Asymmetric Pentafulvene Carbometalation - Access to Enantiopure Titanocene Dichlorides of Biological Relevance

Article abstract of DOI:10.1002/anie.201508034

Unprecedented asymmetric copper-catalyzed addition of ZnEt₂ (ZnBu₂) to the exocyclic C=C bond of pentafulvenes C₅H₄(=CHAr) (Ar=2-MeOPh and related species) results in enantiomerically enriched (up to 93:7 e.r.)

Full text of DOI:10.1002/anie.201508034

Angewandte
Chemie
International Edition: DOI: 10.1002/anie.201508034
German Edition: DOI: 10.1002/ange.201508034
À
C C Coupling
Hot Paper
Asymmetric Pentafulvene Carbometalation—Access to Enantiopure
Titanocene Dichlorides of Biological Relevance
Melchior Cini,* Tracey D. Bradshaw,* Simon Woodward,* and William Lewis
Abstract: Unprecedented asymmetric copper-catalyzed addi-
just one specific example, titanocene dichloride 3, which is an
active anticancer agent in micromolar quantities,[3] and
presently known only as a mixture of stereoisomers, would
become available as single enantiomers, thus facilitating
biological screening and potentially allow clinical trials.
A limited number of stereoselective addition reactions of
organometallic reagents to pentafulvenes are known,[4] but all
these examples involve stoichiometric chiral additives, includ-
ing those of Hayashi using 1a (R = NMe2) and 120 mol%
ArLi/(À)-sparteine (63:37 to 96:4 e.r.), Mintzꢀs hydride
=
tion of ZnEt2 (ZnBu2) to the exocyclic C C bond of
=
pentafulvenes C5H4( CHAr) (Ar = 2-MeOPh and related
species) results in enantiomerically enriched (up to 93:7 e.r.)
cyclopentadienyl ligands (C5H4CHEtAr; abbreviated CpR).
Copper catalyst promotion with both chiral phosphoramidite
ligands and a phosphate additive is vital in realizing both
acceptable enantioselectivities and reaction rates. Enantiomeric
CpR TiCl2 complexes have been prepared; the (S,S) isomer is
2
twice as active towards pancreatic, breast, and colon cancer cell
lines as its (R,R) enantiomer at 24 h.
=
transfer from nBuLi to 1b (using exocyclic CPhMe moieties
and 100 mol% (À)-proline, e.r. < 59:41), Togniꢀs diasteriose-
lective addition of MeLi to the (R) enantiomer of 1c (R =
N(Me)CHcC6H11) (90:10 to 94:6 d.r.), and related work by
Otero using 1d (R = (À)-myrtenyl) (d.r. > 99:1). Apart from
these examples, only nonstereoselective or achiral addition
reactions to fulvenes have been reported (and these are
limited to Me and sp2 C-nucleophiles).[5] The lack of catalytic
methodology is due, in part, to the stability of cyclopentadie-
nide-bound kinetic products (see the putative rest state A).
We found that such rest states could apparently be trans-
metalated with Grignard reagents to allow the closure of
catalytic cycles and effective pentafulvene carbomagnesia-
tion.[3] Unfortunately, CuIL* catalysis using RMgBr provided
only racemic products in our own studies (a library of 13
ligands).[3] We predicted that, because of their higher
covalency, organozinc-derived copper catalysts would max-
imize the chances of attaining the desired enantioselective
A
symmetric copper-catalyzed 1,4-addition reactions of
organozinc reagents, especially ZnEt2, to enones (e.g.,
=
ArCH CHAc) have become commonplace in the last
10 years (Scheme 1a).[1] Although they contain an equally
powerful anion-accepting group (C5H4), the equivalent
copper-catalyzed enantioselective carbozincation of penta-
fulvenes 1 is unknown (Scheme 1b). Such methodology
could, if realized, provide rapid access to enantioenriched
substituted cyclopentadienyl ligands 2 that have many
applications in synthesis, catalysis,[2] and biology.[3] To give
carbocupration. However, the lack of any reported CuI
[1]
=
catalyst for ZnR2 enantioselective C C addition reactions
strongly suggested that intermediates related to A were very
stable and that poor catalyst turnover would have to be
overcome.
Scheme 1. Common CuI-catalyzed asymmetric 1,4-addition (a) vs.
First trials were conducted using ZnEt2 and pentafulvene
1e as an assay to determine the e.r. values of the product 2e is
greatly simplified by rapid exchange of the [1,5]-sigmatropic
a/b tautomers during GC analysis on a chiral support above
1008C. From an initial ligand library (see the Supporting
Information), phosphoramidite L1 in the presence of CuI
precursors was shown to be the highest enantioselective
lead (Table 1). As predicted, the reaction suffered from very
poor activity and conditions leading to the formation of the
Lewis acidic cuprates (runs 1 vs. 2–6) were needed for even
partial turnover. Higher loadings (runs 3 and 6) favored
significant enantioselectivity and a marginal increase in yield.
Additionally, we discovered that MTBE was the optimal
solvent and highly purified phosphoramidite L1 is required as
its degradation product L2 engenders significant ligand
acceleration[6] but with minimal enantioselectivity (run 8).
While AlR3 reagents are known to cleave phosphoramidites
unknown carbozincation (b) reactions, and applications.
[*] Dr. M. Cini, Prof. Dr. S. Woodward, Dr. W. Lewis
School of Chemistry, University of Nottingham, University Park
Nottingham NG7 2RD (UK)
E-mail: simon.woodward@nottingham.ac.uk
Dr. M. Cini, Prof. Dr. T. D. Bradshaw
School of Pharmacy, Centre for Biomolecular Sciences
University of Nottingham, University Park
Nottingham NG7 2RD (UK)
E-mail: cinimelchior@gmail.com
tracey.bradshaw@nottingham.ac.uk
Supporting information for this article (including all synthetic and
catalytic procedures, characterization data for all compounds, NMR
spectra, biological evaluations, and X-ray crystallographic data for
(R,R)- and (S,S)-3) is available on the WWW under http://dx.doi.org/
10.1002/anie.201508034.
Angew. Chem. Int. Ed. 2015, 54, 14179 –14182
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
14179

.
Angewandte
Communications
Table 1: CuI phosphoramidite promoted ZnEt2 addition to pentafulvene
1e.[a]
Table 2: Additive and ligand matching in copper-catalyzed ZnEt2 addi-
tion to pentafulvene 1e.[a]
Run Cu source
([mol%])
L*
Conditions
2e
e.r.
Entry Cu(OTf)2
([mol%])
L*
Additive
([mol%])
Yield
e.r.
[%][b] (2e)[b]
([mol]%)
2e [%][b]
(2e)[b]
([mol%])
1
2
3
4
Cu(TC)[c] (5)
Cu(OTf)2 (5)
Cu(OTf)2 (25)
Cu(OTf)2 (15)
L1 (10)
L1 (10)
L1 (50)
L1 (30)
08C, toluene
08C, toluene
08C, toluene 16
258C, tolu-
ene
2
6
50:50
58:42
71:29
64:36
1
2
3
4
5
6
7
8
9
10
20
20
20
20
20
20
20
12
10
12
(S,R,R)-L1
(40)
(S,R,R)-L1 (S)-L3 (10)
(30)
(S,R,R)-L1 (R)-L3 (10)
(30)
(S,S,S)-L1’
(40)
(S,S,S)-L1’ (S)-L3 (10)
(30)
(S,S,S)-L1’ (R)-L3 (10)
(30)
–
24
1
89:11
86:14
86:14
60:40
90:10
87:13
90:10
88:12
85:15
88:12
10
2
5
Cu(OTf)2 (15)
L1 (30)
258C,
20
88:12
MTBE[c]
–
24
34
81
6
7
8
Cu(OTf)2 (20)
Cu(OTf)2 (20)
Cu(OTf)2 (15)
L1 (40)
L1’[c] (40)
L2 (30)
258C, MTBE 24
258C, MTBE 25
258C, MTBE 76
89:11
63:37
51:49
[a] Cu source and L* in solvent (2.0 mL) stirred for 1 h followed addition
of by 1e (0.5 mmol). After 10 min, ZnEt2 (2.5 equiv) added dropwise and
the mixture stirred (16 h). [b] Yield and e.r. value determined by GC
analysis on a chiral CP-Chirasil-DEXCB column against internal standard.
[c] TC=2-thiophenecarboxylate; MTEB=2-methoxy-2-methylpropane.
L1’ is the (S,S,S)-diastereomer of L1 (see also Table 2).
(S,S,S)-L1’ (Æ)-L3 (10) 58
(30)
(S,S,S)-L1’ (S)-L3 (13)
(18)
(S,S,S)-L1’ (R)-L3 (10)
(15)
36
81
(S,S,S)-L1’ (Æ)-L3 (13) 66
in low-polarity solvents,[7] this behavior is not normally an
issue with ZnR2; we could detect no L1 degradation at the end
of 16 h runs.
(18)
[a] Cu source L* and additive in MTBE (1.2 mL) stirred for 1 h, followed
by addition of 1e (0.3 mmol) in MTBE (0.6 mL). After 10 min, ZnEt2
(2.5 equiv) added dropwise and the mixture stirred (16 h). [b] Yield and
e.r. value determined by GC analysis on a chiral CP-Chirasil-DEX CB
column against tridecane as internal standard. “Variation” in the scheme
indicates change of R/S stereochemistry.
Other phosphoramidite ligands provided a range of e.r.
values but no significant increase in activity ,and non-
phosphoramidite ligand classes were devoid of any enantio-
selectivity (see the Supporting Information). As electron-
deficient Cu(OTf)2 was our most effective precursor (its
derived cuprates are known to favor fast additions in copper
catalysis[1]), we sought a related additive that might improve
or mimic the of behavior of Cu(OTf)2. Bridging ligands are
known to play a critical role in organizing selective transition
states in asymmetric copper(I) catalysis[8] but are seldom, if
ever, modified to chiral units for copper-catalyzed asymmetric
catalysis. To our delight, use of the simple commercially
available phosphoric acid L3 had a profound effect on the rate
of the carbozincation, and, to a more limited extent, its
enantioselectivity (Table 2).
It is clear that L3 is accommodated into the catalyst and
that it affects the rate of turnover—mismatched with the
(S,R,R)-L1 catalyst, but matched with (S,S,S)-L1’ for (R)-L3
(runs 1 vs. 2–3 and 4 vs. 6 and 9). As the e.r. values were not
strongly affected, we trialed the low-cost (Æ)-L3; to our
satisfaction it could provide acceptable catalysis at lower
loadings (runs 7 vs. 10). A small library of phosphoric acid
additives were then tested (see the Supporting Information)
but (Æ)-L3 was still the best coadditive. The hypothesis that
use of (S,S,S)-L1’/(Æ)-L3 results in the formation of (À)-(S)-
2e was validated by formation of the titanocene dichloride 3
and subsequent X-ray crystallographic analysis to confirm
that the (R,R) stereoisomer was formed. We believe that both
the triflate and phosphate L3 are present in the activated
catalyst as Cu(L3)2 sources alone are ineffective when
combined with L1’. Addition of ZnBu2 to 1e (under
conditions of run 10) proceeded analogously to provide the
butyl analogue 2e’ with 90:10 e.r. As we assumed that the
methoxy group within 1e acted as a directing group to the
chiral catalyst, we tested this hypothesis using other 6-
substituted pentafulvene starting materials containing donor
groups on aryl or heteroaryl rings, using the conditions of
Table 2, run 10 (Scheme 2). Use of substrates with 5-alkyl
substituents (2 f and 2g), as well as with an ethoxy substituent
(2h) led to an increase in the enantioselectivity of the
reaction. The requirement of a proximal coordinating group
was confirmed as vital: Ar= 2-MeOPh (2e) gave 86:14 e.r.;
Ar= 3-MeOPh (2j) provided 63:37 e.r., whilst Ar= 4-
MeOPh (2n) led to racemic addition. The methoxy group in
2n is apparently too far away for favorable coordination in
the enantioselective transition state. These conclusions were
supported by thienyl (2k) which gave only a modest e.r. value,
and by related modifications of the OMe to alternative donor
groups (see the Supporting Information). Based on the
configuration determined for 2e, an (S) configuration for
the major enantiomers attained from (S,S,S)-L1 has been
tentatively assigned for 2e’–m.
The utility of the cyclopentadienes (2) was demonstrated
by the preparation of both enantiomers of 3 through
14180 www.angewandte.org
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2015, 54, 14179 –14182

Angewandte
Chemie
The stereoisomer-dependent activity of 3 can only be in
accord with hydrolysis to “Ti4+” species if such processes are
biologically mediated. Hydrolysis of the parent Cp2TiCl2 has
been proposed to be transferrin-controlled, thus fulfilling this
requirement.[10] However, recent studies have shown poor
inhibition of A549 lung cancer cell growth by Cp2TiCl2 either
in the presence or absence of transferrin or Ti-transferrin
itself.[11] Other protein chaperones might well offer
mechanism(s) for the uptake of ligated titanium species into
cells, leading to mechanisms of action dependent on initial
titanium ligation, as has been recently shown for the case of
TiCl2(C5H4CH2C6H4-4-OMe)2 versus salen-based titanium
species.[12]
In conclusion, the first examples of copper-catalyzed
asymmetric carbozincation of pentafulvenes have been dem-
onstrated. This reaction allows access to a range of enan-
tioenriched substituted cyclopentadienes and their metal
complexes. Such species are attracting increasing attention
for use in a wide range of applications.[2] While the enantio-
selectivities and yields of the present system are modest, the
use of dual phosphoamidite/phosphite copper catalysis pro-
vides a new tool for successful asymmetric carbocupration in
what has been a very fallow area.[1,13]
Scheme 2. Scope of catalytic cupration of fulvenes (1) as a function of
6-aryl unit.
complexation of (R)- and (S)-2e to TiCl4.[3] Rapid quantita-
tive deprotonation of (S)-2e by nBuLi (1.1 equiv) in Et2O at
08C led to the formation of the lithium-substituted cyclo-
pentadienide. Transmetalation with titanium tetrachloride in
THF at reflux for 16 h led to the formation of the enantioen-
riched (R,R)-3. The synthesis of (S,S)-3 was carried out in an
analogous way using (R)-2e. After recrystallization enantio-
merically pure samples of (S,S)- and (R,R)-3 were obtained
containing, at worst, traces of the achiral meso diastereomer
(see the Supporting Information). The parent titanocene
dichloride Cp2TiCl2 (Cp = C5H5) is a clinically trialed anti-
cancer agent with lower in vivo tissue toxicity than the more
commonly encountered Pt-based drugs (cisplatin, carbopla-
tin). Substituted titanocenes CpR2TiCl2 (CpR = C5H4R; R = a
wide range of substituents) are much more cytotoxic than the
parent,[5a,b, 9] but the mechanism(s) of action of these agents
remains poorly defined—excessive cellular uptake of Cp-free
“Ti4+” being the most often cited proposal.[10] The antiproli-
ferative activities of the enantiopure titanocenes (R,R)-3,
(S,S)-3, in comparison to the stereoisomeric mixture (rac/
meso)-3, and cisplatin were evaluated in vitro at 24 h against
the carcinoma cell lines: HCT-116 (colorectal), MiaPaCa-2
(pancreatic), and MDA-MB-468 (breast). A 24 h time period
was selected as real-time microscopy studies indicate cancer
cell death reached a maximum by 4–6 h, and activity was
moderated after 24 h. The values given in Table 3 show that
Acknowledgements
We thank the University of Nottingham for a studentship and
the support of the Edith Johnson Bequest (M. Cini). We thank
A. Kuruppu and D. McLean for their assistance with the
biological studies.
Keywords: asymmetric catalysis · carbometalation ·
À
C C coupling · copper · ligand effects
How to cite: Angew. Chem. Int. Ed. 2015, 54, 14179–14182
Angew. Chem. 2015, 127, 14385–14388
[1] Overview of copper-catalyzed asymmetric processes: Copper-
Catalyzed Asymmetric Synthesis (Eds.: A. Alexakis, N. Krause,
S. Woodward), Wiley-VCH, Weinheim, 2014, Chapters 2 and 10.
[2] Overviews of substituted chiral cyclopentadienyls: a) R. L.
Halterman, Chem. Rev. 1992, 92, 965 – 994; b) H. H. Brintzinger,
D. Fisher, R. Mulhaupt, B. Reiger, R. M. Waymouth, Angew.
Chem. Int. Ed. Engl. 1995, 34, 1143 – 1170; Angew. Chem. 1995,
107, 1255 – 1283; c) O. Vassylyev, A. Panarello, J. G. Khinast,
Molecules 2005, 10, 587 – 619; d) J.-B. Sortais, C. Darcel,
ChemCatChem 2013, 5, 1067 – 1068.
Table 3: Stereoisomer-dependent in vitro growth inhibition by 3 vs.
cisplatin.[a]
Cell line
(R,R)-(3)
(S,S)-(3)
(rac/meso)-(3)[b] cisplatin
Mia PaCa-2
42.5Æ0.5 19.9Æ3.5 32.3Æ2.8
35.4Æ1.4
7.4Æ2.8
34.9Æ3.0
MDA-MB-468 22.8Æ2.8 11.5Æ1.2 23.6Æ0.4
HCT-116
31.3Æ1.6 14.8Æ1.7 31.6Æ3.9
[3] M. Cini, T. D. Bradshaw, W. Lewis, S. Woodward, Eur. J. Org.
Chem. 2013, 3997 – 4007.
[a] Cell growth inhibition after 24 determined by MTTassay (GI50 in mm);
GI50 values are represented as mean standard error of mean (SEM) of at
least three independent experiments (n=4 per experiment). [b] 1:1:1:1
mixture of R,R/S,S/R,S/S,R.
[4] a) T. Suzuka, M. Ogasawara, T. Hayashi, J. Org. Chem. 2002, 67,
3355 – 3359; b) H. C. L. Abbenhuis, U. Burckhardt, V. Gramlich,
A. Togni, A. Albinati, B. Müllers, Organometallics 1994, 13,
4481 – 4493; c) E. A. Mintz, PCT Int. Appl., 1992,
WO1992012112A1; d) M. Honrado, A. Otero, J. Fernµndez-
Baeza, L. F. Sµnchez-Barba, A. Lara-Sµnchez, J. Tejeda, M. P.
Carrión, J. Martínez-Ferrer, A. GarcØs, A. M. Rodríguez,
Organometallics 2013, 32, 3437 – 3440.
[5] Nonstereoselective additions to pentafulvenes: a) M. Hogan, J.
Cotter, J. Claffey, B. Gleeson, D. Wallis, D. OꢀShea, M. Tacke,
Helv. Chim. Acta 2008, 91, 1787 – 1797; b) M. Hogan, J. Claffey,
C. Pampillón, R. W. G. Watson, M. Tacke, Organometallics 2007,
statistical (P < 0.05) differential biological activity is observed
in cancer cell lines for the stereoisomers of 3. Additionally,
compared to cisplatin, (S,S)-3 shows more than twice the
activity against colon carcinoma, and almost twice the activity
in pancreatic carcinoma. On the other hand, cisplatin is more
active than (S,S)-3 in breast carcinoma.
Angew. Chem. Int. Ed. 2015, 54, 14179 –14182
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 14181

.
Angewandte
Communications
26, 2501 – 2506; c) E. Kirillov, N. Marquet, M. Bader, A. Razavi,
V. Belia, F. Hampel, T. Roisnel, J. A. Gladysz, J.-F. Carpentier,
Organometallics 2011, 30, 263 – 272; d) G. E. Herberich, U.
Englert, T. Wirth, Eur. J. Inorg. Chem. 2005, 4924 – 4935.
Typically, C-nucleophiles are limited to methyl or sp2 aryl/
heteroaryl systems; with alkyl (ꢀ C2) organometallics, extensive
b-hydride transfer results (e.g. ref. [4c]). This is not the case in
our system.
[11] A. D. Tinoco, H. R. Thomas, C. D. Incarvito, A. Saghatelian,
A. M. Valentine, Proc. Natl. Acad. Sci. USA 2012, 109, 5016 –
5021.
[12] a) J. Schur, C. M. Manna, A. Deally, R. W. Kçster, M. Tacke,
E. Y. Tshuvab, I. Ott, Chem. Commun. 2013, 49, 4785 – 4787;
b) E. Y. Tshuva, J. A. Ashenhurst, Eur. J. Inorg. Chem. 2009,
2203 – 2218.
=
[13] Even in the widest sense (carbometalation of any prochiral C C
[6] D. J. Berrisford, C. Bolm, K. B. Sharpless, Angew. Chem. Int. Ed.
Engl. 1995, 34, 1059 – 1070; Angew. Chem. 1995, 107, 1159 – 1171.
[7] C. Bournaud, C. Falciola, T. Lecourt, S. Rosset, A. Alexakis, L.
Micouin, Org. Lett. 2006, 8, 3581 – 3584.
[8] a) S. Woodward in Copper-Catalyzed Asymmetric Synthesis
(Eds.: A. Alexakis, N. Krause, S. Woodward), Wiley-VCH,
Weinheim, 2014, Chapter 1; b) P. -F. Larsson, P. -O. Norrby, S.
Woodward in Copper-Catalyzed Asymmetric Synthesis (Eds.: A.
Alexakis, N. Krause, S. Woodward), Wiley-VCH, Weinheim,
2014, Chapter 12.
unit), successful asymmetric catalytic systems are rare and
frequently require use of reactive endocyclic alkenes: a) M.
Nakamura, A. Hirai, E. Nakamura, J. Am. Chem. Soc. 2000, 122,
978 – 979; b) K. Kramer, P. Leong, M. Lautens, Org. Lett. 2011,
13, 819 – 821; c) M. Lautens, S. Hiebert, J. Am. Chem. Soc. 2004,
126, 1437 – 1447. Asymmetric alkene carboalumination (85:15 to
98:2 e.r.) is the most developed reaction: d) E. Negishi,
ARKIVOC 2011, 34 – 53.
[9] U. Olszewski, A. Deally, M. Tacke, G. Hamilton, Anti-Cancer
Agents Med. Chem. 2010, 10, 302 – 311.
[10] K. M. Buettner, A. M. Valentine, Chem. Rev. 2012, 112, 1863 –
1881.
Received: August 27, 2015
Published online: October 12, 2015
14182 www.angewandte.org
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2015, 54, 14179 –14182