Photochemical elimination of leaving groups from zwitterionic intermediates generated via electrocyclic ring closure of α,β-unsaturated anilides

Article abstract of DOI:10.1021/jo8017445

(Chemical Equation Presented) Methacrylanilides, ArN(CH₃) COC(CH₂LG)=CH₂, with allylic leaving groups (LG ^- = BocAla, PhCO₂^-, PhCH₂CO ₂^-, PhO^-) u

Full text of DOI:10.1021/jo8017445

Photochemical Elimination of Leaving Groups from Zwitterionic
Intermediates Generated via Electrocyclic Ring Closure of
r,ꢀ-Unsaturated Anilides
Jinli Jia, Majher Sarker, Mark G. Steinmetz,* Ruchi Shukla, and Rajendra Rathore
Department of Chemistry, Marquette UniVersity, Milwaukee, Wisconsin 53201-1881
mark.steinmetz@marquette.edu
ReceiVed August 5, 2008
Methacrylanilides, ArN(CH3)COC(CH2LG)dCH2, with allylic leaving groups (LG- ) BocAla, PhCO2 ,
-
PhCH2CO2 , PhO-) undergo photochemical electrocyclic ring closure to produce a zwitterionic
-
intermediate. Further reaction of the intermediate results in expulsion of the leaving group to give an
R-methylene lactam as the major product. In addition, a lactam product that retains the leaving group is
formed via a 1,5-H shift in the intermediate. Elimination of the leaving group is generally preferred,
even for LG- ) PhO-, although in benzene as the solvent the lactam retaining the phenolate group
becomes the sole photoproduct. The electrocyclic ring closure occurs in the singlet excited-state for the
para-COPh-substituted anilide derivative and is not quenched by 0.15 M piperylene or 0.01 M sodium
2-naphthalenesulfonate (2-NPS) as triplet quenchers. Comparable concentrations of 2-NPS strongly quench
the transient absorption of the triplet excited state observed at 450-700 nm according to laser flash
photolysis experiments. In aqueous media, quantum yields for total products are insensitive to leaving
group ability, and Φtot(para-CO2CH3) ) 0.04-0.06 at 310 nm and Φtot(para-COPh) ) 0.08-0.1 at 365
nm, for which Φisc ) 0.15.
Introduction
The putative zwitterionic intermediates involved in anilide
photocyclizations are of interest because recent studies of R-keto
amide photochemistry have shown7,8 that analogous zwitterionic
intermediates are capable of eliminating leaving group anions
ranging in basicity from carboxylates to phenolates. Zwitterionic
intermediates have also been postulated to account for leaving
group expulsions observed upon photochemical electrocycliza-
tion of enamides and benzanilides.9-11 We have focused upon
exploiting zwitterionic intermediates, generated via such excited-
Nonoxidative photocyclizations of R,ꢀ-unsaturated anilides
such as 1a to give lactams have been known for over 40 years.1,2
The photochemical reaction likely proceeds via an electrocyclic
ring closure to give an intermediate that has zwitterionic
character (Scheme 1).3,4 The zwitterionic intermediate is
converted to a lactam photoproduct via either intramolecular
1,5-H rearrangement or a series of proton transfers, depending
on the solvent and the substituents attached to the amide
nitrogen.3,5,6
(5) Ninomiya, I.; Yamauchi, S.; Kiguchi, T.; Shinohara, A.; Naito, T. J. Chem.
Soc., Perkin Trans. 1 1974, 1747–1751.
(6) Kanaoka, Y.; Itoh, K.; Hatanaka, Y.; Flippen, J. L.; Karle, I. L.; Witkop,
B. J. Org. Chem. 1975, 40, 3001–3003.
(7) (a) Ma, C.; Steinmetz, M. G.; Cheng, Q.; Jayaraman, V. Org. Lett. 2003,
5, 71–74. (b) Ma, C.; Steinmetz, M. G.; Kopatz, E. J.; Rathore, R. Tetrahedron
Lett. 2005, 46, 1045–1048. (c) Ma, C.; Steinmetz, M. G.; Kopatz, E. J.; Rathore,
R. J. Org. Chem. 2005, 70, 4431–4442.
(1) (a) Ninomiya, I.; Naito, T. Heterocycles 1981, 15, 1433–1462. (b) Lenz,
G. R. Synthesis 1978, 489–518. (c) Campbell, A. L.; Lenz, G. R. Synthesis 1987,
421–452.
(2) (a) Nishio, T.; Tabata, M.; Koyama, H.; Sakamoto, M. HelV. Chim. Acta
2005, 88, 78–86. (b) Nishio, T.; Koyama, H.; Sasaki, D.; Sakamoto, M. HelV.
Chim. Acta 2005, 88, 996–1003.
(8) (a) Ma, C.; Steinmetz, M. G. Org. Lett. 2004, 6, 629–632. (b) Chen, Y.;
Steinmetz, M. G. J. Org. Chem. 2006, 71, 6053–6060.
(3) Cleveland, P. G.; Chapman, O. L. Chem. Commun. 1967, 1064–1065.
(4) Ogata, Y.; Takagi, K.; Ishino, I. J. Org. Chem. 1971, 36, 3975–3979.
(9) (a) Lenz, G. R. J. Org. Chem. 1974, 39, 2839–2845. (b) Ninomiya, I.;
Naito, T.; Takasugi, H. J. Chem. Soc., Perkin Trans 1 1975, 1720–1724.
10.1021/jo8017445 CCC: $40.75
Published on Web 10/22/2008
2008 American Chemical Society
J. Org. Chem. 2008, 73, 8867–8879 8867

Jia et al.
SCHEME 1
SCHEME 2
of 4a,b. The total quantum yields for 5a,b + 6a,b in Scheme
2 are insensitive to leaving group basicity, and they are similar
to those found for formation of lactams 3b,c from 1b,c, which
have no leaving groups. These data suggest that expulsion of
the leaving groups does not occur directly in an excited state
of 4a,b.
At the outset of our study, the para substituents, Y ) CO2CH3
and Y ) PhCO in 4a,b, were thought to be necessary because
recent reports2 showed that such para-substituted derivatives
1b,c were capable of overcoming a previously reported4 strongly
adverse polar solvent effect on the photoreactivity of 1a, which
was thought to be due to efficient intersystem crossing in polar
solvents to give an unproductive triplet excited state. We
therefore investigated the multiplicity of the photochemistry of
the benzoyl derivatives 1c and 4b. Our results establish that
the reactive excited-state for both compounds is the singlet
excited-state and that the efficiency of intersystem crossing is
insensitive to solvent polarity.
state electrocyclic ring closure reactions, for the ultimate purpose
of releasing leaving groups that are biologically active. Photo-
chemical generation of zwitterionic intermediates that release
bioeffector leaving groups would constitute a new approach for
the design of caged biological substrates that are generally used
to trigger biological processes under physiological conditions.12
The R,ꢀ-unsaturated anilide photoremovable protecting group
can be synthesized through acylation of aniline derivatives. Since
the aniline group is a common structural motif in long-
wavelength absorbing organic dyes, R,ꢀ-unsaturated anilides
would offer the prospect for achieving the release of biologically
important leaving groups at biologically benign wavelengths.
Substituent effects on efficiencies for excited-state electro-
cyclic ring closure were explored by comparing the quantum
yields for 1b with anilides 8a,b, which should be subject to a
“meta-ortho” effect in the excited-state (eqs 1 and 2).14,15
In this paper, we report in full13 on our mechanistic in-
vestigation of the photochemical elimination reactions of anilides
4a,b bearing various leaving groups (LG-) at the allylic position
of the R-methylacrylamide group. Under aqueous conditions,
expulsion of the leaving groups generally occurs to give R-meth-
ylene lactams 5a,b as the cleavage coproducts (Scheme 2).
Although lactams 6a,b, which retain the leaving groups, are
usually found to be minor products, 6a,b can become the
principal photoproducts if the leaving group is poor (LG-
)
HO-) or if the solvent is changed from an aqueous buffer to a
nonpolar solvent such as benzene, as in the case of a relatively
basic leaving group such as phenoxide (LG- ) PhO-). These
results are consistent with the intermediacy of a ground-state
zwitterionic species, 7a,b, which partitions between 5a,b and
6a,b in the photochemistry of 4a,b (Scheme 2).
The zwitterionic intermediates 7a,b are thought to be formed
by an electrocyclic ring closure step that occurs in the excited
state. We show that the presence of the allylic leaving groups
has little effect on the efficiencies for this ring closure step,
according to quantum yields for products of direct photolysis
Accordingly, the electron-donating substituents in the aromatic
ring of 8a or 8b should transmit electron density to the meta
and ortho positions in the singlet excited state, which would
potentially promote electrocyclization with the electron-
deficient methacrylamide double bond. However, quantum
yields were found to be essentially the same for 8a and 1b
or lowered with 8b.
(10) (a) Ninomiya, I.; Kiguchi, T.; Naito, T. Chem. Commun. 1974, 81–82.
(b) Ninomiya, I.; Kiguchi, T.; Yamamoto, O.; Naito, T. J. Chem. Soc., Perkin
Trans. 1 1979, 1723–1728. (c) Ninomiya, I.; Kiguchi, T.; Yamauchi, S.; Naito,
T. J. Chem. Soc., Perkin Trans. 1 1980, 197–202.
(13) For the preliminary account of part of this work, see: Jia, J.; Steinmetz,
M. G.; Shukla, R.; Rathore, R. Tetrahedron Lett. 2008, 49, 4621–4623.
(14) (a) Zimmerman, H. E.; Sandel, V. R. J. Am. Chem. Soc. 1963, 85, 915–
922. (b) Zimmerman, H. E.; Somasekhara, S. J. Am. Chem. Soc. 1963, 85, 922–
927.
(15) (a) Zimmerman, H. E. J. Am. Chem. Soc. 1995, 117, 8988–8991. (b)
Zimmerman, H. E. J. Phys. Chem. A 1998, 102, 561605621.
(11) Kanaoka, Y.; Itoh, K. Chem. Commun. 1973, 647–648.
(12) (a) Marriot, G., Ed. Methods in Enzymology; Academic Press: San Diego,
1998; Vol. 291. (b) Goeldner, M.; Givens, R., Ed. Dynamic Studies in Biology;
Wiley-VCH: Weinheim, 2005. (c) Mayer, G.; Heckel, A. Angew. Chem., Int.
Ed. 2006, 45, 4900–4921. (d) Pelliccioli, A. P.; Wirz, J. Photochem. Photobiol.
Sci. 2002, 1, 441–458.
8868 J. Org. Chem. Vol. 73, No. 22, 2008

Photochemical Elimination of LeaVing Groups
SCHEME 3
FIGURE 1. Absorption spectra of 0.156 mM 4a (LG- ) PhCH2CO2
----) and 0.169 mM 4b (LG- ) PhCH2CO2-, -) in CH3CN.
,
-
absorption extended to longer wavelengths, and photolyses of
these compounds could routinely be conducted at 365 nm (ε
127 for LG- ) PhCH2CO2-).
Preparative Direct Photolyses. Preparative direct photolyses
of 0.006 M 4a (LG- ) PhCH2CO2-) with unfiltered light from
a medium pressure mercury lamp in N2 saturated 50% aq
CH3CN containing 100 mM phosphate buffer at pH 7 for 7 h
resulted in the release of phenyl acetic acid to give 5a as the
cleavage coproduct. A minor accompanying product was lactam
6a (LG- ) PhCH2CO2-) according to 1H NMR spectroscopy,
which showed ca. 2:1 ratio of 5a/6a (eq 1). The products were
separated chromatographically to obtain pure 5a.
Results
Photochemical Reactants. The synthesis of anilides 4a
involved the acylation of N-methylaniline 10 (Scheme 3) to give
the acrylamide 11a. Baylis-Hillman reaction16 of the acryla-
mide 11a furnished the allylic alcohol 4a (LG- ) HO-) after
48 h of sonification at 50-60 °C. The carboxylate leaving
groups were introduced via acylation of the allylic hydroxyl
group of 4a with PhCH2COCl or PhCOCl in CH2Cl2 with
(C2H5)3N as base or by coupling of BocAla using DCC and
DMAP in CH2Cl2. The allylic alcohol was converted into the
phenolate derivative by Mitsunobu reaction (PhOH, DEAD,
Ph3P, DMF, or THF).17
The synthesis of anilides 4b followed a similar sequence of
reactions as for 4a (Scheme 3). 4-Aminobenzophenone 12
served as the starting material, in which case the tert-amide
11b was obtained by methylation of amide 13.
The minor lactam 6a (LG- ) PhCH2CO2-) was difficult to
separate and therefore was independently synthesized (Scheme
4) via acylation of alcohol 6a (LG- ) HO-), which was
obtained as the sole product from a preparative scale photolysis
of 0.008 M 4a (LG- ) HO-) in N2 saturated benzene containing
a small amount of p-methoxyphenol as a radical inhibitor
Anilides 1b,c, which have no leaving groups, were synthe-
sized for comparisons of their excited-state properties and
photochemistry to those of 4a,b. These amides were readily
obtained by acylation of arylamines 10 and 12 using R-methy-
lacryloyl chloride in CH2Cl2 containing (C2H5)3N, followed by
a further alkylation step (NaH, DMF, then CH3I) for 1c. The
4-methoxy and 3,5-dimethoxyanilides 8a,b were synthesized
from 4-methoxyaniline and 3,5-dimethoxyaniline in >80%
overall yields following the same acylation/methylation sequence
as used for the synthesis of 4b (Experimental Section).
1H NMR analyses showed that anilides 4a,b, bearing various
leaving groups, were stable for at least one week in 50% D2O
in CD3CN containing 100 mM phosphate buffer at pD 7. In
(Experimental Section).
Direct photolyses of solutions of 0.06 M 4a (LG- ) PhCO2
,
-
BocAla) in 50% D2O in CD3CN containing phosphate buffer
at pD 7 in NMR tubes using a Pyrex filter gave R-methylene
lactam 5a upon loss of the leaving group. Lactams 6a (LG- )
BocAla, PhCO2-) were observed as minor products. Samples
of both lactams 6a (LG- ) PhCO2-, BocAla) were obtained
by independent synthesis through acylation of 6a (LG- ) HO-)
(Scheme 4). Similarly, photolysis of the phenolate derivative
4a (LG- ) PhO-) in 50% D2O in CD3CN containing buffer
produced R-methylene lactam 5a upon loss of the leaving group,
along with lactam 6a (LG- ) PhO-) as a minor product. When
the solvent was changed to benzene, photolysis gave 6a (LG-
) PhO-) as the sole product, which was also obtained on a
preparative scale by photolysis of 10-3 M 4a (LG- ) PhO-).
both aqueous CH3CN and neat CH3CN, anilide 4a (LG-
)
PhCH2CO2-) showed a UV absorption maximum below 300
nm, which tailed out into the 300-350 nm region (Figure 1).
For the benzoyl derivative 4b (LG- ) PhCH2CO2-), the UV
(16) (a) Ultrasound16b was used with DABCO as base.16c (b) Feltrin, M. P.;
Almeida, W. P. Synth. Commun. 2003, 33, 1141–1146. (c) Faltin, C.; Fleming,
E. M.; Connon, S. J. J. Org. Chem. 2004, 69, 6496–6499.
(17) Kraus, G. A.; Zhang, N. J. Org. Chem. 2000, 65, 5644–5646.
J. Org. Chem. Vol. 73, No. 22, 2008 8869

Jia et al.
TABLE 2. Chemical Yieldsa for Photolyses of 4b (LG-
PhCH2CO2-, HO-, PhO-) in Various Solvents
)
SCHEME 4
reactant, LG- solvent LG-H, % 5b, % 6b, % unreacted 4b, %
bufferb
CD2Cl2
C6D6
71
71
53
55
57
6.9
9.9
45
25
31
19
87
84
20
85
5.4
11
10
9.5
5.8
36
10
-
PhCH2CO2
70
HO-
bufferb
CD2Cl2
buffer
C6D6
ndc
ndc
ndc
ndc
PhO-
0
a Yields determined by 1H NMR spectroscopy using DMSO as
standard. b 50% D2O in CD3CN containing 100 mM phosphate buffer at
pD 7. c Not determined.
TABLE 3. Quantum Yields for Formation of Products from
Direct Photolyses of 4a and 1ba
leaving group, LG-
solvent
Φ (5a)
Φ (6a)
0.012
0.0078
0.011
BocAla (pKa 4.0b)
bufferc
bufferc
bufferc
bufferc
C6H6
0.052
0.032
0.031
0.037
0
b
-
PhCO2 (pKa 4.2 )
b
-
PhCH2CO2 (pKa 4.3 )
PhO- (pKa 10b)
0.017
0.072
none (1b)
bufferc
C6H6
nad
0.046 (3b)e
0.083 (3b)e
0.077 (3b)e
TABLE 1. Chemical Yieldsa for Photolyses of 4a (LG- ) BocAla,
nad
nad
PhCO2, PhCH2CO2-, HO-, PhO-) in Various Solvents
hexane
reactant, LG- solvent LG-H, % 5a, % 6a, % unreacted 4a, %
a Average of two or more independent runs using ferrioxalate as
actinometer; products were quantified by HPLC using the internal
standard method to calibrate the 254 nm detector. b pKa of the conjugate
acid, see ref 18. c N2 saturated 50% aq CH3CN containing 100 mM
phosphate buffer at pH 7. d Not applicable without a leaving group.
e The product is 3b.
BocAla
PhCO2
PhCH2CO2
bufferb
bufferb
bufferb
CD3CN
CD2Cl2
C6D6
ndd
ndd
24
36
40
18c
19c
22
10
16
31
0
7.3c
70
71
67
51
56
47
0
7.4c
-
-
11
10
9.3
6.8
97
48
HO-
C6D6
ndd
ndd
ndd
coproduct could be isolated chromatographically, the minor
lactam 6b (LG- ) PhCH2CO2-) was difficult to separate and
therefore was obtained, independently, by acylation of allylic
alcohol 6b (LG- ) HO-), which was produced upon preparative
direct photolysis of 4b (LG- ) HO-) (Experimental Section).
Similarly, direct photolysis of the phenolate derivative 4b (LG-
) PhO-) in buffer gave 5b and 6b in a 2:1 ratio, whereas in
benzene, only lactam 6b (LG- ) PhO-) was obtained as a
photoproduct.
PhO-
bufferb
C6D6
30
0
18
90
49
16
a Yields determined by 1H NMR spectroscopy using DMSO as
standard. b 50% D2O in CD3CN containing 100 mM phosphate buffer at
pD 7. c Quantified by HPLC with added biphenyl as a standard and by
1H NMR spectroscopy using DMSO as standard. d Not determined.
Chemical yields for photolyses of 10-2 M 4a (LG-
)
PhCH2CO2-, HO-, PhO-) in various air-saturated solvents using
Pyrex filtered light are collected in Table 1. The yields of 5a
were significantly lower than the yields of the released leaving
group acid (LG- ) PhCH2CO2-), evidently due to secondary
photolysis of 5a, which gave unidentifiable material exhibiting
Chemical yields for photolyses of 10-2 M 4b (LG-
)
PhCH2CO2-, HO-, PhO-) in various air-saturated solvents in
NMR tubes using Pyrex-filtered light are collected in Table 2.
Generally, the photolyses could be taken to very high conver-
sions, although the yields of 5b (LG- ) PhCH2CO2-) were
less than the yields of the released carboxylic acid, evidently
due to secondary photolysis.
1
only broad peaks in H NMR spectra.
Control experiments showed that neither of the lactams 6a
(LG- ) PhCH2CO2-, BocAla) underwent dark reaction in 50%
buffer in CD3CN at pD 7 for over a week. This observation
rules out a “dark” reaction that would diminish the yields of
6a, such as solvolytic elimination of a leaving group, to afford
5a or its isomer with an endocyclic double bond. Control
experiments also showed that solutions of R-methylene lactam
5a were unreactive in aq CD3CN containing buffer at pD 7 in
the dark for a period of one week. In solid form, however, 5a
slowly decomposed to form unidentifiable products, which gave
only broad peaks in 1H NMR spectra. Thus, all results requiring
use of 5a were obtained with fresh samples after chromato-
graphic purification.
When 5b was stored in the dark as a solid, like 5a (vide
supra), a slow reaction was observed to give unknown products,
which showed broad peaks in the 1H NMR spectra. This “dark”
reaction could be slowed substantially by storing 5b as a dilute
solution in 50% aq CH3CN containing buffer. Nevertheless, all
results requiring use of 5b were obtained with fresh samples
after chromatographic purification. A control experiment with
lactam 6b (LG- ) PhCH2CO2-) showed it was stable for at
least a week in 50% D2O in CD3CN containing buffer at pD 7
1
according to H NMR analyses.
Preparative direct photolyses of 10-3-10-2 M 1b,c and 8a,b
in N2-saturated 50% aq CH3CN containing buffer at pH 7 gave
lactams 3b,c2 (Scheme 1), 9a (eq 1), and 9b (eq 2) as the only
product in each case (Experimental Section).
Preparative direct photolyses with Pyrex-filtered light of 0.005
M 4b (LG- ) PhCH2CO2-) in N2 saturated 50% aq CH3CN
containing buffer at pH 7 for 2 h gave 5b as the major
Quantum Yields. Quantum yields for products of 1b,c and
4a,b (Tables 3 and 4) were determined for photolyses in N2
saturated 50% aq CH3CN containing 100 mM phosphate buffer
photoproduct along with minor amounts of 6b (LG-
)
PhCH2CO2-) in an ca. 2: 1 ratio. Whereas the cleavage
8870 J. Org. Chem. Vol. 73, No. 22, 2008

Photochemical Elimination of LeaVing Groups
the excited state,14,15 which, a priori, could result in a more
efficient electrocyclization step. However, for photolyses in 50%
aq CH3CN containing buffer, the quantum yield for the meta-
dimethoxy compound 8a is essentially identical to 1b, whereas
para-methoxy substitution in 8b causes the quantum yield to
decrease.
TABLE 4. Quantum Yields for Formation of Products from
Direct Photolyses of 4b and 1ca
leaving group,
LG-
Φ
Φ
solvent
Additive
none
(5b)
(6b)
bufferc
bufferc
buffer
0.069 0.018
-
PhCH2CO2
6.44 × 10-3 NPSb 0.070 0.017
11.4 × 10-3 NPSb 0.075 0.021
PhO-
bufferc
C6H6
none
none
none
0.061 0.016
0
0.10
none (1c)
bufferc
bufferc
bufferc
20% aq
CH3CN
nad
0.077 (3c)e
0.072 (3c)e
0.070 (3c)e
0.10 (3c)e
6.69 × 10-3 NPSb nad
12.4 × 10-3 NPSb nad
0.15 M trans-
nad
piperylene
a Average of two or more independent runs using ferrioxalate as acti-
nometer; products were quantified by HPLC using the internal standard
method to calibrate the 254 nm detector. b Sodium 2-naphthalene-
sulfonate as the quencher. c N2 saturated 50% aq CH3CN containing 100
mM phosphate buffer at pH 7. d Not applicable without a leaving group.
e The product is 3c.
Quantum yields for direct photolysis of the benzoyl-
substituted anilide 4b (LG- ) PhCH2CO2-, PhO-) (Table 4)
are significantly higher than those for 4a. Regardless of the
leaving group, the total quantum yields for products (5a + 6b)
are nearly the same as the quantum yield for formation of 3c
from 1c, which has no leaving group, consistent with the
electrocyclic ring closure step as governing the efficiency of
the reaction. Like 4a (LG- ) PhO-), the ratio of 5b/6b formed
from 4b (LG- ) PhO-) is solvent dependent, which supports
the involvement of a ground-state intermediate in the photo-
chemistry. The photorelease of the phenolate group to give
R-methylene lactam 5b strongly predominates in buffer, whereas
in nonpolar aprotic solvent the intermediate instead rearranges
to lactam 6b. The overall photoreaction is also somewhat more
efficient in a nonpolar solvent, such as benzene, as compared
to aqueous media.
The photochemical reactions of 1c and 4b are not quenched
by the water-soluble triplet quencher, sodium 2-naphthalene-
sulfonate in buffer, or by trans-piperylene (Table 4) in aq
CH3CN. The longer-wavelength absorptions of 1c and 4b permit
photolyses to be conducted at 365 nm in the presence of the
quenchers, which absorb light at much shorter wavelengths. The
triplet energy, ET, of 1b and 4b should be close to 68 kcal/mol,
while ET of sodium naphthalene sulfonate is assumed to be
similar to naphthalene (ca. 61 kcal/mol20). Thus, the quenching
of the triplet excited states of 1c and 4b is expected to proceed
via diffusion-controlled energy transfer. The laser flash pho-
tolysis experiments with 1c and 4b (vide infra) further show
that the triplet excited states are quenched at similar or lower
concentrations of the sodium 2-naphthalenesulfonate than used
in the steady-state quenching experiments (Table 4). Therefore,
the singlet excited-state is the reactive excited-state in the
formation of products from 1c and 4b.
at pH 7 and in other solvents at low conversions to ascertain
whether the leaving group eliminations occurred directly in the
excited state or whether a ground-state intermediate such as the
zwitterionic species 7a,b was involved, which could partition
between products 5a,b and 6a,b in ratios depending upon
leaving group ability and solvent polarity, as implied by Scheme
2. As shown in Table 3, the quantum yields for 4a are insensitive
to the basicity of the leaving group, which argues against the
elimination of the leaving group as occurring directly in the
excited state. For the series of leaving groups of increasing
basicity (LG- ) BocAla, PhCO2-, PhCH2CO2-, PhO-) the total
quantum yields 5a + 6a measured for buffer as the solvent are
nearly the same within experimental error to the quantum yield
for formation of lactam 3b from 1b, which has no leaving group
(Table 3 and Scheme 1, vide supra). The quantum yields are
consistent with the ring-closure step as taking place in the
excited state, prior to the elimination of the leaving group. The
efficiency for excited-state electrocyclic ring closure is not
expected to be influenced by the basicity of the leaving group.
It is influenced, however, by the solvent. For example, the
quantum yields for formation of lactam 3b from 1b in benzene
or hexane are nearly twice those for aqueous buffer solutions
as the medium.
The ratio of 5a/6a is not markedly sensitive to leaving group
basicity. Although the quantum yields for 4a (LG- ) PhO-)
in buffer are very similar to those substrates with better leaving
groups, when the solvent is changed to benzene, formation of
lactam 6a (LG- ) PhO-) becomes the only reaction and
elimination of the leaving group to form 5a is not observed.
Although we did not determine quantum yields for carboxylate
leaving groups with benzene as solvent, product yields for 4a
(LG- ) PhCH2CO2-) in Table 1 clearly show that, unlike the
phenolate group, the carboxylate leaving group is photochemi-
cally released to give R-methylene lactam 5a under both aqueous
and nonaqueous conditions.
Deuterium Labeling. Deuterium labeling experiments were
performed with 1b to determine whether a ground-state inter-
mediate can be intercepted by protic solvent to form lactam 3b
containing deuterium (Scheme 1). Photolyses of 1b in 50% D2O
in CD3CN containing 100 mM phosphate buffer at pD 7 as the
solvent resulted in incorporation of deuterium into lactam 3b
1
to an extent of 74%, according to H NMR spectroscopy and
GC-MS analysis (eq 4). A control experiment showed that
undeuterated lactam 3b does not undergo significant hydrogen-
deuterium exchange under the reaction conditions. Thus, for
photolyses in protic solvents or aqueous media, much of the
lactam 3b is expected to be produced via a mechanism involving
proton transfer from the protic solvent to an intermediate with
enolate character, e.g., zwitterion 2b, whereas the 1,5-H shift
in zwitterionic intermediate 2b makes a smaller contribution to
the formation of lactam 3b.
Substituent effects on quantum yields for reaction were
explored for 8a,b (eqs 1 and 2) for comparison to 1b (Scheme
1). The electron-donating methoxy substituents and the existant
acrylamide nitrogen should be capable of transmitting electron
density to the ortho and meta positions of the aromatic ring in
(18) (a) pKa values: Rappoport, Z. CRC Handbook of Tables for Organic
Compound Identification, 3rd ed.; Weast, R. C., Ed.; CRC: Cleveland, 1967; pp
429-435. (b) BocAlaOH, pKa 4.02, CA 15761-38-3.
J. Org. Chem. Vol. 73, No. 22, 2008 8871

Jia et al.
Laser Flash Photolyses. Laser flash photolysis experiments
were conducted with 1c and 4b (LG- ) HO-, PhCH2CO2-) at
355 nm in 50% aq CH3CN containing 100 mM phosphate buffer
at pH 7. Each degassed sample under argon gave a transient
absorption with a maximum in the 450-550 nm region with
additional bands extending to 700 nm, which was attributed to
the triplet excited-state of each anilide (Figure 2). The lifetime
of the decay of the transient was essentially identical for the
three compounds: 1c, τ ) 335 ns; 4b (LG- ) HO-), τ ) 358
ns; 4b (LG- ) PhCH2CO2-), τ ) 356 ns. In addition, the
lifetimes for decay were essentially the same at other wave-
lengths over the 450-700 nm wavelength range.
FIGURE 2. Transient triplet absorption spectrum produced upon laser
flash photolysis of 2.5 mM 1c in 50% aq CH3CN containing 0.1 M
phosphate buffer at pH 7 following 10 ns laser excitation at 355 nm.
The spectral bands at 500 nm, 580 nm, and 670 nm showed identical
decay lifetimes (τ ) 335 ns).
The transient absorption spectrum of 1c, 4b (LG- ) HO-),
and 4b (LG- ) PhCH2CO2-) exhibited linear Stern-Volmer
quenching kinetics upon addition of the triplet excited-state
quencher, sodium 2-naphthalene sulfonate (Figure 3). For 1c,
4b (LG- ) HO-), and 4b (LG- ) PhCH2CO2-), the respective
slopes were kq ) 3.77 × 109 M-1 s-1, 2.73 × 109 M-1 s-1, and
4.34 × 109 M-1 s-1, consistent with quenching of the lowest
triplet excited state of each compound by triplet energy transfer
to the quencher.
Compound 1c shows weak fluorescence in the 370-500 nm
region according to steady-state experiments. However, the
fluorescence efficiency in CH3CN is estimated to be low (Φf <
0.01) by comparison to the fluorescence of quinine bisulfate
(Φf ) 0.55) as a standard.21 The low quantum yield is consistent
with the previous report of inefficient fluorescence of analogous
amide derivatives of p-aminobenzophenone.19 Fluorescence is
also observed in the above laser flash photolysis experiments.
For both 1c and 4b (LG- ) PhCH2CO2-), the fluorescence
decay lifetimes were measured at various wavelengths over a
420-470 nm range and found to be less than the duration of
the laser flash (τ < 10 ns).
Triplet Yields. Considering the likelihood that the photore-
activity of 1c and 4b originates from the singlet excited state,
according to the above quenching experiments, it became
important to determine the efficiency of intersystem crossing
as a process that would compete with reaction in the singlet
excited state. Quantum yields for intersystem crossing, Φisc,
were measured for 1c using literature procedures,22 which
involved quenching of its triplet excited-state by trans-piperylene
and determining the extent of trans-cis isomerization of the
1,3-diene quencher for a given amount of light absorbed. The
3
FIGURE 3. Stern-Volmer plot of τ-1 for 1c (b), 4b (LG- ) HO-,
∆), and 4b (LG- ) PhCH2CO2-, O) vs concentration of sodium
2-naphthalenesulfonate used as quencher, Q.
SCHEME 5
(19) Allen, N. S.; Salleh, N. G.; Edge, M.; Corrales, T.; Shah, M.; Weddell,
I.; Catalina, F.; Green, A. J. Photochem. Photobiol. A: Chem. 1996, 99, 191–
196.
(20) Murov, S. L.; Carmichael, I.; Hug, G. L. Handbook of Photochemistry,
2nd ed.; Marcel Dekker: New York, 1993; pp 30-31.
(21) (a) Meech, S. R.; Phillips, D. J. Photochem. 1983, 23, 193–217. (b)
Hamai, S.; Hirayama, F. J. Phys. Chem. 1983, 87, 83–89. (c) Melhuish, W. H.
J. Phys. Chem. 1961, 65, 229–235. (d) Eaton, D. F. In Handbook of Organic
Photochemistry; Scaiano, J. C., Ed.; CRC Press: Boca Raton, FL, 1989; Vol. I ,
Chapter 8.
experiments were performed with benzophenone, compound 1c,
and additionally the N-acetyl derivative of 4-(methylamino)ben-
zophenone 15 (Scheme 5), which does not undergo photochemi-
cal reaction.
In benzene, the acetamide 15 undergoes intersystem crossing
(22) (a) Lamola, A. A.; Hammond, G. S. J. Chem. Phys. 1965, 43, 2129–
2135. (b) See ref, 20 pp 310-312.
(Φisc ) 0.93) almost as efficiently as benzophenone (Φisc
)
8872 J. Org. Chem. Vol. 73, No. 22, 2008

Photochemical Elimination of LeaVing Groups
SCHEME 6
SCHEME 7
1.0020), and the efficiency remains similarly high in going to
the polar solvent, 21% aq CH3CN, for which Φisc (15) ) 0.95.
For the R,ꢀ-unsaturated anilide 1c, the quantum yields are
considerably lower, Φisc ) 0.20 in benzene and Φisc ) 0.15 in
20% aq CH3CN. In the presence of the 0.15 M trans-piperylene,
the quantum yield for photorearrangement of 1c to give 3c was
0.10 in 20% aq CH3CN, and thus the photorearrangement is
not quenched under conditions whereby energy transfer to trans-
piperylene results in trans-cis isomerization. Thus, the triplet
yields do not vary significantly with solvent polarity, and for
1c, intersystem crossing is not a highly efficient decay pathway
of the singlet excited state.
and 4b (LG- ) PhCH2CO2-), which have widely differing types
of allylic substituents.
The results of our study of the photochemistry of 4a,b are
consistent with zwitterions 7a,b as playing an important role
as intermediates involved in the formation of elimination
products 5a,b and lactams 6a,b (Scheme 2). Similarly, zwitte-
rionic intermediates 2b,c would be expected to be the key
intermediates in the formation of lactams 3b,c from anilides
1b,c (Scheme 1). The photochemical ring closure of anilides
1b,c and 4a,b should produce these intermediates via an excited-
state allowed conrotatory electrocyclic ring closure reaction.
Such a ring closure would be consistent with the mesomeric
nature of the amide group, which should have significant double
bond character between the carbonyl group and amide nitrogen.
Although our study provides no information on the stereochem-
istry of the ring closure step itself, previous studies support such
a conrotatory mode of ring closure.5,6 An example is provided
by the photochemistry of 20-d5 (L ) D) under aprotic conditions
(Scheme 7).6 In this case, a conrotatory ring closure step, which
produces 21 (Y ) D), would be needed to account for the
stereochemistry of deuterium in the trans-fused product 22 (Y
) D) since the second step of the reaction should be a thermally
allowed, suprafacial 1,5-D shift.
In protic solvents, the 1,5-H shift may not be the sole
mechanism for the formation of lactams 3a-c and 6a,b from
photolysis of 1a-c and 4a,b. The other mechanism involves a
proton transfer from protic solvent at some stage in the
conversion of the zwitterionic intermediates 2a-c and 7a,b to
lactam products. Such a proton transfer to form enol 24 evidently
occurs upon photolysis of 1b (Scheme 8) in CD3CN containing
50% phosphate buffer in D2O at pD 7, given the observed
extensive incorporation of deuterium into the R-position of 3b.
In CD3CN containing 50% buffer in D2O (pD 7), the lactam
3b does not undergo H/D exchange in the dark. According to
the literature,6 if a protic solvent is used in the photolysis of 20
(L ) H), then the major product becomes the cis-fused isomer
23 rather than 22 (Scheme 7). External deuterium is incorporated
into 23 (Y ) D) from the solvent when undeuterated 20 (L )
H) is photolyzed in 10% D2O in acetonitrile. This deuterated
Discussion
The principal excited-state reaction of 4a,b is thought to be
electrocyclic ring closure to give 7a,b as an intermediate
(Scheme 2 and path A in Scheme 6). The quantum efficiencies
for the excited-state ring closure are not sensitive to the nature
of the remote leaving groups in 4a,b.
The overall quantum yields for reaction of 4a,b to give 5a,b
+ 6a,b are very similar to those observed for the photocycliza-
tion of the para-substituted derivatives 1b,c (Tables 3 and 4),
which have no leaving groups and can only photocyclize to
lactams 3b,c (Scheme 1). These results are contrary to what
would be expected if the leaving groups were expelled as anions
directly in the excited states of 4a,b (path B, Scheme 6). In the
excited-state eliminations, the quantum yields should decrease
with increasing basicity of the leaving group expelled because
the anion expulsions would be competing with rapid decay
processes of the excited state. Homolytic cleavage of the leaving
groups in the excited-state is considered to be unlikely (path C,
Scheme 6) because in the case of 4a,b (LG- ) PhCH2CO2
)
-
radical byproducts, such as toluene or bibenzyl, are not observed
from decarboxylation of the R-phenylacetyloxyl radical,23 which
should also have led to low yields of the photochemically
released phenylacetic acid (Tables 1 and 2). Homolysis might
occur in the triplet excited-state. However, the laser flash
photolysis experiments show that the lifetimes of the transient
absorption attributable to the triplet excited-state are essentially
invariant for the series of compounds 1c, 4b (LG- ) HO-),
(23) (a) Hilborn, J. W.; Pincock, J. A. J. Am. Chem. Soc. 1991, 113, 2683–
2686. (b) Pincock, J. A. Acc. Chem. Res. 1997, 30, 43–49. (c) Banerjee, A.;
Falvey, D. E. J. Am. Chem. Soc. 1998, 120, 2965–2966.
J. Org. Chem. Vol. 73, No. 22, 2008 8873

Jia et al.
SCHEME 8
substituent effects on the excited-state electrocyclic reaction has
not been reported, however. Our attempt to increase the rates
of cyclization and hence quantum yields by introducing meta-
or para-methoxy groups resulted in no change in efficiency or
even decreased efficiency with derivatives 8a,b.
Nonproductive radiationless decay to regenerate reactants is
evidently the principal process for the decay of the singlet
excited states of 1c and 4b. We therefore have considered the
possibility that the radiationless decay process could be due to
a reversible photocyclization step whereby the initially formed
zwitterionic intermediates, 2c and 7b, rapidly revert to ground-
state reactants. If the leaving groups are expelled directly from
the zwitterionic intermediate 7b, rapid regeneration of ground-
state reactant upon disrotatory ring opening of the zwitterionic
intermediates should result in decreased quantum yields for 5b
as well as decreased total quantum yields for products 5b + 6b
as the basicity of the leaving group increases. Experimentally,
these quantum yields are found to be insensitive to leaving group
effects (vide supra). Similar reasoning would argue against a
reversible electrocyclization step to form 7a from 1a since the
total quantum yields for 5a + 6a are unaffected by leaving
groups having a wide range of basicities.
product, formed in 34% yield, is accompanied by 28% of
unlabeled trans-fused product 22, produced via the 1,5-H shift
in zwitterionic intermediate 21 (L ) H, Y ) H). The formation
of 23 upon protonation of 21 probably involves an enol
intermediate. The cis-fused isomer 23 is more stable than trans-
fused compound 22 (Y ) H), which would be consistent with
the base epimerization of 22 (Y ) H) to 23 (Y ) H).
In the case of 4a, the product ratio 6a/5a should increase
with increasing basicity of the leaving group as a consequence
of the partitioning of 7a between these two products, according
to Scheme 2. However, very little variation in product ratio is
observed in polar solvent (50% buffer in CH3CN) as the leaving
group is varied. The insensitivity of product ratio to leaving
group ability is rather similar to what was observed for R-keto
amide photochemistry,8 for which the zwitterionic intermediate
exclusively expelled both carboxylate and the more basic
phenolate group, as long as the photolyses were conducted in
an aqueous medium. As in our earlier study, it proved possible
with LG- ) PhO- to influence the relative rate of leaving group
release vs cyclization of the zwitterion by changing the solvent
from buffer in aq CH3CN to a nonpolar solvent, such as benzene.
In the case of 4a,b (LG- ) PhO-), the product ratio favors
elimination of the leaving group to give 5a,b in polar solvent
but reverses in favor of formation of 6a,b in the nonpolar
solvent, benzene.
The singlet excited-state has experimental support as the
reactive excited-state in the photochemistry of 1c and 4b. The
failure to quench the photochemistry of 1c and 4b suggests that
the triplet excited-state is not involved in the photochemistry
of these two cases. The quantum yields for lactam 3b and for
products 5b and 6b are unchanged, when the photolyses are
conducted with the triplet excited-state quencher, sodium
2-naphthalenesulfonate, whereas the laser flash photolysis
experiments show that the transient absorption assigned to the
triplet excited-state is quenched at the same or lower concentra-
tions of quencher. While these results support the singlet excited-
state in the reaction, the quantum yields for reaction of 1c and
4b are relatively low, Φr < 0.1, and only weak fluorescence is
observed with 4b (Φf < 0.01). Therefore, the singlet excited-
state primarily deactivates to the ground-state via other processes
besides product formation and fluorescence. Other possible
deactivating processes of the singlet excited-state are intersystem
crossing and nonproductive radiationless decay to regenerate
ground-state reactant. Of these processes, the latter makes the
larger contribution, since intersystem crossing accounts for only
15% of the singlet excited-state decay of 1c. This relatively
low triplet yield for 1c is somewhat surprising, considering that
intersystem crossing accounts for 93% of the singlet excited-
state decay of the anilide model compound 15 (Scheme 5).24
The total rate of decay of 1c, as well as 4b (LG- ) HO-,
PhCH2CO2-) is estimated to be at least 108 s-1, since the singlet
lifetimes measured in the laser flash photolysis experiments are
found to be less than the duration of the laser flash of ca. 10 ns.
For a quantum yield of reaction of 0.07-0.09 for 1c and 4b,
the rate constant for electrocyclization is estimated to be at least
ca. 106-107 s-1, if the electrocyclization step is irreversible such
that the initially formed zwitterionic intermediates do not un-
dergo disrotatory ring opening to regenerate ground-state reac-
tants.
Although our previous work7,8 shows that all of the leaving
groups studied herein can be expelled directly from zwitterionic
intermediates that are similar to 7a,b, there is an alternate
mechanism to be considered for 4a,b, which involves initial
loss of a proton from zwitterions 7a,b followed by expulsion
of the leaving group from the enolate 25 (Scheme 9). The
alternate mechanism better accounts for the insensitivity of the
observed 5a,b/6a,b product ratios as the leaving group basicity
(24) (a) The reason for the lower Φisc of unsaturated amide 1c as compared
to the acetanilide 15 is not obvious. It could be due to more rapid competing
radiationless decay of S1 to S0 for 1c, perhaps due to the presence of the
unconstrained double bond. However, the actual reason could be more
complicated.24b The S1 radiationless decay mechanism is not understood.
Alternatively, rate constants for intersystem crossing could be lower due to an
energy mismatch of the S1 excited state and a corresponding triplet excited state,
or the energetically proximate triplet excited state might have the same electronic
configuration.24c There does seem to be some degree of quenching of the singlet
excited state by protic solvent, which has precedent.24d The solvent effects on
the quantum yield data suggest that such quenching is more pronounced for 4a
than 4b, but solvent effects do not account for the difference in Φisc between 1c
and 15. (b) Biczok, L.; Berces, T.; Marta, F. J. Phys. Chem. 1993, 97, 8895–
8899. (c) Schuster, D. I.; Goldstein, M. D.; Bane, P. J. Am. Chem. Soc. 1977,
99, 187–193. (d) Yatsuhashi, T.; Nakajima, Y.; Shimada, T.; Tachibana, H.;
Inoue, H. J. Phys. Chem. 1998, 102, 8657–8663.
It seemed possible to increase the rate of cyclization relative
to singlet excited-state decay by use of substituents attached to
the aromatic ring of 1a-c. Disubstitution of hexatrienes, for
example, has been shown to dramatically enhance rates for
electrocyclic ring closure in the ground state, depending on the
substituents.25 To our knowledge, a similar systematic study of
(25) Yu, T.-Q.; Fu, Y.; Liu, L.; Guo, Q.-X. J. Org. Chem. 2006, 71, 6157–
6164.
8874 J. Org. Chem. Vol. 73, No. 22, 2008

Photochemical Elimination of LeaVing Groups
SCHEME 9
photochemistry derives from the singlet excited state. The singlet
excited-state otherwise undergoes mainly rapid radiationless
deactivation to the ground state and, according to the results
with 1c, intersystem crossing to the triplet excited-state with
Φisc ) 0.15-0.20. The triplet excited-state cannot account for
the photochemistry of 1c and 4b. The triplet excited-state is
strongly quenched by energy transfer to efficient triplet quench-
ers, according to the laser flash photolysis studies, whereas
quantum yields for products remain unaffected at the same or
lower concentrations of the quencher used to quench the
transient absorption of the triplet excited state, observed in the
laser flash photolysis studies.
Experimental Section
varies.26 Accordingly, the 2.2-4.1:1 ratio of 5a,b/6a,b for the
leaving groups would largely be established by the competition
between the 1,5-H shift and the deprotonation of the zwitterion.
It nevertheless seems reasonable to expect that direct elimination
of the leaving groups from 7a,b contributes to varying extent,
depending on leaving group basicity and the solvent, to the
formation of 5a. According to Scheme 9, the exclusive formation
of 6a,b (LG- ) PhO-) from 4a,b (LG- ) PhO-) in benzene
would be explained by a slowed rate of PhO- release and a
suppressed rate of proton release from 7a,b, in which case the
carboxylate leaving groups would be expelled directly from the
zwitterions to form an ion-pair intermediate (Tables 1 and 2).
The mechanism in Scheme 9 makes it possible to reconsider
the possibility that 7a,b are photochemically generated in a
reversible step such that they undergo disrotatory ring opening
in the ground-state to regenerate reactants 4a,b. It is difficult
to say whether such a reversible electrocyclization step could
account for the rapid radiationless decay of the singlet excited
states of 1c and 4b and the relatively low quantum yields for
products. Alternatively, rapid excited-state decay could be due
to rotation about the 1,1-disubstituted double bond of the R,ꢀ-
unsaturated amides, although this remains to be tested exper-
imentally.
Preparative chromatographic separations used a 2.5 cm × 28
cm column packed with 230-400 mesh silica gel (Sorbent
technologies), eluting with the specified solvent at a flow rate of
15 mL/min using a pump. HPLC analysis was performed using a
4.6 × 250 mm column of YMC ODS-AQ S-5 120 Å (Waters) at
a flow rate of 1 mL/min with 30-40% aqueous acetonitrile as the
mobile phase.
Preparation of Methyl 4-[[2-(Hydroxymethyl)acryloyl]methyl-
amino]benzoate (4a, LG- ) HO-). The procedure was adapted
from the literature.16 To a mixture of 35 g (0.16 mol) of 11a and
350 mL of THF was added 350 mL of a 40% aqueous formaldehyde
and 35 g (0.31 mol) of DABCO. The suspension was sonicated
for 48 h at 50-60 °C. The mixture was extracted with ethyl acetate.
The extracts were washed with saturated NaHCO3 and brine and
dried over anhydrous sodium sulfate. Concentration in vacuo
followed by chromatography on silica gel, eluting with 25% ethyl
acetate in hexane, gave 16 g (0.064 mol, 40% yield) of compound
4a (LG- ) HO-) as a colorless crystalline solid, mp 82-83 °C.
1
The spectral data were as follows: H NMR (CDCl3) δ 3.37 (s,
3H), 3.88 (s, 3H), 4.21 (s, 2H), 4.93 (s, 1H), 5.28 (s, 1H), 7.26 (d,
J ) 8.0 Hz, 2H), 7.98 (d, J ) 8.0 Hz, 2H); 13C NMR (CDCl3) δ
37.8, 52.5, 64.3, 120.6, 126.5, 128.7, 130.9, 143.2, 148.7, 166.4,
170.6. Anal. Calcd for C13H15NO4: C, 62.64%; H, 6.07%; N, 5.62%.
Found: C, 62.71%; H, 6.06%; N, 5.62%.
Preparation of Methyl 3-(Hydroxymethyl)-1-methyl-2-oxo-
1,2,3,4-tetrahydroquinoline-6-carboxylate (6a, LG- ) HO-)
by Photolysis of 4a (LG- ) HO-). A solution of 0.37 g (2.0 mmol)
of 4a (LG- ) HO-) in N2 saturated 250 mL of benzene containing
several crystals of p-methoxyphenol as a radical inhibitor was
irradiated with a 450 W Hanovia medium pressure mercury lamp
in an immersion well apparatus for 7.5 h with stirring. The solvent
was removed in vacuo, and the residue was chromatographed on
silica gel, eluting with 20% ethyl acetate in hexane, to obtain 0.31 g
(1.2 mmol, 83% yield) of colorless crystalline product, mp 117-120
°C. The spectral data were as follows: 1H NMR (CDCl3) δ
2.66-2.97 (br m, 3H), 3.41 (s, 3H), 3.92 (br s, 6H), 7.04 (d, J )
8.4 Hz, 1H), 7.88 (s, 1H), 7.97 (dd, J ) 2.1, 8.4 Hz, 1H); 13C
NMR (CDCl3) 27.7, 29.8, 42.2, 52.2, 62.5, 114.5, 124.7, 125.4,
129.3, 129.5, 143.6, 166.5, 172.5. Anal. Calcd for C13H15NO4: C,
62.64%; H, 6.07%; N, 5.62%. Found: C, 62.97%; H, 6.04%; N,
5.70%.
Preparation of Methyl 4-[{2-[(Benzoyloxy)methyl]acryloyl}-
(methyl)amino] Benzoate (4a, LG- ) PhCOO-). To a solution
of 0.80 g (3.2 mmol) of 4a (LG- ) HO-) and 0.45 mL of
triethylamine in 10 mL of CH2Cl2 at 0 °C was added, dropwise
with stirring, 0.38 mL (3.3 mmol) of benzoyl chloride in 10 mL of
CH2Cl2. The reaction was stirred overnight at room temperature.
The reaction mixture was washed with 5% NaHCO3 and brine. After
drying over MgSO4, the solvent was removed in vacuo, and the
residue was chromatographed on silica gel, eluting initially with
10% ethyl acetate in hexane to elute two minor impurities and then
40% ethyl acetate in hexane to elute an impurity and to obtain 0.77 g
(2.2 mmol) of 4a (LG- ) PhCOO-) as a colorless oil. The
Conclusions
R,ꢀ-Unsaturated anilides bearing leaving groups at the allylic
position of the R-methylacrylamide group undergo photochemi-
cal electrocyclic ring closure with release of the leaving group
to form an R-methylene lactam. Under aqueous conditions,
leaving groups spanning a wide range of basicities from
carboxylate groups to phenolate groups can be expelled with
little loss in efficiency as the leaving group is varied. For the
para-carbomethoxy-substituted anilides, the efficiencies for
leaving group release are Φ ) ca. 0.04 at a wavelength of 310
nm. Higher efficiencies (Φ ) ca. 0.07) for leaving group
expulsion are observed for a para-benzoyl derivative of the
anilide, in which case the photolysis wavelength can be routinely
extended to 365 nm. The photochemistry is proposed to involve
photochemically allowed electrocyclization to produce a ground-
state zwitterionic intermediate. Leaving group release could
occur directly from this intermediate or via an enolate produced
upon deprotonation of the zwitterionic intermediate. An ac-
companying minor photoproduct is a lactam, which retains the
leaving group. This lactam is thought to be formed via a 1,5-H
shift of the zwitterionic intermediate and could also be formed
upon protonation of the zwitterion in protic solvent. The
(26) We thank a referee for proposing the alternate mechanism.
J. Org. Chem. Vol. 73, No. 22, 2008 8875

Jia et al.
chromatography was repeated to obtain a colorless solid, mp 62-64
°C. The spectral data were as follows: 1H NMR (CDCl3) δ 3.42 (s,
3H), 3.90 (s, 3H), 4.95 (s, 2H), 5.18 (s, 1H), 5.49 (s, 1H), 7.24 (d,
J ) 8.0 Hz, 2H), 7.47 (d, J ) 7.8 Hz, 2H), 7.59 (t, J ) 7.8 Hz,
1H), 7.95 (d, J ) 8 0.0 Hz, 2H), 8.05 (d, J ) 7.8 Hz, 2H); 13C
NMR (CDCl3) 37.9, 52.5, 65.0, 123.2, 126.4, 128.7, 129.8, 129.9,
130.3, 130.9, 133.5, 139.1, 148.5, 166.1, 166.4, 168.9. Anal. Calcd
for C20H19NO5: C, 67.98%; H, 5.42%; N, 3.96%. Found: C, 67.93%;
H, 5.52%; N, 4.13%.
Preparation of Methyl 3-[(Benzoyloxy)methyl]-1-methyl-2-
oxo-1,2,3,4-tetrahydroquinoline-6-carboxylate (6a, LG- ) Ph-
COO-). To a solution of 0.19 g (0.78 mmol) of 6a (LG- ) HO-)
and 0.15 mL of triethylamine in 10 mL of CH2Cl2 at 0 °C was
added, dropwise with stirring, 0.11 mL (0.78 mmol) of benzoyl
chloride in 10 mL of CH2Cl2. The reaction was stirred overnight
at room temperature. The reaction mixture was washed with 5%
NaHCO3 and brine. After drying over Na2SO4, the solvent was
removed in vacuo and the residue was chromatographed on silica
gel, eluting with 10% ethyl acetate in hexane, to obtain 0.23 g (0.65
mmol, 84% yield) of colorless crystalline product, mp 127-131
°C. The spectral data were as follows: 1H NMR (CDCl3) δ
2.98-3.16 (br m, 3H), 3.40 (s, 3H), 3.88 (s, 3H), 4.57 (dd, J )
6.15, 11.1 Hz, 1H), 4.77 (dd, J ) 3.81, 11.1 Hz, 1H), 7.01 (d, J )
8.52 Hz, 1H), 7.41 (t, J ) 7.8 Hz, 2H), 7.54 (t, J ) 7.8 Hz, 1H),
7.87 (s, 1H), 7.97-7.93 (m, 3H); 13C NMR (CDCl3) 28.7, 30.2,
40.2, 52.3, 63.9, 114.6, 124.8, 124.8, 128.6, 129.5, 129.8, 130.0,
133.3, 144.1, 166.5, 166.6, 169.7. Anal. Calcd for C20H19NO5: C,
67.98%; H, 5.42%; N, 3.96%. Found: C, 68.06%; H, 5.58%; N,
3.99%.
to that reported previously.17 To a solution of 1.0 g (4.0 mmol) of
4a (LG- ) HO-), 0.57 g (6.0 mmol) of phenol, and 1.3 g (4.81
mmol) of PPh3 in 20 mL of dry THF was added 0.76 mL (4.8
mmol) of diethyl azodicarboxylate dropwise at 0 °C under N2. The
reaction mixture was then stirred at room temperature for 24 h and
was concentrated in vacuo. The residue was chromatographed on
silica gel, eluting with 10% ethyl acetate in hexane, to give 0.52 g
(1.6 mmol, 40%) of NMR pure 4a (LG- ) PhO-) as a colorless
oil. The spectral data were as follows: 1H NMR (CDCl3) δ 3.42 (s,
3H), 3.91 (s, 3H), 4.67 (s, 2H), 5.09 (s, 1H), 5.44 (s, 1H), 6.91 (d,
J ) 8.8 Hz, 2H), 6.96 (t, J ) 7.4 Hz, 1H), 7.30 (m, 4H), 7.98 (d,
J ) 8.8 Hz, 2H); 13C NMR (CDCl3) 37.8, 52.5, 68.7, 114.9, 121.4,
121.9, 126.6, 128.7, 129.7, 130.9, 140.0, 148.7, 158.5, 166.5, 169.3.
Anal. Calcd for C19H19NO4: C, 70.14%; H, 5.89%; N, 4.30%.
Found: C, 69.96%; H, 5.85%; N, 4.37%.
Preparation of Methyl 1-Methyl-2-oxo-3-(phenoxymethyl)-
1,2,3,4-tetrahydroquinoline-6-carboxylate (6a, LG- ) PhO-).
A solution of 0.23 g (0.71 mmol) of 4a (LG- ) PhO-) in N2
saturated 250 mL of benzene containing several crystals of
p-methoxyphenol as a radical inhibitor was irradiated with a 450
W Hanovia medium pressure mercury lamp in an immersion well
apparatus for 8 h with stirring. The solvent was removed in vacuo,
and the residue was chromatographed on silica gel, eluting with
10% ethyl acetate in hexane, to give 0.16 g (0.49 mmol, 69% yield)
of 6a (LG- ) PhO-) as a colorless crystals, mp 161-164 °C. The
spectral data were as follows: 1H NMR (CDCl3) δ 3.01-3.06 (m,
2H), 3.24 (dd, J ) 9.96, 21.3 Hz, 1H), 3.42 (s, 3H), 3.91 (s, 3H),
4.19 (dd, J ) 2.94, 9.42 Hz, 1H), 4.49 (dd, J ) 2.94, 9.42 Hz,
1H), 6.92-7.05 (m, 3H), 7.29 (t, J ) 7.47 Hz, 2H), 7.92 (s, 1H),
7.98 (dd, J ) 1.92, 8.46 Hz, 1H); 13C NMR (CDCl3) 28.7, 30.2,
40.7, 52.3, 66.8, 114.6, 114.8, 121.3, 124.9, 125.3, 129.7, 129.8,
129.9, 144.2, 158.8, 166.8, 170.2.
Preparation of Methyl 4-[{2-[(Phenylacetyloxy)methyl]acry-
loyl}(methyl)amino]benzoate (4a, LG- ) PhCH2COO-). To a
solution of 1.0 g (4.0 mmol) of 4a (LG- ) HO-) and 0.56 mL of
triethylamine in 10 mL of CH2Cl2 at 0 °C was added, dropwise
with stirring, 0.56 mL (4.2 mmol) of phenylacetyl chloride in 10
mL of CH2Cl2. The reaction was stirred overnight at room
temperature. The reaction mixture was washed with 5% NaHCO3
and brine. After drying over Na2SO4, the solvent was removed in
vacuo, and the residue was chromatographed on silica gel, eluting
with 10% ethyl acetate in hexane to obtain 1.4 g (3.8 mmol, 95%
yield) of NMR pure 4a (LG- ) PhCH2COO-) as a colorless oil.
Preparation of Methyl 4-[{2-[(Bocalanyloxy)methyl]acryloyl}-
(methyl)amino]benzoate (4a, LG- ) BocAla). To a solution of
1.0 g (4.0 mmol) of 4a (LG- ) HO-), 0.76 g (4.0 mmol) BocAla,
and a tiny bit of DMAP in 20 mL of CH2Cl2 was added 0.83 g
(4.0 mmol) of DCC at room temperature. The reaction mixture was
stirred at room temperature for 24 h and concentrated in vacuo.
Ether was added, the mixture was filtered, and the filtrate was
concentrated in vacuo. The residue was chromatographed on silica
gel eluting with 20% of ethyl acetate in hexane, to give 1.32 g (3.1
mmol, yield 78%) of 4a (LG- ) BocAla) as colorless needles, mp
1
The spectral data were as follows: H NMR (CDCl3) δ 3.27 (s,
3H), 3.62 (s, 2H), 3.85 (s, 3H), 4.63 (s, 2H), 4.97 (s, 1H), 5.22 (s,
1H), 6.95 (d, J ) 7.2 Hz, 2H), 7.19-7.28 (m, 5H), 7.84 (d, J )
7.2 Hz, 2H); 13C NMR (CDCl3) 37.8, 41.6, 52.5, 65.4, 126.4, 127.5,
128.7, 128.9, 128.9, 129.6, 130.9, 133.9, 138.9, 148.4, 166.4, 168.6,
170.9. Anal. Calcd for C21H21NO5: C, 68.65%; H, 5.76%; N, 3.81%.
Found: C, 68.44%; H, 5.82%; N, 4.00%.
1
95-96 °C. The spectral data were as follows: H NMR (CD3CN)
δ 1.31 (d, J ) 7.32 Hz, 3H), 1.37 (s, 9H), 3.32 (s, 3H), 3.85 (s,
3H), 4.19 (br, 1H), 4.68 (s, 2H), 5.08 (s, 1H), 5.36 (s, 1H),
5.71-5.74 (br, 1H), 7.36 (d, J ) 8.82 Hz, 2H), 7.98 (d, J ) 8.82
Hz, 2H); 13C NMR (CDCl3) 17.8, 28.4, 37.8, 50.2, 52.7, 65.6, 79.8,
122.4, 127.6, 129.2, 131.2, 140.1, 149.5, 156.3, 166.9, 169.0, 173.6.
Anal. Calcd for C21H28N2O7: C, 59.99%; H, 6.71%; N, 6.66%.
Found: C, 60.18%; H, 6.68%; N, 6.79%.
Preparation of Methyl 1-Methyl-2-oxo-3-(phenylacetyloxy-
methyl)-1,2,3,4-tetrahydroquinoline-6-carboxylate (6a, LG-
)
)
PhCH2COO-). To a solution of 0.25 g (1.0 mmol) of 6a (LG-
HO-) and 0.14 mL of triethylamine in 10 mL of CH2Cl2 at 0 °C
was added, dropwise with stirring, 0.14 mL (1.1 mmol) of
phenylacetyl chloride in 5 mL of CH2Cl2. The reaction was stirred
overnight at room temperature. The reaction mixture was washed
with 5% NaHCO3 and brine. After drying over Na2SO4, the solvent
was removed in vacuo, and the residue was purified on silica gel,
eluting with 10% ethyl acetate in hexane on silica gel to obtain
0.32 g (0.87 mmol, 87% yield) of colorless crystalline product,
mp 48-51 °C. The spectral data were as follows: 1H NMR (CDCl3)
δ 2.78-2.93 (br m, 3H), 3.37 (s, 3H), 3.62 (s, 2H), 3.92 (s, 3H),
4.33 (dd, J ) 6.12, 11.22 Hz, 1H), 4.56 (dd, J ) 4.08, 11.22 Hz,
1H), 7.00 (d, J ) 8.40 Hz, 1H), 7.26-7.34 (m, 5H), 7.78 (s, 1H),
7.95 (dd, J ) 2.01, 8.04 Hz, 1H); 13C NMR (CDCl3) 28.3, 30.1,
40.0, 41.4, 52.5, 63.5, 114.5, 124.8, 127.4, 128.8, 129.4, 129.4,
129.4, 129.7, 134.0, 144.0, 166.6, 169.6, 171.5. Anal. Calcd for
C21H21NO5: C, 68.65%; H, 5.76%; N, 3.81%. Found: C, 68.86%;
H, 5.74%; N, 3.77%.
Preparation of Methyl 3-[(Bocalanyloxy)methyl]-1-methyl-
2-oxo-1,2,3,4-tetrahydroquinoline-6-carboxylate (6a, LG-
)
BocAla). To a solution of 0.19 g (0.78 mmol) of 6a (LG- ) HO-),
0.16 g (0.82 mmol) of BocAla, and a bit of DMAP in 20 mL of
CH2Cl2 was added 0.19 g (0.94 mmol) of DCC at room temperature.
The reaction mixture was stirred at room temperature for 24 h and
concentrated in vacuo. Ether was added, the mixture was filtered,
and the filtrate was concentrated in vacuo. The residue was
chromatographed on silica gel, eluting with 20% of ethyl acetate
in hexane, to give 0.26 g (0.62 mmol, yield 79%) of 6a (LG-
)
BocAla) as colorless needles, mp 110-115 °C. The spectral data
1
were as follows: H NMR (CDCl3) δ 1.31 (d, J ) 7.2 Hz, 3H),
1.41 (s, 9H), 2.90-3.11 (m, 3H), 3.34 (s, 3H), 3.86 (s, 3H), 4.13
(t, J ) 7.0 Hz, 1H), 4.31-4.52 (m, 2H), 5.66 (br, s, 1H), 7.13 (d,
J ) 8.5 Hz, 1H), 7.84 (s, 1H), 7.92 (dd, J ) 1.9, 8.5 Hz, 1H); 13
C
NMR (CDCl3) δ 1.31 (d, J ) 7.32 Hz, 3H), 1.37 (s, 9H), 3.32 (s,
3H), 3.85 (s, 3H), 4.19 (br, 1H), 4.68 (s, 2H), 5.08 (s, 1H), 5.36 (s,
1H), 5.71-5.74 (br, 1H), 7.36 (d, J ) 8.82 Hz, 2H), 7.98 (d, J )
8.82 Hz, 2H); 13C NMR (CDCl3) 17.7, 28.4, 30.1, 40.3, 40.4, 50.3,
Preparation of Methyl 4-{methyl[2-(phenoxymethyl)acryl-
oyl]amino}benzoate (4a, LG- ) PhO-). The procedure is similar
8876 J. Org. Chem. Vol. 73, No. 22, 2008

Photochemical Elimination of LeaVing Groups
52.4, 64.1, 79.7, 115.5, 125.1, 126.1, 129.6, 129.9, 145.1, 156.3,
167.1, 170.1, 174.0. Anal. Calcd for C21H28N2O7: C, 59.99%; H,
6.71%; N, 6.66%. Found: C, 60.26%; H, 6.81%; N, 6.91%.
Preparation of N-(4-Benzoylphenyl)-2-(hydroxymethyl)-N-
methylacrylamide (4b, LG- ) HO-). The procedure was adapted
from the literature.16 To a mixture of 33 g (0.12 mol) of N-(4-
benzoylphenyl)-N-methyl-2-propenamide 11b in 330 mL of THF
were added 330 mL of a 40% aqueous formaldehyde and 33 g
(0.29 mol) of DABCO. The suspension was sonicated for 48 h at
50-60 °C. The mixture was extracted with ethyl acetate. The
extracts were washed with a saturated solution of NaHCO3, brine,
and dried over anhydrous sodium sulfate. After concentration in
vacuo, the residue was chromatographed on silica gel, eluting with
20% ethyl acetate in hexane, to give 16 g (0.054 mol, 45% yield)
of 4b (LG- ) HO-) as a yellow solid, mp 90-91 °C. The spectral
83% yield) of colorless crystalline product, mp 77-79 °C. The
spectral data were as follows: 1H NMR (CDCl3) δ 2.85-2.68 (m,
3H), 3.33 (s, 3H), 3.55 (s, 2H), 4.28 (dd, J ) 5.20, 8.40 Hz, 1H),
4.50 (dd, J ) 3.33, 8.40 Hz, 1H), 6.96 (d, J ) 6.33 Hz, 1H),
7.12-7.24 (m, 5H), 7.43 (t, J ) 5.52 Hz, 2H), 7.54 (t, J ) 5.56
Hz, 2H), 7.69 (m, 2H); 13C NMR (CDCl3) 28.3, 30.1, 40.0, 41.4,
63.5, 114.4, 124.9, 127.4, 128.6, 128.8, 129.5, 130.0, 130.2, 130.8,
132.3, 132.5, 134.0, 138.0, 143.8, 169.7, 171.6, 195.6. Anal. Calcd
for C26H23NO4: C, 75.53%; H, 5.61%; N, 3.39%. Found: C, 75.37%;
H, 5.64%; N, 3.38%.
Preparation of N-(4-Benzoylphenyl)-N-methyl-2-(phenoxy-
methyl)acrylamide (4b, LG- ) PhO-). The procedure is similar
to that reported previously.17 To a solution of 0.74 g (2.5 mmol)
of 4b (LG- ) HO-), 0.47 g (5.0 mmol) of phenol, and 1.6 g (6.1
mmol) of PPh3 in 10 mL of dry THF was added 0.95 mL (6.0
mmol) of diethyl azodicarboxylate dropwise at 0 °C under N2. The
reaction mixture was stirred at room temperature for 24 h and
concentrated in vacuo, and the residue was chromatographed on
silica gel, eluting with 10% ethyl acetate in hexane, to give 0.30 g
(0.81 mmol, 32%) of NMR pure 4b (LG- ) PhO-) as a colorless
1
data were as follows: H NMR (CDCl3) δ 2.51 (br, 1H), 3.43 (s,
3H), 4.28 (s, 2H), 5.01 (s, 1H), 5.36 (s, 1H), 7.33 (d, J ) 8.7 Hz,
2H), 7.49 (t, J ) 7.8 Hz, 2H), 7.61 (t, J ) 7.8 Hz, 1H), 7.79 (t, J
) 8.7 Hz, 4H); 13C NMR (CDCl3) 37.8, 64.2, 120.4, 126.3, 128.6,
130.1, 131.4, 132.8, 135.8, 137.4, 143.3, 148.2, 170.5, 195.7. Anal.
Calcd for C18H17NO3: C, 73.20%; H, 5.80%; N, 4.74%. Found: C,
73.33%; H, 5.85%; N, 4.77%.
1
oil. The spectral data were as follows: H NMR (CDCl3) 3.45 (s,
3H), 4.70 (s, 2H), 5.16 (s, 1H), 5.49 (s, 1H), 6.92 (m, 3H), 7.27 (t,
J ) 8.7 Hz, 2H), 7.31 (d, J ) 8.7 Hz, 2H), 7.50 (t, J ) 7.6 Hz,
2H), 7.60 (t, J ) 7.6 Hz, 1H), 7.78 (d, J ) 8.7 Hz, 4H); 13C NMR
(CDCl3) 37. 9, 68.7, 114.9, 121.4, 122.0, 126.5, 128.6, 129.7, 130.1,
131.4, 132.8, 135.9, 137.5, 140.0, 148.2, 158.5, 169.3, 195.7. Anal.
Calcd for C24H21NO3: C, 77.61%; H, 5.70%; N, 3.77%. Found: C,
77.39%; H, 5.71%; N, 3.82%.
Preparation 6-Benzoyl-3-(hydroxymethyl)-1-methyl-3,4-di-
hydroquinolin-2(1H)-one of 6b (LG- ) HO-) by Photolysis of
(4b, LG- ) HO-). A solution of 0.38 g (1.3 mmol) of 4b (LG-
) HO-) in N2 saturated 250 mL of benzene containing several
crystals of p-methoxyphenol as a radical inhibitor was irradiated
with a 450 W Hanovia medium pressure mercury lamp in an
immersion well apparatus with a Pyrex filter for 2 h with stirring.
The solvent was removed in vacuo, and the residue was chromato-
graphed on silica gel, eluting with 20% ethyl acetate in hexane, to
obtain 0.35 g (1.2 mmol 92% yield) of 6b (LG- ) HO-) as
colorless crystals, mp 111-113 °C. The spectral data were as
follows: 1H NMR (CDCl3) δ 2.68-2.91 (m, 3H), 3.37 (s, 3H), 3.87
(t, J ) 8.5 Hz, 2H), 7.01 (d, J ) 8.3 Hz, 1H), 7.44 (m, 2H), 7.55
(t, J ) 7.4 Hz, 1H), 7.70 (m, 4H); 13C NMR (CDCl3) 27.9, 29.9,
42.2, 62.6, 114.3, 125.6, 128.4, 129.9, 129.9, 130.5, 132.2, 132.4,
137.8, 143.5, 172.5, 195.5. Anal. Calcd for C18H17NO3: C, 73.20%;
H, 5.80%; N, 4.74%. Found: C, 73.07%; H, 5.79%; N, 4.77%.
Preparation of N-(4-Benzoylphenyl)-N-methyl-2-[(phenyl-
acetyl)oxy]methylacrylamide (4b, LG- ) PhCH2COO-). To a
solution of 2.8 g (9.5 mmol) of 4b (LG- ) HO-) and 1.3 mL of
triethylamine in 20 mL of CH2Cl2 at 0 °C was added, dropwise
with stirring, 1.5 g (9.5 mmol) of phenylacetyl chloride in 10 mL
of CH2Cl2. The reaction was stirred overnight at room temperature.
The reaction mixture was washed with 5% NaHCO3 and brine. After
drying over anhydrous MgSO4, the solvent was removed in vacuo,
and the residue was chromatographed on silica gel, eluting with
10% ethyl acetate in hexane to obtain 2.9 g (7.0 mmol, 74% yield)
of NMR pure 4b (LG- ) PhCH2COO-) as a colorless oil. The
Preparation of 6-Benzoyl-1-methyl-3-(phenoxymethyl)-3,4-
dihydroquinolin-2(1H)-one (6b, LG- ) PhO-). A solution of
0.10 g (0.27 mmol) of 4b (LG- ) PhO-) in N2 saturated 100 mL
of benzene containing several crystals of p-methoxyphenol as a
radical inhibitor in a quartz tube was mounted next to a water-
jacketed immersion well apparatus containing a 450 W Hanovia
medium pressure mercury lamp and a Pyrex filter. The solution
was irradiated for 2 h with stirring. The solvent was removed in
vacuo, and the residue was chromatographed on silica gel, eluting
with 10% ethyl acetate in hexane, to give 0.067 g (0.18 mmol,
67% yield) of NMR pure 6b (LG- ) PhO-) as a colorless oil. The
spectral data were as follows: 1H NMR (CDCl3) δ 2.99-3.12 (m,
2H), 3.26 (dd, J ) 11.6, 21.3 Hz, 1H), 3.44 (s, 3H), 4.19 (dd, J )
7.6, 9.6 Hz, 1H), 4.50 (dd, J ) 3.7, 9.6 Hz, 1H), 6.95 (m, 3H),
7.07 (d, J ) 8.2 Hz, 1H), 7.29 (t, J ) 7.6 Hz, 2H), 7.50 (t, J ) 8.0
Hz, 2H), 7.60 (t, J ) 7.6 Hz, 1H), 7.75-7.79 (m, 4H); 13C NMR
(CDCl3) 28.8, 30.2, 40.6, 66.8, 114.3, 114.8, 121.3, 125.4, 128.5,
129.7, 130.0, 130.3, 130.7, 132.3, 132.5, 138.0, 144.0, 158.8, 170.2,
195.7. Anal. Calcd for C24H21NO3: C, 77.61%; H, 5.70%; N, 3.77%.
Found: C, 77.76%; H, 5.59%; N, 3.78%.
Preparation of Methyl 1,3-Dimethyl-2-oxo-1,2,3,4-tetrahy-
droquinoline-6-carboxylate 3b by Photolysis of 1b. A solution
of 1.0 g (4.3 mmol) of 1b in N2 saturated 50 mL of benzene
containing several crystals of p-methoxyphenol as a radical inhibitor
in a quartz tube was mounted beside a water-jacketed immersion
well apparatus containing a 450 W Hanovia medium pressure
mercury lamp. The solution was irradiated for 8 h with stirring.
The solvent was removed in vacuo, and the residue was chromato-
graphed on silica gel, eluting with 10% ethyl acetate in hexane, to
give 0.6 g (2.6 mmol, 60% yield) of 3b as colorless crystals, mp
135-137 °C. The spectral data were as follows: 1H NMR (CDCl3)
δ 1.26 (d, J ) 6.72 Hz, 3H), 2.62-2.71 (m, 1H), 2.73 (dd, J )
11.0, 14.8 Hz, 1H), 3.00 (dd, J ) 5.04, 14.8 Hz, 1H), 3.38 (s, 3H),
3.91 (s, 3H), 6.99 (d, J ) 8.50 Hz, 1H), 7.84 (s, 1H), 7.94 (d, J )
8.50 Hz, 1H); 13C NMR (CDCl3) 15.8, 30.1, 33.2, 35.5, 52.2, 114.3,
124.4, 125.6, 129.4, 129.6, 144.5, 166.8, 173.4. Anal. Calcd for
C13H15NO3: C, 66.94%; H, 6.48%; N, 6.00%. Found: C, 66.72%;
H, 6.43%; N, 5.97%.
1
spectral data were as follows: H NMR (CDCl3) δ 3.37 (s, 3H),
3.67 (s, 2H), 4.73 (s, 2H), 5.09 (s, 1H), 5.33 (s, 1H), 7.08 (d, J )
8.4 Hz, 2H), 7.26 (m, 2H), 7.28 (m, 4H), 7.51 (d, J ) 7.8 Hz, 2H),
7.62 (d, J ) 8.1 Hz, 1H), 7.68 (d, J ) 8.4 Hz, 2H), 7.77 (d, J )
6.9 Hz, 2H); 13C NMR (CDCl3) 37.9, 41.6, 65.4, 123.1, 126.3,
127.4, 128.6, 128.8, 128.9, 129.5, 130.1, 131.4, 132.8, 135.9, 137.5,
138.9, 148.0, 168.7, 170.9, 195.7. Anal. Calcd for C26H23NO4: C,
75.53%; H, 5.61%; N, 3.39%. Found: C, 75.80%; H, 5.68%; N,
3.49%.
Preparation of 6-Benzoyl-1-methyl-3-[(phenylacetyl)oxymeth-
yl]-3,4-dihydroquinolin-2(1H)-one (6b, LG- ) PhCH2COO-).
To a solution of 0.30 g (1.0 mmol) of 6b (LG- ) HO-) and 0.14
mL of triethylamine in 10 mL of CH2Cl2 at 0 °C was added,
dropwise with stirring, 0.14 mL (1.1 mmol) of phenylacetyl chloride
in 5 mL of CH2Cl2. The reaction was stirred overnight at room
temperature. The reaction mixture was washed with 5% NaHCO3
and brine. After drying over MgSO4, the solvent was removed in
vacuo, and the residue was purified on silica gel, eluting with 10%
ethyl acetate in hexane on silica gel to obtain 0.34 g (0.83 mmol,
Preparation of 6-Benzoyl-1,3-dimethyl-3,4-dihydroquinolin-
2(1H)-one 3c by Photolysis of 1c. A solution of 0.36 g (1.3 mmol)
of 1c in 70 mL of N2 saturated benzene containing several crystals
J. Org. Chem. Vol. 73, No. 22, 2008 8877

Jia et al.
of p-methoxyphenol as a radical inhibitor in a quartz tube was
mounted beside a water-jacketed immersion well apparatus contain-
ing a 450 W Hanovia medium pressure mercury lamp and Pyrex
filter. The solution was irradiated for 2 h with stirring. The solvent
was removed in vacuo, and the residue was crystallized to give
0.36 g (1.3 mmol, 100% yield) of 3c as colorless crystals, mp
117-119 °C. The spectral data were as follows: 1H NMR (CDCl3)
δ 1.27 (d, J ) 8.84 Hz, 3H), 2.62-2.79 (m, 2H), 2.99 (dd, J )
6.32, 19.4 Hz, 1H), 3.39 (s, 3H), 7.02 (d, J ) 11.08 Hz, 2H),
7.45-7.50 (m, 2H), 7.55-7.60 (m, 1H), 7.69-7.78 (m, 4H); 13C
NMR (CDCl3) 15.8, 30.1, 33.2, 35.4, 114.0, 125.7, 128.4, 129.9,
129.9, 130.6, 131.89, 132. 3, 138.0, 144.2, 173.4, 195.6.
ethyl acetate. The extract was dried over anhydrous sodium sulfate
and concentrated in vacuo to obtain an oil. The oil was chromato-
graphed on silica gel, eluting with 10% ethyl acetate in hexane, to
obtain 0.044 g (0.16 mmol, 23% yield) of NMR pure 6-benzoyl-
1-methyl-3-methylene-3,4-dihydroquinolin-2(1H)-one (5b) and a
mixture of 0.021 g (0.068 mmol, 11% yield) of 6b (LG-
)
PhCH2COO-) and 0.12 g (0.29 mmol, 41% yield unreacted) of 4b
(LG- ) PhCH2COO-). The spectral data for 5b were as follows:
1H NMR (CDCl3) δ 3.44 (s, 3H), 3.78 (s, 2H), 5.53 (s, 1H), 6.19
(s, 1H), 7.03 (d, J ) 8.4 Hz, 1H), 7.48 (t, J ) 7.8 Hz, 2H), 7.58
(t, J ) 8.0 Hz 1H), 7.68-7.76 (m, 4H); 13C NMR (CDCl3) 30.3,
34.2, 114.1, 124.0, 124.3, 128.5, 129.5, 129.9, 130.6, 132.0, 132.4,
135.2, 137.9, 143.5, 165.1, 195.6.
Preparation of 5,7-Dimethoxy-1,3-dimethyl-3,4-dihydroquin-
olin-2(1H)-one (9a) by Photolysis of 8a. A solution of 2.0 g (8.5
mmol) of 8a in N2 saturated 200 mL of benzene containing several
crystals of p-methoxyphenol as a radical inhibitor was irradiated
with a 450 W Hanovia medium pressure mercury lamp in an
immersion well apparatus for 6 h with stirring. The solvent was
removed in vacuo, and the residue was chromatographed on silica
gel, eluting with 10% ethyl acetate in hexane, to give 1.0 g (4.2
mmol, 50% yield) of 9a, as colorless crystals, mp 129-130 °C.
General Procedure for Determining Product Yields in Pho-
tolyses of 4a (LG- ) BocAla, PhCO2-, PhCH2CO2-, PhO-,
HO-) and 4b (LG- ) PhCH2CO2-, PhO-, HO-). Samples of
15-20 mg of 4a (LG- ) PhCH2CO2-, PhO-, HO-) or 4b (LG-
) PhCH2CO2-, PhO-, HO-) in 1 mL of 50% aq CD3CN containing
0.1 M phosphate buffer at pD 7 or 0.5 mL of CD3CN or CD2Cl2 or
C6D6 were contained in NMR tubes. Samples were mounted beside
a water-jacketed 450 W medium pressure mercury lamp equipped
with a Pyrex filter. The samples were at room temperature thro-
ughout the photolyses. Yields were determined by 1H NMR
spectroscopy using DMSO as the internal standard.
1
The spectral data were as follows: H NMR (CDCl3) δ 1.24 (d, J
) 6.69 Hz, 3H), 2.40-2.60 (m, 2H), 3.00 (dd, J ) 5.22, 14.97 Hz,
1H), 3.32 (s, 3H), 3.82 (s, 6H), 6.20 (dd, J ) 2.16, 9.60 Hz, 2H);
13C NMR (CDCl3) 16.0, 25.6, 30.3, 35.4, 55.6, 55.8, 92.6, 94.1,
106.3, 142.2, 157.3, 159.8, 173.8. Anal. Calcd for C13H17NO3: C,
66.36%; H, 7.28%; N, 5.95%. Found: C, 66.35%; H, 7.21%; N,
5.96%.
General Procedure for Testing Hydrolytic Stability of 6a
-
(LG- ) PhCH2CO2-, BocAla) and 6b (LG- ) PhCH2CO2
)
in the Dark. The stabilities of 6a (LG- ) PhCH2CO2-, BocAla)
and 6b (LG- ) PhCH2CO2-) were determined by HPLC analyses.
Samples of 100-150 mg of 6a (LG- ) PhCH2CO2-, BocAla) or
6b (LG- ) PhCH2CO2-) were dissolved in 15 mL of CH3CN
containing 0.021 M biphenyl as an internal standard and 15 mL of
0.1 M phosphate buffer at pH 7. After various time periods, the
ratio between biphenyl and 6a (LG- ) PhCH2CO2-, BocAla) or
6b (LG- ) PhCH2CO2-) was checked by HPLC. In each case, the
analyses showed no detectable reaction had occurred at room
temperature in the dark for at least one week.
Preparation of 6-Methoxy-1,3-dimethyl-3,4-dihydroquinolin-
2(1H)-one (9b) by Photolysis of 8b. A solution of 2.6 g (13 mmol)
of 8b in N2 saturated 250 mL of benzene containing several crystals
of p-methoxyphenol as a radical inhibitor was irradiated with a
450 W Hanovia medium pressure mercury lamp in an immersion
well for 10 h with stirring. The solvent was removed in vacuo, and
the residue was chromatographed on silica gel, eluting with 10%
ethyl acetate in hexane, to obtain 1.8 g (8.8 mmol, 68% yield) of
9b as a colorless solid, mp 41-44 °C. The spectral data were as
follows: 1H NMR (CDCl3) δ 1.21 (d, J ) 6.63 Hz, 3H), 2.53-2.66
(m, 2H), 2.83 (dd, J ) 4.20, 13.8 Hz, 1H), 3.29 (s, 3H), 3.75 (s,
3H), 6.68-6.75 (m, 2H), 6.84 (d, J ) 8.7 Hz, 1H); 13C NMR
(CDCl3) 15.8, 30.0, 33.6, 35.5, 55.6, 111.8, 114.1, 115.4, 127.2,
134.1, 155.3, 172.8.
General Procedure for Product Quantum Yield Determina-
tions. A semimicro optical bench was used for quantum yield
determinations, similar to the apparatus described by Zimmerman.27
Light from a 200 W high-pressure mercury lamp was passed
through an Oriel monochromator, which was set to 310 or 365 nm
wavelengths. The light was collimated through a lens. A fraction
of the light was diverted 90° by a beam splitter to a 10 × 3.6 cm
side quartz cylindrical cell containing an actinometer. The pho-
tolysate was contained in a 10 × 1.8 cm quartz cylindrical cell of
25 mL volume. Behind the photolysate was mounted a quartz
cylindrical cell containing 25 mL of actinometer. Light output was
monitored by ferrioxalate actinometery28 using the splitting ratio
technique. Products were analyzed by HPLC using biphenyl as an
internal standard.
Preparative Photolysis of Methyl 4-[N-Methyl-N-(2-phenyl-
acetyloxymethyl-1-oxo-2-propenyl)amino]benzoate (4a, LG-
)
)
PhCH2COO-). A solution of 0.21 g (0.57 mmol) of 4a (LG-
PhCH2COO-) in N2 saturated 250 mL of acetonitrile and buffer
(1:1) was irradiated with a 450 W Hanovia medium pressure
mercury lamp in an immersion well apparatus for 5 h with stirring.
The photolysate was concentrated in vacuo and then was extracted
by ethyl acetate. The extract was dried over anhydrous sodium
sulfate and concentrated in vacuo to obtain an oil, and the oil was
chromatographed on silica gel, eluting with 10% ethyl acetate in
hexane, to obtain 0.037 g (0.16 mmol, 28% yield) of NMR pure
methyl 1-methyl-3-methylene-2-oxo-1,2,3,4-tetrahydroquinoline-6-
carboxylate (5a) and a mixture of 0.025 g (0.068 mmol, 12% yield)
of 6a (LG- ) PhCH2COO-) and 0.12 g (0.33 mmol, 57% yield
unreacted) of 4a (LG- ) PhCH2COO-). The spectral data for 5a
General Procedure for Determining Triplet Yields.22 Sol-
utions of 0.05 M of 1c, 15, or benzophenone and 0.15 M of trans-
piperylene in either benzene or 20% aq CH3CN were photolyzed
at 365 nm for 2-5 h while performing actinometry, as in the
quantum yield determinations. The trans- and cis-piperylenes were
quantified on a 20 ft × 0.25 in column packed with 15%
1,2-bis(cyanoethoxy)ethane on Chromosorb P (45-60 mesh) at
40 °C.
1
were as follows: H NMR (CDCl3) δ 3.48 (s, 3H), 3.78 (s, 2H),
3.91 (s, 3H), 5.53 (s, 1H), 6.19 (s, 1H), 7.01 (d, J ) 8.8 Hz, 1H),
7.85 (s, 1H), 7.93 (d, J ) 8.8 Hz, 1H); 13C NMR (CDCl3) 30.4,
34.2, 52.3, 114.4, 123.9, 124.2, 124.6, 129.0, 129.6, 135.3, 143. 8,
165.2, 166.7.
General Procedure for Nanosecond Laser Flash Photolysis.
The flash photolysis was performed with an Edinburgh Instruments
LP920 system equipped with a Continuum Surelite Nd:YAG laser,
which gives 10 ns pulses at 355 nm. The sample concentrations
were between 2 and 3 mM to give absorbance of 0.7-0.8. The
solvent was 50% aq CH3CN containing 0.1 M phosphate buffer at
pH 7. Samples were flushed initially with argon for 30 min, then
Preparative Photolysis of N-(4-Benzoylphenyl)-N-methyl-2-
phenylacetyloxymethyl-2-propenamide (4b, LG- ) PhCH2-
COO-). A solution of 0.29 g (0.70 mmol) of 4b (LG-
)
PhCH2COO-) in N2 saturated 300 mL of acetonitrile and buffer
(1:1) was irradiated with a 450 W Hanovia medium pressure
mercury lamp with a Pyrex filter for 2 h with stirring. The
photolysate was concentrated in vacuo and then was extracted by
(27) Zimmerman, H. E. Mol. Photochem. 1971, 3, 281–92.
(28) Hatchard, C. G.; Parker, C. A. Proc. R. Soc London 1956, 235, 518–
36.
8878 J. Org. Chem. Vol. 73, No. 22, 2008

Photochemical Elimination of LeaVing Groups
degassed by three freeze-pump-thaw cycles, and then kept under
argon. The compounds used for flash photolysis studies were 1c
and 4b (LG- ) PhCH2CO2-, HO-). The Stern-Volmer quenching
studies used the same techniques but with various amounts of
sodium 2-naphthalenesulfonate added as the triplet quencher.
Supporting Information Available: Synthesis procedures
for 1b,c, 8a,b, 11a,b, 13, N-(4-benzoylphenyl)-2-methylacryl-
amide, and N-(3,5-dimethoxyphenyl)-2-methylacrylamide. NMR
spectra for 1b,c, 3b,c, 4a (LG- ) BocAla, PhCO2
PhCH2CO2-, PhO-, HO-), 6a (BocAla, PhCO2-, PhCH2CO2
PhO-, HO-), 4b (PhCH2CO2-, PhO-, HO-), 6b (PhCH2CO2
,
,
,
-
-
-
Acknowledgment. We thank Ms. Ja Eun Lee and Dr. Yugang
Chen for technical assistance. Acknowledgment is made to the
donors of the Petroleum Research Fund, administered by the
American Chemical Society, for support of this research
(M.G.S.) and to the National Science Foundation for a Career
Award (R.R.) for support of this research.
PhO-, HO-), 5a,b, 11a,b, 13, N-(3,5-dimethoxyphenyl)-2-
methylacrylamide, and 8a,b. This material is available free of
charge via the Internet at http://pubs.acs.org.
JO8017445
J. Org. Chem. Vol. 73, No. 22, 2008 8879