α,n-didehydrotoluenes by photoactivation of (chlorobenzyl) trimethylsilanes: An alternative to enyne-allenes cyclization

Article abstract of DOI:10.1002/anie.201202794

Doubly radical: A novel entry to ?,n-didehydrotoluene (DHT) diradicals is disclosed and proceeds through the photochemical activation of (chlorobenzyl)trimethylsilanes with chloride loss and elimination of the SiMe3+ group (see scheme). The products formed in solution are indicative of the intermediacy of the three isomers of the ?,n-DHT

Full text of DOI:10.1002/anie.201202794

Angewandte
Chemie
DOI: 10.1002/anie.201202794
Radicals
a,n-Didehydrotoluenes by Photoactivation of (Chlorobenzyl)-
trimethylsilanes: An Alternative to Enyne–Allenes Cyclization**
Stefano Protti, Davide Ravelli, Barbara Mannucci, Angelo Albini, and Maurizio Fagnoni*
Chemotherapeutic agents must be able to kill tumor cells
while being inactive against healthy cells. Typically, a prodrug
is activated in situ by some mild mechanism and generates an
aggressive intermediate that may for example, abstract
hydrogen atoms from DNA and induce DNA cleavage and
subsequent apoptosis. Among the few chemical classes that
come close to this paradigm are highly unsaturated hydro-
carbons, such as enediynes and enyne–allenes which are
present in some natural compounds known as antibiotics (e.g.
Neocarzinostatin or Dynemicin).[1–4] These moieties cyclize
under physiological conditions by converting two p bonds
into a s bond through the Bergman[5] and the Myers–Saito[5–7]
reaction to give a p-benzyne and a a,3-didehydrotoluene (a,3-
DHT), respectively.[8]
Scheme 1. Thermal generation and reactivity of a,3-DHT in a MeOH/
water mixture.
The chemistry of heterosymmetric (p1s1) a,3-DHT di-
radicals has been explored on simple models.[5,9] Thus, when
the enyne–allene A was heated at 1008C in aqueous
methanol, products from a diradical path (phenethyl alcohol
and 1,2-diphenylethane) and from a polar path (methyl
benzyl ether and benzyl alcohol; Scheme 1) were formed.[6]
The product distribution changed in deuterated solvents. For
example, in CD3OH [D3]methyl benzyl ether became the
exclusive product (70% yield). The detailed mechanism and
the dual polar/diradical reactivity of the DHT intermediate
are still discussed.[10] Models that mimic cyclization to DHT
diradicals[11,12] are limited to a,3-DHT. The a,3-DHT is the
only isomer accessible by the Myers–Saito cyclization through
the C2-C7 path (Scheme 1), which is in competition with the
C2-C6 cyclization (discovered by Schmittel).[13] Light-induced
formation of a,3-DHT and related diradicals has been
likewise reported.[14]
structures of this type, and potentially to all of the three
isomeric a,n-DHTs. There is sparse indication of the gen-
eration of such intermediates from halotoluenes. These
findings include the collision-induced halide elimination
from gas-phase halobenzyl anions in a flow afterglow-triple
quadrupole reported by Squires and co-workers,[15] the EPR
characterization of a,2- and a,4-DHTs by photolysis in
a matrix of the corresponding iodobenzyl iodides reported
by Sander and co-workers,[16] and possibly the formation of
a a,2-DHT by thermolysis of some phthalide derivatives.[17]
However, to the best of our knowledge, the generation of
DHTs from aromatic compounds in solution has no prece-
dent.
The rationale behind our choice is that electron-rich aryl
chlorides have been shown to undergo heterolytic cleavage of
It occurred to us that a double elimination process from
toluene derivatives would offer a simple alternative entry to
ꢀ
the Ar Cl bond upon irradiation. Phenyl cations (in the
3
triplet state, Ar+ having a p5s1 structure) are thus smoothly
generated (Scheme 2a).[18,19] The efficiency depends on the
nature of the substituents more than on their relative position
with respect to the nucleofugal group, and is maximal when
a strong electron-donating group (EDG; for example, NMe2
and OR) is present. In neat solvents triplet phenyl cations are
mainly reduced to the corresponding aromatic hydrocar-
bons,[20] whereas in the presence of p-bond nucleophiles
(NuE) arylation takes place with the concomitant release of
the E+ group (Scheme 2a).
We reasoned that introducing an electrofugal group in the
benzylic position would cause a second cleavage in the cation
and generate the same DHTs which are formed from the
enyne–allenes cycloaromatization (Scheme 2b). (Chloroben-
zyl)trimethylsilanes (1–3) appeared to be suitable models for
testing the hypothesis, since the CH2SiMe3 group facilitates
initial Clꢀ loss (as an electron donor, comparable to
[*] Dr. S. Protti, Dr. D. Ravelli, Prof. A. Albini, Prof. M. Fagnoni
PhotoGreen Lab, Department of Chemistry, University of Pavia
V. Le Taramelli 12, 27100 Pavia (Italy)
E-mail: fagnoni@unipv.it
Homepage: http://www.unipv.it/photochem
Dr. B. Mannucci
Centro Grandi Strumenti (CGS), University of Pavia
V. Bassi 21, 27100 Pavia (Italy)
[**] S.P. acknowledges MIUR, Rome (FIRB-Futuro in Ricerca 2008
project RBFR08J78Q) for financial support. This work has been
supported by the Fondazione Cariplo (grant no. 2011-1839). We are
grateful to D. Sbarbada, D. Dondi, F. Corana, and D. Merli for the
precious help. This work was funded by the CINECA Supercomputer
Center, with computer time granted by ISCRA ARCLEAV
(HP10CCXGOF) and COMPDHT (HP10CZEHG6) projects.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201202794.
Angew. Chem. Int. Ed. 2012, 51, 1 – 5
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1
These are not the final page numbers!

.
Angewandte
Communications
ꢀ
The likelihood of a heterolytic cleavage of the CAr Cl
bond in these aromatic derivatives was first checked compu-
tationally. Calculations carried out in bulk MeOH at the
CPCM-UB3LYP/6-311 + G(2d,p)//UB3LYP/6-311 + G(2d,p)
level of theory for 31–33 clearly supported the hypothesis.
Thus, excitation of the silane to the triplet state resulted in the
ꢀ
heterolytic cleavage of the CAr Cl bond (see the Supporting
Information for further details).
The diagnostic arylation reaction in the presence of p-
bond nucleophiles could be used to prove the intermediacy of
triplet phenyl cations (Scheme 2a). If the expected process
delivers DHT intermediates, then end products such as those
depicted in Scheme 1 would be expected.
Turning to the experiment, the photoreactivity of
1 (0.025m) at l = 254 nm was examined (see Table 1, 1 h
irradiation). The course of the reaction depended on the
choice of medium. In ethyl acetate or acetonitrile less than
23% of the reagent was consumed and the only product was
the dechlorinated benzyltrimethylsilane 4. In protic meth-
anol, the consumption was almost complete (the quantum
Scheme 2. Photogeneration of triplet phenyl cations (a) and of a,n-
DHTs (b). ISC=intersystem crossing.
yield was one of the highest among aryl chlorides, F
=
ꢀ1
a OMe)[21] and then should split off the SiMe3 cation,
0.71)[19a,23] and, apart from a little 4, only silicon- and
chlorine-free products were formed, that is, toluene (6),
phenethyl alcohol 7a, benzyl ether 8a, and dibenzyl 9
+
a process promoted by the electron-deficient nature of the
p system[22] in triplet phenyl cations.
Table 1: Photoreactivity of compounds 1–3.[a]
[b]
ꢀ1
Entry Halide Solvent
F
Conv. [%]
Yield [%]
7
DHT/
4
5
6
8
9
cation[c]
1
2
3
4
5
6
7
1
1
1
1
1
1
1
CH3COOEt (MeCN)
MeOH
MeOH/H2O (4:1)
EtOH
EtOH/H2O (2:1)
MeOH, ATMS (0.2m)
MeOH, ATMS (1m)
–
23 (20)
91
100
46
100
100
100
51 (37)
4
2
14
1
–
–
–
–
–
2
4
6
3
2
7
–
–
–
13
14
11
10
10
–
0
0.71
0.73
–
–
–
7a: 39
7a: 39
7b: 27
7b: 21
7a: 20
–
8a: 2
8a: 4
8c: 2
8b: <1
8b: 2
8c: 2
14.0
31.5
3.14
38
1.60
0.19
–
6
2
14[d]
35[d]
–
8
9
10
2
2
2
MeOH
MeOH/H2O (4:1)
CD3OH/H2O (4:1)
0.90
0.82
–
98
100
92
44
18
[D1]-4: 16
–
–
–
2
5
7a: 11
7a: 14
8a: 11
8a: 38; 8c: 20
<1
2
<1
0.54
4.39
4.19
[D1]-6: 2 [D3]-7a: 2 [D3]-8a: 44;
[D1]-8c: 19
11
12
2
2
EtOH
EtOH/H2O (2:1)
–
80
100
69
22
–
–
4
4
7b: 11
7b: 17
8b: 5
8b: 12; 8c: 21
1
3
0.30
2.59
13
14
15
16
3
3
3
3
MeOH
MeOH/H2O (4:1)
EtOH
0.56
0.27
–
91
98
74
98
61
23
69
19
–
–
–
–
8
7
9
4
7a: 13
7a: 25
7b: 6
8a: 4
8a: 4; 8c: 2
8b: 1
–
4
<1
1
0.41
1.83
0.23
1.05
EtOH/H2O (2:1)
–
7b: 13
8b: 1; 8c: 1
[a] Irradiation conditions: 0.025 m of the substrate in the chosen solvent irradiated at l=254 nm (four lamps) for 1 h. Yields were based on consumed
1–3 and determined by GC/MS analysis. [b] The decomposition quantum yields (Fꢀ1) of 1–3 (10-2 m) were determined by irradiation at l=254 nm.
[c] Ratio between DHT-derived products and phenyl cation adducts. [d] A partial secondary desilylation of 5 to the corresponding 4-allyltoluene has
been observed.
2
www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 1 – 5
These are not the final page numbers!

Angewandte
Chemie
(entry 2). Going to a MeOH/H2O mixed solvent (Fꢀ1 = 0.73)
caused a further decrease in the yield of 4 and the formation
of some benzyl alcohol 8c (entry 3). In ethanol the reaction
was slower than in methanol and led to an analogous product
mixture (6, 7b, 8b, 9; entry 4). The addition of water to the
solvent (entry 5) caused a marked increase in the fraction of
DHT-derived products to essentially the same value as that
obtained in MeOH/H2O. In contrast, in the presence of
allyltrimethylsilane (ATMS) some of the allylated 5 was
formed,[19a] the yield of which increased with an increase in
the ATMS concentration (entries 6 and 7).
has a marginal effect on the quantum yield, Fꢀ1 = 0.82 in
MeOH/H2O) reversed the distribution with the latter prod-
ucts predominating four to one in MeOH/H2O and three to
one in EtOH/H2O (entries 9 and 12). Gratifyingly, their
distribution closely corresponds to that obtained by cyclo-
aromatization of enyne–allenes (compare the thermally
induced cyclization of (Z)-1,2,4-heptatrien-6-yne (A) in
MeOH/water 9:1 mixture, Scheme 1).[6] Irradiation of 2 in
a CD3OH/H2O mixture led to an increased yield of [D3]-8a
(from 38 to 44%) at the expense of [D3]-7a (from 14 to 2%
yield), thus supporting a primary isotope effect[6] and the key
role of hydrogen abstraction by DHT 14 (entry 10).
The above results fit nicely with the presence of cation
310+ and the a,4-DHT 13 as intermediates (Scheme 3). The
The ortho-substituted 3 was the least reactive of the series
and water in this case decreased the photoreactivity (compare
Fꢀ1 = 0.56 in MeOH and Fꢀ1 = 0.27 in aqueous methanol;
3
efficiency of the 1! 10+ conversion is affected, as one may
expect, by the polarity of the medium. The fate of the cation
310+ depends on the competition between reaction with the
entries 13 and 14). Reduction to 4 was the main process in
neat alcohol, but the presence of water markedly increased
the formation of DHT-derived products 6–9 (entries 13–16).
The above results fully support the contention that DHTs
are smoothly photogenerated from all of the three isomeric
(chlorobenzyl)trimethylsilanes. The main mechanistic fea-
tures are detailed in Scheme 3. Thus, derivatives 1–3 hetero-
litically cleave via the triplet state to afford the triplet phenyl
3
cations 10+–312+ (Scheme 3), just as other chlorobenzenes
having electron-donating substituents[19a] with a convenient
quantum yield in protic solvents (Fꢀ1 0.56–0.90 in MeOH), as
predicted by calculations. Desilylation producing DHTs does
not occur in nonprotic solvents (see Table 1, entry 1 and
Scheme 3, path a), but ion stabilization favors this process in
protic solvents, thereby resulting in DHTs with a structure-
dependent efficiency.
Contrary to the Myers–Saito cyclization, wherein the two
radical centers are formed simultaneously, the radical centers
are formed sequentially with the present approach. Thus,
desilylation competes with other reactions at the phenyl
cation stage (H abstraction from the solvent, more efficient
with EtOH than with MeOH; addition to alkenes). A protic,
nonreducing solvent such as water obviously shifts the
balance toward DHT formation. Solvent-assisted desilylation
of 310+ in a protic solvent is extremely fast as judged from the
competition with ATMS trapping (see below). These general-
izations are valid throughout the series, but the size of the
effect depends on the substrate structure and conditions. In
neat MeOH, desilylation to give DHT competes with reaction
at the cation level, thus resulting in product ratios of 14, 0.54,
and 0.41 for the para-, meta-, and ortho-substituted isomers,
respectively (Table 1). These values increase to 31.5, 4.4, and
1.8 in MeOH/H2O (4:1), but decrease when an efficient trap
Scheme 3. Generation of a,n-didehydrotoluenes 13–15 from silanes 1–
3.
solvent (H abstraction) and desilylation to 13, which is highly
favored in water. The other products result from the loss of
the SiMe3 group as well as that of chlorine, that is, via 13. The
role of the first cationic intermediate 310+ on the way to 13 is
further supported by the concentration-dependent trapping
by ATMS.
The examination was extended to the other isomeric
(chlorobenzyl)silanes. The solvent dependence with the meta-
substituted 2 (quite efficient reaction, Fꢀ1 = 0.90; Table 1,
entry 8) was larger than with 1. Reduction to 4 was the main
process in alcohols, but DHT-derived compounds (6–9)
accounted for one-third to one-fourth of the products in
neat alcohols (entries 8 and 11). The addition of water (which
3
such as ATMS is present for 10+. Isomeric DHTs exhibit
a partially different chemistry, as judged from the different
distribution of the products arising from them (6 to 9). Thus,
with a,4- and a,2-DHTs hydrogen atom abstraction to yield
6,[24] 7, and 9 predominates, whereas with the a,3-DHT
derivative the large proportion of the polar addition product 8
points to nucleophilic addition.[25]
Although this point requires further investigation, it is
tempting to attribute this difference to the fact that the
lowest-energy state of a,3-DHT is a singlet, whereas the
lowest-energy state is a triplet for the other two isomers.[15]
Angew. Chem. Int. Ed. 2012, 51, 1 – 5
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3
These are not the final page numbers!

.
Angewandte
Communications
[8] I. Alabugin, B. Breiner, M. Manoharan, Adv. Phys. Org. Chem.
2007, 42, 1 – 33.
[9] A. G. Myers, C. A. Parrish, Bioconjugate Chem. 1996, 7, 322 –
331.
[10] M. E. Cremeens, T. S. Hughes, B. K. Carpenter, J. Am. Chem.
Soc. 2005, 127, 6652 – 6661.
[11] I. A. Maretina, Russ. J. Gen. Chem. 2008, 78, 223 – 257.
[12] M. Kar, A. Basak, Chem. Rev. 2007, 107, 2861 – 2890.
[13] M. Schmittel, C. Vavilala, M. Emin Cinar, J. Phys. Org. Chem.
2012, 25, 182 – 197.
In conclusion, we have presented a new photochemical
approach to a,n-DHTs and it is based on the fragmentation of
easily available (chlorobenzyl)silanes in mixed aqueous
solutions, thus overcoming the limitation of their generation
from enyne–allenes. In the latter approach, aromaticity is
achieved at the same time as diradical formation, whereas in
the present approach the starting substrate is aromatic and is
transformed into the final intermediate through a double
elimination process by absorption of a photon.
[14] a) B. Breiner, K. Kaya, S. Roy, W.-Y. Yang, I. V. Alabugin, Org.
Biomol. Chem. 2012, 10, 3974 – 3987, and references therein;
b) I. V. Alabugin, W.-Y. Yang, R. Pal, CRC Handbook of
Organic Photochemistry and Photobiology, 3rd ed. (Eds: A.
Griesbeck, M. Oelgemoeller, F. Ghetti), CRC, Boca Raton,
2012, pp. 549 – 592; c) A. Polukhtine, G. Karpov, D. R. Pandi-
thavidana, A. Kuzmin, V. V. Popik, Aust. J. Chem. 2010, 63,
1099 – 1107.
[15] P. G. Wenthold, S. G. Wiershke, J. J. Nash, R. R. Squires, J. Am.
Chem. Soc. 1994, 116, 7378 – 7392.
[16] P. Neuhaus, S. Henkel, W. Sander, Aust. J. Chem. 2010, 63, 1634 –
1637.
Contrary to the Myers–Saito protocol, which is valid only
for a,3-DHT, all isomers are generated in this way, thus
allowing appreciation of the varied chemical behavior of
these intermediates. Notice further that the photochemical
mode of activation allows the spatial and temporal control of
the reaction.[14]
Additional work is currently ongoing in our lab to gain
more insights into the generation of the three a,n-DHT
isomers and their reactivities.
[17] U. E. Wiersum, T. Nieuwenhuis, Tetrahedron Lett. 1973, 14,
2581 – 2584.
[18] P. J. Stang in Dicordinated Carbocations (Eds: Z. Rappoport,
P. J. Stang), Wiley, New York, 1997, p. 451.
Experimental Section
General procedure for the irradiation of compounds 1–3: A solution
of the (chlorobenzyl)silane 1–3 (0.025m), in the chosen medium was
purged with nitrogen in quartz tubes and irradiated using 4 Hg lamps
(15 W, emission centered at l = 254 nm) for 1 h. The solution was then
analyzed using GC/MS analysis. The photoproducts were identified
by comparison with authentic samples, and quantified by using
calibration curves. Quantum yields were measured by irradiation of
a 10ꢀ2 m solution of the (chlorobenzyl)silanes 1–3 at l = 254 nm (1 Hg
lamp, 15 W).
[19] a) V. Dichiarante, M. Fagnoni, Synlett 2008, 787 – 800, and
references therein; b) S. Lazzaroni, D. Dondi, M. Fagnoni, A.
Albini, J. Org. Chem. 2008, 73, 206 – 211; c) S. Lazzaroni, D.
Dondi, M. Fagnoni, A. Albini, J. Org. Chem. 2010, 75, 315 – 323.
[20] V. Dichiarante, M. Fagnoni, A. Albini, Green Chem. 2009, 11,
942 – 945.
[21] C. Hansch, A. Leo, R. W. Taft, Chem. Rev. 1991, 91, 165 – 195.
[22] a) R. Tacke, H. Linoh, The Chemistry of Organic Silicon
Compounds (Eds.: S. Patai, Z. Rappoport), Wiley, Chichester,
1989, pp. 1143 – 1206; b) J. P. Dinnocenzo, S. Farid, J. L. Good-
man, I. R. Gould, S. L. Mattes, W. P. Todd, J. Am. Chem. Soc.
1989, 111, 8973 – 8975; c) We thank a referee for making this
point.
Received: April 11, 2012
Revised: June 8, 2012
Published online: && &&, &&&&
Keywords: arenes · photochemistry · radicals · silanes ·
[23] M. Freccero, M. Fagnoni, A. Albini, J. Am. Chem. Soc. 2003, 125,
13182 – 13190.
[24] The formation of toluene has been reported for the Myers–Saito
cyclization of A only in the presence of good H-donors, such as
1,4-cyclohexadiene. See Ref. [6].
.
synthetic methods
[1] Enediyne Antibiotics as Antitumor Agents (Eds.: D. B. Borders,
[25] An alternative view is that the cations 310+–312+ abstract
a hydrogen atom and the resulting radical cation split off
SiMe3+. However, the nature and distribution of the products
obtained do not fit with the intermediacy of the benzyl radical
(as suggested by a referee), which would be the common
intermediate from either of the three precursors, even if it is
difficult to exclude any contribution by this path.
T. W. Doyle), Marcel Dekker, New York, 1995.
ˇ
ˇ
[2] M. Gredicak, I. Jeric, Acta Pharm. 2007, 57, 133 – 150.
[3] U. Galm, M. H. Hager, S. G. Van Lanen, J. Ju, J. S. Thorson, B.
Shen, Chem. Rev. 2005, 105, 739 – 758.
[4] A. L. Smith, K. C. Nicolaou, J. Med. Chem. 1996, 39, 2103 – 2117.
[5] K. K. Wang, Chem. Rev. 1996, 96, 207 – 222.
[6] A. G. Myers, P. S. Dragovich, E. Y. Kuo, J. Am. Chem. Soc. 1992,
114, 9369 – 9386.
[7] R. Nagata, H. Yamanaka, E. Okazaki, I. Saito, Tetrahedron Lett.
1989, 30, 4995 – 4998.
4
www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 1 – 5
These are not the final page numbers!

Angewandte
Chemie
Communications
Radicals
Doubly radical: A novel entry to a,n-
didehydrotoluene (DHT) diradicals is
disclosed and proceeds through the
photochemical activation of (chloroben-
zyl)trimethylsilanes with chloride loss
and elimination of the SiMe3+ group (see
scheme). The products formed in solu-
tion are indicative of the intermediacy of
the three isomers of the a,n-DHT.
S. Protti, D. Ravelli, B. Mannucci,
A. Albini, M. Fagnoni*
&&&& — &&&&
a,n-Didehydrotoluenes by
Photoactivation of (Chlorobenzyl)-
trimethylsilanes: An Alternative to
Enyne–Allenes Cyclization
Angew. Chem. Int. Ed. 2012, 51, 1 – 5
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5
These are not the final page numbers!