Palladium-catalyzed completely linear-selective negishi cross-coupling of allylzinc halides with aryl and vinyl electrophiles

Article abstract of DOI:10.1002/anie.201308585

Completely linear: The title reaction provides an effective means to access a wide range of prenylated arenes and skipped dienes in a completely linear-selective fashion, as demonstrated by a concise synthesis of the anti-HIV natural product siamenol. DFT calculations shed light on the origin of the excellent regioselectivity observed with the current Pd-based catalyst system. Copyright

Full text of DOI:10.1002/anie.201308585

.
Angewandte
Communications
DOI: 10.1002/anie.201308585
Selective Cross-Coupling
Hot Paper
Palladium-Catalyzed Completely Linear-Selective Negishi Cross-
Coupling of Allylzinc Halides with Aryl and Vinyl Electrophiles**
Yang Yang, Thomas J. L. Mustard, Paul Ha-Yeon Cheong,* and Stephen L. Buchwald*
Prenylated arenes are found in a broad spectrum of naturally
occurring bioactive compounds (Scheme 1).[1] Studies have
revealed that the inclusion of the prenyl side chain could
enhance both the bioactivity and bioavailability of natural
products, in part due to the increased protein binding affinity
and membrane permeability caused by the lipophilicity of the
prenyl group.[2] In nature, prenylation processes usually
involve enzymatic reactions mediated by a range of sub-
strate-specific prenyltransferases (PTases),[3] giving rise to the
corresponding prenylated natural products in a highly selec-
tive fashion. However, the ability of synthetic chemists to
directly introduce prenyl and related 3,3-disubstituted allyl
groups onto functionalized aromatic compounds is generally
hampered by poor regioselectivity with respect to the unsym-
metrical allyl nucleophile.[4,5] Therefore, a general, reliable,
and practical method for regioselective prenylation that
emulates the efficiency of natureꢀs biosynthetic machinery is
highly desirable.[6]
Recently, Organ et al.[4] and we[5] have developed con-
ditions for the linear-selective Suzuki–Miyaura coupling of
3,3-disubstituted allylboronates using N-heterocyclic carbene
(NHC)- and phosphine-based catalysts, respectively. While
both methods were highly selective, relatively high temper-
atures and extended reaction times were required, and the
yields of the desired prenylated products were moderate due
to the inevitable formation of homocoupling products.[4,5] In
an effort to develop milder and more efficient prenylation
methods, we sought to utilize alternative prenyl nucleophiles.
Prenyl-type organozinc reagents were our first choice, as we
had previously demonstrated the high reactivity as well as
functional group compatibility of organozinc reagents in
a variety of Negishi coupling processes.[7,8] To date, the cross-
coupling of 3,3-disubstituted allylzinc reagents with aryl
halides remains rare, presumably due to issues of regioselec-
tivity often experienced with this class of nucleophile.[4,5,9]
Herein we report the first general and completely linear-
selective conditions for the Negishi coupling of 3,3-disubsti-
tuted allylzinc reagents with aryl halides and the application
of this methodology in the concise and convergent synthesis
of anti-HIV natural product siamenol (1). Computational
studies were also carried out to gain a deeper insight into the
high level of selectivity observed with the current catalyst
system.[10,11]
Scheme 1. Top: Representative biologically active prenylated natural
products. Bottom: Regioselective Negishi cross-coupling: rapid access
to prenylated compounds.
Using prenylZnBr·LiCl prepared by Knochelꢀs proto-
col[12,13] and 1-bromo-4-butylbenzene as model substrates, we
commenced our study by examining our recently developed
easily activated palladacycle precatalysts derived from biar-
ylphosphine ligands (Table 1).[14] While SPhos (L1) and
RuPhos (L2)-based catalysts were effective for the linear-
selective prenylation, only moderate conversion was observed
(Table 1, entries 1 and 2). Both the XPhos (L3)- and the
CPhos (L5)-based catalyst furnished full conversion of the
aryl halide component, affording the a-isomer (4) in good
yield with minimal amounts of the g-isomer or other side
products (entries 3 and 5). Ultimately, the catalyst generated
from CPhos[15] was identified as the optimal choice for this
coupling reaction, affording the a-coupling product in
a highly selective manner.[16] Under the optimized reaction
conditions, treatment of the aryl bromide with prenylzinc
bromide in the presence of 2 mol% CPhos-based catalyst
afforded the prenylated product in 94% yield in 30 min at
[*] Y. Yang, Prof. Dr. S. L. Buchwald
Department of Chemistry
Massachusetts Institute of Technology
77 Massachusetts Avenue, Cambridge, MA 02139 (USA)
E-mail: sbuchwal@mit.edu
T. J. L. Mustard, Prof. Dr. P. H.-Y. Cheong
Department of Chemistry, Oregon State University
135 Gilbert Hall, Corvallis, OR 97331 (USA)
E-mail: paul.cheong@oregonstate.edu
[**] We thank the National Institutes of Health (GM46059 for S.L.B.),
Oregon State University (P.H.-Y.C.), Vicki and Patrick F. Stone
Scholar Funds (P.H.-Y.C.), and David P. Shoemaker Memorial
Fellowship (T.J.L.M.) for financial support. MIT has patents on
some of the ligands and precatalysts used in this study from which
S.L.B. as well as current or former co-workers receive royalty
payments.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201308585.
14098
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 14098 –14102

Angewandte
Chemie
Table 1: Ligand effect on the palladium-catalyzed Negishi cross-coupling
of prenylzinc bromide and 1-bromo-4-butylbenzene.
Entry
Ligand
Conversion[a]
Yield[a]
a/g ratio[a]
1
2
3
4
L1
L2
L3
L4
L5
L5
L5
L5
54%
52%
100%
84%
98%
100%
100%
100%
47%
41%
92%
74%
94%
92%
94%
94%
>99:1
>99:1
>99:1
>99:1
>99:1
>99:1
>99:1
>99:1
5
6[b]
7[c]
8[d]
[a] Conversions and yields were determined by GC analysis of the crude
reaction mixture using dodecane as an internal standard. [b] 708C for
only 3 min. [c] 1-Butyl-4-iodobenzene was used in lieu of 2. [d] 4-
Butylphenyl triflate was used in lieu of 2.
room temperature (entry 5). Complete conversion could also
be achieved in ꢀ 3 min at 708C, further demonstrating the
excellent activity of this catalyst system (entry 6). Aryl
triflates (entry 7) and iodides (entry 8) were also suitable
coupling partners using this protocol. Interestingly, the well-
tailored NHC-supported PEPPSI precatalyst[4] (PEPPSI-
IPent, 6) proved to be less effective for this transformation
(see the Supporting Information). Other commercially avail-
able and frequently used catalysts or ligands such as [Pd-
(PPh3)4], dppm, dppp, and dppb were ineffective for this
prenylation reaction (see the Supporting Information for
details).
We next investigated the substrate scope of the method
(Scheme 2). It was found that with the CPhos-based catalyst,
both electron-donating (7a) and electron-withdrawing (7b,
7c) substituents were tolerated with no decrease in the
observed regioselectivity. An extremely sterically hindered
substrate (7d) could also be successfully transformed. Sub-
strates containing acidic functional groups, such as an
acetamide (7e) and a phenol (7 f), were also transformed to
the desired products. Given the importance of heterocycles in
medicinal chemistry,[17] we examined a variety of sterically
and electronically diverse heteroaryl halides (Scheme 2B).
We found that pyridine (7g), pyrimidine (7h), indoles (7i, 7j),
azaindole (7k), and pyrazole (7l) all underwent smooth
transformation using this protocol.
Scheme 2. Substrate scope of aryl and heteroaryl halides. Reaction
conditions: Ar-X (1.0 mmol), prenylZnBr·LiCl (1.3 mmol), 3
(0.02 mmol), L5 (0.02 mmol), RT, THF, 2 h. Yields are of isolated
1
product as an average from two runs. [a] Determined by H NMR
spectroscopy due to the volatility of coupling product. [b] 2.3 equiv
prenylzinc was used.
a number of biologically active natural products,[19] in
a completely regioselective manner (Scheme 3). Notably,
mono- (8a), di- (8b–8d), and trisubstituted (8e, 8 f) vinyl
electrophiles could all be applied in this reaction without
noticeable erosion of regioselectivity. Five- (8e) and six-
membered (8b–8d) cyclic vinyl triflates represented compat-
ible coupling partners as well.
In an effort to expand the utility of this method, we
examined the coupling of various 3,3-unsymmetrically disub-
stituted allylzinc halides (Scheme 4). In all cases examined,
the allylation proceeded smoothly furnishing the linear-
coupling product exclusively in excellent yields. While both
geranyl- (9a) and farnesylzinc bromides (9b) afforded 75:25
mixtures of olefin stereoisomers, allylzinc halides bearing two
substituents of greater steric difference reacted with
improved stereoselectivity with respect to the trisubstituted
olefin moiety (9c–9e). For example, while the use of 3-
methyl-3-cyclohexylallylzinc bromide (9c) furnished coupling
product as stereoisomeric mixtures (E/Z ratio = 85:15), the
coupling of 3-methyl-3-tert-butylallylzinc bromide provided
the trisubstituted olefin exclusively with E geometry (9d).
To further showcase the utility of this prenylation
methodology in a complex setting, we performed a concise
synthesis of siamenol (1), a prenylated natural product
isolated from Murraya siamensis and exhibiting anti-HIV
activity.[20] Beginning with 4-bromotoluene (10), palladium-
One limitation of our previously developed Suzuki–
Miyaura coupling was its ability to effectively engage vinyl
electrophiles. Lower levels of regioselectivity were often
observed with regard to vinyl bromides,[18] and in the case of
vinyl triflates, competitive hydrolysis resulted in low yields of
the desired coupling products. The current Negishi coupling
protocol successfully addressed these problems. Vinyl bro-
mides and triflates were converted to the corresponding
“skipped dienes”, which represent key structural motifs in
Angew. Chem. Int. Ed. 2013, 52, 14098 –14102
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 14099

.
Angewandte
Communications
computation of intermediates and transition state structures
(Figure 1).[21,22] The catalytic cycle begins with the initial
complexation of the Pd0 catalyst with the aryl bromide (I).
À
Oxidative addition into the C Br bond (TS-II) affords the
resting state intermediate III. The two possible transmetala-
tion processes involving a-prenylzinc were next investigated,
namely a four-membered transition state (TS-IV-a-4-mem)
and a six-membered transition state (TS-IV-g-6-mem) leading
to a- and g-prenyl palladium intermediates V, respectively.
Given that the prenylzinc species undergoes rapid 1,3-shift at
room temperature,[23] we also evaluated two additional
mechanisms utilizing g-prenylzinc bromide as the transmeta-
lating agent (Figure 1, TS-IV-g-4-mem and TS-IV-a-6-mem,
respectively). We found that there is an energetic preference
for the a-4-membered transition state over the a-6, but this
was reversed in the g case presumably due to the steric bulk
À
around the forming Pd C bond in TS-IV-g-4-mem. The
minimum-energy pathway involves the TS-IV-a-4-mem, gen-
erating the a-prenyl palladium intermediate V, which then
undergoes product-determining h1-a reductive elimination
(TS-VI) to afford the product–catalyst complex VII. We also
investigated g reductive elimination leading to the g-isomer
and found that the h3-g elimination was preferred over the h1-
g. The h3-g process was found to be disfavored by 12.2 kcal
molÀ1. In a model system in which the prenyl substrate is
replaced with an allyl, the h1 reductive elimination is
preferred over h3 by 2.2 kcalmolÀ1.[24] In the prenyl case,
Scheme 3. Substrate scope of vinyl halides and pseudohalides. Reac-
tion conditions: vinyl halide (0.5 mmol), prenylZnBr·LiCl (0.65 mmol),
3 (0.01 mmol), L5 (0.01 mmol), RT, THF, 2 h. Yields are of isolated
product as an average from two runs.
À
the steric bulk around the forming C C bond precludes the
g reductive elimination from achieving the more stable h1
geometry. The a isomer is favored universally for all steps,
including the a-4-mem transmetalation. The high level of
linear-selectivity observed with the current catalyst system
is controlled by the preference for the a-isomer-forming
reductive elimination via TS-VI-h1-a.
In summary, we have developed the first general,
practical, and completely linear-selective Negishi coupling
of 3,3-disubstituted allyl organozinc reagents with aryl,
heteroaryl, and vinyl halides. This reaction features excep-
tionally mild reaction conditions and is broad in scope with
respect to both aryl/vinyl halide and allylzinc coupling
partners. This linear-selective allylation methodology pro-
vides an effective means for accessing highly functional-
ized aromatic and heteroaromatic compounds bearing
Scheme 4. Substrate scope of 3,3-disubstituted allylzinc reagents. Reaction
conditions: Ar-X (0.50 mmol), allylZnX’·LiCl (0.65 mmol), 3 (0.01 mmol),
L5 (0.01 mmol), RT, THF, 2 h. Yields are of isolated product as an average
from two runs.
catalyzed amination proceeded smoothly to deliver the
unsymmetrical diarylamine 12 (Scheme 5). Subse-
À
quently, palladium-catalyzed intramolecular C H acti-
vation and bromination at C3 furnished the carbazole
14, which in turn, underwent the completely linear-
selective Negishi cross-coupling to afford the prenylated
carbazole 15. In the final step, demethylation employing
methylmagnesium iodide furnished the natural product,
siamenol (1). Overall, the strategic applications of
a series of palladium-catalyzed cross-coupling reactions
facilitated by the use of dialkylbiarylphosphine ligands
developed in our laboratory have enabled the rapid
assembly of a prenylated carbazole natural product.
To gain further understanding into the regiocontrol
of this reaction, we analyzed the reaction coordinate by Scheme 5. Concise synthesis of siamenol (1).
14100 www.angewandte.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 14098 –14102

Angewandte
Chemie
Figure 1. Top: Reaction coordinate of the Pd0-catalyzed cross-coupling of prenylZnBr (3) with p-bromotoluene. Bottom: Transition structures for
and transmetalation and reductive elimination transition structures.[22] Energies are given in kcalmolÀ1 and distances in ꢀ. Computed structures
are rendered in Pymol.[23]
[6] For recent reviews on prenylation methods, see: T. Lindel, N.
prenyl-type side chains, as illustrated by the concise synthesis
Marsch, S. K. Adla, Top. Curr. Chem. 2012, 309, 67 – 130; For
additional recent examples, see: C. A. Kuttruff, H. Zipse, D.
Trauner, Angew. Chem. 2011, 123, 1438 – 1441; Angew. Chem.
Int. Ed. 2011, 50, 1402 – 1405.
of anti-HIV natural product siamenol. Finally, theoretical
calculations provided insights into the regiochemical-control-
ling parameters of these coupling processes, demonstrating
the critical choice of the catalyst and the transmetalating
reagent in achieving good selectivity for coupling reactions
involving 3,3-disubstituted allyl nucleophiles.
[7] For palladium-catalyzed sp2–sp2 Negishi cross-coupling reac-
tions, see: a) J. E. Milne, S. L. Buchwald, J. Am. Chem. Soc. 2004,
126, 13028 – 13032; b) Y. Yang, N. J. Oldenhuis, S. L. Buchwald,
Angew. Chem. 2013, 125, 643 – 647; Angew. Chem. Int. Ed. 2013,
52, 615 – 619.
Received: October 2, 2013
Published online: November 8, 2013
[8] For palladium-catalyzed sp2 – sp3 Negishi cross-coupling reac-
tions, see: C. Han, S. L. Buchwald, J. Am. Chem. Soc. 2009, 131,
7532 – 7533.
[9] For an example where prenylzinc reagents were used in a non-
regioselective sense, see: Y. Mitamura, Y. Asada, K. Murakami,
H. Someya, H. Yorimitsu, K. Oshima, Chem. Asian J. 2010, 5,
1487 – 1493.
Keywords: cross-coupling · heterocycles · organozinc reagents ·
.
palladium · prenylation
[1] For recent reviews, see: a) R. M. Williams, E. M. Stocking, J. F.
Sanz-Cervera, Top. Curr. Chem. 2000, 209, 97 – 173; b) S. Li, Nat.
Prod. Rep. 2010, 27, 57 – 78.
[2] a) B. Botta, G. Delle Monache, P. Menendez, A. Boffi, Trends
Pharmacol. Sci. 2005, 26, 606 – 608; b) B. Botta, A. Vitali, P.
Menendez, D. Misiti, G. Delle Monache, Curr. Med. Chem. 2005,
12, 717 – 739.
[3] a) M. Tello, T. Kuzuyama, L. Heide, J. P. Noel, S. B. Richard,
Cell. Mol. Life Sci. 2008, 65, 1459 – 1463; b) C. J. Balibar, A. R.
Howard-Jones, C. T. Walsh, Nat. Chem. Biol. 2007, 3, 584 – 592.
[4] J. L. Farmer, H. N. Hunter, M. G. Organ, J. Am. Chem. Soc.
2012, 134, 17470 – 17473.
[10] For cross-coupling reactions involving allylzinc halides, see:
a) E. Nakamura, Tetrahedron Lett. 1988, 29, 5155 – 5156; b) Y.
Baba, A. Toshimitsu, S. Matsubara, Synlett 2008, 2061 – 2063;
c) Z. Xu, E.-i. Negishi, Org. Lett. 2008, 10, 4311 – 4314; d) C.
Wang, T. Tobrman, Z. Xu, E.-i. Negishi, Org. Lett. 2009, 11,
4092 – 4095.
[11] For palladium-catalyzed coupling reactions involving other
allylmetals, see: Si: a) Y. Hatanaka, Y. Ebina, T. Hiyama, J.
Am. Chem. Soc. 1991, 113, 7076 – 7077; b) Y. Hatanaka, K.
Goda, T. Hiyama, Tetrahedron Lett. 1994, 35, 1279 – 1282; c) Y.
Hatanaka, K. Goda, T. Hiyama, Tetrahedron Lett. 1994, 35,
6511 – 6514; d) S. E. Denmark, N. S. Werner, J. Am. Chem. Soc.
2010, 132, 3612 – 3620; Sn: e) B. M. Trost, E. Keinan, Tetrahe-
dron Lett. 1980, 21, 2595 – 2598; f) A. M. Echavarren, J. K. Stille,
[5] Y. Yang, S. L. Buchwald, J. Am. Chem. Soc. 2013, 135, 10642 –
10645.
Angew. Chem. Int. Ed. 2013, 52, 14098 –14102
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 14101

.
Angewandte
Communications
J. Am. Chem. Soc. 1987, 109, 5478 – 5486; g) V. Farina, B.
Krishnan, J. Am. Chem. Soc. 1991, 113, 9585 – 9595; h) Y. Obora,
Y. Tsuji, M. Kobayashi, T. Kawamura, J. Org. Chem. 1995, 60,
4647 – 4649.
[17] a) J. A. Joule, K. Mills in Heterocyclic Chemistry, 5th ed., Wiley,
Chichester, 2010; b) R. Leurs, R. A. Bakker, H. Timmerman,
I. J. P. de Esch, Nat. Rev. Drug Discovery 2005, 4, 107 – 120.
[18] For example, when we used our previously developed Suzuki–
Miyaura coupling protocol, 8a was obtained as a mixture of a-
and g-isomers (a/g = 93:7) in 67% yield as determined by
1H NMR spectroscopy.
[19] T. K. Macklin, G. C. Micalizio, Nat. Chem. 2010, 2, 638 – 643, and
references therein.
[20] a) K. M. Meragelman, T. C. McKee, M. R. Boyd, J. Nat. Prod.
2000, 63, 427; b) M. P. Krahl, A. Jꢁger, T. Krause, H.-J. Knçlker,
Org. Biomol. Chem. 2006, 4, 3215—3219; c) W. Kong, C. Fu, S.
Ma, Chem. Eur. J. 2011, 17, 13134 – 13137.
[12] a) A. Krasovskiy, V. Malakhov, A. Gavryushin, P. Knochel,
Angew. Chem. 2006, 118, 6186 – 6190; Angew. Chem. Int. Ed.
2006, 45, 6040 – 6044; b) H. Ren, G. Dunet, P. Mayer, P. Knochel,
J. Am. Chem. Soc. 2007, 129, 5376 – 5377.
[13] We found that the addition of LiCl accelerated the rate of these
Negishi cross-coupling reactions. For the role of added LiCl, see:
L. C. McCann, H. N. Hunter, J. A. C. Clyburne, M. G. Organ,
Angew. Chem. 2012, 124, 7130 – 7133; Angew. Chem. Int. Ed.
2012, 51, 7024 – 7027, and references therein.
[14] Previous studies showed palladacycle precatalysts (5) allowed
the most rapid and quantitative generation of the catalytically
active L1Pd0 species, see: N. C. Bruno, M. T. Tudge, S. L.
Buchwald, Chem. Sci. 2013, 4, 916 – 920 and Ref. [7b].
[15] CPhos (CAS# 1160556-64-8) is now commercially available from
Aldrich (catalog# 759171).
[21] a) The B3LYP hybrid functional with the 6-31G* (for H, C, N,
and P) and LANL2DZ + ECP (for Zn, Br and Pd) basis sets
were used; b) M. J. Frisch, et al. Gaussian, Inc.: Pittsburgh, PA,
2004.
[22] The PyMOL Molecular Graphics System, version 1.3; Schrç-
dinger, LLC.
[16] For some substrates, the catalyst generated from XPhos (L3)
afforded isomeric mixtures of a- and g-coupling products and
thus did not represent the most general system for this reaction.
For example, the XPhos-based catalyst provided a mixture of
linear and branched coupling products for 5-bromoindole,
whereas the CPhos (L5)-based catalyst remained highly linear-
selective:
[23] a) Figure and energies for this transformation are given in the
Supporting Information; b) R. W. Hoffmann, Angew. Chem.
1982, 94, 569 – 580; Angew. Chem. Int. Ed. Engl. 1982, 21, 555 –
566.
[24] Figures and energies for both the allyl h1 and h3 reductive
elimination are given in the Supporting Information.
14102 www.angewandte.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 14098 –14102