Enantioselective Rh-catalyzed domino transformations of alkynylcyclohexadienones with organoboron reagents

Article abstract of DOI:10.1021/ol400363f

A new enantioselective rhodium-catalyzed domino reaction is described that gives access to fused heterocycles by desymmetrization of alkyne-tethered cyclohexadienones. Two new C-C bonds and two stereocenters are formed in one step with good enantioselecti

Full text of DOI:10.1021/ol400363f

ORGANIC
LETTERS
2013
Vol. 15, No. 5
1148–1151
Enantioselective Rh-Catalyzed
Domino Transformations of
Alkynylcyclohexadienones with
Organoboron Reagents
Juliane Keilitz, Stephen G. Newman, and Mark Lautens*
Department of Chemistry, University of Toronto, 80 St. George Street, Toronto,
Ontario M5S 3H6, Canada
mlautens@chem.utoronto.ca
Received February 6, 2013
ABSTRACT
A new enantioselective rhodium-catalyzed domino reaction is described that gives access to fused heterocycles by desymmetrization of alkyne-
tethered cyclohexadienones. Two new CÀC bonds and two stereocenters are formed in one step with good enantioselectivity. In contrast to prior
reports, it was found that a vinylidene is not involved in the product formation but that syn-addition of the rhodium-aryl species onto the alkyne
takes place.
The formation of multiple CÀC bonds in transition-
metal-catalyzed domino reactions is an attractive means
to construct complex structures starting from relatively
simple starting materials.1 In the past decade, the addition
of arylboronic acids to unsaturated organic functional
groups has proven to be a useful synthetic tool for CÀC
bond formation.2 In particular, the addition of Rh(I)Àaryl
complexes to alkynes, also referred to as hydroarylation,
has found widespread application.3
either C-atom of the triple bond (Scheme 1). Depending on
the respective flanking groups, moderate to excellent
regioselectivities can be obtained with internal alkynes.
Hayashi and our group were early contributors,3a,4 and
since then hydroarylations have been investigated exten-
sively. For terminal alkynes another catalytic pathway has
beenproposed.5 Formal1,1-carborhodationvia formation
of an intermediate vinylidene and subsequent R-migration
would give the 1,1-addition product.6
Rh-catalyzed hydroarylation proceeds via transmetala-
tion of the catalyst with the boronic acid. Subsequent syn-
addition onto an alkyne results in 1,2-carborhodation and
can occur in two orientations with rhodium ending up on
(4) (a) Lautens, M.; Yoshida, M. Org. Lett. 2002, 4, 123–125. (b)
Lautens, M.; Yoshida, M. J. Org. Chem. 2003, 68, 762–769.
(5) Chen, Y.; Lee, C. J. Am. Chem. Soc. 2006, 128, 15598–15599.
(6) Prior to their work, rhodium vinylidene intermediates were only
observed in stoichiometric systems by using Rh(I)Àphosphine com-
plexes: (a) Wolf, J.; Werner, H.; Serhadli, O.; Ziegler, M. L. Angew.
Chem., Int. Ed. Engl. 1983, 22, 414–416. (b) Wiedemann, R.; Steinert, P.;
(1) (a) Tietze, L. F. Chem. Rev. 1996, 96, 115–136. (b) Negishi, E.-i.;
ꢀ
Coperet, C.; Ma, S.; Liou, S.-Y.; Liu, F. Chem. Rev. 1996, 96, 365–393.
(c) Montgomery, J. Angew. Chem., Int. Ed. 2004, 43, 3890–3908.
(d) Boyer, A.; Lautens, M. Angew. Chem., Int. Ed. 2011, 50, 7346–
7349. (e) Liu, H.; El-Salfiti, M.; Chai, D. I.; Auffret, J.; Lautens, M. Org.
Lett. 2012, 14, 3648–3651.
(2) (a) Fagnou, K.; Lautens, M. Chem. Rev. 2003, 103, 169–196.
(b) Hayashi, T.; Yamasaki, K. Chem. Rev. 2003, 103, 2829–2844. (c)
Edwards, H. J.; Hargrave, J. D.; Penrose, S. D.; Frost, C. G. Chem. Soc.
Rev. 2010, 39, 2093–2105.
(3) (a) Hayashi, T.; Inoue, K.; Taniguchi, N.; Ogasawara, M. J. Am.
Chem. Soc. 2001, 123, 9918–9919. (b) Arcadi, A.; Aschi, M.; Chiarini,
M.; Ferrara, G.; Marinelli, F. Adv. Synth. Catal. 2010, 352, 493–498.
(c) Tsui, G. C.; Lautens, M. Angew. Chem., Int. Ed. 2010, 49, 8938–8941.
(d) Panteleev, J.; Zhang, L.; Lautens, M. Angew. Chem., Int. Ed. 2011,
50, 9089–9092.
€
Schafer, M.; Werner, H. J. Am. Chem. Soc. 1993, 115, 9864–9865. But
several examples in catalytic processes have been published, proposing
vinylidene intermediates when using phosphine ligands on rhodium: (c)
Kim, H.; Lee, C. J. Am. Chem. Soc. 2006, 128, 6336–6337. (d) Fukumoto,
Y.; Asai, H.; Shimizu, M.; Chatani, N. J. Am. Chem. Soc. 2007, 129,
13792–13793. (e) Fukumoto, Y.; Kawahara, T.; Kanazawa, Y.; Chatani,
N. Adv. Synth. Catal. 2009, 351, 2315–2318. (f) Mizuno, A.; Kusama, H.;
Iwasawa, N. Angew. Chem., Int. Ed. 2009, 48, 8318–8320. (g) Joo, J. M.;
Yuan, Y.; Lee, C. J. Am. Chem. Soc. 2006, 128, 14818–14819. (h) Joo,
J. M.; David, R. A.; Yuan, Y.; Lee, C. Org. Lett. 2010, 12, 5704–5707.
Additionally, Rh(cod)Àvinylidene complexes have been suggested as
intermediates: (i) Moran, G.; Green, M.; Orpen, A. G. J. Organomet.
Chem. 1983, 250, C15–C20. (j) Komatsu, H.; Suzuki, Y.; Yamazaki, H.
Chem. Lett. 2001, 998–999.
r
10.1021/ol400363f
Published on Web 02/20/2013
2013 American Chemical Society

Scheme 1. Carborhodation Pathways and Products Resulting
from Subsequent Intramolecular Cyclization5
Scheme 3. Proposed Catalytic Cycle
After addition of Rh(I)-aryl to the alkyne, the resulting
Rh(I) species can undergo cyclization onto an internal
functional group in a domino fashion.2c,7 In the case of
1,2-carbometalation, the emerging ring bears an endo-
cyclic double bond, while after 1,1-carbometalation an
exo-cyclic double bond is formed (Scheme 1).
With cyclohexadienone as an internal functional
group, a desymmetrization of this structural motif can be
achieved. Since cyclohexadienones can be easily derived
from phenols, it only takes a few steps to generate com-
plexity from simple and commercially available starting
materials.8,9
We now report the domino transformation of boronic
acids with a cyclohexadienone-tethered alkyne 1 which can
be easily obtained from p-cresol and propargyl alcohol in
moderate yields. This scaffold has been used before in
desymmetrization reactions,9 and applying it to the desired
reaction allows for the formation of a fused heterocycle,
two CÀC bonds, and two stereocenters. Application of
reaction conditions similar to those reported by Lee et al.
([Rh(cod)Cl]2, 1.1 equiv of p-methoxy phenylboronic acid
(2a), 1.5 equiv of Et3N, MeOH) led to the formation of
product 3a in 46% yield (Scheme 2). In addition, 37% of a
side product 5a was isolated. 2D-NMR, MS, and also
deuterium labeling did not help to clarify the structure of
the isolated product 3a, but finally, we were able to obtain
the X-ray structure of the 4-Br-derivative 3b, which proved
the formation of a five-membered ring (Figure 1). The
1H NMR spectrum of 5a showed that it contains not only
1 equiv of the boronic acid 2a but also 2 equiv of the starting
alkyne 1 with one of the cyclohexadienone systems intact.
The formation of both 3a and 5a can be explained
by syn-addition of the [Rh]Àaryl complex A onto the
alkyne 1 (Scheme 3). When the aryl group adds to the
terminal carbon of the starting alkyne 1 [Rh]-complex, B
is generated. Upon cyclization, B forms 3 with a five-
memberedringand anexo-cyclic double bond (see product
formation).
Addition of the aryl group to the internal carbon of the
starting alkyne 1 leads to [Rh]-complex C, which cannot
cyclize onto the cyclohexadienone moiety and instead
reacts with another alkyne 1. syn-Addition of the [Rh]
complex C takes place with opposite regioselectivity, and
the alkyl residue ends up at the terminal carbon. The
resulting rhodium-complex D can undergo cyclization
onto the cyclohexadienone moiety, which yields the side
product 5 and reforms the catalyst A.
Scheme 2. Domino Reactions of Alkyne 1 with Boronic Acid 2a
(PMP = p-methoxy phenyl)
(7) (a) Trost, B. M.; Rhee, Y. H. J. Am. Chem. Soc. 2003, 125, 7482–
7483. (b) Miura, T.; Murakami, M. Chem. Commun. 2007, 217–224.
(8) (a) Studer, A.; Schleth, F. Synlett 2005, 20, 3033–3041. (b) Liu, Q.;
Rovis, T. J. Am. Chem. Soc. 2006, 128, 2552–2553. (c) Gu, Q.; Rong,
Z.-Q.; Zheng, C.; You, S.-L. J. Am. Chem. Soc. 2010, 132, 4056–4057. (d)
Takizawa, S.; Nguyen, T. M.-N.; Grossmann, A.; Enders, D.; Sasai, H.
Angew. Chem., Int. Ed. 2012, 51, 5423–5426.
Figure 1. Crystal structure of 3b.
Due to the similarity between Lee’s and our work, the
question arises which mechanism is operative with their
(9) Tello-Aburto, R.; Harned, A. M. Org. Lett. 2009, 11, 3998–4000.
Org. Lett., Vol. 15, No. 5, 2013
1149

substrate. The absence of a vinylidene intermediate under
similar conditions has already been observed and reported
by other groups,10 who alsoexclusively obtainedexo-cyclic
double bonds. We therefore repeated the cyclization reac-
tion with compound 6 as reported by Lee et al (Scheme 4).
[Rh(coe)2Cl]2 did not catalyze the reaction well on its
own, but when norbornadiene (L1) was added, 68% of
3a and 28% 5a were observed in the crude 1H NMR.
Chiral and achiral phosphine ligands were found to be
inefficient and mainly led to decomposition of the starting
material, which is in agreement with previous findings.5,13
We next tested several chiral diene ligands (see Figure 2).
Using L2 gave 19% of 5a and predominantly the desired
product 3a in 72% yield with 76% ee. Other types of chiral
diene ligands L3ÀL5 in combination with [Rh(coe)2Cl]2
were not efficient in catalyzing the title reaction, requiring
significantly longer reaction times and giving low to no ee.
The absolute stereochemistry of the products was deter-
mined by an X-ray structure of the 4-Br-derivative 3b
(Figure 1).
Scheme 4. Domino Reaction of 65 and Oxidative Cleavage
A thorough optimization of the conditions with regard
to the amount of catalyst, ligand, base, solvent, and
boronic acid as well as the temperature and the kind of
base and solvent was undertaken, but did not lead to an
improvement in yield or ee.12 Instead, modification of the
phenyl substituent on the diene ligand proved to have a
significant effect on the ee, and a number of different diene
ligands was synthesized12 and tested (Table 1).
Since standard analytical techniques (e.g., 2D-NMR) can-
not explicitly rule out either of the possible products 7 or 8,
an oxidative cleavage was performed. Instead of a single
product containing an aldehyde and ketone, we isolated
two products, one of which was benzaldehyde and the
other had an IR-stretch at 1767 cmÀ1 and a 13C NMR
signal at 209.2 ppm which indicate the formation of
cyclobutanone 9, whose molecular mass was also con-
firmed by MS. These results support that, instead of
vinylidene formation, a syn-addition forms the interesting
and unexpected strained four-membered ring 8 with an
exo-cyclic double bond.11
Table 1. Ligand Screeninga
entry
R =
yield [%]b
ee [%]c
1
C6H5 (L2)
72
13
72
67
75
76
66
60
62
15
67
20
75
76
22
75
79
76
84
83
55
64
10
75
30
75
2
2-Me-C6H4 (L6)
3
4-Me-C6H4 (L7)
4
3-MeO-C6H4 (L8)
4-MeO-C6H4 (L9)
3,4-di-MeO-C6H3 (L10)
3,4,5-tri-MeO-C6H2 (L11)
4-NO2-C6H4 (L12)
4-CF3-C6H4 (L13)
2-CF3-C6H4 (L14)
4ÀF-C6H4 (L15)
5
6
7
8
9
10
11
12
13
1-naphthyl (L16)
2-naphthyl (L17)
a Reaction conditions: see Table 1. b Yields determined by 1H NMR
of the crude mixture. c The ee was determined by chiral HPLC.
Figure 2. Ligand optimization.
In order to achieve an enantioselective variant, a catalyst
optimization with different precursors and ligands was
performed.12 It was found that the amount of side product
strongly depends on the ligand with [Rh(cod)Cl]2, giving
37% of 5a along with 46% of 3a. The precursor
Substitution of the phenyl group in the 2-position
(entries 2, 10, and 12) leads to a decrease of activity and
selectivity. This result and the fact that diene ligands with
more than one substituent on the alkenes gave poor results
(see Figure 2) are interesting because, in contrast to previous
findings, increasing the steric bulk of the ligand is highly
detrimental to the enantioselectivity.14,15 Electron-donating
(10) (a) Miura, T.; Takahashi, Y.; Murakami, M. Org. Lett. 2007, 9,
€
5075–5077. (b) Artok, L.; Kus-, M.; Urer, B. N.; Turkmen, G.; Aksin-
€
Artok, O. Org. Biomol. Chem. 2010, 2010, 2060–2067.
(11) Chen, Y.; Lee, C. J. Am. Chem. Soc. 2012, 134, 20564–20564.
(12) For a more detailed discussion, see Supporting Information.
(13) Shintani, R.; Tsurusaki, A.; Okamoto, K.; Hayashi, T. Angew.
Chem., Int. Ed. 2005, 44, 3909–3912.
1150
Org. Lett., Vol. 15, No. 5, 2013

Table 2. Boronic Acid Scopea
entry
boronic acid R=
product
yield [%]b
ee [%]c
1
H (2c)
3c
3d
3a
3e
3f
69
67
67
67
71
58
62
58
74
52
62
49
59
71
85
88
84
88
88
82
64
82
71
62
79
66
86
78
2
2-MeO (2d)
4-MeO (2a)
2,6-di-MeO (2e)d
4-Me (2f)
3
4
5
6
2-Br (2g)
3g
3b
3h
3i
Figure 3. Boronic acid scope.
7
4-Br (2b)
8
2-Cl (2h)e
4-Cl (2i)
9
10
11
12
13
14
4-CF3 (2j)
4-CO2Me (2k)
4-CN (2l)f
4-OH (2m)
4-CHO (2n)
3j
3k
3l
3m
3n
a Reaction conditions: [Rh(coe)2Cl]2 (5 mol % Rh), L5 (5.5 mol %),
alkyne 1 (1 equiv), arylboronic acid 2 (1.1 equiv), Et3N (1.5 equiv),
methanol (0.5 M), rt, 30À120 min. b Isolated yields. c The ee was
determined by chiral HPLC. d Time: 180 min. e Time: 240min. f Time: 7 h.
Figure 4. Alkyne scope.
groups in the 3- or 4-position were found to improve the
performance, and the best ligand for the title reaction is the
3,4-dimethoxy-phenyl derivative L10, which gives a 76%
NMR-yield of 3a in 84% ee. In all cases the amount of side
product is in the range of 10À15%. With the optimized reac-
tion conditions in hand, the reaction scope with regard to the
substitution pattern on the boronic acids was investigated
using the 3,4-dimethoxy-substituted ligand L6 (Table 2).
The reaction shows a high functional group tolerance,
and electron-neutralor-richboronicacids workwell, while
electron-poor boronic acids give slightly lower yields.
Electron-rich (cf. entries 3 and 7, Table 2) and sterically
hindered boronic acids (entries 6 and 7) result in higher
ee-values than electron-poor and sterically less demanding
boronic acids. With 1-naphthylboronic acid 2o the corre-
sponding product 3o is obtained with only a 36% yield but
a 90% ee, while the sterically less demanding 2-naphthyl-
substituted cyclization product 3p is obtained in 71% yield
and 84% ee (Figure 3). Heteroaromatic boronic acids such as
3-furylboronic acid (2q), 3-thiopheneboronic acid (2r), or the
benzofused derivatives 2s and 2t yield the corresponding
products in moderate yields and high ee’s (Figure 3, 3oÀ3t).
Under the standard reaction conditions, 3- and 4-pyridyl-
boronic acids and alkylboronic acids did not react.
substituted ligand L8 worked best.12 The products 10À13
were obtained in moderate yields and with moderate to
good ee’s (Figure 4).
In conclusion, we developed a new enantioselective
Rh-catalyzed domino reaction that gives access to fused
heterocycles by desymmetrization of alkyne-tethered cy-
clohexadienones. In this process two new CÀC bonds and
two new stereocenters are formed with good enantioselec-
tivities. Wehaveevidencethat analkylidene is notinvolved
in the product formation but that the rhodium-aryl species
reacts via a syn-addition with the alkyne.
Acknowledgment. We thank Merck, the Natural Science
and Engineering Research Council (NSERC) for an Indus-
trial Research Chair, and the University of Toronto for
financial support. We also thank Dr. Alan Lough (X-ray
analysis), Dr. Darcy Burns (2D NMR-analyses), and
Dr. Harald Weinstabl from the University of Toronto
for fruitful discussions. J.K. thanks the Deutsche
Forschungsgemeinschaft (DFG) for a postdoctoral
fellowship. S.G.N. thanks NSERC for a postgraduate
scholarship.
Lastly, the cyclohexadienone was modified near the alkyne.
Preliminary experiments showed that the 3-methoxyphenyl-
Supporting Information Available. Experimental pro-
cedures, spectroscopic and analytical data for starting
materials and products, copies of NMR spectra, and
X-ray data for compound 3b. This material is available
free of charge via the Internet at http://pubs.acs.org.
(14) Gendrineau, T.; Chuzel, O.; Eijsberg, H.; Genet, J.-P.; Darses, S.
Angew. Chem., Int. Ed. 2008, 47, 7669–7672.
(15) Darses and Fillion found that the presence of ortho-substituents on
the aryl ring of the diene ligand is essential to obtain high enantioselectivities
in Rh-catalyzed 1,4-addition reactions: Reference 14; Mahoney, S. J.;
Dumas, A. M.; Fillion, E. Org. Lett. 2009, 11, 5346–5349.
The authors declare no competing financial interest.
Org. Lett., Vol. 15, No. 5, 2013
1151