Asymmetric synthesis of QUINAP via dynamic kinetic resolution

Article abstract of DOI:10.1021/ja409383f

A palladium-catalyzed, atroposelective C-P coupling process has been developed for the asymmetric synthesis of QUINAP and its derivatives in high enantiomeric excess. Bromide, triflate (OTf) and 4- methanesulfonylbenzenesulfonate (OSs) precursors were stu

Full text of DOI:10.1021/ja409383f

Communication
pubs.acs.org/JACS
Asymmetric Synthesis of QUINAP via Dynamic Kinetic Resolution
Vikram Bhat, Su Wang, Brian M. Stoltz,* and Scott C. Virgil*
Caltech Center for Catalysis and Chemical Synthesis, Warren and Katharine Schlinger Laboratory for Chemistry and Chemical
Engineering, Division of Chemistry and Chemical Engineering, California Institute of Technology, Pasadena, California, 91125,
United States
S
* Supporting Information
QUINAP. Moreover, the recovered enriched precursor 2 could
ABSTRACT: A palladium-catalyzed, atroposelective C−P
coupling process has been developed for the asymmetric
synthesis of QUINAP and its derivatives in high
enantiomeric excess. Bromide, triflate (OTf) and 4-
methanesulfonylbenzenesulfonate (OSs) precursors were
studied, leading in the case of the triflate to a novel
dynamic kinetic resolution involving isomerization of an
arylpalladium intermediate. The operationally simple
methods described in this communication afford these
important ligands in good to high yields and selectivity
using low catalyst loading (≤3 mol % Pd).
be used for further manipulations. Alternatively, if the substrate
(2) could undergo racemization under typical reaction
conditions while one of the enantiomers selectively underwent
phosphination, asymmetric synthesis of QUINAP would be
achieved via dynamic kinetic resolution (DKR).8 Finally, if the
substrate racemization is too slow to achieve a standard DKR,
isomerization of the arylpalladium intermediate (3) along the
reaction pathway could prove to be a unique application of DKR
in the asymmetric synthesis of QUINAP.
Preliminary phosphination experiments involved the treat-
ment of bromide 2a with commercially available chiral
bis(phosphine) ligands and Pd[P(o-tol)3]2. We were delighted
to find that at 5 mol % Pd(0), 10 mol % ligand, and 1.5 equiv of
diphenylphosphine, several ligands afforded high levels of
enantioenrichment for both the recovered bromide 2a as well
as the product QUINAP (1a), suggesting high selectivity factors9
for the desired reaction (Table 1). Early results indicated that
bidentate ligands bearing dialkylphosphino groups were the best
choices for this reaction and proceeded at 80−90 °C (entries 4
and 8) while the less electron-rich ligands having diary-
s one of the most useful chiral P,N-ligands in asymmetric
1
Acatalysis, QUINAP (1a) has found applications in several
enantioselective reactions.2 However, after many years since its
discovery, synthetic methods for securing QUINAP in high
enantiomeric excess still involve costly resolving agents1a,b or
extended synthetic procedures.3,4 We were interested in
developing concise enantioselective methods that would provide
QUINAP in optically pure form and be amenable to the synthesis
of derivatives so as to support further advances in the field of P,N-
ligand mediated catalysis.
a
Table 1. Kinetic Resolution of Bromide 2a
Our strategy for the enantioselective synthesis of QUINAP
(1a) involved investigation of atroposelective reactions of the
known bromide3 and triflate1a precursors (2a and 2b) with
diphenylphosphine in the presence of commercially available
chiral bis(phosphine) ligands and a suitable palladium precursor
(Scheme 1).5,6 We hypothesized that in the case where neither
the substrate (2) nor the intermediate (3) were subject to
isomerization under the reaction conditions and one of the two
atropisomers of 2 reacted faster than the other, a kinetic
resolution7 pathway would be feasible for the synthesis of
Scheme 1. Kinetic Resolution and DKR Syntheses of
QUINAP (1a) (2a−c, 3a−c: X = Br, OTf, OSs)
a
Reactions were performed with indicated amounts of Pd and ligand,
b
c
2.0 equiv Ph2PH, and 4.0 equiv of DIPEA. See ref 10. Determined
d
by UHPLC-MS analysis. Determined by chiral SFC methods.
e
f
Negative ee values represent (S)-2a and (R)-1a predominating. See
g
ref 9. 10 mol % Pd and 20 mol % ligand was required. NR = no
reaction; ND = not determined.
Received: September 12, 2013
Published: October 23, 2013
© 2013 American Chemical Society
16829
dx.doi.org/10.1021/ja409383f | J. Am. Chem. Soc. 2013, 135, 16829−16832

Journal of the American Chemical Society
Communication
lphosphino groups were often weakly reactive and required
higher temperature (entries 1−2). Moreover, at elevated
temperatures, side reactions including hydrodebromination of
2a led to lower yields and were therefore not further pursued.
Ligand (S,S)-4 displayed unique reactivity for this reaction
(Table 1, entry 4) and improved dramatically both in rate and
selectivity upon the addition of tetra-n-butylammonium
bisulfate.11 Thus, under the optimized conditions (Table 2,
the optically pure bromide (R)-2a produced by the aforemen-
tioned kinetic resolution. In practice, we found that achiral or
racemic ligands suffered low stereofidelity in the C−P coupling
reaction with (R)-2a (Table 3, entries 1−2). Only (R,R)-4 was
effective in producing (R)-1a in 99% ee and also offered an
improvement in the preparation of derivatives (R)-1c and (R)-1d
(entries 3−5).
a
Table 3. C−P Coupling with Resolved (R)-2a
a
Table 2. Preparative Kinetic Resolution of 2a
a
Reactions were performed under the conditions used in Table 2,
b
c
d
entry 1. See ref 10. Determined by chiral SFC methods. 3.0 mol %
Pd and 4.5 mol % (R,R)-4 was used.
a
Reactions were performed with 1.0 equiv nBu4NHSO4, 1.5 equiv
b
Ar2PH, and 4.0 equiv of DIPEA. Determined by UHPLC-MS
analysis. ee determined by chiral SFC methods. See ref 9. After
recrystallization. 1b−c were found to be foams and could not be
further enriched via recrystallization. 3.0 mol % Pd and 4.5 mol %
(S,S)-4 was used; 1d and 2a were inseparable by chromatography and
crystallization methods.
Encouraged by the efficiency of the kinetic resolution of 2a at
70 °C terminating near 50% conversion, we were eager to
investigate the possibility of developing a DKR by extending the
reaction time and/or elevating the temperature. Upon measure-
ment of racemization rates for the bromide 2a and the product
1a, it was apparent that the bromide 2a required a higher
temperature for racemization than the product QUINAP (1a)
(Table 4, entries 1−2).14 Further analysis of the racemization
c
d
e
f
g
entry 1),12 a preparative gram-scale reaction of racemic bromide
2a required only 0.5 mol % Pd(0) catalyst and afforded (S)-1a
(95% ee) with recovered bromide (R)-2a (96% ee), both of
which were easily recrystallized to >99.5% ee. A selectivity factor
of 154 is consistent with these results.
Table 4. Racemization Rates of 1a and 2a−c
We believe that this highly selective kinetic resolution operates
by an atroposelective oxidative addition of the aryl bromide to
the C2-symmetric bis(phosphine)-Pd(0) complex.13 In the case
of (S,S)-4, the (S)-atropisomer of 2a would be the preferred
antipode to generate an arylpalladium bromide complex with
minimal steric interactions between the aryl group and ligand as
shown in Figure 1.
rates of sulfonate esters revealed that the 4-methanesulfonyl-
benzenesulfonate15 derivative 2c offered some opportunity for
the DKR synthesis of 1a.
In contrast to the unreactivity of tosylate and mesylate
derivatives in this reaction, we were pleased to find that sosylate
2c reacted similarly to bromide 2a in the standard kinetic
resolution with (R,R)-4 at 80 °C (Table 5, entry 1). However,
securing a useful DKR with 2c was found to be very difficult.
Reaction of sosylate 2c in dioxane at 90 °C required 4 days for
complete consumption of the starting material. Upon isolation
Figure 1. Proposed intermediate 3a from bromide (S)-2a and Pd
catalyst with (S,S)-4.
a
Table 5. Kinetic Resolution and DKR with 2c
Application of these conditions to the synthesis of QUINAP
derivatives 1b−d afforded less favorable outcomes (Table 2,
entries 2−4). Although bis(p-tolyl)phosphine (entry 2)
produced results similar to entry 1, bis(p-trifluoromethyl-
phenyl)phosphine reacted with lower selectivity (entry 3).
Additionally, neither of the two derivatives (1b−c) was found to
be crystalline and hence could not be further enriched. Despite
the higher catalyst loading required in the case of bis(o-
tolyl)phosphine, the corresponding QUINAP derivative 1d was
obtained with poor selectivity (entry 4).
a
Reactions were performed with Pd to (R,R)-4 ratio = 1:1.5, 1.5 equiv
b
of Ph2PH, and 4.0 equiv of DMAP. Determined by UHPLC-MS
analysis. ee determined by chiral SFC methods. ND = not
determined.
c
We therefore hoped that access to the derivatives 1b−d in
higher enantiomeric excess would be achievable by coupling to
16830
dx.doi.org/10.1021/ja409383f | J. Am. Chem. Soc. 2013, 135, 16829−16832

Journal of the American Chemical Society
Communication
a
and purification, (R)-1a was obtained with only 43% yield and
56% ee (Table 5, entry 2). We observed that the extended
reaction time at 90 °C led to nucleophilic deprotection of the
sulfonate group of 2c, resulting in lower yield.16 The
enantioselectivity of the C−P coupling suffered erosion and
the product 1a was subject to racemization at 90 °C.
Nevertheless, we hope that the as yet unreported reactivity of
sosylate esters such as 2c in palladium-catalyzed cross-coupling
reactions might assist other researchers in cases where a need for
an alternative to the triflate ester arises.
Table 7. C−P Coupling with Resolved Triflate 2b
a
Reactions were carried out using 2.0 mol % Pd[(o-tol)3P]2, 3.0 mol %
(R,SFc)-5, 2.0 equiv Ph2PH and 4.0 equiv of DMAP in dioxane.
Determined by UHPLC-MS analysis. Determined by chiral SFC
methods.
b
c
Initial studies with the triflate 2b using several commercially
available ligands afforded poor to moderate enantioselectivities
and conversions (Table 6). Once again, as seen for 2a, electron-
The rapid reaction rates observed with (R,SFc)-5 at 80 °C
requires that the lifetime of the arylpalladium triflate
intermediate would necessarily be short. This indicates that a
dramatically accelerated rate of isomerization (>4600-fold
relative to racemization of 2b) is operative during the lifetime
of the arylpalladium intermediate. The labile nature of the triflate
group in intermediate 3b appears to facilitate the isomerization
strongly favoring one diastereomeric structure (Scheme 2).
a
Table 6. DKR with Triflate ( )-2b
Scheme 2. DKR via Isomerization of Arylpalladium
Intermediate
a
Reactions were performed with 1.5 equiv of Ph2PH, and 4.0 equiv of
b
c
DMAP. See ref 10. Determined by UHPLC-MS analysis.
d
e
Determined by chiral SFC methods. Negative ee values represent
f
g
(R)-1a predominating. Reaction was complete within 1 h. The
reaction was performed with 1 equiv of Bu4NBr. NR = no reaction.
n
rich bidentate phosphine ligands, especially ferrocenyldialkl-
phosphine ligands, showed remarkable reactivity for this reaction
(entries 5−11). Ligand (S,S)-4 proved to be a highly active
catalyst for this substitution reaction but afforded QUINAP (1a)
with 0% ee (entry 5). Neither decreasing the temperature nor
selection of bulkier variants of ligand (S,S)-4 offered significant
improvement (entry 6). Among the several Josiphos ligands that
were tested, the enantiomeric excesses of the product 1a were
uniformly low (entries 7−9), except for SL-J003-1 [(R,SFc)-5]
which has both phosphines bearing cyclohexyl groups (entry 10).
The C−P coupling reaction with (R,SFc)-5 proceeded with
complete conversion affording (S)-1a in 60% ee. Interestingly,
addition of tetra-n-butylammonium bromide (TBAB) to the
reaction had the effect of reducing the selectivity to 8% ee (entry
11).
Given the slow racemization rate of 2b at 80 °C in dioxane
(Table 4, entry 7) and the fact that the reaction was complete in
only 1 h (Table 6, entry 10), we believe that the reaction
conditions are mediating a rapid isomerization somewhere along
the reaction course. In order to learn more about the nature of
this remarkable, albeit modest, enantioselectivity, the reaction
was conducted using Josiphos (R,SFc)-5 and the resolved triflate
isomers (S)-2b and (R)-2b at 40 and 80 °C. Gratifyingly, these
experiments verified the high stereofidelity of the matched
reaction (Table 7, entries 1 and 3) and the stereocorrecting
nature of the unmatched reactions without racemization of the
starting material (entries 2 and 4). In contrast to the kinetic
resolution of bromide 2a, the comparable rates of the two
isomers of triflate 2b in Table 7 indicate a nonselective oxidative
addition by the catalyst system with (R,SFc)-5.
Indeed, this ion-accelerated effect is supported by the fact that
the observed enantiomeric excess of the product 1a is lowered in
the presence of added TBAB (Table 6, entry 11). In order to
draw light to the possible origin of this favorable isomerization,
we considered the recent work of Hartwig that demonstrated
significant agostic interactions present in T-shaped
arylpalladium(II) halide complexes using structures solved by
X-ray diffraction.17 In addition to noticing shorter M−H
distances for the more electrophilic metal centers, Hartwig and
co-workers postulated that the formally trivalent triflate complex
could be depicted in a square planar geometry with an agostic
M−H interaction occupying the otherwise vacant fourth
coordination site.18 It is likely that these effects are at play in
our system also and can explain the favorable transition state for
the isomerization of intermediate (RBiaryl)-3b where a stabilizing
agostic interaction develops as the quinoline ring peri-hydrogen
passes the cationic palladium atom in the chelated structure 9.
Our observations in this intriguing DKR presumably point to a
kinetic manifestation of the agostic interactions of palladium
intermediates.
Based on this mechanistic postulate, the overall effectiveness of
the DKR was greatly improved by allowing more time for the
isomerization of (RBiaryl)-3b to proceed before its subsequent
reaction with diphenylphosphine. This was achieved by slow
addition of diphenylphosphine over 4 h to the reaction mixture
containing triflate ( )-2b, 3.0 mol % Pd(0) and 4.5 mol %
(R,SFc)-5 to obtain 86% yield of (S)-1a with 90% ee (Scheme 3).
In conclusion, an atroposelective kinetic resolution and DKR
strategy has been developed for the asymmetric synthesis of
QUINAP (1a). Furthermore, the kinetic resolution pathway
16831
dx.doi.org/10.1021/ja409383f | J. Am. Chem. Soc. 2013, 135, 16829−16832

Journal of the American Chemical Society
Communication
(5) Recent reviews on transition-metal-catalyzed C−P coupling
reactions: (a) Li, Y.-M.; Yang, S.-D. Synlett 2013, 24, 1739.
(b) Tappe, F. M. J.; Trepohl, V. T.; Oestreich, M. Synthesis 2010,
3037. (c) Schwan, A. L. Chem. Soc. Rev. 2004, 33, 218. Reviews on
asymmetric synthesis of phosphines: (d) Kolodiazhnyi, O. I.
Tetrahedron: Asymmetry 2012, 23, 1. (e) Glueck, D. S. Chem.Eur. J.
2008, 14, 7108. (f) Grabulosa, A.; Granell, J.; Muller, G. Coord. Chem.
Rev. 2007, 251, 25.
Scheme 3. Optimized DKR of Triflate 2b
(6) Recent reviews on atroposelective synthesis of axially chiral biaryl
compounds: (a) Bringmann, G.; Mortimer, A. J. P.; Keller, P. A.;
Gresser, M. J.; Garner, J.; Breuning, M. Angew. Chem., Int. Ed. 2005, 44,
5384. (b) Baudoin, O. Eur. J. Org. Chem. 2005, 4223.
(7) Discussions on kinetic resolution: (a) Kagan, H. B.; Fiaud, J. C. In
Topics in Stereochemistry; Eliel, E. L., Ed.; Wiley: NewYork, 1988; Vol.
18, p 249; (b) Keith, J. M.; Larrow, J. F.; Jacobsen, E. N. Adv. Synth.
Catal. 2001, 343, 5. (c) Vedejs, E.; Jure, M. Angew. Chem., Int. Ed. 2005,
44, 3974.
provides convenient access to enriched bromide 2a and sosylate
2c precursors. Our future endeavors in this area will investigate
the structural nature of intermediates along the reaction course of
the arylpalladium isomerization and explore the scope of
atroposelective C−X coupling reactions for the synthesis of
related P,N- and N,N-ligands.
(8) (a) Extensive review on DKR: Pellissier, H. In Chirality from
Dynamic Kinetic Resolution; Royal Society of Chemistry: Cambridge,
UK, 2011; (b) Huerta, F. F.; Minidis, A. B. E.; Backvall, J.-E. Chem. Soc.
Rev. 2001, 30, 321. (c) Recent example of atroposelective DKR:
Gustafson, J. L.; Lim, D.; Miller, S. J. Science 2010, 328, 1251. (d) While
the present manuscript was under review for publication, a report
appeared on a C(sp2)−C(sp2) cross-coupling DKR strategy for the
ASSOCIATED CONTENT
* Supporting Information
Experimental procedures and characterization data. This material
is available free of charge via the Internet at http://pubs.acs.org.
■
S
synthesis of axially chiral scaffolds: Ros, A.; Estepa, B.; Ramírez-Lop
́
ez,
AUTHOR INFORMATION
Corresponding Authors
stoltz@caltech.edu
́
■
P.; Alvarez, E.; Fernan
́
dez, R.; Lassaletta, J. M. J. Am. Chem. Soc. 2013,
135, 15730.
(9) Selectivity factor, s, was calculated from ee of 2a and 1a using the
equation: s = kfast/kslow = ln[(1 − c)(1 − eesm)]/ln[(1 − c)(1 + eesm)]
where the value of c was taken as c = eesm/(eesm + eepdt). See ref 7a.
(10) See SI for more information on ligands used in this study.
(11) Among the several nBu4N+ salts that were surveyed as additives,
bromide and bisulfate salts were found to be most effective.
(12) DIPEA, DMAP, and DABCO were superior bases than DBU,
pyridine, 2,6-lutidine, N-methylmorpholine, and Cs2CO3. 1a itself was
found to be a poor ligand for this reaction both in terms of rate and
selectivity. Pd[P(o-tol)3]2 alone gave no background reaction.
(13) Examples of atroposelective oxidative addition: (a) Hayashi, T.;
Niizuma, S.; Kamikawa, T.; Suzuki, N.; Uozumi, Y. J. Am. Chem. Soc.
1995, 117, 9101. (b) Cho, Y.-H.; Kina, A.; Shimada, T.; Hayashi, T. J.
Org. Chem. 2004, 69, 3811. (c) Atroposelective activation of C−H
bonds: Kakiuchi, F.; Le Gendre, P.; Yamada, A.; Ohtaki, H.; Murai, S.
Tetrahedron: Asymmetry 2000, 11, 2647.
(14) See SI for a detailed study on racemization rates of 1a and 2a−c.
(15) We have assigned the name sosylate (abbreviated as −OSs) to
denote the 4-methanesulfonylbenzenesulfonate group. To the best of
our knowledge, this group has not previously been used in transition-
metal-catalyzed reactions. Sosylate 2c was prepared using commercially
available 4-methansulfonylbenzenesulfonyl chloride.
(16) The DKR reaction of 2c at 90 °C required rigorous exclusion of
water. Replacing DMAP with less nucleophilic bases, like DIPEA, or the
addition of 4Å MS offered no discernible improvement.
svirgil@caltech.edu
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank the Gordon and Betty Moore Foundation and Caltech
for financial support and Prof. Robert H. Grubbs, Prof. Sarah E.
Reisman, and Prof. Gregory C. Fu for helpful discussions. Dr.
Jacob Y. Cha, Mr. Andrew D. Lim, and Mr. Boram D. Hong are
acknowledged for contributing samples of 2a and 2b.
REFERENCES
■
(1) (a) Alcock, N. W.; Brown, J. M.; Hulmes, D. I. Tetrahedron:
Asymmetry 1993, 4, 743. (b) Lim, C. W.; Tissot, O.; Mattison, A.;
Hooper, M. W.; Brown, J. M.; Cowley, A. R.; Hulmes, D. I.; Blacker, A. J.
Org. Process Res. Dev. 2003, 7, 379. For QUINAP derivatives: (c) Valk, J.
M.; Claridge, T. D. W.; Brown, J. M. Tetrahedron: Asymmetry 1995, 6,
2597. (d) Doucet, H.; Brown, J. M. Tetrahedron: Asymmetry 1997, 8,
3775. (e) Chapoulaud, V. G.; Audoux, J.; Ple,
́
H.; Turck, A.; Queg
́
uiner,
G. Tetrahedron Lett. 1999, 40, 9005. (f) Knopfel, T. F.; Aschwanden, P.;
̈
Ichikawa, T.; Watanabe, T.; Carreira, E. M. Angew. Chem., Int. Ed. 2004,
43, 5971. (g) Kloetzing, R. J.; Knochel, P. Tetrahedron: Asymmetry 2006,
17, 116. (h) Fekner, T.; Muller-Bunz, H.; Guiry, P. J. Org. Lett. 2006, 8,
̈
(17) (a) Stambuli, J. P.; Incarvito, C. D.; Buhl, M.; Hartwig, J. F. J. Am.
̈
5109. (i) Feng, J.; Dastgir, S.; Li, C. J. Tetrahedron Lett. 2008, 49, 668.
(2) Recent applications of QUINAP (1a) in asymmetric catalysis:
(a) Lim, A. D.; Codelli, J. A.; Reisman, S. E. Chem. Sci. 2013, 4, 650.
(b) Miura, T.; Yamauchi, M.; Kosaka, A.; Murakami, M. Angew. Chem.,
Int. Ed. 2010, 49, 4955. (c) Li, X.; Kong, L.; Gao, Y.; Wang, X.
Tetrahedron Lett. 2007, 48, 3915. (d) Taylor, A. M.; Schreiber, S. L. Org.
Lett. 2006, 8, 143. (e) Gommermann, N.; Knochel, P. Chem.Eur. J.
2006, 12, 4380. (f) Trudeau, S.; Morgan, J. B.; Shrestha, M.; Morken, J.
P. J. Org. Chem. 2005, 70, 9538. (g) Black, A.; Brown, J. M.; Pichon, C.
Chem. Commun. 2005, 5284. (h) Daura-Oller, E.; Segarra, A. M.; Poblet,
Chem. Soc. 2004, 126, 1184. (b) Roy, A. H.; Hartwig, J. F.
Organometallics 2004, 23, 194. (c) Shen, Q.; Ogata, T.; Hartwig, J. F.
J. Am. Chem. Soc. 2008, 130, 6586.
(18) For a related structural study, see: Walter, M. D.; Moorhouse, R.
A.; Urbin, S. A.; White, P. S.; Brookhart, M. J. Am. Chem. Soc. 2009, 131,
9055.
́
J. M.; Claver, C.; Fernandez, E.; Bo, C. J. Org. Chem. 2004, 69, 2669.
(i) Maeda, K.; Brown, J. M. Chem. Commun. 2002, 310. (j) Faller, J. W.;
Grimmond, B. J. Organometallics 2001, 20, 2454.
(3) Thaler, T.; Geittner, F.; Knochel, P. Synlett 2007, 2655.
(4) Formal asymmetric synthesis of 1a: Clayden, J.; Fletcher, S. P.;
McDouall, J. J. W.; Rowbottom, S. J. M. J. Am. Chem. Soc. 2009, 131,
5331.
16832
dx.doi.org/10.1021/ja409383f | J. Am. Chem. Soc. 2013, 135, 16829−16832