Photogenerated α,n-didehydrotoluenes from chlorophenylacetic acids at physiological pH

Article abstract of DOI:10.1021/jo502318v

Aromatic diradicals are recognized as promising intermediates for DNA cleavage, but their formation has thus far been limited to the Bergman and Myers-Saito cycloaromatizations. We report here the phototriggered generation of all isomers of the potential DNA-cleaving α,n-didehydrotoluene diradicals at physiological pH, accomplished by the irradiation of chlorophenylacetic acids under mild conditions. The desired diradicals were formed upon photolysis of the chosen aromatic in aqueous phosphate buffer solution (pH = 7.3), with the consecutive elimination of biologically compatible chloride ion and carbon dioxide. Theoretical simulations reveal that the efficient decarboxylation of the primarily generated phenyl cations involves a previously not known diradical structure.

Full text of DOI:10.1021/jo502318v

Article
pubs.acs.org/joc
Photogenerated α,n‑Didehydrotoluenes from Chlorophenylacetic
Acids at Physiological pH
Davide Ravelli, Stefano Protti, and Maurizio Fagnoni*
Photogreen Lab, Department of Chemistry, University of Pavia, Viale Taramelli 12, 27100 Pavia, Italy
*
S Supporting Information
ABSTRACT: Aromatic diradicals are recognized as promising inter-
mediates for DNA cleavage, but their formation has thus far been limited
to the Bergman and Myers−Saito cycloaromatizations. We report here the
phototriggered generation of all isomers of the potential DNA-cleaving
α,n-didehydrotoluene diradicals at physiological pH, accomplished by the
irradiation of chlorophenylacetic acids under mild conditions. The desired
diradicals were formed upon photolysis of the chosen aromatic in aqueous
phosphate buffer solution (pH = 7.3), with the consecutive elimination of
biologically compatible chloride ion and carbon dioxide. Theoretical
simulations reveal that the efficient decarboxylation of the primarily
generated phenyl cations involves a previously not known diradical
structure.
INTRODUCTION
Scheme 1. (a) Generation of α,n-Didehydrotoluene
■
Diradicals (α,n-DHTs, III) via the One-Photon Double
Elimination from (n-Chlorobenzyl)trimethylsilanes (I) and
(b) Photochemical Pathways for Phenylacetic Acids (IV)
Diradicals are emerging species that find application in several
fields, including synthesis, the chemistry of materials, and
1
medicine. With regard to the last field, naturally occurring
2
enyne antibiotics show broad anti-neoplastic activity due to
the formation of a 1,5-didehydroindene or a 1,4-didehydro-
benzene diradical via a cycloaromatization process (via a
3
,4
Myers−Saito or a Bergman reaction, respectively). Such
diradicals abstract hydrogen atoms from deoxyribose residues
in DNA. Accordingly, many efforts have been devoted to the
time-consuming synthesis of cyclic enyne-allenes or enediynes
that mimic the reactivity of natural compounds, albeit these
5,6
models have found no practical applications so far.
At any rate, the scope of cycloaromatization is limited since
only α,3-didehydrotoluenes (α,3-DHTs) are formed from
5
,6
enyne model compounds via the Myers−Saito reaction.
7
,8
However, we recently demonstrated that all of the three
isomeric α,n-DHT (n = 2−4) diradicals (III, Scheme 1a) are
generated by a one-photon-induced double elimination from
(
chlorobenzyl)trimethylsilanes (I). The process involves the
consecutive detachment of the chloride anion (forming a triplet
3
phenyl cation, II) and of the trimethylsilyl cation (as an
7
,8
electrofugal group). The chemistry of the generated phenyl
cation and the ensuing diradical depended on the medium. In
particular, formation of the latter was favored when the polarity
of the protic solvent increased, for example, by shifting from
neat alcohol to aqueous mixtures. Thus, both the fact that the
particular, physiological) environment to the precursor and
explored whether aggressive diradicals could be formed by
consecutive dechlorination/decarboxylation rather than dech-
lorination/desilylation. In fact, the carboxylic group serves as an
electrofugal group in biosynthetic and other biological
processes, as in the isopentenyl pyrophosphate synthesis, in a9
way functioning as nature’s substitute for a silyl group.
7
presence of water improves the generation of DHTs and the
obvious interest in this solvent for a medicinal application
prompted us to investigate precursors more soluble in water
than chlorobenzylsilanes.
We then reasoned that an alternative electrofugal group
could be chosen to impart solubility in an aqueous (in
Received: October 9, 2014
Published: December 5, 2014
©
2014 American Chemical Society
852
dx.doi.org/10.1021/jo502318v | J. Org. Chem. 2015, 80, 852−858

The Journal of Organic Chemistry
Article
a
Table 1. Irradiation of Chlorophenylacetic Acids 1a−c
a
−3
Irradiation conditions: 1a−c (5 × 10 M in the chosen solvent), irradiated at 254 nm (4 × 15 W Hg lamps), tirr = 20 min (>70% 1a−c
b
−1
c
consumption; yields based on consumed 1). Quantified by HPLC analysis, eluant: MeCN/water (pH = 1.6) 25/75, flux 0.8 mL min . Cs CO (5
×
2
3
−3
d
10 M) added. Buffer = 0.05 M aqueous phosphate buffer solution at pH = 7.3.
Recently, the use of a carboxylic moiety as a leaving group in
synthesis has gained much attention in metal-mediated cross-
generally low when FG = H or an electron-donating
9
,18−21
group,
where it proceeded via the first singlet excited
10−14
coupling reactions,
compounds induced by enzymes (e.g., decarboxylase), and in
in the defunctionalization of organic
state (S ). The decarboxylaton efficiency dramatically increased
1
15
when FG = −NO or −COCH , where it proceeded via the
2
3
1
6,17
photoinduced electron transfer processes.
first triplet excited state (T ) and yielded a strongly stabilized
1
We thus decided to test n-chlorophenylacetic acids (IV, FG =
Cl, Scheme 1b) in the search for a new class of promising,
readily prepared antitumor compounds able to form diradicals
under physiological conditions when triggered by light.
Interestingly, harmless and stable fragments (chloride ion and
benzyl anion.
Our surmise was that the presence of a chlorine atom on the
aromatic ring would increase the intersystem crossing (ISC)
24
efficiency, preventing any photoreaction from the S state
1
−
and, in addition, that the CH COOH/CH COO substituent
in chlorophenylacetic acids would be sufficiently electron-
2
2
CO ) would be expelled. Obviously, this would be a pH-
2
dependent process, in which the elimination of carbon dioxide
from the phenylacetate anion (IV′) is expected to be more
favorable than from the free acid.
donating to allow the photoheterolysis of the Ar−Cl bond. In
−
fact, the CH COO group has an electron-donor ability
2
25
comparable to a Me group, and since a triplet phenyl cation
26
This approach (Scheme 1b, path i) would be successful only
if alternative photodecarboxylation pathways do not compete
with dechlorination. Actually, several largely used NSAIDs
was smoothly formed by photolysis of 4-chlorotoluene, the
generation of the corresponding cation from photoheterolysis
of the Ar−Cl bond in the 4-chlorophenylacetate anion should
be feasible. There is an early report in the literature that
irradiation in water of chlorophenylacetic acid isomers, used as
models for the herbicide Fenac, led mainly to dechlorinated
18−21
(
non-steroidal anti-inflammatory drugs; e.g., ibuprofen)
containing the ArCH(R)COOH moiety are known to cause
adverse photosensitivity responses when susceptible patients
22
are exposed to sunlight. In this case, homolytic cleavage of the
27
products.
CH −COOH bond was found to be the main pathway for the
2
2
3
decarboxylation of phenylacetic acid. The photoreactivity of
these acids was found to be enhanced in the anionic form, to an
extent that varied with the substituents present on the aromatic
RESULTS
■
In this work, we report the pH-dependent photochemical
behavior in protic aqueous media of the three isomeric n-
chlorophenylacetic acids 1a−c (Table 1) and the correspond-
ing carboxylate anions 1′a−c, supported by theoretical
1
8,19
ring.
Thus, photolysis of carboxylates IV′ proceeded via the
direct photoelimination of carbon dioxide to give a benzyl
anion (Scheme 1b, path ii), and the consumption of IV′ was
8
53
dx.doi.org/10.1021/jo502318v | J. Org. Chem. 2015, 80, 852−858

The Journal of Organic Chemistry
Article
3
Figure 1. (a) Optimized geometry at the CASSCF/6-31G(d) level and relevant parameters for 1a(α). ESP charge (qESP) for the Cl atom at the
1
3
equilibrium geometry is shown, as derived from single-point calculation in bulk water. (b) Potential energy curves for 1a, 1a, and 1a(α) at the
CASSCF/6-31G(d) level of theory in bulk water (see Table S2 for details). The values reported refer to the energy change for the last point of each
curve with respect to the equilibrium geometry; in parentheses, the ESP charges for the Cl atom at the same point are reported.
calculations aimed at the rationalization of the observed
reactivity.
found to be an almost exclusive product along with small
amounts of 9 (1%) and of 4-hydroxyphenylacetic acid 5a (3%).
In addition, the consumption of 1a under the latter conditions
(Φ = 0.26) was markedly enhanced with respect to neat
Experimental Study. First notice that no significant
differences in the UV absorption spectra of 1a in methanol
and aqueous phosphate buffer were observed (see Supporting
Information Figure S1). The irradiation experiments performed
are gathered in Table 1, along with the obtained results.
−
1
methanol.
The photochemistry observed for the isomeric acids 1b and
1c is quite similar to that of 1a. Photoreduction to 2 and
photosolvolysis were the preferred paths for both the meta- and
the ortho-isomers in both basic methanol and 9/1 methanol/
water (entries 9, 10 and 13, 14). When switched to the 1/9
methanol/buffer mixture and neat buffer with no methanol
(entries 11, 12 and 15, 16), photoreductive dechlorination to 2
was dwarfed in favor of photosolvolysis (with the formation of
3- or 2-hydroxyphenyl acetic acids 5b,c in up to 53% yield) and
of the generation of the corresponding α,3- and α,2-DHTs, as
evidenced by the formation of benzyl alcohol 6 in high yield.
The use of neat buffer as the reaction medium leveled the
photoreactivity of the three chlorophenylacetic acid isomers
1a−c, and they exhibited comparable quantum yields for
photodecomposition (Φ = 0.2−0.3).
Dechlorination took place when 4-chlorophenylacetic acid 1a
3
(
5 × 10− M) was irradiated (λ = 254 nm, 4 lamps) in neat
methanol, yielding a mixture of phenylacetic acid (2, 44%) and
4-methoxyphenylacetic acid (3a, 25%).
We then examined the photoreactivity of carboxylate 1′a. In
basic methanol (Cs CO = 0.05 M, entry 1), significant
2
3
amounts of products resulting from both dechlorination and
decarboxylation (phenethyl alcohol 8 and 1,2-diphenylethane
9
9
) were formed at the expense of 2 and 3a. Compounds 8 and
are diagnostic for the formation of an α,4-DHT intermediate,
7
,8
as previously reported by our group. The process was quite
inefficient, however, with a disappearance quantum yield (Φ )
−1
of 0.076. The presence of water in the reaction medium
−
1
(
MeOH/water 9/1; entry 2) did not change the nature of the
As a mechanistic test, the irradiation of 1a was carried out in
products formed but had a marked impact on their distribution,
making the products formed through α,4-DHT (8 + 9)
the presence of allyltrimethylsilane (ATMS, 0.5 M, entries 7
7,30
and 8).
In this case, irradiation caused the exclusive
predominant. No effect resulted from replacing CH OH with
formation of 4-(2-propenyl)phenylacetic acid 4a (95% yield)
at the expense of 2 and 3a. Likewise, 4a was the exclusive
product when the reaction was carried out in the presence of
ATMS and Cs CO , in which case the carboxylate anion 1′a
3
CD OH (entry 3). The presence of a −COOH group allowed
3
the photoreactivity to be tested in a 1/9 MeOH/aqueous
phosphate buffer (pH = 7.3) mixture, in which compound 1a
was completely soluble (Table 1, entry 4). In this medium, the
results changed dramatically and irradiation led to the exclusive
formation of dechlorinated and decarboxylated compounds,
that is 8, 9, and benzyl alcohol 6 in comparable yields, along
with a minor amount (3%) of methyl benzyl ether 7. Irradiation
2
3
was the prevalent form.
Irradiations of phenylacetic acid (2) and 4-methoxyphenyl-
acetic acid (4) in basic methanol and in buffer have been
likewise carried out to exclude competitive pathways in the
formation of α,n-DHTs. As a result, the photorelease of carbon
dioxide from dechlorinated acids 2 and 4 is a side process that
can take place but did not markedly affect the final product
distribution (see Supporting Information for details).
of 1a in a 1/9 CD OH/buffer mixture (entry 5) led to the
3
disappearance of 8 from the products and to a significant
decrease of 9-d (down to 5% yield) in favor of 6 (from 17 up
2
to 41%) and of deuterated methyl benzyl ether 7-d . This
Computational Study. Calculations were then carried out
in bulk water (C-PCM solvation model) for modeling the
dechlorination of acid 1a (taken as a model) to give a phenyl
cation. The level of theory chosen involved a multiconfigura-
tional approach, namely, the complete active space self-
consistent field (CASSCF; also capable of handling excited
3
supported the operation of a primary isotope effect, as
29
suggested by Myers and the key role of hydrogen abstraction
by the DHT in determining the product distribution. Finally,
when the same reaction was carried out in neat buffer to mimic
physiological conditions (entry 6), benzyl alcohol 6 (72%) was
8
54
dx.doi.org/10.1021/jo502318v | J. Org. Chem. 2015, 80, 852−858

The Journal of Organic Chemistry
Article
states), which was further refined by corrections at the MP2
level. A detailed description of computational analysis has been
included in the Supporting Information (see Chapter 2). The
choice of the active space depended on the structure
considered.
Cl bond cleavage. In particular, with a small stretching of the
3
bond (up to a 2.3 Å length) in 1′a(β), the structure was best
3
Z
described as a “zwitterion” (hereafter named as 10′a , Table
S5 and Figure 2), where a marked charge separation in the
As for the C−Cl bond cleavage from the precursor, the
modeling focused on the role of the aromatic moiety and of the
substituents attached to the ring. Thus, a (10,10) active space
was adopted, including the 3π and the 3π* orbitals of the
aromatic as well as the two σ/σ* pairs of the C−Cl and the
C −COOH bonds. We first investigated the geometry of the
α
1
ground state (1a; Figure S2a), the lowest lying singlet ( 1a;
Figure S2b), and triplet ( 1a; Figures 1 and S2c,d) excited
3
states. A marked distortion from planarity of the aromatic ring
3
was observed only for 1a. Two different conformations
exhibited nonplanarity, one with the chlorine atom and the
carboxylic group on the same side of the aromatic ring
(
designed as “α”; Figures 1a and S2c), and the other with the
chlorine and carboxylic acid on opposite sides (“β”; Figure
S2d). The two conformations are close in energy (ca. 69 kcal
−1
mol above the ground state), with the α structure slightly
3
more stable than the β (see Table S1). It may be noted that in
Figure 2. Energy profile upon chloride anion removal from 1′a(β) at
3
1
a(α) C sticks out of the aromatic system, with a C −C −
the CASSCF/6-31G(d) level of theory in bulk water (see Table S5
and section 3.1 in Supporting Information for details).
4
4
3
C −C dihedral angle of 6°, and the C −Cl bond is bent further
upward by an angle of 28° with respect to the C −C −C plane
2
1
4
4
5
3
and is significantly elongated, from 1.78 Å (for 1a) to 1.81 Å. A
negative charge likewise develops at the chlorine atom (−0.19).
The potential energy curves reported in Figure 1b (see also
Table S2) show the energy variation occurring upon stretching
organic fragment is observed. Further stretching of the C−Cl
bond, however, led to a surface crossing to a diradical structure
3
D
(hereafter named as 10′a ). As for the energetic profile, this
−1
process confronts a very small barrier (a few kcal mol ) and
then proceeds downhill after the surface crossing.
the C −Cl bond in the three different states considered above.
4
3
It is apparent that in triplet 1a an increase in the bond length
As for the decarboxylation of 1′a, the computational analysis
−
up to 3 Å causes only a modest increase of energy (ca. 12 kcal
predicted a large barrier for the cleavage of the C −COO
α
−1
−1
mol ) and a marked charge separation. By contrast, imposing
bond (ca. 29 kcal mol ) from the triplet state to give the
the same deformation to the C−Cl bond in either ground (1a)
corresponding benzyl anion (see section 3.2 in Supporting
Information).
1
or excited ( 1a) singlet states results in a much higher energy
increase, with no significant negative charge localization at the
chlorine atom. This is consistent with our surmise that
heterolytic C−Cl bond cleavage will occur from the lowest
With the aim of understanding the actual pathway involved
in the processes observed, we modeled computationally the
intermediates arising after loss of the chloride anion from 1a
(hereafter named 10a) adopting a (8,9) active space, including
the aromatic system (3π and 3π* orbitals), the σ/σ* couple of
28
triplet state of 4-chlorophenylacetic acid.
The analysis was then extended to phenylacetates 1′a−c,
where two paths were considered, involving loss of chloride
the C −COOH bond, and the orbital at the dicoordinated
α
anion and CO . As for the former, when adopting the same
carbon (see Supporting Information for further details).
Likewise, the structures of 10′a−c were modeled considering
(as in precursor 1′a) the inclusion of the carboxylate group
orbitals in the active space. Thus, an overall (12,12) active
space was used (see again Supporting Information), comprising
the same orbitals adopted for 10a and the whole carboxylate
moiety.
2
active space used for 1a, calculations on 1′a indicate no
difference with the ionic state of the carboxyl function, with
similar results in terms of geometry of the different states
(
Figure S3), as well as for the potential energy curves with
respect to the C−Cl bond stretching and for charge localization
on the chlorine atom (Tables S3 and S4 and Figure S4a). The
same holds for the isomeric carboxylates 1′b and 1′c (see
Supporting Information Tables S3 and S4 and Figures S4b, S5,
and S6).
We wondered, however, whether the lone pair at the
carboxylate moiety might affect the dechlorination step from
the triplet state of 1′a (with the β conformer here more stable
than the α). Unfortunately, the computational cost prevented
inclusion of the lone pair into the above-defined active space
1
,3
Noteworthy, the molecular structures of
10a closely
26
resemble those of alkyl-substituted phenyl cations. The
calculations predict that in both spin species the phenyl ring
is planar, with 10a showing a sort of cumulene character at the
dicoordinated carbon (with a π σ electronic structure), while
in 10a, the ring is more close to a regular hexagon (π σ
1
6
0
3
5 1
electronic structure, see Figure 3a, where the shape of the two
SOMOs is reported).
(
which would result in an overall (12,11) active space). A
The ground state of 10a is the singlet, with a rather large S−
−1
satisfactory analysis was, however, obtained through a simplified
computational approach, where only the lone pair at the
carboxylate group and the relevant orbitals of the aromatic
system were included in the active space (see section 3.1 in
Supporting Information). This computational study revealed
T gap (ca. 12 kcal mol ; see Figure 4a and Tables S7 and S8
for details). The computational investigation established,
however, that the situation is completely different for
intermediates 10′a−c that are diradical structures with the
two SOMOs located predominantly at the dicoordinated
carbon and at the carboxylic group, regardless of the isomer
that two different states concur with the description of the C −
4
8
55
dx.doi.org/10.1021/jo502318v | J. Org. Chem. 2015, 80, 852−858

The Journal of Organic Chemistry
Article
small S−T gaps (see Figure 4b and Tables S9 and S10 for
1
,3
details). As for the decarboxylation of 10′a−c (see Tables
S11 and S12), it was shown to occur via a well-defined
transition state (tagged TSC), describing the cleavage of the
·
C −COO bond during an overall exoergonic process (see
α
Tables S13−S15 and Figures S9−S11 for details). The
observed energy changes are reported in Table S12 and
demonstrate that loss of CO from 10′a−c occurs with very
2
−
1
low activation energies (4.10−8.42 kcal mol ), with
parameters similar to those observed for phenylacetoxy radical
(4.82 kcal mol at the same level of theory; see Table S16
·
−1
2
and Figure S12), which is well-known to undergo facile
31
decarboxylation.
DISCUSSION
■
The above experimental observations are corroborated by
computational results, and the competing pathways in the
photochemistry of carboxylic acids 1a−c and of carboxylates
1
′a−c are represented in Schemes 2 and 3 for the para-
derivatives 1a and 1′a, respectively.
3
Figure 3. Singly occupied molecular orbitals (SOMOs) for (a) 10a
1
,3
and (b) 10′a, as determined from CPCM-CASSCF/6-31G(d)
Scheme 2. Photoreactivity of 4-Chlorophenylacetic Acid 1a
calculations in bulk water.
(
ATMS = Allyltrimethylsilane)
In the framework of the known photochemistry of aryl
chlorides, these results suggest that photolysis of 1a generates
3
triplet phenyl cation 10a as the initial intermediate (by
3
heterolytic fragmentation of triplet 1a). This intermediate
abstracts a hydrogen atom from the solvent to give phenylacetic
acid 2 (Scheme 2, path a). A competitive pathway is ISC, which
1
yields the more stable singlet cation 10a (path b), from which
4
-methoxyphenylacetic acid 3a is formed by solvent addition
(
path c). On the other hand, in the presence of ATMS, trapping
3
of 10a takes place faster than both reduction and ISC, and 4a
is obtained exclusively (path d). Irradiation of carboxylate 1′a
caused again the heterolysis of the Ar−Cl bond in the triplet
state (Scheme 3, path a′), but now yielding a triplet zwitterion
intermediate that is closely associated with chloride ion
3
Z
(
10′a ).
Figure 4. Relative energy of (a) 1,310a and (b) 1,310′a−c according to
3
The close similarity with the behavior of 10a fully supports
the isodesmic equations reported as the inset.
3
Z
the role of intermediate 10′a (with radical/radical cation
character at the aromatic moiety), showing the typical reactivity
of triplet phenyl cations, viz. hydrogen abstraction from the
solvent and trapping by ATMS (path b′), in competition with
or of the spin state (whether singlet or triplet, Figure 3b). The
relative energies of these intermediates have been carefully
1
,3
1
3
1
Z
evaluated. Contrary to cations 10a, 10′a−c and 10′a−c are
intersystem crossing (path c′), to give 10′a and compounds
3a and 5a from it. However, computational data also support
almost isoenergetic, resulting in ground state triplets with very
8
56
dx.doi.org/10.1021/jo502318v | J. Org. Chem. 2015, 80, 852−858

The Journal of Organic Chemistry
Article
Scheme 3. Chemical Pathways Involved in the
Photoreactivity of the 4-Chlorophenylacetate Anion (1′a)
readily available substituted phenylacetic acids. Calculations
give insight in the sequence of events involved, confirming that
−
the first step is dechlorination. Nucleophiles such as Cl and
water stabilize the phenyl cation, giving room to competing
processes, but the absence of hydrogen donors made
chlorophenylacetates excellent precursors for DHTs (or at
least for the para-isomer). This fact, along with the solubility in
water, militates in favor of the choice of these derivatives as the
34
active moiety of a new class of photoactivatable therapeutics,
and it is hoped that this work will lay the foundations for the
development of sound pharmaceutically active leads.
EXPERIMENTAL SECTION
■
The photochemical reactions were performed by using nitrogen-
purged solutions in quartz tubes in a multilamp reactor equipped with
four Hg lamps (15 W each, emission centered at 254 nm) for the
irradiation. Quantitative determination for compounds of interest was
carried out by means of HPLC (2−3, 5) and GC (6−9) calibration
curves. Quantum yields were measured at 254 nm (1 Hg lamp, 15 W).
Synthesis of 4-(2-Propenyl)phenylacetic acid (4a). A solution
of 1a (85 mg, 0.5 mmol, 0.005 M), Cs CO (156 mg, 0.5 mmol, 0.005
2
3
M), allyltrimethylsilane (8.05 mL, 50 mmol, 0.5 M) in MeOH was
irradiated at 254 nm for 1 h. The solvent was evaporated, and the
resulting residue was dissolved in water, acidified at pH = 3 by acetic
acid, and extracted with ether (3 × 20 mL). Organic phases were
combined, dried over Na SO , and the solvent evaporated under
2
4
3
D
vacuum. Purification by column chromatography of the residue
the role of surface crossing that makes the diradical form 10′a
1
afforded 76 mg of 4a (viscous oil, 86% yield): H NMR (δ, CDCl )
3
the most stable one when chloride is at a larger distance (path
3
.40−3.45 (d, 2H, J = 7 Hz), 3.65 (s, 2H), 5.10−5.15 (m, 2H), 5.95−
d′). The diradical structure is obviously the ideal precursor for
13
6
.00 (m, 1H), 7.20−7.30 (AA′BB′, 4H), 10.80 (br s, 1H); C NMR
CO detachment, and indeed, only a low-energy barrier is
2
(
(
δ, CDCl ) 40.8 (CH ), 41.6 (CH ), 116.8 (CH ), 129.6 (CH), 130.3
3 2 2 2
−1
required for this process (less than 5 kcal mol , with an
expected rate around 10 s , as previously determined for
). Thus, access to 10′a plays a key role in the ensuing
−1
CH), 131.9, 138.5 (CH), 140.1, 179.1; IR (ν/cm ) 3050, 2955,
1
0
−1
1701, 1408, 909, 837. Anal. Calcd for C H O : C, 74.98; H, 6.86.
11 12 2
·
31
3
D
2
Found: C, 75.0; H, 6.9. GC-MS (m/z): 177 (5), 176 (45), 131 (100),
117(45), 91 (45).
formation of the corresponding α,4-DHT.
Water is known to improve the efficiency of photoinduced
Ar−Cl cleavage and also to stabilize the triplet phenyl cation
ASSOCIATED CONTENT
Supporting Information
Experimental procedures, sample spectra; optimized geo-
metries, energies and CASSCF output data for all of the
structures involved in this work. This material is available free
of charge via the Internet at http://pubs.acs.org.
■
3
2
formed. Stabilization of triplet phenyl cation may increase the
chances of competing paths, such as ISC to the corresponding
S
*
singlets in some cases (particularly in 10′b,c), preventing CO2
loss due to the high reactivity of these singlets with the reaction
medium. Accordingly, solvolysis products accounted for ca.
5
0% of the resulting mass balance arising from 1b and 1c in
3
Z
neat buffer. On the other hand, cations 10′a−c are easily
reduced in MeOH (a hydrogen-donating solvent), yielding
dechlorinated phenylacetic acids as the main products in all
cases. The generation of α,n-DHTs is maximized when the
leaving group and the benzylic function are para to each other
AUTHOR INFORMATION
Corresponding Author
Fax: +39 0382 987323. Tel. +39 0382 987198. E-mail:
fagnoni@unipv.it.
■
*
Notes
(
compound 1a), as was previously observed in the case of
7,8
The authors declare no competing financial interest.
chlorobenzylsilanes.
As for the chemistry of the formed α,n-DHTs, the dual
ACKNOWLEDGMENTS
behavior (diradical and ionic) of such species is apparent from
■
3
3
Table 1. The fastest reaction is hydrogen abstraction from
This work was supported by the Fondazione Cariplo (Grant
2011-1839). S.P. acknowledges MIUR, Rome (FIRB-Futuro in
Ricerca 2008 project RBFR08J78Q), for financial support. We
thank Prof. A. Albini (University of Pavia) and P. Hoggard
(Santa Clara University) for fruitful discussions. We are grateful
to B. Mannucci and C. Nicola (Centro Grandi Strumenti-
Pavia) for the valuable assistance. This work was funded by the
CINECA Supercomputer Center, with computer time granted
by ISCRA projects (code: HP10CZEHG6, HP10C6PCC2, and
HP10C3CPWN).
methanol to yield a benzyl radical and the final products from it
alcohol 8 and bibenzyl 9 via a diradical pathway). This
reaction is subjected to a clear primary isotope effect (see
entries 4 and 5, Table 1). Thus, product distribution arising
(
from the α,n-DHT in CD OH buffer (1:9) resulted in the
3
preferential formation of polar products such as alcohol 6 and
ether 7 at the expense of the radical product 8.
CONCLUSION
In conclusion, this work demonstrates that aggressive diradicals
α,n-DHTs) can be easily obtained via a double elimination
■
REFERENCES
(1) Abe, M. Chem. Rev. 2013, 113, 7011−7088.
(
■
sequence by absorption of a photon from inexpensive and
8
57
dx.doi.org/10.1021/jo502318v | J. Org. Chem. 2015, 80, 852−858

The Journal of Organic Chemistry
Article
(
(
(
(
(
2) Maeda, H. Anticancer Res. 1981, 1, 175−186.
3) Gredicak, M.; Jeric, I. Acta Pharm. 2007, 57, 133−150.
̌
̌
4) Smith, A. L.; Nicolaou, K. C. J. Med. Chem. 1996, 39, 2103−2117.
5) Kar, M.; Basak, A. Chem. Rev. 2007, 107, 2861−2890.
6) Breiner, B.; Kaya, K.; Roy, S.; Yang, W.-Y.; Alabugin, I. V. Org.
Biomol. Chem. 2012, 10, 3974−3987.
(
7) (a) Protti, S.; Ravelli, D.; Mannucci, B.; Albini, A.; Fagnoni, M.
Angew. Chem., Int. Ed. 2012, 51, 8577−8580. (b) Protti, S.; Ravelli, D.;
Fagnoni, M.; Albini, A. Pure Appl. Chem. 2013, 85, 1479−1486.
(
c) Raviola, C.; Ravelli, D.; Protti, S.; Fagnoni, M. J. Am. Chem. Soc.
2
(
7
(
1
(
2
(
014, 136, 13874−13881.
8) Ravelli, D.; Protti, S.; Fagnoni, M.; Albini, A. J. Org. Chem. 2013,
8, 3814−3820.
9) Meiggs, T. O.; Grossweiner, L. I.; Miller, S. I. J. Am. Chem. Soc.
972, 94, 7981−7986.
10) Gooßen, L. J.; Rodriguez, N.; Gooßen, K. Angew. Chem., Int. Ed.
008, 47, 3100−3120.
11) Rodríguez, N.; Gooßen, L. J. Chem. Soc. Rev. 2011, 40, 5030−
048.
5
(
12) Bonesi, S.; Fagnoni, M. Chem.Eur. J. 2010, 16, 13572−13589.
13) Song, Q.; Feng, Q.; Zhou, M. Org. Lett. 2013, 15, 5990−5993.
14) Feng, Q.; Song, Q. J. Org. Chem. 2014, 79, 1867−1871.
15) Kourist, R.; Guterl, J.-K.; Miyamoto, K.; Sieber, V.
(
(
(
ChemCatChem 2014, 6, 689−701.
16) See for example: Oelgemoller, M.; Griesbeck, A. G. J. Photochem.
Photobiol. C 2002, 3, 109−127.
17) Photochemically-Generated Intermediates in Synthesis; Albini, A.,
Fagnoni, M., Eds.; John Wiley & Sons: Hoboken, NJ, 2013.
18) Lukeman, M. Photodecarboxylation of Arylacetic Acids. In CRC
(
̈
(
(
Handbook of Organic Photochemistry and Photobiology, 3rd ed.;
Griesbeck, A., Oelgemoeller, M., Ghetti, F., Eds.; CRC Press: Boca
Raton, FL, 2012; pp 715−726.
(
̀
19) Bosca, F.; Marin, M. L.; Miranda, M. A. Photodecarboxylation of
Acids and Lactones: Antiinflammatory Drugs. In CRC Handbook of
Organic Photochemistry and Photobiology, 2nd ed.; Horspool, W., Lenci,
F., Eds.; CRC Press: Boca Raton, FL, 2004; pp 64-1−64-10.
(
20) Castell, J. V.; Gomez-L, M. J.; Miranda, M. A.; Morera, I. M.
Photochem. Photobiol. 1987, 46, 991−996.
(
21) Moore, D. E.; Chappuis, P. P. Photochem. Photobiol. 1988, 47,
1
(
(
(
73−180.
22) Moore, D. E. Drug Safety 2002, 25, 345−372.
23) Epling, G. A.; Lopes, A. J. Am. Chem. Soc. 1977, 99, 2700−2704.
24) Previtali, C. M.; Ebbesen, T. W. J. Photochem. 1985, 30, 259−
267.
(
(
25) Hansch, C.; Leo, A.; Taft, R. W. Chem. Rev. 1991, 91, 165−195.
26) Qrareya, H.; Raviola, C.; Protti, S.; Fagnoni, M.; Albini, A. J.
Org. Chem. 2013, 78, 6016−6024.
27) Crosby, D. G.; Leitis, E. J. Agric. Food Chem. 1969, 17, 1036−
040.
(
1
(
28) The occurrence of a heterolytic cleavage of the Ar−Cl bond
from triplet state aryl halides has found computational and
experimental evidences. See for instance: Raviola, C.; Protti, S.;
Ravelli, D.; Mella, M.; Albini, A.; Fagnoni, M. J. Org. Chem. 2012, 77,
9
(
094−9101 and ref 26.
29) Myers, A. G.; Dragovic, P. S.; Kuo, E. Y. J. Am. Chem. Soc. 1992,
1
(
(
14, 9369−9386.
30) Dichiarante, V.; Fagnoni, M. Synlett 2008, 787−800.
31) Abel, B.; Assmann, J.; Buback, M.; Grimm, C.; Kling, M.;
Schmatz, S.; Schroeder, J.; Witte, T. J. Phys. Chem. A 2003, 107, 9499−
510.
32) Manet, I.; Monti, S.; Bortolus, P.; Fagnoni, M.; Albini, A.
Chem.Eur. J. 2005, 11, 4274−4282.
33) (a) For reviews on the diradical/zwitterion dichotomy in
9
(
(
synthesis, see: Mohamed, R. K.; Peterson, P. W.; Alabugin, I. V. Chem.
Rev. 2013, 113, 7089−7129. (b) Peterson, P. W.; Mohamed, R. K.;
Alabugin, I. V. Eur. J. Org. Chem. 2013, 2505−2527.
(
34) Velema, W. A.; Szymanski, W.; Feringa, B. L. J. Am. Chem. Soc.
2014, 136, 2178−2191.
8
58
dx.doi.org/10.1021/jo502318v | J. Org. Chem. 2015, 80, 852−858