.
598
•
Vol. 13, No. 3, 2009 / Organic Process Research & Development
10.1021/op900015g CCC: $40.75  2009 American Chemical Society
Published on Web 04/22/2009

Scheme 2. Synthesis of bis-aryl halides 7a and 7b
compound 5 (M ) I, Zn/Al, respectively). We anticipated that
the well-known stereocontrolling element of the coordination
of the alkoxide to an appropriate metal (M in 5) could be
exploited in the subsequent cross-coupling reaction in order to
define the stereochemistry of the double bond.5 Such coordina-
tion could be engineered by a hydro-alumination reaction of
unprotected propargyl alcohols, such as 6, the latter being a
product of a Sonogashira reaction of the commercially available
7a or 7b and propargyl alcohol.
Preparation of Propargyl Alcohol 6. For commercial-scale
production, both 7a and 7b are available at relatively low cost.
However, for research purposes they were conveniently syn-
thesized as shown in Scheme 2.
Thus, Suzuki coupling of p-bromo-iodobenzene with 2-furyl
boronic acid 8 was readily accomplished using Pd/C in
THF-water mixtures although a small amount of PPh3 was
needed for the reaction to proceed at a reasonable rate.6 The
product 7a, can be isolated in ∼93% yield with >99% purity
by crystallization from ethanol-water. Although the Pd-content
in solution was not measured at this juncture, the isolated
product contained <10 ppm of Pd. It is worth noting that this
purification is crucial as crude 7a (97% purity) performed
erratically in the subsequent Sonogashira reaction. The iodide
7b was obtained in high yield by following the Buchwald
protocol.7
Since its discovery, the Songashira reaction has been a
valuable tool for the functionalization of aromatic and het-
eroaromatic halides with the versatile acetylene function.8 Apart
from the normal acetylene manipulations, these derivatives can
undergo cyclization reactions to prepare indoles, benzo- and
heteroarylfurans, and other useful pharmacophores.2 One of the
difficulties of the reaction is the propensity of acetylenes to
dimerize under the reaction conditions particularly when the
cross-coupling reaction is slow as in the case of aromatic
bromides and chlorides which normally require elevated tem-
peratures. In our case the need to employ propargyl alcohol as
the acetylene complicates the problem further as this reagent
polymerizes rather easily at higher temperatures.9
In order to avoid these higher temperatures we started our
investigation by utilizing the protocol developed by Buchwald
and Fu (eq 1) for performing the Sonogashira reaction of
acetylenes with aryl bromides and chlorides at ambient tem-
perature.10 Unfortunately, these conditions11 afforded only 50%
conversion along with visible polymerization products. Exami-
nation of alternative methods to effect this coupling, for example
employing the corresponding silyl protected acetylene, were also
not successful.12
To find a catalyst system that would allow complete
conversion while minimizing polymerization of propargyl
alcohol, we undertook an extensive screening of catalysts and
reaction conditions. Initial experiments showed that polar aprotic
solvents such as DMF and NMP (but not DMSO) were much
more effective than THF, toluene, or dioxane. In addition, it
was established that inorganic bases were not useful for this
transformation.
These initial observations allowed for a more focused
screening of ∼60 experiments with the most pertinent “hits”
shown in Table 1.
The trends observed for this reaction can be summarized as
follows: PPh3 was the best ligand (Table 1, entries 1-7), while
BuNH2 or DMPU were the preferred solvents for this reaction
(Table 1, entries 1, 8, and 10). We have confirmed earlier
observations that primary amine bases were preferred to
(5) (a) Sato, F.; Kodama, H.; Sato, M. J. Organomet. Chem. 1978, 157,
C30. (b) Baba, S.; Van Horn, D. E.; Negishi, E. Tetrahedron Lett.
1976, 17, 1927. (c) Ziegler, F. E.; Mikami, K. Tetrahedron Lett. 1984,
25, 131. (d) Review: Hudrlik, P. F.; Hudrlik, A. M. The Chemistry of
the Carbon-Carbon Triple Bond; John Wiley and Sons: New York,
1978; pp 199-273.
(6) (a) Lipshutz, B. H.; Blomgren, P. A. J. Am. Chem. Soc. 1999, 121,
5819. (b) Eichen, K.; Gebhart, J.; Rack, M.; Schafer, P. Int. Patent
WO 97/733846, 1997.
(7) Klapars, A.; Buchwald, S. L. J. Am. Chem. Soc. 2002, 124, 14844.
(8) Sonogashira, K. Metal-Catalyzed Reactions; Wiley-VCH: New York,
1998; pp 203-229.
Vol. 13, No. 3, 2009 / Organic Process Research & Development
•
599

Table 1. Initial screening of the Sonogashira coupling of 7a with propargyl alcohola
entry base (equiv) Pd/L (additive) Pd/L (mol %) solvent time [h]
7ab
6b
Σ restc in situ yieldd
1
2
3
4
5
6
7
8
n-BuNH2
Cy2NMe
n-BuNH2
n-BuNH2
n-BuNH2
n-BuNH2
n-BuNH2
n-BuNH2
(excess)
Pd(OAc)2/PPh3
2/8
2/8
2/8
2/8
2/8e
2/8
2/8
2/8
NMP
NMP
NMP
NMP
NMP
NMP
NMP
n-BuNH2
16
16
4
18
18
4
4
4
18
18
35
90
51
<10
13
-
-
Pd(OAc)2/PPh3
-
Pd(OAc)2/P(furyl)3
>80
>80
>80
>80
>80
Pd(OAc)2/P(t-Bu)2-norbornylxHBF4
Pd(OAc)2/P(t-Bu)2-norbornylxHBF4
Pd(OAc)2/P(t-Bu)2-biphenyl
Pd(OAc)2/PCy3
22
23
19
16
80
86
71
Pd(OAc)2/PPh3
12
>30
76
75
12
∼15
9
n-BuNH2
(excess)
n-BuNH2
Pd(OAc)2/PPh3
1,2-diaminobenzene (8%)
Pd(OAc)2/PPh3
2/8
2/8
n-BuNH2
10
DMPU
1
4
18
16
22
29
7
3
11
92
59
83
90
82
8
12
10
7
68
96
95
74
11
12
n-BuNH2
n-BuNH2
(2.0 equiv)
n-BuNH2
(1.5 equiv)
n-BuNH2
(2.0 equiv)
n-BuNH2
(5.0 equiv)
n-BuNH2
n-BuNH2
Pd/C (5%)/PPh3
(TPPTS)2PdCl2
2/8 f
1/2
DMPU
DMPU
7
-
-
13
14
15
Na2PdCl4/PPh3
Na2PdCl4/PPh3
Na2PdCl4/PPh3
1/2.5f
1/3 f
1/4 f
DMPU
DMPU
DMPU
22
21
22
>15
>10
65
74
93
1
0
97
>95
-
-
16
17
Pd2(dba)3/PPh3
Pd2(dba)3/PPh3
2 vol % H2O
1/3 f
1/3 f
DMPU
DMPU
16
4
16
-
-
61
73
29
2
68
94
3
3
100
a These experiments were performed by dissolving 1 mmol of 7a in 5 mL of the appropriate solvent, followed by 1.1-1.3 equiv of amine (unless otherwise mentioned), 4
mol % CuI (unless otherwise mentioned) 2-8 mol % of phosphine ligand, 1-2 mol % of Pd-precursor and using 1.1-1.3 equiv of propargylic alcohol added in one portion).
Finally the blue-green homogeneous mixture was heated 70 °C. b Determined by GLC analysis (area %). c Sum of the total remaining impurities determined by GLC analysis
(area %). d Yield of 6 determined by GLC analysis using biphenyl as the internal standard. e CuBr was used instead of CuI. f 2 mol % CuI was used.
secondary or tertiary,13 while bulky alkyl phosphines were
surprisingly ineffective ligands (Table 1, entries 3-7 and eq
Scheme 3. Observation of exothermic behavior in the
reaction of 7a with propargyl alcohol
1). The first success with PPh3 (Table 1, entry 8) provided 86%
in situ yield of 6; however, the reaction did not proceed to
completion, and a number of impurities were observed as well
as obvious polymerization products. The Pd/L ratio and the
amount of n-BuNH2 also seems crucial (entries 13-15),
whereas the Pd-precursor does not seem to have an important
influence on the reaction.
DMPU appears to be the best solvent, and it is worth noting
that thorough degassing of the solvent is essential. Finally, the
presence of water does not adversely affect the outcome of the
reaction and may indeed be helpful (Table 1, entry 17).
highly exothermic polymerization of the acetylene was deter-
However, we were surprised to find that when the reaction
mined not surprisingly to be due to CuI.8 Control experiments
was attempted on larger scale (∼100 g) a significant exotherm
established that the copper cocatalyst is needed in this trans-
was observed when applying the protocol used in the screening
formation (Scheme 3).
The optimal reaction conditions were established by limiting
the concentration of propargyl alcohol in the medium. Thus,
reactions. Specifically, mixing the components at room tem-
perature in a 2 L vessel followed by slow heating over 45 min
to 70 °C caused an abrupt temperature rise to 145 °C along
with the visible formation of dark polymeric material. This
heating
a dark-green solution of a mixture of 7a,
Pd2(dba)3 ·CHCl3 (0.5 mol %),14 PPh3 (4 mol %), and CuI (2
mol %) in DMPU or DMI, followed by slow addition of
propargyl alcohol, afforded 6 in >90% assay yield with little
polymerization (eq 2). The isolation of 6 from the reaction
mixture was problematic and could be accomplished in one of
two ways: (a) filtration through a pad of silica gel (silica:6 )
3:1, elution with toluene; Pd level <50 ppm) or (b) treatment
(9) Li, J.-H.; Liang, Y.; Wang, D.-P.; Liu, W.-J.; Xie, Y.-X.; Yin, D.-L.
J. Org. Chem. 2005, 70, 2832, and reference 4d therein.
(10) Hundermark, T.; Litke, A. F.; Buchwald, S. L.; Fu, C. G. Org. Lett.
2000, 2, 1729.
(11) Reasonable variations of the prescribed conditions were attempted,
such as temperature, Pd:ligand ratio, base and catalyst loading, without
significant success.
(12) Use of the silyl protected acetylene in the presence of N-heterocyclic
carbenes was also not successful in this transformation: Yang, C.;
Nolan, S. P. Organometallics 2002, 21, 1020.
(13) Repeating the xperiment in Table 1, entry 10, with Et3N or i-Pr2NH
afforded no conversion to the desired product. Houpis, I. N.; Choi,
W.-B.; Reider, P. J.; Molina, Churchill, A. H.; Lynch, J.; Volante,
R. P. Tetrahedron Lett. 1994, 35, 9355.
(14) Lower catalyst loads seemed to afford “cleaner” (less colored) reaction
mixtures; however, reducing the pre-catalyst load below the 0.5 mol
% level led to incomplete reactions.
600
•
Vol. 13, No. 3, 2009 / Organic Process Research & Development

The hydroxyl-directed hydroalumination17 reaction provides
an excellent vehicle to accomplish both goals as it affords almost
complete stereocontrol due to the formation of the stable
intermediate 5 which directs both the regio- and diastereose-
lectivity of the addition to the triple bond.18 Stereocomplemen-
tary results can be obtained if the alcohol function is protected.
In the original reports of the hydroalumination reaction,
pyrophoric DiBALH or LiAlH4 was used; however, this
problem has recently been overcome by the introduction of a
safer reagent, Red-Al [NaH2Al(OCH2CH2OMe)2],3b which is
Scheme 4. Reaction of the iodide 7b with propargyl alcohol
readily available on large scale under the trade name VitrideTM
.
As our plan was to use 5 for further elaboration via a Pd- or
Ni-catalyzed cross-coupling reaction (e.g., with bromo-thiophene
4b) we decided to study the hydroalumination reaction and the
stability of the vinyl aluminate species 5. We were quite hopeful
that this strategy would succeed since earlier work by the
Dvorˇa´k group indicated convenient access to the desired coupled
product (eq 3).19
with Silica-Bond-thiol15 and activated carbon (Norit A supra,
10 weight %) followed by with toluene-heptane crystallization
to afford 74% yield of the desired compound (Pd content
120-150 ppm). Both of these treatments were needed to
remove Pd contaminants and “gummy” byproduct that interfere
with the crystallization. Fortunately Cu was much less of an
issue with two ammonia treatments, during workup, effectively
reducing the Cu levels to <20 ppm.
Thus, purified 6 was dissolved in THF in a MultiMax
equipped with a calorimeter and an IR probe and cooled to
-20 °C followed by addition of Red-Al in toluene (Scheme
5). During addition of the first 0.5 mol equiv of the reagent,
there was vigorous gas evolution and significant exothermic
behavior due to reaction of the aluminum hydride with the free
hydroxyl group. The acetylene signals were still present in the
IR, whereas no olefinic resonance could be observed. From
these observations, we are postulating the formation of the
tetraalkoxy intermediate 5a. Upon further addition of Red-Al,
a second large and rapid exotherm occurred, followed by the
appearance of the olefinic bands of 5, albeit at a slower rate.
From the magnitude and rate of the observed exotherm and
the slower appearance of 5 we postulate initial rapid formation
of 5b, followed by a somewhat slower intramolecular hydroalu-
mination. An intermolecular delivery of the hydride cannot be
conclusively ruled out but does appear inconsistent with the
thermal and IR data and further studies are needed to establish
the course of this reaction.
Finally it is worth noting that in some cases,16 evaporation
of n-BuNH2 or propargyl alcohol were observed which neces-
sitated further addition of these reagents to effect complete
conversion.
Predictably, using the iodide 7b (Scheme 4) allowed the
reaction to proceed smoothly at 40 °C with no polymerization
and thus a more facile workup. It is worth noting, however,
that the iodide is not commercially available and would have
to be prepared from the bromide. Thus, the bromide was
selected as our starting material for further processing.
Stereospecific Synthesis of the Trisubstituted Allylic
Alcohol 3. Our strategy for effecting stereoselective formation
of 3 is based on the selective formation of a reactive aluminate
intermediate 5 which would have the dual role of defining
the geometry of the double bond, via the internal chelation,
while providing the appropriate reactivity for further reaction
(Scheme 1).
Next we investigated the reactivity and stability of 5 and in
particular its thermal stability as the cross-coupling reactions
were expected to take place above room temperature (Scheme
6). The effect of the equivalents of Red-Al was also examined
(15) SilicaBond-thiol is an efficient method for removing both Pd(II) and
Pd(0) dissolved in reaction mixtures. It is marketed by Silicycle
Corporation, and information and specification including efficiency
of removal of Pd and other transition metal catalysts can be found at
http://www.silicycle.com.
(16) The initial screening experiments were conducted in autoclave-type
vessels where evaporation of the components was not an issue.
However, when typical “bench-top” experimental setups were used,
some evaporation was observed (despite the presence of a condenser)
that necessitated further addition of amine and propargyl alcohol to
drive the reaction to completion.
(17) Corey, E. J.; Katzenellenbogen, J. A.; Gilman, N. W.; Roman, S. A.;
Erickson, B. W. J. Am. Chem. Soc. 1968, 90, 5618.
(18) The Jamison group (ref 3b) has elegantly demonstrated a comple-
mentary strategy where the resulting aluminate species is transmet-
allated to a Cu reagent and then functionalized via alkylation to afford
stereodefined trisubstituted olefins.
(19) Havra´nek, M.; Dvorˇa´k, D. J. Org. Chem. 2002, 67, 2125.
Vol. 13, No. 3, 2009 / Organic Process Research & Development
•
601

addition of ZnCl2. A screening exercise was initiated21 however,
most Pd:L combinations afforded low conversions primarily
due to the immediate catalyst decomposition as indicated by
the precipitation of Pd-black. Even the more robust Ni catalysts
failed to afford the desired product. Surprisingly, using 3-iodo
thiophene afforded significant amounts of the iodo olefin 9 in
an apparent Al-I exchange.
Scheme 5. Hydroalumination reaction: mechanism and
iodine quench
When the same screen was performed in the presence of
ZnCl2 it was found that Pd(OAc)2/P(t-Bu)3 combination afforded
90% conversion to the desired product 3 along with 10% of
the disubstituted olefin 10 (Scheme 8).
Furthermore, although the reaction initially proceeded rapidly
at ambient temperature, significant amounts of Pd black could
be seen precipitating out of the reaction within 20-30 min thus
preventing complete conversion. Moreover, multiple repetitions
of the reaction without any “apparent” change in the procedure,
afforded irreproducible results with up to 20% of 10 forming
and premature catalyst decomposition. It is worth noting that
in these reactions the vinyl aluminum or zinc species had not
decomposed as shown by quenching portions of the mixture
with I2. Only the vinyl iodo compound, 9, was formed with
only traces of olefin 10.
In order to solve the relative instability of the catalyst with
bulky electron rich phosphines, we turned our attention to the
N-heterocyclic carbene species that have been highly popular
the last several years.22 Some initial screens with generally
successful carbene23 catalysts gave no evidence of the desired
product 3 under a variety of conditions. However, in these
reactions, we did not observe any catalyst decomposition as in
the case of the phosphine-based catalysts. Interestingly, the best
catalyst to date, achieving sufficient reactivity and stability at
ambient temperature was the commercially available PEPPSI
catalyst (Scheme 9 and Table 2).24 Following this initial success,
several experiments were dedicated to the understanding of
solvent effects, the source of Zn (II) and the equivalents of Zn
needed for a successful transformation (Table 2).
From these results it was shown that quenching the excess
Red-Al with ethyl acetate is important for the success of the
reaction (entries 1, 2). Furthermore (entries 5-8) 0.5 equiv of
ZnCl2 appears to be the optimal amount, consistent with
Dvorˇa´k’s observations. Other commonly used sources of Zn
(entries 9-12) did not give satisfactory results, while metal salts
suchasMgBr2-Et2Oalsodidnoteffectthedesiredtransformation.
Further optimization brought to light several other essential
parameters: the optimal solvent for the hydroalumination/
coupling reaction is 2-methyltetrahydrofuran and polar coor-
dinating solvents such as NMP were not appropriate solvents
as excess reagent might interfere with the subsequent cross-
coupling reaction. The stability of the aluminate was determined
by quenching the mixture with I2 and using the vinyl iodide 9
as a surrogate to the aluminate 5.20
From these studies, we can draw the following conclusions:
(a) the vinyl aluminum compound, 5, is stable at -20 °C for
several hours; (b) as described by Dvorˇa´k et al., it is also stable
in the presence of EtOAc at -20 °C; (c) when the temperature
rises above 0 °C, up to 16% of olefin 10 is observed particularly
in the presence of excess (1.3 equiv) Red-Al; (d) a solvent
screen showed that, in terms of yield, 2-Me-THF > THF,
toluene > DME; (e) even without excess Red-Al, 5 decomposes
up to 20% upon heating to 50 °C; (f) diisobutyl aluminum
hydride and, surprisingly, lithium aluminum hydride are not
successful in this reaction, giving no reaction and inferior yields,
respectively.
Direct Coupling of the Vinyl Aluminum Species. Use of
the Dvorˇa´k protocol with our substrate (Scheme 7) unfortunately
provided low conversion to the desired 3 even under scrupu-
lously anhydrous and oxygen-free conditions. Thus, these
conditions, effective for electron-deficient iodides, had to be
modified to accommodate the reactivity of our substrate
3-bromothiophene (4b).
(21) The screen consisted of combinations of the following: Solvents: NMP,
THF, Dioxane, Toluene; Pd source: Pd(OAc)2, Pd2(dba)3-CHCl3;
Ligands: [OMe3Ph]3P, [OMe2Ph]3P t-BuPh2P, t-Bu3P, CTQPHOS,
DPEPHOS, dppf, P(o-tolyl)3, Degussa catCXium kit.
(22) Bourissou, D.; Guerret, O.; Gabbai, F. P.; Bertrand, G. Chem. ReV.
2000, 100, 39.
(23) Several of the NHC catalysts pioneered by Nolan (some now
commercially available) were investigated: (a) Marion, N.; Navarro,
O.; Mei, J.; Stevens, E. D.; Scott, N. M.; Nolan, S. P. J. Am. Chem.
Soc. 2006, 128, 4020.
Initially, in an attempt to boost the reactivity of our
nucleophilic component, we sought to use directly the more
reactive aluminate species (5, M ) AlOR2), rather than the
presumed, less reactive, zinc derivative resulting from the
(24) Organ, M. G.; Avola, S.; Dubovyk, I.; Hadei, N.; Assen, E.; Kantchev,
B.; O’Brien, C. J.; Valente, C. Chem. Eur. J. 2006, 12, 4749. O’Brien,
C. J.; Assen, E.; Kantchev, B.; Valente, C.; Hadei, N.; Chass, G. A.;
Lough, A.; Hopkinson, A. C.; Organ, M. G. Chem. Eur. J. 2006, 12,
4743.
(20) Control experiments had established that this transformation proceeds
in almost quantitative yield.
602
•
Vol. 13, No. 3, 2009 / Organic Process Research & Development

Scheme 6. Stability study of the vinyl aluminum intermediate
Scheme 7. Attempts to prepare 3 using the literature protocol
Scheme 8. Initial success in the synthesis of 3 using t-Bu3P as the ligand
for this reaction. The preferred temperature of the hydroalu-
In the coupling step, the optimal amount of PEPPSI catalyst
mination was -30 °C although satisfactory results were
obtained also at -20 °C. On the other hand, at 0 °C, the ratio
of 3:10 was 9-10:1 as opposed to 15:1 at -30 °C.
was determined to be ∼3 mol %. With 1% catalyst the reaction
proceeded to ∼86% conversion even after prolonged reaction
times.
Vol. 13, No. 3, 2009 / Organic Process Research & Development
•
603

Scheme 9. Efficient Coupling Using PEPPSITM to Form 3
It is worth noting that the coupling reaction is most likely
completely selective for the desired product and the amount of
10 observed was actually formed during the hydroalumination
reaction (entry 7). This was shown by quenching a portion of
the reaction with EtOAc followed by iodine. There, the ratio
of 9 to 10 was 94:6, similar to the amount of 10 observed during
the preparation of 3. Finally, the allylic alcohol 3 was shown
to be sensitive to acid workup, and thus a basic workup (K, Na
tartrate) is needed for the removal of the aluminum salts to avoid
decomposition of the product.
amine in CH2Cl2 cooled to 0 °C, and MsCl was added dropwise.
After aqueous workup and addition of oxalic acid, the desired
compound 11 was isolated in 56% yield as the oxalate salt.
The DIEA adduct 12 (∼25%) was removed during the aqueous
workup. Hydrolysis of 11 afforded the API, 1, in almost
quantitative yield. The Pd content of the API was <10 ppm,
thus validating our initial synthetic strategy.
Filtration of the reaction mixture through a silica pad or
treatment with SilicaBond-Thiol followed by EtOH-H2O
crystallization succeeded in achieving low metal content in the
final intermediate. Thus Pd, Cu, and Al levels were <50, 10,
and 3 ppm, respectively, in the isolated 3 and were <10 ppm
in the final active pharmaceutical ingredient (API).
Final Coupling. The transformation of 3 to 11 was
significantly more challenging than first anticipated due to the
reactivity of the corresponding electron-rich allylic system and
the reduced nucleophilicity of the sarcosine nitrogen. The
preliminary results are shown below. Further work will be
needed to optimize the final steps of synthesis.
Initial attempts to repeat established procedures, from a
previous route, produced only low yields of the desired product
11. The remainder of the mass balance was identified as the
impurities shown in Figure 1.
Although work in this area continues, we have been able to
produce the desired 1 with a low-temperature reaction (eq 4)
Figure 1
Experimental Section
All experiments were carried out using standard glassware,
and the used reagents and solvents were standard technical grade
reagents and solvents used without any purification. The
phosphine ligands used in the screening experiments were
purchased from Acros, Aldrich, or Strem and were used without
any purification. HPLC analyses were performed on reversed-
phase columns (Waters XTerra RP 3.5 µm, 4.6 mm × 50 mm)
in gradient mode (ammonium acetate buffer to acetonitrile) with
UV detection (254 nm). GLC analyses were performed on
Hewlett-Packard 5890 series II (capillary column HP 5, 10 m)
in temperature gradient mode (50 to 300 °C, rate 10 °C/min.)
with FID detection. High-resolution mass spectra were per-
formed on Jeol-JMS-T100LP (DART ionization). NMR spectra
were recorded on a Bruker AV 400 and 360 MHz instruments.
Palladium determinations were performed by AAS on a Varian
SpectrAA-880.
Synthesis of 2-(4-Bromophenyl)furan (7a). A 10 L vessel
under N2 was charged with 4-bromo(iodo)benzene (301 g; 1.06
mol), 2-furanboronic acid (148.81 g; 1.25 mol), triphenylphos-
phine (13.95 g; 0.05 mol), tetra-n-butylammonium bromide
by managing the order of addition of the reagents. Thus, 3 was
mixed with sarcosine methyl ester-HCl and diisopropyl ethyl
604
•
Vol. 13, No. 3, 2009 / Organic Process Research & Development

Table 2. Evaluation of reaction parameters in the direct coupling reaction of aluminate 5 with 3-bromothiophene
entry
quench with EtOAca
Zn source
Zn (mol %)
solvent
ratio of 3:10 (yield)
1
no
ZnCl2 (0.5 M THF)
ZnCl2 (0.5 M THF)
ZnCl2 (0.5 M THF)
ZnCl2 (0.5 M THF)
ZnCl2 (0.5 M THF)
ZnCl2 (0.5 M THF)
ZnCl2 (0.5 M THF)
ZnCl2 (0.5 M THF)
ZnCl2 (0.5 M THF)
ZnCl2-TMEDA
30
30
100
0
30
50
50
60
70
100
50
50
THF
THF
THF
THF
THF
THF
80:20
2
no
66:34
3
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
74:26
4
50:50
5
88:12
6
92:8
7
Me-THF
94:6 (75%)
88:12
8
THF
THF
DME-NMP (1:1)
DME-NMP (1:1)
THF
9
85:15
10
11
12
64:36 20% E-isomer
63:37 20% E-isomer
no reaction
ZnCl2-TMEDA
Zn(OAc)2
a yes: the reaction was first quenched with ethyl acetate to decompose excess RedAl before ZnCl2 was added. no: ZnCl2 was added directly to the reaction mixture.
(377.29 g; 1.1 mol), sodium carbonate (225.53 g; 2 mol),
palladium on carbon 10% (60.20 g; 56.57 mmol), tetrahydro-
furan (3.72 L), and water (3.72 L). The reaction mixture was
heated at 60 °C for 20 h (until gas chromatography showed no
unreacted starting material). After cooling to room temperature,
the reaction mixture was filtered over decalite. The organic layer
was separated and washed with a saturated NaCl solution
(2 L), concentrated under reduced pressure, and exchanged with
ethanol to a final volume of 2.27 L. The product was crystallized
by slow addition of water at ambient temperature to afford 266 g
of 7a (94% yield; Pd content <10 ppm).
130.83 (1 C), 123.59 (2 C), 121.32 (1 C), 111.76 (1 C), 105.93
(1 C), 88.08 (1 C), 85.69 (1 C), 51.57 (1 C). M+ found )
198.0687; calcd ) 198.0681.
Synthesis of (Z)-3-[4-(2-Furyl)phenyl]-3-(3-thienyl)-2-pro-
pen-1-ol (3) (Method 2: Direct Coupling Using the Peppsi
Catalyst). A round-bottom flask was charged with intermediate
6 (15 g, 75 mmol) and degassed 2-methyl-tetrahydrofuran (90
mL) and cooled to 0 °C. Sodium bis(2-methoxyethoxy)alumi-
num hydride (19.8 g, 99 mmol) was added, and the mixture
was stirred for 15 min. Ethyl acetate (270 mL) was added to
quench the reaction, and 3-bromothiophene was added (neat,
16 g, 99 mmol). The reaction vessel was removed from the
cooling bath and treated with zinc dichloride (72 g, 39 mmol)
and [1,3-bis(2,6-diisopropylphenyl)imidazol-2-ylidene](3-chlo-
ropyridyl palladium(II) dichloride (PEPPSI) under an Ar
blanket. The reaction mixture was quenched after 3 h at room
temperature by addition of saturated aqueous K, Na, tartrate
(100 mL) and extracted with 2-methyltetrahydrofuran (300 mL).
The organic layer was washed with water, filtered, and
concentrated in vacuo. The residue was dissolved in dichlo-
romethane and filtered through a pad of silica gel (90 g). The
fractions containing the reaction product were collected and
concentrated, yielding crude (Z)-3-(4-(2-furyl)phenyl)-3-(3-
thienyl)-prop-2-en-1-ol (17.5 g). The product can be further
purified by dissolution in ethanol (30 mL), and precipitation
by addition of water (24 mL). The precipitate was washed with
ethanol-water (1:1, 15 mL) and dried: Yield of 3: 15.4 g (Pd
content < 50 ppm).
Synthesis of 3-[4-(2-Furyl)phenyl]-2-propyn-1-ol (6). A
500 mL flask equipped with a mechanical stirrer was charged
with N-methylpyrrolidone (100 mL) and degassed by bubbling
nitrogen at 60 °C for 15 min. The solvent was cooled to ambient
temperature, and the reagents were added in the following order:
2-(4-bromophenyl)furan 7a (22.31 g; 100 mmol), copper(I)
iodide (761.8 mg, 4 mmol), 1-butanamine (11.88 mL, 120
mmol), and tetrakis-(triphenylphosphine) palladium (2.31 g, 2
mmol). The reaction mixture was heated to 50 °C, and a solution
of 2-propyn-1-ol (6.99 mL, 120 mmol) in N-methylpyrrolidone
(10 mL) was added slowly over 1 h (syringe pump: 17 mL/h).
The reaction mixture was transferred to a separation funnel and
treated with water (400 mL) and isopropyl acetate (IPAc) (100
mL). The water layer was extracted a second time with
isopropyl acetate (100 mL). The combined organic layers were
washed twice with an ammonia solution (20 mL of saturated
ammonia diluted in 100 mL of water) until colorless. The
organic layer was concentrated in vacuo to afford 16 g of crude
product which can be further purified by chromatography
through a pad of silica gel (Silica:6 ) 3:1, elution with toluene).
Alternatively, the IPAc solution was azeotropically dried (KF
∼200 ppm) and the mixture treated with SilicaBond-Thiol (10
g/mmol of Pd) and activated carbon (Norit A supra, 10 wt %)
The solvent was switched to toluene (5 mL/g) and heptane was
added (5 mL/g), to afford 74% yield of the desired compound
(silica gel; dichloromethane) or crystallization (1:3 toluene-
heptane). Yield 6: 14.6 g (74%; Pd content 120-150 ppm; Cu
< 20 ppm).
1H NMR (600 MHz, CDCl3): δ 7.57 (dt, J ) 8.40, 2.03 Hz,
2 H), 7.44 (d, J ) 1.51 Hz, 1 H), 7.30 (dd, J ) 4.91, 3.02 Hz,
1 H), 7.27 (dt, J ) 8.69, 1.89 Hz, 2 H), 7.16 (dd, J ) 2.83,
1.32 Hz, 1 H), 6.88 (dd, J ) 4.91, 1.13 Hz, 1 H), 6.62 (d, J )
3.02 Hz, 1 H), 6.45 (dd, J ) 3.21, 1.70 Hz, 1 H), 6.20 (t, J )
6.80 Hz, 1 H), 4.29 (d, J ) 6.80 Hz, 2 H), 1.94 (br s, 1 H); 13C
NMR (151 MHz, CDCl3): δ 153.60 (1 C), 142.12 (1 C), 140.39
(1 C), 138.88 (1 C), 138.56 (1 C), 130.11 (1 C), 129.03 (1 C),
127.97 (1 C), 127.76 (2 C), 125.26 (1 C), 124.71 (1 C), 123.50
(2 C), 111.67 (1 C), 105.18 (1 C), 60.52 (1 C). MH+ found )
283.0783; calcd ) 283.0793.
Synthesis of Methyl N-[(2Z)-3-[4-(2-Furyl)phenyl]-3-(3-
thienyl)-2-propen-1-yl]-N-methylglycinate Compound with
Oxalic Acid (1:1) (11). Allylic alcohol 3 (35 g, 124 mmol) in
dichloromethane (372 mL) was cooled to 0 °C and treated with
methanesulfonyl chloride (10.1 mL, 130 mmol). Diisopropyl-
1H NMR (400 MHz, CDCl3): δ 7.57 (dt, J ) 8.37, 1.73 Hz,
2 H), 7.44 (d, J ) 1.51 Hz, 1 H), 7.42 (ddd, J ) 8.31, 1.51,
1.26 Hz, 2 H), 6.63 (d, J ) 3.27 Hz, 1 H), 6.44 (dd, J ) 3.27,
1.76 Hz, 1 H), 4.48 (s, 2 H), 2.00 (s, 1 H); 13C NMR (101
MHz, CDCl3): δ 153.36 (1 C), 142.49 (1 C), 132.03 (2 C),
Vol. 13, No. 3, 2009 / Organic Process Research & Development
•
605

ethylamine (23.8 mL, 136.3 mmol) was added dropwise so that
the temperature did not exceed 5 °C. The mixture was stirred
for 30 min at 0 °C. Solid sarcosine methyl ester hydrochloride
salt (19.8 g, 136 mmol) was added in one portion followed by
diisopropylethylamine (48 mL, 273 mmol), maintaining the
temperature between 0 and 4 °C. The mixture was allowed to
stir overnight at this temperature, and then it was quenched with
water. The organic layer was washed twice with water, and
during the final wash the pH was adjusted to ∼10. The organic
layer was separated, the solvent was evaporated, and the residue
was dissolved in 270 mL of ethanol and heated to 60 °C. Oxalic
acid (14.8 g) was added in 180 mL of ethanol, and the mixture
was stirred until a precipitate was observed. The latter was
stirred overnight, filtered, and dried to afford the oxalate salt
was separated, and the organic was concentrated in vacuo. The
residue was dissolved in methanol (160 mL) and treated with
NaOH (3 mL, 50% aqueous solution). The mixture was warmed
to 60 °C and stirred for 4 h. Upon completion, the mixture was
treated with water (120 mL), and formic acid was added (19.6
g, 219 mmol) at 60 °C until precipitation of 1 began after 30
min; the slurry was stirred for 2 h, cooled to room temperature,
and stirred overnight. The solid was filtered and washed with
methanol-water (100 mL, 1:1 ratio). The filtrate was dried to
afford 1. Yield of 1: 14.7 g, 95% 1H NMR (400 MHz, DMSO-
d6): δ 2.41 (s, 3 H) 3.21 (s, 2 H) 3.42 (d, J ) 6.77 Hz, 2 H)
3.46 (br s, 1 H) 6.19 (t, J ) 6.67 Hz, 1 H) 6.60 (dd, J ) 3.43,
1.79 Hz, 1 H) 6.91 (dd, J ) 4.89, 1.24 Hz, 1 H) 6.94 (dd, J )
3.43, 0.46 Hz, 1 H) 7.29 (d, J ) 8.53 Hz, 2 H) 7.44 (dd, J )
2.93, 1.24 Hz, 1 H) 7.62 (dd, J ) 4.89, 2.93 Hz, 1 H) 7.66 (d,
J ) 8.53 Hz, 2 H) 7.75 (dd, J ) 1.79, 0.46 Hz, 1 H) MH+
found ) 354.1171; calcd ) 354.1164. Pd content < 10 ppm.
1
11. Yield of 11: 31.7 g (56%). H NMR (360 MHz, DMSO-
d6): δ 2.50 (s, 3 H) 3.53 (d, J ) 6.70 Hz, 2 H) 3.63-3.64 (m,
2 H) 3.63 (s, 3 H) 5.95 (br s, 2 H) 6.21 (t, J ) 6.70 Hz, 1 H)
6.60 (dd, J ) 3.36, 1.85 Hz, 1 H) 6.91 (dd, J ) 4.89, 1.10 Hz,
1 H) 6.96 (d, J ) 3.36 Hz, 1 H) 7.31 (d, J ) 8.46 Hz, 2 H)
7.45 (dd, J ) 2.90, 1.10 Hz, 1 H) 7.63-7.66 (m, J ) 4.89,
2.90 Hz, 1 H) 7.67 (d, J ) 8.46 Hz, 2 H) 7.76 (d, J ) 1.85 Hz,
1 H). MNa+ found ) 390.1135; calcd ) 390.1140.
Synthesis of N-[(2Z)-3-[4-(2-Furyl)phenyl]-3-(3-thienyl)-
2-propen-1-yl]-N-methylglycine (1). The oxalate salt 11 (20.2
g, 43.8 mmol) was dissolved in MTBE (160 mL) and treated
dropwise with a solution of K2CO3 (saturated) in water (160
mL). Stir for 1 h at ambient temperature. The aqueous layer
Supporting Information Available
Spectra of the key compounds described herein. This
material is available free of charge via the Internet at
http://pubs.acs.org.
Received for review January 24, 2009.
OP900015G
606
•
Vol. 13, No. 3, 2009 / Organic Process Research & Development