A Trialkylphosphine-Derived Palladacycle as a Catalyst in the Selective Cross-Dimerization of Terminal Arylacetylenes with Terminal Propargyl Alcohols and Amides

Article abstract of DOI:10.1021/acscatal.6b01541

A method for the selective cross-dimerization of terminal aryl alkynes with propargyl alcohols to afford linear (E)-enynol products is reported. The complex [Pd(μ-κ²-O,O-OAc)(κ²-C,P-(t-Bu)₂PCH₂C(Me)₂CH₂)]₂ selectively affords (E)-5-aryl-2-en-4-yn-1-ol products in good yields under mild conditions with high chemo-, regio-, and stereoselectivity. In contrast, previously reported examples of this reaction afford the branched 4-aryl-2-hydroxymethanol-1-buten-3-yne. Propargyl amides are also selectively cross-dimerized, but with lower regioselectivity for the linear enyne. The method has been applied to the synthesis of (E)-5-phenyl-2-penten-4-yn-1ol, which is a precursor to type 2 diabetes drug candidate NNC 61-4655, in 72% yield from phenylacetylene and propargyl alcohol. The palladacycle precatalyst reacts with aryl alkynes to afford the first example of a dimeric palladacycle complex with a μ-κ²-C¹,C¹-bound acetylide ligand. This complex is observed during the catalytic reaction and is a competent precatalyst.

Full text of DOI:10.1021/acscatal.6b01541

Research Article
pubs.acs.org/acscatalysis
A Trialkylphosphine-Derived Palladacycle as a Catalyst in the
Selective Cross-Dimerization of Terminal Arylacetylenes with
Terminal Propargyl Alcohols and Amides
Matthew G. Lauer,† Benjamin R. Headford,† Olivia M. Gobble,† Michelle B. Weyhaupt,†
Deidra L. Gerlach,† Matthias Zeller,‡,§ and Kevin H. Shaughnessy*,†
†Department of Chemistry, The University of Alabama, Box 870336, Tuscaloosa, Alabama 35487, United States
‡Department of Chemistry, Youngstown State University, One University Plaza, Youngstown, Ohio 44555, United States
S
* Supporting Information
ABSTRACT: A method for the selective cross-dimerization of
terminal aryl alkynes with propargyl alcohols to afford linear (E)-
enynol products is reported. The complex [Pd(μ-κ2-O,O-
OAc)(κ2-C,P-(t-Bu)2PCH2C(Me)2CH2)]2 selectively affords
(E)-5-aryl-2-en-4-yn-1-ol products in good yields under mild
conditions with high chemo-, regio-, and stereoselectivity. In
contrast, previously reported examples of this reaction afford the
branched 4-aryl-2-hydroxymethanol-1-buten-3-yne. Propargyl
amides are also selectively cross-dimerized, but with lower
regioselectivity for the linear enyne. The method has been
applied to the synthesis of (E)-5-phenyl-2-penten-4-yn-1ol,
which is a precursor to type 2 diabetes drug candidate NNC 61-4655, in 72% yield from phenylacetylene and propargyl
alcohol. The palladacycle precatalyst reacts with aryl alkynes to afford the first example of a dimeric palladacycle complex with a
μ-κ2-C1,C1-bound acetylide ligand. This complex is observed during the catalytic reaction and is a competent precatalyst.
KEYWORDS: palladacycle, alkyne dimerization, hydroalkynylation, enyne, μ-acetylide complex
precursors in a variety of synthetic methods.4 Conjugated linear
INTRODUCTION
■
enynes are commonly synthesized via Sonogashira reactions of
Linear trans-enynes are an important motif found in a number
vinyl halides,5 olefination of conjugated ynals,3,4,6 or the
of biologically active compounds and natural products,
elimination of propargyl alcohols.7
including oxamflatin (1),1 bioactive constituents isolated from
Asparagus Cochinchinensis (2 and 3),2 and diabetes treatment
The selective catalytic dimerization of terminal alkynes is an
efficient method for the formation of enyne structures.8
candidate NNC 61-4655 (4) (Chart 1).3 Notably, each of these
Homodimerization of alkynes to form enynes can result in
examples contains an oxygen functionality in the allylic
position. 2-En-4-yn-1-ols and related compounds are useful
the formation of either linear (E/Z-5) or branched (6) enynes
(Scheme 1) depending on the choice of catalyst. Selectively
producing a single product becomes even more challenging
when coupling two different alkynes. In this case, the catalyst
must selectively produce a single product from 12 possible
stereoisomers and constitutional isomers through precise
control of chemoselectivity, regioselectivity, and stereoselectiv-
ity.
Chart 1. Biologically Active Compounds Containing a
Linear Enyne Motif
Although a number of methods have been published for the
homodimerization of alkynes to give either linear (5) or
branched (6) enyne products, far fewer methods are known for
the cross-dimerization of two different terminal alkynes. Silyl
alkynes are selectively coupled with alkyl and aryl alkynes to
give E or Z head-to-head enynes (E/Z-7, R1 = R3Si)9 or head-
to-tail enynes (8, R1 = R3Si)10 depending on the choice of
Received: June 1, 2016
Revised: July 19, 2016
© XXXX American Chemical Society
5834
DOI: 10.1021/acscatal.6b01541
ACS Catal. 2016, 6, 5834−5842

ACS Catalysis
Research Article
Scheme 1. Possible Enyne Isomers in Alkyne Dimerization
Reactions
Scheme 2. Palladacycles Derived from Bulky Electron-Rich
Phosphines
catalyst. Oshovsky and de Bruin reported a Ti(III) metallocene
catalyst for the selective addition of aryl alkynes across alkyl
alkynes to give head-to-tail enyne products (8, R1 = aryl, R2 =
alkyl).11 Propargyl alcohols and amines provide high selectivity
in cross-dimerization reactions to give type 8 enynes, where R2
is the − CH2X group (X = OH, NHR), using both Ti(III)11
and late transition metals.12 There are no examples of cross-
dimerization of alkynes with propargyl alcohols or amines to
afford linear type 7 enynes relevant to the structures in Chart 1.
Herein, we report the first example of a catalyst that cross-
dimerizes aryl acetylenes with propargyl alcohols or amides to
afford linear 2-en-4-yn-1-ols and 2-en-4-yn-1-yl amides (E-7)
with high selectivity under mild conditions.
Scheme 3. Proposed Solution Equilibrium of Complex 10
RESULTS
■
DTBNpP-Derived Palladacycle. During our investigations
of the use of neopentyl phosphine ligands in Heck couplings an
unknown broad peak in the phosphorus NMR spectra at 94
ppm was often observed in the reaction mixture when di-tert-
butylneopentylphosphine (DTBNpP) was used as a ligand.13
We hypothesized that this species could be a cyclometalated
complex on the basis of the known ability of neopentylphos-
phines to form metallacyclic compounds.14 Platinum15 and
iridium16 metallacycles derived from neopentylphosphines have
been reported. Surprising, no neopentylphosphine-derived
palladacycles have been reported. In contrast, the palladacycle
derived from tri-tert-butylphosphine (TTBP) is known.17
The DTBNpP-derived palladacycle 10 (Scheme 2) was
prepared using conditions reported by Stambuli and co-
workers17a for palladacycle 9 derived from TTBP and
Pd(OAc)2 (Scheme 2). Mixing Pd(OAc)2, with 1 equiv of
DTBNpP in THF at ambient temperature produced the
cyclometalated dimer 10 in 94% yield as an air-stable material
giving a broad 31P NMR resonance at 94 ppm. The fluxional
nature of acetate-bridged phosphapalladacycles has been
previously noted.18 At low temperature, the broad feature at
94 ppm resolves into three resonances at 98, 94.5, and 90 ppm
(Figure S1 in the Supporting Information). These species likely
correspond to cis and trans isomers of the dimer plus a
monomeric palladacycle with a κ2-O,O-bound acetate (Scheme
3). Since the potential cyclometalated complex formed in the
Heck couplings of aryl bromides would have bromide as ligands
rather than acetate, the bromide dimer 11 was prepared.
Treatment of crude product 10 with 20 equiv of lithium
bromide in acetone afforded dimer 11 in 99% yield in a one-pot
reaction. The peak in the 31P NMR spectrum for 11 matched
that observed in the Heck coupling reaction mixtures.
Despite the fact that 9 has found limited use as an active
catalyst in organic synthesis,17b,19 palladacycles are widely used
as both precatalysts and active catalysts in organic synthesis.20
Due to the ease of synthesis and stability of complex 10, uses in
palladium(II) catalysis were explored. A 2006 report by
Tenaglia and co-workers on the Herrmann−Beller palladacycle
(12) catalyzed addition of alkynes to norbornadienes led us to
test complex 10 in similar reactions (eq 1).21 Similarly to the
Herrmann−Beller palladacycle, complex 10 was able to catalyze
the addition of phenylacetylene to norbornadiene. Attempts to
couple phenylacetylene with other alkenes such as styrene,
butyl acrylate, and norbornene resulted in dimerization of
phenylacetylene, however. Inspired by recent reports of the
selective coupling of aryl acetylenes with propargyl alcohols or
amines,11,12 we examined the coupling of phenylacetylene with
propargyl alcohol catalyzed by 10.
Coupling of Acetylenes with Propargyl Alcohols. The
reaction of phenylacetylene with 2.5 equiv of propargyl alcohol
5835
DOI: 10.1021/acscatal.6b01541
ACS Catal. 2016, 6, 5834−5842

ACS Catalysis
Research Article
in acetonitrile with 2 mol % of 10 led to a 60% yield of linear
enyne product 7a (eq 2) and a 63% yield of propargyl alcohol
palladacycle precatalyst played a key role in determining the
regioselectivity of the coupling reaction. DTBNpP in
combination with Pd(OAc)2 also gave linear enyne product
7a as the major product, but with lower selectivity and yield
than for palladacycle 10 (entry 6). In the case of TTBP, the
yield and product distribution of the in situ catalyst (entry 7)
were nearly identical with those of complex 9. This result may
reflect the lower stability of palladacycle 9, resulting in the κ1-
DTBNpP complex being the active species when 9 is used.. Tri-
o-tolylphosphine (P(o-tol)3) in combination with Pd(OAc)2
gave 8a as the major product, as for the Herrmann−Beller
palladacycle (12), but with higher selectivity and yield (entry
8). Trineopentylphosphine (TNpP) also gave 8a as the major
product (entry 9). Tris(2,6-dimethoxyphenyl)phosphine
(TDMPP), which Trost and co-workers12b,22 used in many of
their pioneering alkyne dimerization studies, gave only trace
amounts of product under the conditions optimized for
palladacycle 10 (entry 10) but exclusively formed branched
isomer 8a. Palladium(II) acetate without added ligand failed to
give any coupled products (entry 11).
Under the optimized conditions, enyne 7a was isolated in a
72% yield (Table 2). The linear propargyl alcohol dimer side
product 12 can be easily separated via column chromatography
due to its higher polarity. It can also be removed via aqueous
extraction prior to column chromatography. Therefore,
formation of 5a does not impede isolation of product 7a.
Enyne 7a is an intermediate in Novo Nordisk’s GMP-kilo-
Table 1. Ligand Effect on Selectivity of Alkyne
Dimerization
a
b
yield (%)
entry
Pd/L
7a
8a
5a
6a
5b
c
1
10
10
11
9
56
74
0
2
2
63
150
0
5
9
8
4
0
8
0
4
8
0
0
0
0
Table 2. Dimerization of Arylacetylenes with Propargyl
Alcohol
2
3
0
0
4
36
10
44
35
2
10
18
18
11
36
51
2
76
16
118
78
4
20
49
18
18
77
88
10
0
5
12
d
6
DTBNpP
d
7
TTBP
d
8
P(o-tol)3
d
9
TNpP
17
0
32
0
d
10
11
TDMPP
d
no L
0
0
0
a
Reaction run using conditions in eq 1 unless noted otherwise.
Product yields determined via 1H NMR spectroscopy with an internal
b
standard. All percentages are based on moles of phenylacetylene used.
c
d
2.5 equiv of propargyl alcohol used. Ligand (4 mol %) used in
combination with Pd(OAc)2 (4 mol %).
dimer 5a (Table 1, entry 1). Notably, the phenylacetylene
homodimer 5b and the branched enyne products 6a and 8a
were produced in small amounts. An extensive solvent screen
showed that nitrile solvents were most effective for this
reaction, with acetonitrile providing the highest yield of enyne
7a (Table S1 in the Supporting Information). Increasing the
amount of propargyl alcohol to 5 equiv improved the yield of
7a to 74% and resulted in minimal phenylacetylene
homodimerization products (5b, entry 2). Using a larger
excess of propargyl alcohol results in a larger amount of
homodimer 5a being produced (2:1 5a:7a), however.
Palladacycle dimer 11 with bridging bromides failed to
catalyze the reaction, which shows the key role of the basic
acetate ligands (Table 1, entry 3). The TTBP-derived
palladacycle dimer 9 gave 7a as the major product, but with
lower selectivity and yield (entry 4). The Herrmann−Beller
palladacycle 12 gave the branched enyne 8a as the major
product, but in low yield and selectivity (entry 5). The
a
Reaction performed at 80 °C.
5836
DOI: 10.1021/acscatal.6b01541
ACS Catal. 2016, 6, 5834−5842

ACS Catalysis
Research Article
Table 3. Optimization of Alkyne Dimerization with 2.5 Equiv of Propargyl Alcohol
a
yield (%)
entry
additive (equiv)
atm
7a
8a
5a
5b
6a
1
2
N2
air
N2
N2
N2
N2
air
N2
air
N2
air
N2
air
O2
56
55
0
2
0
63
55
0
8
8
5
0
b
3
AcOH (1)
0
0
0
b
4
b
K2CO3 (1)
35
2
0
54
1
20
0
0
5
CuI (0.04)
0
0
6
H2O (1)
55
56
57
60
55
54
58
0
54
58
61
61
58
52
37
64
47
6
1
7
H2O (1)
0
6
0
8
EtOH (1)
10
0
3
10
0
9
EtOH (1)
5
10
11
12
13
14
1,3-propanediol (1)
1,3-propanediol (1)
trimethylolethane (1)
trimethylolethane (1)
trimethylolethane (1)
1
8
7
0
6
0
1
4
6
75 (72)
1
0
0
67
1
0
0
a
b
Yields of products determined via 1H NMR analysis with an internal standard. Yields in parentheses are of isolated product. Reaction conducted
with 5 equiv of propargyl alcohol.
laboratory synthesis of the PPAR-agonist NNC 61-4655 (4).3
Novo Nordisk produced enyne 7a in four steps with an overall
yield of 19%. The synthesis involved conversion of phenyl-
acetylene to 3-phenylpropiolaldehyde followed by Wittig−
Horner olefination and reduction of the resulting enynoate
ester. The direct coupling of phenylacetylene and propargyl
alcohol produces the same product in one step and in much
higher yield. The reaction uses inexpensive starting materials
and occurs with high atom economy.
The cross-dimerization of aryl acetylenes and propargyl
alcohol tolerates a variety of aryl acetylene coupling partners
(Table 2). Enyne 7b derived from electron-rich (4-
methoxyphenyl)acetylene was isolated in 61% yield. 2-
Ethynylacetanilide gave a lower yield of enyne 7c, presumably
due to the steric demand of the substrate. Enynes 7d,e derived
from electron-poor aryl acetylenes were produced in lower
yields primarily due to lower selectivity for the aryl acetylene to
act as the donor acetylene, which resulted in formation of the
homodimer of the aryl acetylene. Enyne 7d was isolated as a
mixture along with the isomer 7d′, in which propargyl alcohol
acts as the donor acetylene. The more sterically hindered 2-
ethynyl-1,3,5-trimethylbenzene gave only 45% conversion after
24 h at 60 °C. 2-Ethynyl-1,3,5-trimethylbenzene does not self-
dimerize under the reaction conditions, which has been
previously observed with aryl acetylenes containing two ortho
substituents.23 Increasing the temperature to 80 °C resulted in
full consumption of the aryl acetylene, and enyne 7f was
isolated in 44% yield. The heterocyclic aryl acetylenes 3-
ethynylthiophene and 5-ethynyl-1H-indole gave good yields of
the linear enyne products 7g,h. However, 2-ethynylpyridine
failed to react under the optimized reaction conditions.
propargyl alcohols (Table 3). Since we hypothesized that
deprotonation of phenylacetylene by acetate plays a key role in
the reaction, the pH of the reaction was modified by addition of
acetic acid or potassium carbonate. Both additives greatly
hindered the reaction, resulting in low yields (entries 3 and 4).
Copper(I) salts often play key roles in the reaction of alkynes,
but surprisingly a catalytic amount of copper(I) iodide shut
down the reaction (entry 5). The addition of water, ethanol,
1,3-propandiol, or trimethylolethane had very little effect on the
reaction (entries 6, 8, 10, and 12) on running under inert
conditions. Interestingly, reactions performed in air (entries 2,
7, 9, 11, and 13) showed improved regioselectivity for linear
products 7a and 8a, with no branched products being observed
(6a, 8a). With trimethylolethane as an additive in the presence
of air, the yield of 7a increased to 75% with only 2.5 equiv of
the propargyl alcohol (entry 13) with no other byproducts
other than 5a. Running the reaction under 1 atm of oxygen
increased the yield relative to a nitrogen atmosphere but
resulted in a lower yield in comparison to that in the presence
of air (entry 14).
Under the newly optimized conditions, enyne 7a could be
isolated in a 72% yield (Table 3). Propargyl alcohols with an
sp2-hybridized substituent on C1 gave good yields of enyne
products (Table 4). Reacting phenylacetylene with 1-(4-
methoxyphenyl)-2-propyn-1-ol afforded 7i in 73% yield.
NMR analysis of the crude product showed 22% of the
homodimer of the propargyl alcohol and no other enyne
byproducts. The improved selectivity for cross-coupling over
dimerization of the propargyl alcohol may be due to the
increased steric hindrance of the propargyl aryl group. Other 1-
aryl-substituted propargyl alcohols gave the resulting enyne
products 7j−l in good yields (61−73%). The presence of
electron-donating or -accepting substituents on the propargyl
Further optimization was desired in order to decrease the
amounts of noncommercially available or more expensive
5837
DOI: 10.1021/acscatal.6b01541
ACS Catal. 2016, 6, 5834−5842

ACS Catalysis
Research Article
products (<1%) were observed using only 2.5 equiv of N-
propargyl phthalimide. N-Propargylphthalimide dimers 5c
(120%, relative to phenylacetylene) and 6c (60%) were
produced in significant amounts with a similarly low
regioselectivity observed for the cross-coupling reaction.
Again, the more polar homocoupling products could be easily
separated by column chromatography. The reason for the
decreased regioselectivity for the N-propargyl amide in
comparison to that for propargyl alcohol is unclear.
Table 4. Dimerization of Arylacetylenes with Substituted
Propargyl Alcohols
Propargyl amides reacted effectively with a range of aryl
acetylenes to afford generally high yields of mixtures of linear
and branched products (Table 5). Products 7o and 8o derived
from phenylacetylene were isolated in 82% yield as a 77:23
mixture favoring 7o. N-Propargylphthalimide produced good
yields of the linear enyne products with electron-rich (7p) and
electron-poor (7q) aryl acetylenes as well as 3-ethynylth-
Table 5. Dimerization of Arylacetylenes with Propargyl
Amines
a
n.d. = not determined.
alcohol had little effect on the isolated yield. 1-Phenyl-1,4-
heptadien-6-yn-3-ol gave dienyne 7m in 70% yield. An aliphatic
substituent on the propargyl carbon resulted in low conversion
1
to 7n with multiple isomeric products observed by H NMR
analysis.
Coupling of Alkynes with Propargyl Amides. Next,
propargylamines were examined as coupling partners with
arylacetylenes. Coupling of N-propargylphthalimide with
phenylacetylene produces a mixture of linear enyne 7o (70%)
and branched enyne 8o (18%) as the cross-coupling products
at 60 °C (eq 3). Little to no phenylacetylene dimerization
a
Value in parentheses is the ratio of regioisomers 7 and 8 determined
1
by H NMR spectroscopy.
5838
DOI: 10.1021/acscatal.6b01541
ACS Catal. 2016, 6, 5834−5842

ACS Catalysis
Research Article
iophene (7r). Regioselectivities for these products ranged from
3:1 to 4:1 of 7:8. Enyne 7s containing a propargylacetamide
was produced with slightly higher regioselectivity in compar-
ison to the propargyl phthalimide substrates. The use of tosyl-
and benzyl-protected amines was also explored but resulted in
unsatisfactory selectivities or yields.
Mechanistic Studies. Monitoring the reaction of phenyl-
acetylene with propargyl alcohol via 31P NMR spectroscopy
shows two major peaks after 3 h, a broad peak at 94.8 ppm
(63%) and a sharper peak at 82.1 ppm (13%) ppm, along with
minor peaks, including a broad feature around 81 ppm (18%)
and small sharp peaks at 87.1 ppm (4%) and 44.2 ppm (2%).
The broad peak at 94.8 ppm corresponds to 10, and the other
peaks above 80 ppm are likely palladacyclic species. The peak at
44 ppm corresponds to Pd0(DTBNpP)2. We hypothesized that
one of the peaks in the 80 ppm region might be an acetylide
complex derived from 10. This complex was independently
synthesized by mixing 10 with 2 equiv of phenylacetylene in
acetonitrile for 5 min (eq 4). Cooling the reaction mixture to 0
Figure 1. ORTEP diagram of one molecular unit of complex 14
(ellipsoids at 50% probability). Hydrogen atoms have been omitted for
clarity. Selected bond distances (Å) and angles (deg): Pd1−P1,
2.2818(8); Pd1−C1, 2.039(2); Pd1−Pd2, 2.9093(9); Pd2−P2,
2.2358(9); Pd2−C16, 2.039(2); Pd1−C29, 2.071(2); Pd2−C29,
2.159(2); C29−C30, 1.211(3); P1−Pd1−C1, 82.12(6); P2−Pd2−
C16, 82.00(6); Pd1−C29−C30, 155.3(2); Pd2−C29−C30, 117.5(2);
C29−C30−C31, 178.4(2). Distances for the other molecule in the
asymmetric unit can be found in the Supporting Information.
In solution, complex 14 appears symmetrical on the basis of
31P, H, and 13C NMR data. A single set of κ2-P,C-DTBNpP
1
°C and filtering the resulting yellow solid allowed for isolation
and characterization of acetylide complex 13. 2-Ethynyl-1,3,5-
trimethylbenzene also reacts with 10 to form the mixed dimer
14. Characterization of complexes 13 and 14 showed the
presence of acetylide, acetate, and κ2-P,C-DTBNpP ligands in a
1:1:2 ratio, suggesting a dipalladacyclic complex.
Complex 14 afforded X-ray-quality crystals that allowed the
substitution pattern around the palladium center to be
definitively determined. Complex 14 (Figure 1) crystallized in
the monoclinic space group P21/c with two molecular units in
the asymmetric unit. The two molecules in the asymmetric unit
have similar structural parameters. Data for one of the
molecular structures are discussed here. A comparison of the
two molecular structures can be found in Table S3 in the
Supporting Information.
1
ligand resonances is observed in the 31P, H, and 13C NMR
spectra. In the 13C NMR spectrum, the carbons of the acetylide
moiety both appear as triplets due to coupling with the
symmetrical phosphorus centers (3JC−P = 77.6 Hz, 4JC−P = 13.8
Hz). Similar coupling patterns are seen for complex 13. The
solution data would suggest that the acetylide either is
symmetrically coordinated or is rapidly isomerizing between
two unsymmetric structures analogous to the solid-state
structure.
Palladacycle 13 is a competent catalyst for the reaction of
phenylacetylene with propargyl alcohol, but the yield of 7a is
about 10% lower in comparison to that using 10 as the catalyst
(Table 6, entries 1 and 2). Hypothesizing that the decrease in
yield was due to the lower amount of acetate anion present, 2
mol % of potassium acetate was added along with 13. Doing so
resulted in a yield of 7a similar to that when 10 was used as the
precatalyst (entry 3). The addition of 2 mol % of acetic acid
The solid-state structure of 14 (Figure 1) shows a μ-κ2-C1,C1
binding mode for the acetylide anion. Bridging acetylide ligands
can adopt a variety of bonding modes, with κ1,η2 being the
typical structure.24 The μ-κ2 bonding mode is less common but
has been reported for a variety of metals, including Pt,25 Cu,26
Ru,27 and Mn.28 There are no examples of palladium dimers
with a μ-κ2-C1,C1-bound acetylide ligand in the literature,
however. As in other reported examples, the acetylide is
unsymmetrically coordinated to the two palladium centers in
the solid-state structure. The Pd1−C29 bond (2.071(2) Å) is
shorter than the Pd2−C29 bond (2.159(2) Å), and the Pd1−
C29−C30 angle (155.3(2)°) is larger than the Pd2−C29−C30
angle (117.5(2)°). The lack of symmetry does not appear to
indicate an η2 binding mode with Pd2, however. C30 of the
acetylide is much farther (2.923(2) Å) from the Pd2 center
than C29 (2.159(2) Å). In addition, the C29−C30 bond
(1.211(3) Å) is typical for a carbon−carbon triple bond and
remains essentially linear (C29−C30−C31 = 178.4(2)°).
Table 6. Evaluation of Complex 13 as a Precatalyst
a
b
entry
catalyst
additive
yield of 7a (%)
1
2
3
4
10
13
13
13
74
65
73
51
KOAc
AcOH
a
b
2 mol % of additive. Yields were determined via 1H NMR
spectroscopy with an internal standard.
5839
DOI: 10.1021/acscatal.6b01541
ACS Catal. 2016, 6, 5834−5842

ACS Catalysis
Research Article
further decreased the yield of 7a to only 51%. The catalytic
competence of complex 13 suggests that it is involved in the
catalytic cycle for this reaction.
observations. First, performing the reaction with a mixture of
DTBNpP and P(OAc)2 affords lower conversion and decreased
regioselectivity. Although palladacycle 10 likely forms from
DTBNpP and Pd(OAc)2 during the catalytic reaction, the
lower regioselectivity suggests that the κ1-P PdII(DTBNpP)
complex catalyzes the reaction less regioselectively than does
palladacycle 10. Second, 31P NMR analysis of the reaction in
progress shows that 98% of the phosphorus species present are
palladacycles, primarily 10 and 13 (Figure S2 in the Supporting
Information). Although Pd(DTBNpP)2 accounts for 2% of the
phosphine species, we do not think this plays a role in the
selective product of 7 (vide infra). We cannot completely rule
out the possibility of a small concentration of active species,
however. Finally, the preparation of μ-acetylide 13 by reaction
with 10 is consistent with a base-assisted mechanism to
generate the acetylide intermediate, rather than oxidative
addition of the C−H bond. Furthermore, complex 13 is
catalytically competent in this reaction, showing that it is likely
part of the catalytic system.
A number of mechanisms have been proposed for the alkyne
coupling reaction, including (1) base-assisted formation of a
metal acetylide followed by carbometalation of the second
alkyne, (2) oxidative addition of the alkyne C−H bond to a
metal followed by hydrometalation and reductive elimination,
and (3) generation of a metal vinylidene species followed by α
migration of an acetylide.8b On the basis of the observations
above, we propose a mechanism involving generation of a
palladium acetylide followed by carbopalladation of the
propargyl alkyne (Scheme 4). This mechanism is similar to
Scheme 4. Proposed Catalytic Cycle
The chemoselectivity of this system is consistent with the
systems reported by Trost,12b Oshovsky,11 and Xu.12a The
initial formation of the palladium acetylide II presumably
occurs with modest selectivity, since both product 7 and the
propargyl alcohol homodimer 5a are formed in similar amounts
in the reaction. Although propargyl alcohol is present in higher
concentration, the higher of acidity of phenylacetylene allows it
to compete effectively to form complex II. Complex II, or
possibly 13, then reacts selectively with propargyl alcohol over
phenylacetylene to form complex III. This selectivity is shown
by the lack of phenylacetylene homodimer 5b formed in the
reaction. The inductive effect of the propargyl oxygen or
nitrogen is thought to lower the alkyne HOMO−LUMO gap,
which makes it a better ligand.11,12 In addition, the inductive
withdrawing effect would make migratory insertion occur more
readily.
At this time we can only speculate about the reasons for the
unique nature of the regioselectivity of complex 10. The results
in Table 1 show that palladacycles derived from electron-rich
and sterically demanding ligands (DTBNpP and TTBP) afford
selectivity for the linear enyne product 7a, with complex 10
providing the highest selectivity. In contrast, the Herrmann−
Beller palladacycle derived from the less electron rich but
sterically demanding tri-o-tolylphosphine preferentially affords
the branched enyne isomer 8a by a 2:1 ratio. Catalysts
generated in situ from a phosphine and Pd(OAc)2 appear to
favor the branched enyne product 8. In the case of DTBNpP,
using a mixture of the ligand and Pd(OAc)2 in place of 10
affords a lower level of regiocontrol, although linear product 7
is still favored. Tri-o-tolylphosphine affords branched enyne 8a
in higher selectivity (95:5) in comparison to the palladacycle
(64:36 8a:7a).
The regioselectivity in this reaction would be determined by
the conformation of the alkyne that is favored for migratory
insertion (Scheme 5). Steric bulk at the metal center would be
expected to favor insertion to place the metal at the terminal
carbon, leading to the branched enyne product 8. A common
feature of the results in Table 1 is that the palladacyclic
complexes provide a higher proportion of the linear enyne 7 in
comparison to the nonpalladacyclic catalyst systems. We do not
know the stereochemistry of the proposed palladium acetylide
acetylene intermediate. The structure with the acetylide trans to
the Pd−C bond would be expected to be more stable. The
isomer of complex III shown in Scheme 5 may be favored
that proposed by Tenaglia and co-workers for their Herrmann−
Beller palladacycle catalyzed addition of alkynes to norborna-
dienes.21 The palladacycle dimer 10 first dissociates to form
monomer I. Phenylacetylene then undergoes acetate-assisted
C−H activation with monomer I to form the Pd(II) acetylide
intermediate II and release acetic acid. The stereochemistry of
this species is unknown, but the more stable form would likely
have the acetylide trans to the phosphorus. The critical role of
acetate as a base can be seen by the lack of reactivity of
bromide-bridged palladacycle 11. Intermediate II is in
equilibrium with the off-cycle mixed bridging acetylide dimer
13. Intermediate II then coordinates propargyl alcohol to form
III. Again, the stereochemistry of III is unknown, but we
propose that II may isomerize to place the acetylide cis to the
phosphorus, allowing the incoming alkyne to coordinate cis to
the less sterically demanding carbon of the palladacycle.
Complex III then undergoes migratory insertion to afford
vinyl species IV. This insertion is the regioselectivity-
determining step. Complex 13 could also potentially react
directly with propargyl alcohol to give III. Intermediate IV then
undergoes protonolysis with acetic acid generated in the initial
step to form 7a and regenerate monomer I. Alternatively IV
could react directly with phenylacetylene to form 7a and II.
We hypothesize that the palladacycle remains intact
throughout the catalytic cycle and that no redox processes
are involved. This hypothesis is supported by a number of
5840
DOI: 10.1021/acscatal.6b01541
ACS Catal. 2016, 6, 5834−5842

ACS Catalysis
Research Article
phenyl-2-penten-4-yn-1-ol, a key intermediate in the synthesis
of type 2 diabetes treatment candidate NNC 61-4655, in good
yield in one step from inexpensive precursors. In comparison,
the reported process-scale synthesis requires four steps in 19%
overall yield.3 Mechanistic insights were found by isolating and
obtaining a crystal structure of the first example of a μ-κ2-C1,C1-
acetylide palladacycle complex (14). μ-Acetylide complex 13 is
observed under the catalytic conditions and serves as a
competent precatalyst for the alkyne cross-dimerization
reaction. We propose that this species serves as an off-cycle
resting state in the catalytic mechanism. Although we can not
definitively account for the unique regioselectivity observed
with complex 10, it appears that the palladacycle provides the
necessary steric and electronic properties to favor formation of
linear enyne products over the more commonly observed
branched enynes.
Scheme 5. Regioselectivity of Migratory Insertion
ASSOCIATED CONTENT
* Supporting Information
■
sterically, however, as it places the smaller acetylide ligand next
to the larger phosphorus group. It is possible that palladacycles
provide a relatively open coordination site cis to the Pd−C
bond of the palladacycle. Coordination of the alkyne in this
position may provide a higher concentration of rotomer IIIa,
which leads to the linear enyne product. This steric effect
cannot solely explain the regioselectivity effects seen; however,
as the P(o-tol)3-based palladacycle favors branched product 8a
in contrast to the palladacycles derived from bulky
alkylphosphines. The electron-donating ability of DTBNpP
and TTBP appears to further enhance the selectivity for the
linear product. A more electron deficient metal center would be
expected to have a stronger preference to insert with the metal
at the less substituted carbon. The more electron rich metal
center in 10 may weaken the electronic preference, allowing the
steric preference to take over.
Additives also play a role in the selectivity of the alkyne
coupling reaction, although we do not fully understand these
effects at this time. One interesting observation was that the
presence of air suppressed the formation of branched enynes 8a
and 6a in the coupling of phenylacetylene and propargyl
alcohol. It is not clear how the presence of oxygen affects the
coupling reaction. One possible hypothesis is that air serves to
reoxidize any Pd(0) species formed during the reaction. A small
of amount of Pd0(DTBNpP)2 is observed during the catalytic
reaction. It is possible that this species, or something related,
promotes formation of the branched isomer. Further
mechanistic studies are needed to test this hypothesis.
S
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acscatal.6b01541.
Full experimental procedures and characterization data,
solvent screen results, VT-NMR of complex 10, in situ
31P NMR of the catalytic reaction, crystal structure data
and refinement details for 14, and 1H, 13C, and 31P NMR
spectra for the palladium complexes and compounds
reported in Tables 2, 4, and 5 (PDF)
Crystallographic data for 14 (CIF)
AUTHOR INFORMATION
Corresponding Author
*E-mail for K.H.S.: kshaughn@ua.edu.
■
Present Address
§Department of Chemistry, Purdue University, 560 Oval Dr.,
West Lafayette, IN 47902-2084.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We are grateful to the National Science Foundation (CHE-
1058984) for financial support of this work, FMC, Lithium
Division, for donation of DTBNpP and TNpP, and Johnson-
Matthey for donation of palladium compounds.
Lower regioselectivity is seen for propargyl amide substrates
in comparison to propargyl alcohol. It is possible that the
increased steric demand of the phthalimide substituent results
in an increased preference for insertion to occur with the
palladium at the terminal carbon of the propargyl amide to give
branched product 8. This steric preference may partially offset
the inherent preference for complex 10 to afford the linear
product 7. The details of these selectivity effects will require
further study to fully elucidate.
REFERENCES
■
(1) Sonoda, H.; Nishida, K.; Yoshioka, T.; Ohtani, M.; Sugita, K.
Oncogene 1996, 13, 143−149.
(2) Zhang, H.-J.; Sydara, K.; Tan, G. T.; Ma, C.; Southavong, B.;
Soejarta, D. D.; Pezzuto, J. M.; Fong, H. H. S. J. Nat. Prod. 2004, 67,
194−200.
(3) Deussen, H.-J.; Jeppessen, L.; Scharer, N.; Junager, F.; Bentzen,
̈
B.; Weber, B.; Weil, V.; Mozer, S. J.; Sauerberg, P. Org. Process Res.
Dev. 2004, 8, 363−371.
(4) (a) Kuroda, H.; Hanaki, E.; Izawa, H.; Kano, M.; Itahashi, H.
Tetrahedron 2004, 60, 1913−1920. (b) Faizi, D. J.; Issaian, A.; Davis,
A. J.; Blum, S. A. J. Am. Chem. Soc. 2016, 138, 2126−2129. (c) Song,
F.; Gaul, M. D.; Murray, W. V. Tetrahedron Lett. 2015, 56, 5413−5415.
CONCLUSION
■
In conclusion, the first selective synthesis of linear (E)-enynols
7, found in a number of biologically active compounds, from
terminal aryl alkynes and propargyl alcohols has been reported.
The reaction proceeds with low catalyst loadings, operates
under mild conditions, and does not require the addition of
stoichiometric acids or bases. This methodology affords (E)-4-
(d) Karagoz, E. S
2014, 79, 9222−9230.
̧
.; Kus,
̧ M.; Akpınar, G. E.; Artok, L. J. Org. Chem.
̈
(5) (a) Negishi, E.-i.; Anastasia, L. Chem. Rev. 2003, 103, 1979−2017.
(b) Sonogashira, K.; Tohda, T.; Hagihara, N. Tetrahedron Lett. 1975,
16, 4467−4470. (c) Takeuchi, R.; Tanabe, K.; Tanaka, S. J. Org. Chem.
5841
DOI: 10.1021/acscatal.6b01541
ACS Catal. 2016, 6, 5834−5842

ACS Catalysis
Research Article
2000, 65, 1558−1561. (d) Thongsornkleeb, C.; Danheiser, R. L. J. Org.
Chem. 2005, 70, 2364−2367.
(6) Lebel, H.; Davi, M. Adv. Synth. Catal. 2008, 350, 2352−2358.
(7) Yan, W.; Ye, X.; Akhmedov, N. G.; Petersen, J. L.; Shi, X. Org.
Lett. 2012, 14, 2358−2361.
(8) (a) García-Garrido, S. E. In Modern Alkyne Chemistry: Catalytic
and Atom-Economic Transformations, 1st ed.; Trost, B. M., Li, C.-J.,
Eds.; Wiley-VCH: Weinheim, Germany, 2015; pp 301−334. (b) Trost,
B. M.; Masters, J. T. Chem. Soc. Rev. 2016, 45, 2212−2238.
(9) (a) Ogata, K.; Oka, O.; Toyota, A.; Suzuki, N.; Fukuzawa, S.-i.
Synlett 2008, 2008, 2663−2666. (b) Katagiri, T.; Tsurugi, H.; Satoh,
T.; Miura, M. Chem. Commun. 2008, 3405−3407. (c) Rivada-
Wheelaghan, O.; Chakraborty, S.; Shimon, L. J. W.; Ben-David, Y.;
Milstein, D. Angew. Chem., Int. Ed. 2016, 55, 6942−6945.
́ ́
(10) Azpíroz, R.; Rubio-Perez, L.; Castarlenas, R.; Perez-Torrente, J.
J.; Oro, L. A. ChemCatChem 2014, 6, 2587−2592.
(11) Oshovsky, G. V.; Hessen, B.; Reek, J. N. H.; de Bruin, B.
Organometallics 2011, 30, 6067−6070.
(12) (a) Xu, H.-D.; Zhang, R.-W.; Li, X.; Huang, S.; Tang, W.; Hu,
W.-H. Org. Lett. 2013, 15, 840−843. (b) Trost, B. M.; McIntosh, M. C.
Tetrahedron Lett. 1997, 38, 3207−3210.
(13) Lauer, M. G.; Thompson, M. K.; Shaughnessy, K. H. J. Org.
Chem. 2014, 79, 10837−10848.
(14) Albrecht, M. Chem. Rev. 2010, 110, 576−623.
(15) Mason, R.; Textor, M.; Al-Salem, N.; Shaw, B. L. J. Chem. Soc.,
Chem. Commun. 1976, 292−293.
(16) (a) Yamashita, M.; Moroe, Y.; Yano, T.; Nozaki, K. Inorg. Chim.
Acta 2011, 369, 15−18. (b) Yano, T.; Moroe, Y.; Yamashita, M.;
Nozaki, K. Chem. Lett. 2008, 37, 1300−1301. (c) Dahlenburg, L.;
Hock, N. Inorg. Chim. Acta 1985, 104, L29−L30. (d) Dahlenburg, L.;
̈
Yardimcioglu, A. J. Organomet. Chem. 1985, 291, 371−386.
(17) (a) Henderson, W. H.; Alvarez, J. M.; Eichman, C. C.; Stambuli,
J. P. Organometallics 2011, 30, 5038−544. (b) Nagaki, A.; Takabayashi,
N.; Moriwaki, Y.; Yoshida, J.-i. Chem. - Eur. J. 2012, 18, 11871−11875.
(c) Wu, L.; Hartwig, J. F. J. Am. Chem. Soc. 2005, 127, 15824−15832.
(18) Herrmann, W. A.; Brossmer, C.; Reisinger, C.-P.; Riermeier, T.
̈
H.; Ofele, K.; Beller, M. Chem. - Eur. J. 1997, 3, 1357−1364.
(19) (a) Duric, S.; Sypaseuth, F. D.; Hoof, S.; Svensson, E.;
Tzschucke, C. C. Chem. - Eur. J. 2013, 19, 17456−17463. (b) Shiota,
A.; Malinakova, H. C. J. Organomet. Chem. 2012, 704, 9−16. (c) Tan,
Y.; Barrios-Landeros, F.; Hartwig, J. F. J. Am. Chem. Soc. 2012, 134,
3683−3686.
(20) (a) Beletskaya, I. P.; Cheprakov, A. V. J. Organomet. Chem.
2004, 689, 4055−4082. (b) Dupont, J.; Pfeffer, M. Palladacycles:
Synthesis, Characterization and Applications; Wiley-VCH: Weinheim,
Germany, 2008. (c) Mo, D.-L.; Zhang, T.-K.; Ge, G.-C.; Huang, X.-J.;
Ding, C.-H.; Dai, L.-X. Synlett 2014, 25, 2686−2702.
(21) Tenaglia, A.; Giordano, L.; Buono, G. Org. Lett. 2006, 8, 4315−
4318.
(22) Trost, B. M.; Chan, C.; Ruhter, G. J. Am. Chem. Soc. 1987, 109,
3486−3487.
(23) (a) Rubina, M.; Gevorgyan, V. J. Am. Chem. Soc. 2001, 123,
11107−11108. (b) Pu, X.; Li, H.; Colacot, T. J. J. Org. Chem. 2013, 78,
568−581.
(24) Berenguer, J. R.; Lalinde, E.; Moreno, M. T. Coord. Chem. Rev.
2010, 254, 832−875.
(25) Hutton, A. T.; Shebanzadeh, B.; Shaw, B. L. J. Chem. Soc., Chem.
Commun. 1984, 549−51.
(26) Bruce, M. I.; Zaitseva, N. N.; Skelton, B. W.; Somers, N.; White,
A. H. Inorg. Chim. Acta 2007, 360, 681−685.
(27) Takemoto, S.; Ito, T.; Yamazaki, Y.; Tsujita, M.; Matsuzaka, H. J.
Organomet. Chem. 2016, 812, 158−166.
(28) Kheradmandan, S.; Fox, T.; Schmalle, H. W.; Venkatesan, K.;
Berke, H. Eur. J. Inorg. Chem. 2004, 2004, 3544−3554.
5842
DOI: 10.1021/acscatal.6b01541
ACS Catal. 2016, 6, 5834−5842