Photochemical cleavage and release of carboxylic acids from α-keto amides

Article abstract of DOI:10.1021/jo050246s

In aqueous media, α-keto amides LGCH₂COCON(R)CH(R′) CH₃ (1a, R = Et, R′ = H; 1b, R = ⁱPr, R′ = Me; 1c, R = Ph, R′ = H) with various carboxylate leaving groups (LG) at the C-3 position undergo photocleavage and release of carboxylic acids with formation of diastereomeric 5-hydroxyoxazolidin-4-ones 2a,c in the cases of 1a,c or 5-methyleneoxazolidin-4-ones 3b in the case of 1b. For 1a,b, Φ(photocleavage) = 0.24-0.38, whereas Φ(photocleavage) = ca. 0.05 for 1c. The proposed mechanism involves transfer of hydrogen from an N-alkyl group to the keto oxygen to produce zwitterionic intermediates 4a-c that eliminate carboxylate anions. The resultant imminium ions, H₂C=C(OH)CON ⁺(R)=C(R′)CH₃ 5a-c, cyclize intramolecularly to 3b or undergo intermolecular addition of water followed by tautomerization and cyclization to give 2a,c. These inter- or intramolecular trapping reactions of 5 release protons that decrease the pH and cause bleaching of the 620 nm band of the pH indicator, bromocresol green. Determination of the bleaching kinetics by laser flash photolysis methods in the case of 1a gives time constants of 18-137 μs, depending on the leaving group ability of the carboxylate anion, whereas amides 1b show only a small leaving group effect. For 1a, the large leaving group effect is consistent with rate-limiting carboxylate elimination from 4a, whereas the proton release step would be largely rate determining for 1b. Photolyses of 1a (LG = CH₃CO₂^-, PhCH ₂CO₂^-) in neat CH₃CN results in carboxylate elimination to form imminium ion 5a, followed by internal return to give aminals.

Full text of DOI:10.1021/jo050246s

Photochemical Cleavage and Release of Carboxylic Acids from
r-Keto Amides
Chicheng Ma, Mark G. Steinmetz,* Erica J. Kopatz, and Rajendra Rathore
Department of Chemistry, Marquette University, Milwaukee, Wisconsin 53201-1881
mark.steinmetz@marquette.edu
Received February 7, 2005
i
In aqueous media, R-keto amides LGCH2COCON(R)CH(R′)CH3 (1a, R ) Et, R′ ) H; 1b, R ) Pr,
R′ ) Me; 1c, R ) Ph, R′ ) H) with various carboxylate leaving groups (LG) at the C-3 position
undergo photocleavage and release of carboxylic acids with formation of diastereomeric 5-hydroxy-
oxazolidin-4-ones 2a,c in the cases of 1a,c or 5-methyleneoxazolidin-4-ones 3b in the case of 1b.
For 1a,b, Φ(photocleavage) ) 0.24-0.38, whereas Φ(photocleavage) ) ca. 0.05 for 1c. The proposed
mechanism involves transfer of hydrogen from an N-alkyl group to the keto oxygen to produce
zwitterionic intermediates 4a-c that eliminate carboxylate anions. The resultant imminium ions,
H2CdC(OH)CON+(R)dC(R′)CH3 5a-c, cyclize intramolecularly to 3b or undergo intermolecular
addition of water followed by tautomerization and cyclization to give 2a,c. These inter- or
intramolecular trapping reactions of 5 release protons that decrease the pH and cause bleaching
of the 620 nm band of the pH indicator, bromocresol green. Determination of the bleaching kinetics
by laser flash photolysis methods in the case of 1a gives time constants of 18-137 µs, depending
on the leaving group ability of the carboxylate anion, whereas amides 1b show only a small leaving
group effect. For 1a, the large leaving group effect is consistent with rate-limiting carboxylate
elimination from 4a, whereas the proton release step would be largely rate determining for 1b.
Photolyses of 1a (LG ) CH3CO2-, PhCH2CO2-) in neat CH3CN results in carboxylate elimination
to form imminium ion 5a, followed by internal return to give aminals.
Recent studies have shown that R-keto amides 1a
undergo photochemical cleavage and release of carboxy-
late leaving groups (eq 1).1 Photochemical elimination of
substituted phenolate leaving groups has also been
observed.2
chemical yields of the released leaving groups are nearly
quantitative, and carboxylate release occurs on the
microsecond time scale.3 Zwitterionic intermediates such
as 4a are thought to be involved in the eliminations.
Analogous intermediates, generated photochemically via
hydrogen transfer from an N-alkyl group to the keto
oxygen, have often been proposed to account for the
photorearrangement products of R-keto amides, which
lack leaving groups.4,5 The elimination of leaving groups
(3) Ma, C.; Steinmetz, M. G.; Kopatz, E. J.; Rathore, R. Tetrahedron
Lett. 2005, 46, 1045-1048.
(4) (a) Aoyama, H.; Sakamoto, M.; Kuwabara, K.; Yoshida, K.;
Omote, Y. J. Am. Chem. Soc. 1983, 105, 1958-1964. (b) Aoyama, H.;
Sakamoto, M.; Omote, Y. J. Chem. Soc., Perkin Trans. 1 1981, 1357-
1359. (c) Aoyama, H.; Hasegawa, T.; Omote, Y. J. Am. Chem. Soc. 1979,
101, 5343-5347. (d) Aoyama, H.; Hasegawa, T.; Watabe, M.; Shiraishi,
H.; Omote, Y. J. Org. Chem. 1978, 43, 419-422.
These photoelimination reactions are highly efficient, the
(1) Ma, C.; Steinmetz, M. G.; Cheng, Q.; Jayaraman, V. Org. Lett.
2003, 5, 71-74.
(2) Ma, C.; Steinmetz, M. G. Org. Lett. 2004, 6, 629-32.
10.1021/jo050246s CCC: $30.25 © 2005 American Chemical Society
Published on Web 04/29/2005
J. Org. Chem. 2005, 70, 4431-4442
4431

Ma et al.
SCHEME 1
FIGURE 1. R-Keto amides 1a-c and leaving groups LG-.
from zwitterionic intermediates has warranted detailed
studies, because it could serve as a mechanistic basis for
developing a new class of photoremovable protecting
groups and phototriggers (“cage compounds”).6,7
the R-keto group. A number of mechanisms can be
envisioned for the hydrogen transfer, including (1) hy-
drogen atom abstraction, (2) transfer of an electron
followed by a proton,5 and (3) net transfer of hydride
(1,5-H shift). Although the hydrogen transfer mechanism
is not well understood, previous studies4c,5 have shown
that when the carboxamide group favors an unreactive
conformation with respect to hydrogen transfer, then the
photochemistry is inefficient. The carboxylate photoelimi-
nations are no different in this regard. Low quantum
yields are found for the N-ethyl-N-phenyl amides 1c (LG
) PhCO2- and PhCH2CO2-) (Figure 1), which otherwise
give similar photoproducts as 1a.
We recently determined the rate constants for release
of carboxylate groups from 1a by a pH jump method that
used laser flash photolysis.3 In this paper, we report
complete results of this study and additional results for
1b. When R-keto amides 1a,b are photolyzed, carboxylate
anions and protons are released. The pH of the unbuf-
fered aqueous medium decreases, which causes a change
in the color of a pH indicator. The kinetics for the color
change (bleaching) due to the released protons can be
determined on the microsecond time scale by laser flash
photolysis techniques (pH jump method).8 Establishing
the relationship between proton release rates and car-
boxylate anion release rates requires detailed knowledge
of the mechanism, which for purposes of discussion is
taken to be that depicted in Scheme 1. The proton release
being monitored by the pH jump method is thought to
occur after carboxylate elimination from zwitterionic
intermediate 4 and would occur either via intermolecular
addition of water to imminium ion 5 (path a) to ulti-
mately give 2 or via intramolecular cyclization of 5 to
give 3 (path b). These proton release rates would be
subject to effects governing rates of carboxylate elimina-
tion, depending on which step is rate determining, the
carboxylate elimination step, or the steps involving paths
a or b.
In this paper, we report full details of our photochemi-
cal studies of the photoelimination of a variety of car-
boxylate leaving groups (LG) from a series of N,N-diethyl
R-keto amides 1a (Figure 1). Comparisons are made to
an analogous series of N,N-diisopropyl R-keto amides 1b
(Figure 1 and eq 2). Although the photocleavages of 1a
and 1b appear to be similar, in the case of 1a we find
that the release rates are sensitive to the leaving group
ability of the carboxylate groups, whereas only a small
leaving group effect is observed for photolyses of 1b.
Furthermore, the photolytic eliminations of carboxylic
acids from N,N-diethyl and N,N-diisopropyl R-keto amides
1a,b are accompanied by the formation of different
cleavage coproducts for the two amides. Hemiacetal 2a
is strongly favored in the case of 1a (eq 1), whereas 3b
is exclusively formed from 1b (eq 2). These observations
point to some distinct differences in the mechanisms for
formation of photoproducts from 1a,b, which will be
discussed in this paper.
The R-keto amide photocleavages require transfer of a
hydrogen from an N-alkyl substituent to the oxygen of
(5) Chesta, C. A.; Whitten, D. G. J. Am. Chem. Soc. 1992, 114, 2188-
2197.
Results
(6) (a) Marriott, G., Ed. Methods in Enzymology; Academic Press:
San Diego, 1998; Vol. 291. (b) Kahl, J. D.; Greenberg, M. M. J. Am.
Chem. Soc. 1999, 121, 597-604. (c) Pirrung, M. C.; Fallon, L.; Lever,
D. C.; Shuey, S. W. J. Org. Chem. 1996, 61, 2129-2136. (d) Lloyd-
Williams, P.; Albericio, F.; Giralt, E. Tetrahedron 1993, 49, 11065-
11133.
Synthesis of R-Keto Amides 1a-c. In the syntheses
of 1a-c (Scheme 2), ethereal suspensions of dry potas-
(8) (a) Barth, A.; Corrie, J. E. T. Biophys. J. 2002, 83, 2864-2871.
(b) Choi, J.; Hirota, N.; Terazima, M. J. Phys. Chem. 2001, 105, 12-
18. (c) Viappiani, C.; Bonetti, G.; Carcelli, M.; Ferrari, F.; Sternieri,
A. Rev. Sci. Instrum. 1998, 69, 270-276. (d) Walker, J. W.; Reid, G.
P.; McCray, J. A.; Trentham, D. R. J. Am. Chem. Soc. 1988, 110, 7170-
7177. (e) Gutman, M.; Huppert, D.; Pines, E. J. Am. Chem. Soc. 1981,
103, 3709-3713. (f) Clark, J. H.; Shapiro, S. L.; Campillo, A. J.; Winn,
K. R. J. Am. Chem. Soc. 1979, 101, 746-748.
(7) (a) Morrison, J.; Wan, P.; Corrie, J. E. T.; Papageorgiou, G.
Photochem. Photobiol. Sci. 2002, 1, 960-969. (b) Givens, R. S.; Lee,
J.-I. J. Photosci. 2003, 10, 37-48. (c) Rajesh, C. S.; Givens, R. S.; Wirz,
J. J. Am. Chem. Soc. 2000, 122, 611-618. (d) Cheng, Q.; Steinmetz,
M. G.; Jayaraman, V. J. Am. Chem. Soc. 2002, 124, 7676-7677. (e)
Corrie, J. E. T.; Barth, A.; Munasinghe, V. R. N.; Trentham, D. R.;
Hutter, M. C. J. Am. Chem. Soc. 2003, 125, 8546-8554.
4432 J. Org. Chem., Vol. 70, No. 11, 2005

Photochemical Cleavage and Release of Carboxylic Acids
SCHEME 2a
a Conditions: (a) (COCl)2, CH2Cl2; HNRR′, Et3N. (b) Aq HOAc.
(c) Acylation methods, see text. (d) PCC, CH2Cl2.
sium salt 89 were converted to the acid chloride using
oxalyl chloride, which was reacted with diethylamine,
diisopropylamine, or N-ethyl-N-phenylamine in the pres-
ence of Et3N to obtain amides 9a-c in 65-70% yields.
Deprotection of the vicinal diols 10 in ca. 90% yield was
effected by refluxing with 70% aqueous acetic acid. The
carboxylate leaving groups were then introduced in 60-
85% yields by acylation of diols 10 using the correspond-
ing acid chlorides in the presence of triethylamine at 0
°C (LG ) PhCO2-, PhCH2CO2-, 4-NCC6H4CO2-) or acetic
anhydride at room temperature (LG ) CH3CO2-) or by
DCC coupling of the carboxylic acids with catalytic
DMAP (LG ) BocGABA, BocGlu, BocAla). After chro-
matographic purification of the monoacylated products
11, PCC oxidation of the secondary alcohols furnished
FIGURE 2. Absorption spectra of 1a (LG ) CH3CO2-) at
0.070, 0.0070, and 0.00070 M in 50% aqueous acetonitrile.
in pure D2O and 26 h in 25 mM phosphate buffer at pD
5.4. None of the hydrolyses produced 2a or 3a.
Absorption Properties. In water and aqueous ac-
etonitrile as solvents, R-keto amides 1a,b show UV
absorption maxima below 300 nm, and absorption tails
out into the 300-350 nm region (Figure 2). For N,N-
dialkyl R-keto amides, the n,π* transition has been
reported to occur at 315 nm in protic solvents.10 At this
wavelength, the corresponding extinction coefficient was
ꢀ 49 L mol-1 cm-1 for 1a (LG ) CH3CO2-). The absorption
properties of 1a-c allowed routine photolyses of dilute
10-2 M solutions above 300 nm with Pyrex-filtered light
or at 310 nm for quantum yield determinations. However,
the R-keto amides absorbed only very weakly (ꢀ < 10 L
mol-1 cm-1) at 355 nm, which necessitated high concen-
trations (>0.1 M) of compound to conduct laser flash
photolysis experiments at this wavelength.
Photoproducts in Aqueous Media. Under aqueous
conditions, the photolyses of 10-2-10-1 M 1a,b with
Pyrex-filtered light (>300 nm) produced high yields of
carboxylic acids along with one or more cleavage coprod-
ucts. In the case of the N,N-diethyl R-keto amides 1a,
the principal cleavage coproduct was the hemiacetal 2a,
which was obtained as a 3:2 mixture of diastereomers
(eq 1), and R-methyleneoxazolidinone 3a was only a
minor product or was not observed (Table 1). In contrast,
R-methyleneoxazolidinone 3b was the sole cleavage co-
product obtained in photolyses of the N,N-diisopropyl
R-keto amides 1b, and hemiacetal 2b was never observed
(eq 2 and Table 2).
the carboxylate-substituted R-keto amides 1a-c in 80-
-
85% yields. At 16.8 °C, compound 1c (LG ) PhCH2CO2
)
appeared to be an 8:1 mixture of conformers with respect
to rotation about the OdC-N carboxamide bond, accord-
ing to 1H NMR integration of signals of the O-CH2 group
at δ 5.16 and 5.37.
The BocGABA derivatives of 1a,b and BocGlu deriva-
tive of 1a were deprotected in ca. 85% yields by treatment
with CF3CO2H in CH2Cl2 at room temperature and
purified on Sephadex LH-20. For 1a (LG ) Glu), further
purification was effected by adding 1 equiv of triethyl-
amine to a suspension in EtOAc, followed by filtration
of the product.
-
-
-
Compounds 1a,b (LG ) CH3CO2 , PhCH2CO2 , PhCO2 ,
BocAla) and 1a (LG ) 4-NCC6H4CO2-) were stable for
at least 2 weeks in 50% D2O in CD3CN. This was also
true for 1a (LG ) CH3CO2-) in D2O, although in 25 mM
phosphate buffer in D2O hydrolysis occurred with a half-
life of 136 h at room temperature. The GABA and
glutamate derivatives of 1a also underwent hydrolytic
decomposition, and their half-lives under buffered and
unbuffered conditions were determined by 1H NMR
spectroscopy. For the GABA derivative, the half-life was
38 h in 25 mM phosphate buffer at pD 5.4 and 8 h at pD
7.1. For the glutamate derivative, the half-life was 70 h
The above photolyses used 50% aqueous CH3CN as the
solvent. However, with 1a (LG ) CH3CO2-, GABA, Glu),
solubilities allowed a fully aqueous 25 mM phosphate
buffer to be used as the medium. The product yields were
1
typically determined by H NMR spectroscopy for pho-
tolyses using the corresponding deuterated solvents
(Tables 1 and 2). For 1a (LG ) CH3CO2-), product yields
were similar for photolyses conducted in 50% D2O in CD3-
CN (air or N2 saturated), 100% D2O, and 25 mM
phosphate buffer at pD 7.
(9) (a) Earle, M. J.; Abdur-Rashid, A.; Priestley, N. D. J. Org. Chem.
1996, 61, 5697-5700. (b) Gluchowski, C.; Bischoff, T. E.; Garst, M.
E.; Kaplan, L. G.; Dietrich, S. W.; Aswad, A. S.; Gaffney, M. A.; Aoki,
K. R.; Garcia, C.; Wheeler, L. A. J. Med. Chem. 1991, 34, 392-397. (c)
Jefford, C. W.; Wang, Y. J. Chem. Soc., Chem. Commun. 1988, 634-
635. (d) Tanaka, A.; Yamashita, K. Agric. Biol. Chem. 1980, 44 (1),
199-202. (e) Andurkar, S. V.; Stables, J. P.; Kohn, H. Bioorg. Med.
Chem. 1999, 2381-2389. (f) Handa, S.; Hawes, J. E.; Pryce, R. J. Synth.
Commun. 1995, 25 (18), 2837-2845.
Photolyses of N-ethyl-N-phenyl amide 1c (LG )
PhCH2CO2-) in 30% aqueous CH3CN gave phenylacetic
(10) Fischer, G.; Oehme, G.; Schellenberger, A. Tetrahedron 1971,
27, 5683-5695.
J. Org. Chem, Vol. 70, No. 11, 2005 4433

Ma et al.
TABLE 1. Chemical Yields of Products from
TABLE 3. Chemical Yields of Products from
Steady-State Photolyses of r-Keto Amides 1aa,b
Steady-State Photolyses of r-Keto Amides 1aa in CD3CN
without Added Water
yield, %
yield, %
LG-
LG-H
2a
3a
1a
LG-
LG-H
12a
3a
1a
-c
CH3CO2
82
86
93
89
81
75
nd
75
70
75
66
88
nd
78
77
85
83
54
4.5
10
<5
nd
0
19
12
7
11
18
24
13
20
22
-
-
PhCH2-CdO2
CH3CO2
0
9
59
68
<5
8
32
25
-
PhCO2-c,e
PhCH2CO2
PhCO2
BocAla
a DMSO as the internal standard.
-
CH3CONHCH2CO2
4-NCC6H4CO2
0
-
0
0
15
according to 1H NMR spectroscopic analyses with DMSO
added as an NMR standard. These results indicated that,
under the above conditions, 2a and 3a did not intercon-
vert. In addition, 3b did not give 2b.
GABA
Glu
a Yields determined by 1H NMR spectroscopy using DMSO,
DMF, or glycine as a standard, unless noted otherwise. b Solvent
was air-saturated 50% D2O in CD3CN, except for GABA and Glu,
which used air-saturated D2O containing 25 mM phosphate buffer
as a solvent at pD 5 or 6, respectively. c Similar yields were
obtained under nitrogen. d No standard was used, and run was
performed under N2. e Yields of PhCO2H and unreacted 1a (LG )
PhCO2-) determined by HPLC analysis.
Deuterium Labeling. Information about the mech-
anisms for formation of hemiacetals 2a,c and R-methyl-
eneoxazolidinone 3a under aqueous conditions in pho-
tolyses of N,N-diethyl- and N-ethyl-N-phenyl R-keto
amides 1a,c was provided by deuterium labeling studies.
Such studies would also further address the possibility
of interconversion of 2a and 3a and, in the case of N,N-
diisopropyl R-keto amide 1b, would provide the means
to determine whether the unobserved hemiacetal 2b
TABLE 2. Chemical Yields of Products from
Steady-State Photolyses of r-Keto Amides 1ba,b
yield, %
could produce R-methyleneoxazolidinone 3b.
LG-
LG-H
2b
3b
unreacted 1b
When photolyses of 1a,c (LG ) PhCO2-, PhCH2CO2
)
-
-
CH3CO2
55
86
0
0
0
0
61
82
60
86
44
17
33
10
were conducted in 50% D2O in CD3CN, diastereomeric
hemiacetals 2a and 2c were found to be monodeuterated
at the CH3 group R to the carboxamide carbonyl group.
The presence of deuterium was established by integration
-
PhCH2-CO2
PhCO2
BocAla
ndb
88
a Yields determined by 1H NMR spectroscopy using DMSO,
DMF, or glycine as the standard. b Solvent was air-saturated 50%
D2O in CD3CN; the product acid could not be determined due to
overlapping of peaks with those of reactant 1b (LG ) PhCO2-).
1
of the H NMR signals at δ 1.54 and 1.47 and from the
overlapping 1:1:1 patterns of peaks centered at δ 23.6
and 24.2 in the 13C NMR spectrum.
acid and hemiacetal 2c as a 11:7 mixture of diastereo-
mers (eq 3). The R-methyleneoxazolidinone 3c was not
1
observed as a product. H NMR analyses of photolyses
conducted in the corresponding deuterated solvent showed
30% yield of 2c and 40% of yield acid at 44% conversion.
Photolyses of 1c (LG ) PhCO2-) were qualitatively
similar to those of 1c (LG ) PhCH2CO2-).
The minor product of 1a, R-methyleneoxazolidinone 3a,
was unlabeled at the R-methylene CH2 group, whereas
3c was not observed as a photoproduct of 1c. As with
3a, R-methyleneoxazolidinone 3b contained no detectable
deuterium at its R-methylene CH2 group or elsewhere
in the molecule when photolysis of 1b was conducted in
50% D2O in CD3CN.
These deuterium labeling results show that the mecha-
nistic pathway that leads to the labeling of 2a does not
produce 3a, since such a pathway would also result in
labeling of this latter product at the R-methylene position.
The unlabeled R-methylene group of 3a also indicates
that 2a and 3a do not interconvert. For the same reason,
3b is not formed from the unobserved hemiacetal 2b as
a precursor, which like 2a and 2c, would have been
expected to be monodeuterated at its R-methyl group.
No evidence was found that the photolyses of 1a-c (LG
) PhCH2CO2-) produced radical pairs. In these cases,
the phenylacetyloxyl radicals produced by such radical
cleavages would likely undergo rapid decarboxylation to
give benzyl radicals.11 Radical recombination products of
benzyl radicals, including bibenzyl, were not observed.
To determine whether the observed products, hemi-
acetals 2 and R-methyleneoxazolidinones 3, were stable
under the aqueous acidic conditions used for the pho-
tolyses of 1a and 1b, control experiments were performed
with each photoproduct in 50% D2O in CD3CN with
added acetic acid (pD 2.8) in the dark. During a period
of 14 days, 2a, 3a, and 3b each remained unchanged
Photoproducts in Acetonitrile with No Added
-
Water. Photolyses of 1a (LG ) CH3CO2-, PhCH2CO2
)
with Pyrex-filtered light in pure CH3CN gave aminals
12a (LG ) CH3CO2-, PhCH2CO2-) (eq 4, Table 3). The
major conformational diastereomer of 12a (LG ) CH3CO2-,
PhCH2CO2-) showed a distinctive quartet in the δ 6.3
1
(11) (a) DeCosta, D. P.; Pincock, J. A. J. Am. Chem. Soc. 1993, 115,
2180-2190. (b) Hilborn, J. W.; MacKnight, E.; Pincock, J. A.; Wedge,
P. J. J. Am. Chem. Soc. 1994, 116, 3337-3346. (c) Pincock, J. A. Acc.
Chem. Res. 1997, 30, 43-49.
region of the H NMR spectrum, and the minor diaste-
reomer showed a downfield quartet in the δ 6.8 region.
The aminals 12a could be isolated in pure form by silica
4434 J. Org. Chem., Vol. 70, No. 11, 2005

Photochemical Cleavage and Release of Carboxylic Acids
) CH3CO2-, BocAla, the quantum yields for formation
of 3b were similar to the quantum yields for disappear-
ance of the reactants. Compared to 1a,b, the photochem-
istry of compound 1c (LG ) PhCH2CO2-, PhCO2-) was
substantially less efficient, Φ ) ca. 0.05 for formation of
the carboxylic acids.
Quantum yields for appearance of PhCH2CO2H from
1a-c (LG ) PhCH2CO2-) were similar to those for
photocleavages that produce other carboxylic acids. No
decreases in quantum yields were found due to decar-
boxylation11 of phenylacetyloxyl radicals produced by a
radical cleavage mechanism.
TABLE 4. Quantum Yields for Photocleavage and Time
Constants for Bleaching of Bromocresol Green for r-Keto
Amides 1a-ca
reactant
LG-
Φb
lifetime (µs)c
-
1a
CH3CO2
0.31d
0.36e
0.36d
0.38e
0.31e
0.32d
0.38d
0.30d
0.24d,f
0.25e
0.24e
0.31f
137
55.5
29.7
18.2
ndg
nd
-
PhCH2CO2
BocAla
4-NCC6H4CO2
-
-
PhCO2
GABA
Glu
nd
-
CH3CONHCH2CO2
nd
-
1b
1c
CH3CO2
132
145
124
99
-
PhCH2-CO2
Photoinduced pH Changes. During photolyses in
unbuffered 50% aqueous acetonitrile, the photolysates
became acidic due to the presence of the photoreleased
carboxylic acids. Starting at a pH of 6.1-6.6, the pH of
PhCO2
BocAla
-
PhCH2-CO2
0.055e
0.058e
nd
nd
PhCO2
a Conducted in 50% aqueous CH3CN, except for LG ) GABA
or Glu, which were determined in 25 mM phosphate buffer (pH )
5.4). b Average of two to four determinations. c Average error was
less than 5% for 1a and less than 10% for 1b, except for LG )
BocAla, which had 16% error. d Quantum yield for disappearance
of reactant. e Quantum yield for appearance of carboxylic acid.
f Quantum yield for appearance of 3b. g Not determined.
the photolysates decreased by 2.6-3.3 units in the case
-
of 0.04-0.05 M 1a,b (LG ) CH3CO2-, PhCH2CO2
,
BocAla). These pH jumps were sufficient to cause change
in the color of the pH indicator bromocresol green (pKa
4.90, 20 °C12). By monitoring the photolyses using
absorption spectroscopy, the color change was found to
correspond to decreased intensity (bleaching) of the 620
nm band of the anionic form of the indicator, and there
was a corresponding increase in absorption at 420 nm.
Time-Resolved pH Jump Studies. To obtain infor-
mation on the rates of release of protons and, indirectly,
carboxylate anions from 1a,b, the kinetics of bleaching
of 15-20 µM bromocresol green pH indicator at 620 nm
were studied by 355 nm nanosecond laser flash photoly-
ses of 0.1-0.5 M 1a,b in air-saturated 50% aqueous CH3-
CN starting at a pH of 6-7 under unbuffered conditions.
The samples had to be flowed in order to observe a signal.
Each experiment represented a series of 10-30 shots
from a Nd:YAG laser spaced at 30 or 60 s time intervals.
The observed ∆OD versus time plots for each shot were
summed together and then fit to a monoexponential
function. This procedure furnished the time constants
and hence, rate constants for the bleaching of the
bromocresol green upon flash photolysis of 1a,b (Table
4). Representative kinetic plots are shown for 1a,b (LG
) PhCH2CO2-) in Figure 3A,B. Control experiments with
R-keto amide 1a (LG ) H), which did not have a leaving
group, showed no observable bleaching of the dye under
the conditions used in the flash photolysis experiments.
For 1a, the time constants for bleaching of bromocresol
green strongly decreased with increasing leaving ability
of the carboxylate groups (Table 4). The time constants
gel chromatography. In contrast to 1a, the N,N-diisopro-
-
pyl amides, 1b (LG ) CH3CO2-, PhCH2CO2-, PhCO2
,
BocAla, BocGABA) did not give aminals when photolyzed
in pure CH3CN, and only the carboxylic acids and the
R-methyleneoxazolidinones 3b were observed as the
products. As with photolyses under aqueous conditions,
no recombination products derived from benzyl radicals
were observed in the cases of 1a,b (LG ) PhCH2CO2-).
When dissolved in aqueous CH3CN, aminals 12a
1
hydrolyzed quantitatively to hemiacetals 2a. H NMR
studies using 50% D2O in CD3CN as the solvent showed
that product 2a was deuterated only at the OH(D) group
and nowhere else in the molecule. The disappearance of
aminal 12a (LG ) PhCH2CO2-) followed pseudo-first-
order kinetics, and the half-life was ca. 3.1 h.
Quantum Yields. For the series of N,N-diethyl amides
1a, quantum yields for reaction at 310 nm ranged from
0.24 to 0.38 in 50% aqueous CH3CN and showed little
variation for the series of carboxylate leaving groups
(Table 4). In the case of 1a (LG ) CH3CO2-) under fully
aqueous conditions (H2O or 25 mM phosphate buffer),
quantum yields were 0.31-0.32, i.e., identical to those
obtained in 50% aqueous CH3CN (Table 4); furthermore,
the quantum yields were essentially unchanged (Φ )
0.32) in D2O containing 25 mM phosphate buffer (pD 5.1).
Quantum yields were also similar (Φ ) 0.32, 0.38) for 3
and 5 h respective photolysis times for disappearance of
1a (LG ) GABA, Glu) in 25 mM phosphate buffer at pH
5.4, after correction for the contributions due to hydroly-
sis of the reactants in the dark (vide supra).
-
-
τ followed the order CH3CO2 (pKa 4.7) > PhCH2CO2
(pKa 4.3) > BocAla (pKa 4.02) > 4-CNC6H4CO2- (pKa 3.6).
A plot of log k versus -pKa13 gave a slope of 0.74 ( 0.07
σ (Figure 4). For 1b, the time constants were insensitive
to the basicity of the leaving group (Table 4).
Solvent isotope effects (H2O/D2O) for 1a (LG )
-
PhCH2CO2 and 1b (LG ) PhCH2CO2-) were found to
be opposite for the two compounds. An inverse solvent
isotope effect of 0.74 ( 0.02 was observed for 1a
(PhCH2CO2-), whereas a normal solvent isotope effect of
2.2 was observed for 1b (LG ) PhCH2CO2-).
(12) CRC Handbook of Chemistry and Physics, 76th ed.; Lide, D.
R., Ed.; CRC Press: Boca Raton, FL, 1995-1996; pp 8-17.
(13) (a) pKa values: Rappoport, Z. In CRC Handbook of Tables for
Organic Compound Identification, 3rd ed.; Weast, R. C., Ed.; CRC:
Cleveland, 1967; pp 429-433. (b) BocAla, pKa ) 4.02, CA 15761-38-3.
Quantum yields for reaction of N,N-diisopropyl amide
1b were 0.24-0.31 and similar to those for 1a. For LG
J. Org. Chem, Vol. 70, No. 11, 2005 4435

Ma et al.
SCHEME 3
FIGURE 3. Kinetics of bleaching of bromocresol green at 620
-
17 °C (see Results). Such a conformational dependence
for excited-state hydrogen transfer has been previously
shown for N-phenyl-N-ethyl R-keto amides that lack
leaving groups.4c,5 Nevertheless, an alternate rationale
for the low efficiencies would invoke an electron-
withdrawing effect of the N-phenyl group that retards
or makes reversible the initial step of the reaction, which
can be considered a net hydride transfer.
nm upon laser flash photolysis of (A) 1a (LG ) PhCH2CO2
)
and (B) 1b (LG ) PhCH2CO2-). Insets show the fit to a single-
exponential decay function made after the transient spike due
to cuvette luminescence in each case.
The mechanism for the initial, excited-state hydrogen
transfer step of the reaction (Scheme 3) has not been
definitively established, nor is the multiplicity of the
excited-state known, although previous studies5 have
shown that the reactive excited state is nonemissive and
nonquenchable via energy transfer, which would be
consistent with the involvement of a short-lived singlet
excited state. The hydrogen transfer could entail an
excited-state hydrogen atom abstraction from an N-alkyl
group, which would be consistent with a lowest energy
n,π* excited state.10 Alternatively, a previous study5 has
proposed an initial electron transfer from amide nitrogen
to the photoexcited keto group, followed by a proton
transfer. A third mechanistic possibility would involve
net hydride transfer from the amide alkyl group, which
can be viewed as an excited-state 1,5-H shift, given the
substantial double-bond character between amide car-
bonyl carbon and nitrogen in the ground state.
The photochemical hydrogen transfer would ultimately
produce a ground-state intermediate that can be de-
scribed as having zwitterionic character, as shown by
structures 4a-c (vide supra). Such zwitterionic inter-
mediates are thought to be responsible for the release of
the carboxylate leaving groups. Potentially competing
with elimination are cyclizations of intermediates 4 to
give oxazolidinones that retain the carboxylate leaving
groups. Such cyclizations are well-precedented in the
photochemistry of R-keto amides that have no leaving
groups.4,5,14 We have only observed oxazolidinone forma-
tion in the case of R-keto amide 14,2 which has a poor
alkoxide leaving group (eq 6). For less basic leaving
FIGURE 4. Log k vs -pKa plot for bleaching of bromocresol
green by 1a upon release of the carboxylic acids corresponding
-
to LG ) CH3CO2-, PhCH2CO2-, BocAla, and 4-NCC6H4CO2
.
Discussion
Photolyses of 1a,b in aqueous solutions gave high
yields of carboxylic acids for all of the carboxylate leaving
groups tested (Figure 1). The cleavage coproducts 2a, 2a
plus 3a, or 3b were formed in yields comparable to those
of the carboxylic acids. Since these coproducts do not
absorb light effectively above 300 nm, high photochemical
conversions of 1a,b were generally achievable. In most
respects, the photochemistry of 1c was similar to that of
1a except that longer photolysis times were necessary
to achieve significant conversions because of its low
quantum yield for reaction.
The photocleavage reactions of 1a,b are efficient (Φ >
0.2). The high quantum yields reflect, in part, efficient
hydrogen transfer from an N-alkyl group to the keto
oxygen upon generation of the excited state. The 5-7-
fold lower efficiency of photocleavage of 1c (Φ ) ca. 0.05)
may be attributable to the R-keto amide adopting mainly
an s-cis conformation in the ground state (eq 5), which
would be unreactive with respect to excited-state hydro-
(14) (a) Zehavi, U. J. Org. Chem. 1977, 42, 2821-2825. (b) Johans-
son, N. G.; Akermark, B.; Sjoberg, B. Acta Chem. Scand. B 1976, B30,
383-390.
1
gen transfer. H NMR spectroscopic data suggest that
1c exists as an 8:1 mixture of conformational isomers at
4436 J. Org. Chem., Vol. 70, No. 11, 2005

Photochemical Cleavage and Release of Carboxylic Acids
groups such as phenolates, 4-ZC6H4O- (Z ) H, CF3, or
CN),2 or carboxylates, elimination is invariably observed
as the sole reaction under aqueous conditions.
SCHEME 4
A second potential cyclization reaction, which could
compete with elimination, would be cyclization to form
â-lactam products.4,5,14b,15 â-Lactam formation, however,
is thought to occur via a nonplanar, equilibrium con-
former of 4 that has been described as a 1,4-diradical in
electronic configuration.5 We have never observed â-lac-
tam photoproducts in photolyses of 1a-c in aqueous
media or in pure CH3CN.
Initial studies to measure the rates of photochemical
release of the carboxylic acids from 1a utilized time-
resolved difference FT-IR spectroscopy as a method1 to
directly monitor the 1560 cm-1 absorption of photore-
leased carboxylate anions such as acetate and GABA in
D2O under buffered conditions. Use of a rapid scan
technique showed that the photochemical formation of
GABA occurred within 30 ms after the 355 nm laser
pulse. Later experiments that used a step-scan technique
with flowed samples of 1a (LG ) Glu) to obtain time
constants on the microsecond time scale were unsuccess-
ful, because of the weak absorption of the R-keto amide
at 355 nm.
intermediate imminium ion 5a, since pseudo-first-order
rate constants ranging from 106 to 108 s-1 have been
reported for a few examples in the literature.18 Once
proton release has occurred via path a, subsequent steps
in the mechanism (Scheme 4) are needed to account for
the formation of diastereomeric hemiacetals 2a. These
steps likely involve tautomerization19 of 6a followed by
cyclization of 7a. However, these acid-catalyzed reactions
involve relatively stable compounds and are not consid-
ered to contribute to the rates of proton release and
bleaching of the bromocresol green.
The formation of aminal products 12a (eq 4, vide supra)
in photolyses of 1a (LG ) CH3CO2-, PhCH2CO2-) in pure
CH3CN can be attributed to internal return of the ion
pair comprised of the initially released carboxylate anion
and imminium ion 5a.20 Such ion pair return would be
expected in the absence of water. Under aqueous condi-
tions, compounds 12a hydrolyze to diastereomeric hemi-
acetals 2a with release of the corresponding carboxylic
acids. Since this hydrolytic release of acid occurs slowly,
such aminals are unlikely to be intermediates in the
release of protons from 1a in the pH jump experiments
performed under aqueous conditions.
Permanent decreases in pH of aqueous photolysates
were observed during steady-state photolyses of 1a,b,
which caused a change in the color of a pH indicator such
as bromocresol green. The color change was due to a
decrease in intensity of the 620 nm absorption of the
bromocresol green, and there was a concomitant increase
in absorption at 420 nm. These observations suggested
that proton release rates could be obtained by a time-
resolved pH jump method8 using laser flash photolysis,
which would determine the kinetics of the bleaching of
the 620 nm absorption band of the bromocresol green.
Under aqueous conditions, products 2a and 3a are
formed via two mechanistically distinct pathways (paths
a and b, Scheme 4), according to deuterium labeling
studies and control experiments. Compounds 2a are not
the source of the minor amounts of 3a observed, because
-
-
For 1a (LG ) CH3CO2 , PhCH2CO2 , BocAla, 4-NCC6H4-
CO2-), the time constants for bleaching of the 620 nm
absorption of bromocresol green decrease with increasing
leaving group ability of the carboxylate groups (Table 4).
The leaving group effect (Figure 4) is substantial and
would imply that the rate-determining step in the release
of protons involves the elimination of carboxylate anions.
According to Scheme 4, such carboxylate eliminations
would occur in zwitterionic intermediates 4a, prior to the
proton release step that leads to hemiacetal 2 along path
a. The observed inverse isotope effect of 0.74 would be
consistent with the enolate group of 4a being a stronger
base in D2O than H2O, such that the elimination occurs
more rapidly in the deuterated solvent,16 whereas a
normal solvent isotope effect would be expected if the
subsequent proton release step along path a were rate
determining.17
(16) (a) An inverse isotope effect usually signifies the existence of a
preequilibrium protonation step that is followed by a rate-determining
step.16b,c We suspect that the isotope effect could be due to weaker
solvation of the intermediate in D2O than in H2O, although the
interpretation of the isotope effect is complicated by the fact that the
pH rapidly changes in our unbuffered experiments. (b) Eldin, S.; Digits,
J. A.; Huang, S.-T.; Jencks, W. P. J. Am. Chem. Soc. 1995, 117, 6631-
6632. (c) Eldin, S.; Jencks, W. P. J. Am. Chem. Soc. 1995, 117, 4851-
4857.
(17) Schowen, R. L. Progress in Physical Organic Chemistry; Wiley-
Interscience: New York, 1972; Vol. 9, pp 275-332.
(18) (a) Keefe, J. R.; Jencks, W. P. J. Am. Chem. Soc. 1983, 105,
265-279. (b) Keefe, J. R.; Jencks, W. P. J. Am. Chem. Soc. 1981, 103,
2457-2459.
(19) Chiang, Y.; Kresge, A. J.; Santaballa, J. A.; Wirz, J. J. Am.
Chem. Soc. 1988, 110, 5506-5510.
(20) A referee has suggested a different view, namely, that 1,5-
hydride shift to give 4a in the N,N-diethyl compound 1a depends on
the solvent. Since the plus charge in zwitterions 4 would not be
stabilized as much in 1a as in the N,N-diisopropyl compound 1b, the
product difference for the two compounds in neat CH3CN could be due
to 1a undergoing hydrogen atom abstraction to give a 1,4-diradical
followed by loss of the leaving group as a radical. Cage recombination
would then give 12a. In the case of 1a (LG ) PhCH2CO2-), the
phenylacetyloxyl radical produced via radical cleavage would be
expected to decarboxylate to benzyl radicals.11 Although we observed
no bibenzyl or other products of benzyl radicals, it is not guaranteed
that the decarboxylation would effectively compete with rapid recom-
bination of the caged radicals to give 12a.
In the case of 1a, the major pathway for proton release,
path a, is thought to involve addition of water to an
(15) (a) Asahi, T.; Nakamura, M.; Kobayashi, J.; Toda, F.; Miyamoto,
H. J. Am. Chem. Soc. 1997, 119, 3665-3669. (b) Natarajan, A.; Wang,
K.; Ramamurthy, V.; Scheffer, J. R.; Patrick, B. Org. Lett. 2002, 4,
1443-1446.
J. Org. Chem, Vol. 70, No. 11, 2005 4437

Ma et al.
SCHEME 5
flash photolytic techniques. These time constants range
from 18 to 137 µs for the N,N-diethyl amides and reflect
the leaving group ability of the carboxylate groups. In
contrast, for the N,N-diisopropyl amides the time con-
stants are less sensitive to the nature of the leaving
groups and range from 99 to 145 µs. The large leaving
group effect and an inverse solvent isotope effect (H2O/
D2O) on bleaching rates for the N,N-diethyl amides
support a rate-limiting elimination of the carboxylate
group in the postulated zwitterionic intermediate, whereas
for the N,N-diisopropyl amides a subsequent step involv-
ing intramolecular cyclization and proton release is
evidently significantly rate limiting.
The major cleavage coproducts, diastereomeric hemi-
acetals, which are observed for N,N-diethyl amides, are
formed via intermolecular addition of water to an inter-
mediate enolic imminium ion. On the other hand, in-
tramolecular cyclization of an enolic hydroxyl group in
an imminium ion intermediate produces exclusively the
cleavage coproduct of the N,N-diisopropyl amides, an
R-methyleneoxazolidinone. The postulated imminium ion
produced initially upon carboxylate elimination can be
trapped by the carboxylate anions when the photolyses
of N,N-diethyl amides 1a are conducted in CH3CN
without added water. Hydrolysis of these trapping prod-
ucts furnishes exclusively diastereomeric hemiacetals.
the latter product is found to contain no deuterium at
the R-methylene group or elsewhere in the molecule,
whereas the formation of 2a results in incorporation of
deuterium from D2O at the methyl group R to the
carboxamide carbonyl. The deuterium incorporation into
2a would be indicative of a prior tautomerization step in
the mechanism for its formation (Scheme 4). We also find
that compound 3a does not convert to 2a under unbuf-
fered aqueous photolysis conditions. At pD 2.8, compound
3a remains unchanged over extended periods of time (14
days).
Although path a is the major pathway for product
formation in the case of N,N-diethyl amides 1a, only path
b is apparently operative in the case of N,N-diisopropyl
amides 1b, since the sole photoproduct observed is 3b
(Scheme 5). Deuterium is not incorporated into 3b from
D2O, and therefore, 3b is not formed from 2b, which
would be an unobserved product of path a. Evidently,
intramolecular trapping of the imminium ion group by
the enolic hydroxyl of 5b is more rapid than intermo-
lecular addition of water to this intermediate.
In the case of 1b, the time constants for bleaching of
bromocresol green (Table 4) are much less sensitive to
the nature of the leaving group as compared to those
determined for 1a. The small leaving group effect sug-
gests that the intramolecular cyclization of 5b (path b,
Scheme 1) may be significantly rate determining. The
greater importance of this step in determining the
bleaching rates could be due to stabilization of both 4b
and 5b relative to product 3b by the extra methyl group
at the imminium ion moiety. Rate-determining cycliza-
tion of 5b would be consistent with the observed normal
solvent isotope effect (H2O/D2O) on bleaching rates for
1b (LG ) PhCH2CO2-) of 2.2.
Experimental Section
HPLC analyses were performed on a 14.6 × 250 mm, 10
µm ODS-2 column eluting with 70% acetonitrile and 30% 25
mM phosphate ion buffer at 1.5 mL min-1; UV (254 nm)
detector response was calibrated using standard mixtures.
Preparative chromatographic separations (MPLC) used a 2.5
× 28 cm column of 230-400 mesh silica gel eluting with 50%
ethyl acetate in hexane at 15 mL min-1 with a pump, unless
otherwise noted. The 25 mM phosphate ion buffer was
prepared from 100 mM KH2PO4 and 150 mM NaCl in D2O,
adjusted to pD 7.4 with KOH, and then diluted 4-fold with
deionized water.
Preparation of Acetic Acid, 3-(Diethylamino)-2-hy-
droxy-3-oxopropyl Ester 11a (LG ) CH3CO2-). A solution
of 0.50 g (3.1 mmol) of N,N-diethyl-2,3-dihydroxypropanamide
10a and 0.20 g (2.0 mmol) of acetic anhydride in 10 mL of
CH2Cl2 was stirred at room temperature for 36 h. The reaction
mixture was diluted by ethyl acetate, washed with 5% NaH-
CO3 and saturated NaCl, and dried over Na2SO4. After
concentration in vacuo, the residue was purified by MPLC,
eluting with 50% EtOAc in hexane to give 0.4 g (60% yield) of
an oil. The spectral data were as follows: 1H NMR (CDCl3) δ
1.12 (t, J ) 7.2 Hz, 3 H), 1.21 (t, J ) 7.2 Hz, 3 H), 2.18 (S,
3H), 3.48 (dq, J ) 14.2, 7.2 Hz, 1H), 3.29 (dq, J ) 14.4, 7.5
Hz, 1H), 3.56 (dq, J ) 14.4, 6.9 Hz, 1 H), 3.26 (dq, J ) 14.2,
7.5 Hz, OH), 3.95 (ddd, J ) 7.8, 4.5, 15.6 Hz, 2 H), 4.27 (dd, J
) 3.0, 11.7, 1 H), 4.52 (dt, J ) 3.0,7.5 Hz, 1 H); 13C NMR
(CDCl3) δ 13.5, 14.8, 21.5, 41.0, 41.5, 67.0, 67.3, 169.1, 170.4.
Anal. Calcd for C9H17NO4: C, 53.19; H, 8.43; N, 6.89. Found:
C, 52.89; H, 8.37; N, 6.79.
Conclusion
R-Keto amides with various carboxylate leaving groups
undergo photocleavage to form carboxylic acids and
diastereomeric hemiacetals or R-methyleneoxazolidino-
nes. The photoreaction is efficient, with quantum yields
of 0.2-0.4 for N,N-diethyl and N,N-diisopropyl amides,
whereas for N-ethyl-N-phenyl amides the photocleavage
is inefficient, because the conformer required for hydro-
gen transfer from the N-ethyl group to the keto group in
the excited state is disfavored.
The photoreleased carboxylic acids cause a decrease
in pH of 2-3 units in aqueous solutions. The time
constants for proton release were determined by monitor-
ing the kinetics of bleaching of the 620 nm absorption of
the pH indicator, bromocresol green, using nanosecond
Preparation of Benzoic Acid, 3-(Diethylamino)-2-hy-
droxy-3-oxopropyl Ester 11a (LG ) PhCO2-). To a solution
of 4.0 g (25 mmol) of N,N-diethyl-2,3-dihydroxypropanamide
10a in 20 mL of benzene was added, dropwise with stirring,
5.4 g (24 mmol) of neat benzoic anhydride at room tempera-
ture. The reaction was stirred for 36 h. The benzene was
removed in vacuo, and ethyl acetate was added, followed by
washing with 5% NaHCO3 and saturated NaCl. After drying
over Na2SO4, the solvent was removed in vacuo to give the
crude product, which was purified by MPLC to give 2.2 g (35%
yield) of 11a (LG ) PhCO2-) as a colorless oil. The spectral
4438 J. Org. Chem., Vol. 70, No. 11, 2005

Photochemical Cleavage and Release of Carboxylic Acids
data were as follows: 1H NMR (CDCl3) δ 1.14 (t, J ) 7.0 Hz,
3 H), 1.25 (t, J ) 7.0 Hz, 3 H), 3.27 (dq, J ) 13.5, 7.0 Hz 1 H),
3.31 (dq, J ) 14.8, 7.0 Hz 1 H), 3.57 (dq, J ) 14.8, 7.0 Hz 1
H), 3.61 (dq, J ) 13.5, 7.0 Hz, 1 H), 4.12 (d, J ) 6.3 Hz, OH),
4.25 (dd, J ) 6.3,11.7 Hz, 1 H), 4.56 (dd, J ) 3.0,11.7, 1 H),
4.68 (dt, J ) 3.0, 7.0 Hz, 1 H), 7.42 (t, J ) 7.5 Hz, 2 H), 7.55
(t, J ) 7.5 Hz, 1 H), 8.03 (d, J ) 7.2 Hz, 2 H); 13C NMR (CDCl3)
δ 13.5, 14.9, 41.0, 41.6, 67.3, 128.1, 129.1, 129.4, 132.9, 165.9,
169.2. Anal. Calcd for C14H19NO4: C, 63.38; H, 7.22; N, 5.28.
Found: C, 63.31; H, 7.13; N, 5.20.
Preparation of Phenylacetic Acid, 3-(Diethylamino)-
2-hydroxy-3-oxopropyl Ester 11a (LG ) PhCH2CO2-). To
a solution 5.4 g (34 mmol) of N,N-diethyl-2,3-dihydroxypro-
panamide 10a, 1 mL of pyridine, and a catalytic amount
DMAP in 40 mL of CH2Cl2 was added dropwise 5.0 g (32 mmol)
of phenylacetyl chloride in 10 mL of CH2Cl2 over a 30 min
period while stirring in an ice bath under N2. After 24 h at
room temperature, the reaction mixture was washed with
water, 5% NaHCO3, and saturated NaCl, dried over Na2SO4,
and concentrated in vacuo. The resulting oil was purified by
MPLC to give 5.0 g (75% yield) of the product as an oil. The
spectral data were as follows: 1H NMR (CDCl3) δ 1.13 (t, J )
7.0 Hz, 3 H), 1.20 (t, J ) 7.0 Hz, 3 H), 3.78 (s, 2 H), 3.23 (dq,
J ) 13.5, 7.0 Hz 1 H), 3.25 (dq, J ) 14.8, 7.0 Hz 1 H), 3.45
(dq, J ) 13.5, 7.0 Hz 1 H), 3.55 (dq, J ) 14.8, 7.0 Hz 1 H),
4.01 (dd, J ) 7.2, 11.7 Hz, 1H), 4.35 (dd, J ) 2.7, 11.4 Hz,
1H), 4.56 (d,t, J ) 3.0, 7.5 Hz, 1H), 7.50 (m., 5 H); 13C NMR
(CDCl3) δ 13.5, 14.9, 41.1, 41.6, 42.6, 67.3, 67.4, 127.0, 128.3,
129.0, 133.2, 169.1, 171.1. Anal. Calcd for C15H21NO4: C, 64.50;
H, 7.58; N, 5.01. Found: C, 64.22; H, 7.69; N, 5.20.
Preparation of Acetic Acid, 3-(Diethylamino)-2,3-di-
oxopropyl Ester 1a (LG ) CH3CO2-). To a stirred solution
of 1.5 g (7.1 mmol) of acetic acid 3-(diethylamino)-2-hydroxy-
3-oxopropyl ester 11a (LG ) CH3CO2-) in 70 mL of CH2Cl2
was added 3.9 g (18 mmol) of PCC at room temperature. After
24 h reaction, 100 mL of ether and 5 g of Florisil (60-100mesh)
were added, followed by suction filtration. The filtrate was
washed with 5% NaHCO3 and saturated NaCl and dried over
Na2SO4. After concentration in vacuo, the residue was purified
by MPLC to give 1.0 g (68% yield) of product. The spectral
data were as follows: 1H NMR (CDCl3) δ 1.16 (t, J ) 7.2 Hz,
3 H), 1.20 (t, J ) 7.2 Hz, 3 H), 2.15 (s, 3 H), 3.35 (q, J ) 7.2
Hz, 2 H), 3.88 (q, J ) 7.2 Hz, 2 H), 5.0 (s, 2 H); 13C NMR
(CDCl3) δ 13.3, 15.2, 21.1, 40.5, 42.7, 67.1, 163.2, 169.9, 192.1.
Anal. Calcd for C9H15NO4: C, 53.72; H, 7.51; N, 6.96. Found:
C, 53.74; H, 7.53; N, 6.72.
Preparation of Benzoic Acid, 3-(Diethylamino)-2,3-
dioxopropyl Ester 1a (LG ) PhCO2-). The same method
for synthesis of acetic acid 3-(diethylamino)-2,3-dioxopropyl
ester 1a (LG ) CH3CO2-) was used to prepare 2.3 g (70% yield)
of R-keto amide 1a (LG ) PhCO2-) as a colorless oil from 3.3
g (12 mmol) of benzoic acid, 3-(diethylamino)-2-hydroxy-3-
oxopropyl ester 11a (LG ) PhCO2-). The spectral data were
as follows: 1H NMR (CDCl3) δ 1.16 (t, J ) 7.2 Hz, 3 H), 1.23
(t, J ) 7.2 Hz, 3 H), 3.42 (q, J ) 7.2 Hz, 4 H), 5.28 (s, 2 H),
7.46 (m, 2 H), 7.59 (m, 1 H), 8.09 (m, 2 H); 13C NMR (CDCl3)
δ 12.3, 14.2, 39.7, 42.0, 66.9, 128.1, 128.5, 129.6, 133.2, 163.5,
165.7, 192.6. Anal. Calcd for C14H17NO4: C, 63.87; H, 6.51; N,
5.32. Found: C, 64.02; H, 6.63; N, 5.36.
Calcd for C15H19NO4: C, 64.97; H, 6.91; N, 5.05. Found: C,
65.11; H, 7.06; N, 5.01.
Preparation of N-BOC-Alanine, 3-(Diethylamino)-2,3-
dioxopropyl Ester 1a (LG ) BocAla). To a solution of 2.2
g (14 mmol) of N,N-diethyl-2,3-dihydroxypropanamide 10a and
a catalytic amount of DMAP was added 2.0 g (10 mmol) of
2-(tert-butoxycarbonylamino)propionic acid (N-BocAla). After
the carboxylic acid dissolved, 13 mL (1.0 M in CH2Cl2) of DCC
was added under nitrogen over 15 min, and the reaction was
stirred for 24 h at room temperature. The dicyclohexylurea
was removed by suction filtration, and the filtrate was washed
twice with water (10 mL), 5% NaHCO3, and saturated NaCl,
dried over Na2SO4, and concentrated in vacuo to give 2.7 g of
N-BOC-alanine 3-(diethylamino)-2-hydroxy-3-oxopropyl ester
11a (LG ) BocAla) as an oil. Without further purification, a
solution of 2.7 g (8.3 mmol) of the oil in 30 mL of CH2Cl2 was
stirred while adding 2.7 g (12 mmol) of PCC at room temper-
ature. After 12 h of reaction, 10 mL of ether and 3 g of Florisil
(60-100 mesh) were added, followed by suction filtration. The
filtrate was washed with 5% NaHCO3 and saturated NaCl and
dried over Na2SO4. After concentration in vacuo, the residue
was purified by MPLC to give 1.8 g (69% yield) of crystalline
product, which was crystallized from ethyl acetate in hexane
to give material with mp 69-72 °C. The spectral data were as
follows: 1H NMR (CDCl3) δ 1.20 (t, J ) 6.9 Hz, 3 H), 1.23 (t,
J ) 6.9 Hz, 3 H), 1.45 (s, 9 H), 1.49 (d, J ) 6.9 Hz, 1 H), 3.38-
(q, J ) 7.2 Hz, 2 H), 3.42 (q, J ) 6.9 Hz, 2 H), 4.43 (qd, J )
6.9, 5.7 Hz, 1 H), 5.05 (d, J ) 5.7 Hz, 1 H), 5.06 (d, J ) 17.1
Hz, 1 H), 5.17 (d, J ) 17.1 Hz, 1 H); 13C NMR (CDCl3) δ 13.3,
19.3, 28.9, 40.6, 42.7, 49.5, 67.5, 80.2, 154.5, 163.0, 172.3, 191.4.
Anal. Calcd for C15H26N2O6: C, 54.53; H, 7.93; N, 8.48.
Found: C, 54.83; H, 7.88; N, 8.42.
Preparation of 4-Cyanobenzoic Acid, 3-(Diethylamino)-
2,3-dioxopropyl Ester 1a (LG ) 4-CNC6H4CO2-). To a
solution of 1.6 g (9.9 mmol) of N,N-diethyl-2,3-dihydroxypro-
panamide 7a and 1 mL of triethylamine in 20 mL of CH2Cl2
in an ice bath under nitrogen was added 1.3 g (8.0 mmol) of
4-cyanobenzoyl chloride over 15 min. The reaction was stirred
for 24 h at room temperature and then washed twice with
water (10 mL), 5% NaHCO3, and saturated NaCl, dried over
Na2SO4, and concentrated in vacuo to give 1.4 g of 4-cyanoben-
zoic acid 3-(diethylamino)-2-hydroxy-3-oxopropyl ester 11a (LG
) 4-CNC6H4CO2-) as a colorless solid. Without further puri-
fication, a solution of 1.4 g (4.8 mmol) of the oil in 30 mL of
CH2Cl2 was stirred while adding 1.5 g (7.2 mmol) of PCC at
room temperature. After 12 h of reaction, 10 mL of ether and
3 g of Florisil (60-100 mesh) were added, followed by suction
filtration. The filtrate was washed with 5% NaHCO3 and
saturated NaCl and dried over Na2SO4. After concentration
in vacuo, the residue was purified by MPLC to give 1.0 g (75%
yield) of product as colorless crystals. Crystallization from
ethyl acetate in hexane gave material with mp 90-91 °C. The
spectral data were as follows: 1H NMR (CDCl3) δ 1.18 (t, J )
7.2 Hz, 3 H), 1.24 (t, J ) 7.2 Hz, 3 H), 3.43(q, J ) 7.2 Hz, 2
H), 3.44 (q, J ) 7.2 Hz, 2 H), 5.35 (s, 2 H), 7.77 (d, J ) 8.1 Hz,
2 H), 8.20 (d, J ) 8.1 Hz, 2 H); 13C NMR (CDCl3) δ 12.9, 14.8,
40.6, 42.6, 68.0, 117.1, 118.0, 130.5, 132.4, 132.9, 163.5, 164.6,
192.1. Anal. Calcd for C15H16N2O4: C, 62.49; H, 5.59; N, 9.72.
Found: C, 62.62; H, 5.65; N, 9.62.
Preparation of Acetylglycine, 3-(Diethylamino)-2,3-
dioxopropyl Ester 1a (LG ) CH3CONHCH2CO2-). To a
solution of 1.8 g (11 mmol) of N,N-diethyl-2,3-dihydroxypro-
panamide 10a and a catalytic amount of DMAP was added
1.0 g (8.5 mmol) N-acetylglycine. After the carboxylic acid
dissolved, 10 mL (1.0 M in CH2Cl2) of DCC was added under
nitrogen over 15 min, and the reaction was stirred for 24 h at
room temperature. The dicyclohexylurea was removed by
suction filtration; the filtrate was washed twice with water
Preparation of Phenylacetic Acid, 3-(Diethylamino)-
2,3-dioxopropyl Ester 1a (LG ) PhCH2CO2-). The same
method for synthesis of acetic acid 3-(diethylamino)-2,3-
dioxopropyl ester 1a (LG ) CH3CO2-) was used to prepare 1.7
g (68% yield) of R-keto amide 1a (LG ) PhCH2CO2-) as a
colorless oil from 2.5 g (8.9 mmol) of phenylacetic acid,
3-(diethylamino)-2-hydroxy-3-oxopropyl ester 11a (LG )
PhCH2CO2-). The spectral data were as follows: 1H NMR
(CDCl3) δ 1.18 (t, J ) 7.2 Hz, 3 H), 1.21 (t, J ) 7.2 Hz, 3 H),
3.35 (q, J ) 7.2 Hz, 2H), 3.40 (q, J ) 7.2 Hz, 2H), 5.07 (s, 2 H),
7.2-7.4 (m, 5 H); 13C NMR (CDCl3) δ 13.2, 15.1, 40.4, 41.0,
42.6, 67.3, 126.9, 128.2, 129.0, 132.8, 163.2, 170.5, 191.9. Anal.
(10 mL), and the water was removed by distillation at reduced
-
pressure to obtain crude ester 11a (LG ) CH3CONHCH2CO2
)
as an oil. Without further purification, ester 11a in 20 mL of
CH2Cl2 was stirred while adding 2.4 g (12 mmol) of PCC at
J. Org. Chem, Vol. 70, No. 11, 2005 4439

Ma et al.
room temperature. After 12 h of reaction, 10 mL of ether and
3 g of Florisil (60-100 mesh) were added, followed by suction
filtration. The filtrate was concentrated in vacuo, and the
residue was purified by MPLC to give 0.1 g (4% yield) of an
oil. The spectral data were as follows: 1H NMR (CDCl3) δ 1.22
(t, J ) 7.2 Hz, 3H), 1.18 (t, J ) 7.2 Hz, 3H), 2.06 (s, 3H), 3.35
(q, J ) 7.2 Hz, 2H), 3.40 (q, J ) 7.2 Hz, 2H), 4.18 (br, 2H),
5.12 (s, 2H), 6.15 (br, 1H); 13C NMR (CDCl3) δ 12.8, 14.8, 13.2,
40.5, 41.3, 42.5, 67.6, 163.4, 169.7, 170.4, 192.1; MS (m/z) 230
(0.6), 158 (3.3), 141 (5.6), 100 (100), 72 (57.5). Anal. Calcd for
C11H18N2O5: C, 51.16; H, 7.02; N, 10.85. Found: C, 51.30; H,
7.07; N, 10.57.
Preparation of N-BOC-GABA, 3-(Diethylamino)-2,3-
dioxopropyl Ester 1a (LG ) BOC-GABA). To a solution of
0.64 g (4.0 mmol) of N,N-diethyl-2,3-dihydroxypropanamide
10a and a catalytic amount of DMAP in 20 mL of CH2Cl2 was
added 0.71 g (3.5 mmol) of 4-(tert-butoxycarbonylamino)butyric
acid (N-BOC-GABA). After the carboxylic acid dissolved, a
solution of 0.78 g (3.5 mmol) of DCC in 10 mL of CH2Cl2 was
added under nitrogen over 15 min. The reaction was stirred
for 24 h at room temperature. The mixture was washed with
water, 5% NaHCO3, and saturated NaCl, dried over Na2SO4,
and concentrated in vacuo to give 0.50 g (46% yield) of N-BOC-
GABA, 3-(diethylamino)-2-hydroxy-3-oxopropyl ester (11a, LG
) BocGABA), as an oil. The hydroxyamide was oxidized to
R-keto amide 1a (LG ) BocGABA) in 80% yield by the same
procedure used to prepare the acetic acid ester derivative 1a
(LG ) CH3CO2-). Crystallization from ether in hexane gave
the product as white crystals, mp 68-70 °C. The spectral data
were as follows: 1H NMR (CDCl3) δ 1.21 (t, J ) 7.2 Hz, 3H),
1.24 (t, J ) 7.2 Hz, 3 H), 1.46 (s, 9 H), 1.88 (qui, J ) 7.2 Hz,
2 H), 2.51 (t, J ) 7.2 Hz, 2 H), 3.31 (br q, J ) 7.2 Hz, 2 H),
3.39 (q, J ) 7.2 Hz, 2 H), 3.43 (q, J ) 7.2 Hz, 2 H), 4.69 (br, 1
H, NH), 5.07 (s, 2 H); 13C NMR (CDCl3) δ? 13.4, 15.3, 25.9,
29.1, 31.6, 40.2, 40.5, 42.7, 67.1, 79.2, 156.5, 163.2, 172.1, 192.1.
Anal. Calcd for C16H28N2O6: C, 55.79; H, 8.19; N, 8.13.
Found: C, 55.44; H, 7.93; N, 7.78.
spectral data below were obtained in 25 mM pD 7.4 phosphate
buffer prepared in D2O.
Trifluoroacetate Salt of r-Keto Amide 1a (LG ) GABA).
The spectral data were as follows: 1H NMR (D2O) δ 1.16 (t, J
) 7.2 Hz, 3 H), 1.20 (t, J ) 7.2 Hz, 3 H), 2.25 (ddt, J ) 21.0,
7.2, 7.2 Hz, 1 H), 2.75 (td, J ) 7.2, 2.9 Hz, 2 H), 3.38 (q, J )
7.2 Hz, 2 H), 3.42 (q, J ) 7.2 Hz, 2 H), 3.98 (t, J ) 7.2 Hz, 1
H), 5.18 (s, 2 H); 13C NMR (D2O) δ 15.0, 16.7, 28.4, 32.5, 43.4,
46.0, 55.7, 70.1, 116.4 (q, J ) 383 Hz), 162.38 (q, J ) 46.5
Hz), 167.0, 174.3, 175.3, 197.5.
r-Keto Amide 1a (LG ) Glu). Colorless solid, mp 103-
105 °C. The spectral data were as follows: 1H NMR (CDCl3) δ
1.13 (t, J ) 7.2 Hz, 3 H), (t, J ) 7.2 Hz, 3 H), 1.97 (qui, J )
7.5 Hz, 2 H), 2.62 (t, J ) 7.2 Hz, 2 H), 3.03 (br t, J ) 7.5 Hz,
2 H), 3.35 (q, J ) 7.2 Hz, 2 H), 3.39 (q, J ) 7.2 Hz, 2 H), 5.14
(s, 2 H); 13C NMR (CDCl3) δ 12.6, 14.3, 22.9, 31.0, 39.3, 40.9,
43.5, 67.6, 116.4 (q, J ) 383 Hz), 162.4 (q, J ) 46.5 Hz), 164.5,
173.2, 195.0. Anal. Calcd for C14H21N2O8‚0.26CF3CO2H: C,
47.29; H, 6.42; N, 8.81. Found: C, 47.27; H, 6.31; N, 8.84.
Preparation of Acetic Acid, 3-(Diisopropylamino)-2,3-
dioxopropyl Ester 1b (LG ) CH3CO2-). The same method
for the synthesis of benzoic acid 3-(diethylamino)-2-hydroxy-
3oxopropyl ester 11a (LG ) PhCO2-) was used to prepare 0.4
g (60% yield) of acetic acid, 3-(diisopropylamino)-2-hydroxy-
3-oxopropyl ester 11b (LG ) CH3CO2-), as a colorless oil from
0.53 g (2.8 mmol) of N,N-diisopropyl-2,3-dihydroxypropan-
amide 10b. Oxidation of 0.81 g (3.5 mmol) of acetic acid,
3-(diisopropylamino)-2-hydroxy-3-oxopropyl ester 11b (LG )
CH3CO2-) by PCC gave 0.6 g (70% yield) of R-keto amide 1b
(LG ) CH3CO2-) after MPLC purification. The spectral data
were as follows: 1H NMR (CDCl3) δ 1.23 (d, J ) 6.9 Hz, 6 H),
1.44 (d, J ) 6.9 Hz, 6 H), 2.17 (s, 3 H), 3.52 (sept, J ) 6.9 Hz,
1 H), 3.94 (sept, J ) 6.9 Hz, 1 H), 4.93 (s, 2 H); 13C NMR
(CDCl3) δ 20.8, 21.1, 21.4, 46.7, 50.2, 66.7, 164.4, 169.9, 192.6.
Anal. Calcd for C11H19NO4: C, 57.62; H, 8.35; N, 6.11. Found:
C, 57.65; H, 8.41; N, 6.06.
Preparation of Benzoic Acid, 3-(Diisopropylamino)-
2,3-dioxopropyl Ester 1b (LG ) PhCO2-). The same
method for synthesis of acetic acid 3-(diethylamino)-2,3-
dioxopropyl ester 1a (LG ) CH3CO2-) was used to prepare 1.3
g (67% yield) of R-keto amide 1b (LG ) PhCO2-) as a colorless
oil from 2.0 g (6.6 mmol) of benzoic acid, 3-(diisopropylamino)-
2-hydroxy-3-oxopropyl ester 11b (LG ) PhCO2-). The spectral
data were as follows: 1H NMR (CDCl3) δ 1.25 (d, J ) 6.9 Hz,
6 H), 1.43 (d, J ) 6.9 Hz, 6 H), 3.53 (sept, J ) 6.9 Hz, 1 H),
4.03 (sept, J ) 6.9 Hz, 1 H), 5.20 (S, 2 H), 8.10 (d, J ) 7.2 Hz,
2 H), 7.46 (t, J ) 7.2 Hz, 2 H), 7.60 (t, J ) 7.5 Hz, 1 H); 13C
NMR (CDCl3) δ 20.8, 21.4, 46.7, 50.3, 67.1, 128.2, 128.6, 129.6,
133.3, 164.4, 165.6, 192.7. Anal. Calcd for C16H21NO4: C, 65.96;
H, 7.22; N, 4.81. Found: C, 66.23; H, 7.42; N, 4.82.
Preparation of Phenylacetic Acid, 3-(Diisopropylami-
no)-2,3-dioxopropyl Ester 1b (LG ) PhCH2CO2-). To a
solution of 1.1 g (5.8 mmol) of N,N-diisopropyl-2,3-dihydroxy-
propanamide 10b, 5 mL (1.0 M in CH2Cl2) of DCC, and a
catalytic amount DMAP in 10 mL of CH2Cl2 was added
dropwise with stirring 0.53 g (3.9 mmol) of phenylacetic acid
at room temperature. After 12 h of reaction, the dicyclohexyl-
urea was removed by filtration, and the filtrate was washed
twice with water and saturated NaCl and then dried over Na2-
SO4. The solvent was removed in vacuo to give the crude
product, which was purified by MPLC to give 0.96 g (80% yield)
of the ester as a colorless oil. To a solution of 0.96 g (3.1 mmol)
of the ester in 10 mL of CH2Cl2 was added 1.0 g (4.5 mmol) of
PCC at room temperature. After 12 h of reaction, 10 mL of
ether and 3 g of Florisil were added followed by suction
filtration. The filtrate was washed by saturated NaCl and then
dried over Na2SO4. The solvent was removed in vacuo to give
the crude product, which was purified by MPLC to give the
ester as a colorless oil. The spectral data were as follows: 1H
NMR (CDCl3) δ 1.19 (d, J ) 6.6 Hz, 6 H), 1.40 (d, J ) 6.9 Hz,
6 H), 3.49 (sept, J ) 6.6 Hz, 1 H), 3.74 (s, 2H), 3.90 (sept, J )
6.6 Hz, 1 H), 4.96 (s, 2 H), 7.30 (m, 5H); 13C NMR (CDCl3) δ
Preparation of N-BOC-Glutamate, 3-(Diethylamino)-
2,3-dioxopropyl Ester 1a (LG ) BocGlu). The same method
used to prepare N,N-diethyl R-keto amide 1a (LG ) BocGABA)
was used to prepare 1.3 g (60% yield) of R-keto amide 1a (LG
) BocGlu) as an oil, starting from 2.2 g (5.0 mmol) of N-BOC-
glutamate, 3-(diethylamino)-2-hydroxy-3-oxopropyl ester 11a
(LG ) BocGlu). The spectral data were as follows: 1H NMR
(CDCl3) δ 1.13 (t, J ) 7.2 Hz, 3 H), 1.20 (t, J ) 7.2 Hz, 3 H),
1.38 (s, 9 H), 1.41 (s, 9 H), 1.90 (m, 1 H), 2.12 (m, 1 H), 2.46 (t,
J ) 7.2 Hz, 1 H), 2.50 (t, J ) 7.2 Hz, 1 H), 3.32 (q, J ) 7.2 Hz,
2 H), 3.36 (q, J ) 7.2 Hz, 2 H), 4.16 (m, 2 H), 4.98 (s, 2 H); 13
C
NMR (CDCl3) δ 15.2, 15.4, 28.8, 29.2, 29.3, 30.7, 40.8, 40, 42.9,
53.9, 67.4, 80.1, 82.6, 155.1, 163.5, 170.8, 171.9, 192.2. Anal.
Calcd for C21H36N2O8: C, 56.74; H, 8.16; N, 6.30. Found: C,
56.29; H, 7.93; N, 6.23.
General Procedure for Deprotection of 1a (LG )
BocGABA, BocGlu). To an ice-cooled, stirred solution of 0.50
mmol of 1a (LG ) BocGABA, BocGlu) in 10 mL of CH2Cl2
under argon was added dropwise a solution of 1.5 mmol of TFA
in 5 mL of CH2Cl2. After 30 min, the reaction mixture was
allowed to warm to room temperature and was stirred for 24
h, followed by concentration in vacuo. The crude trifluoro-
acetate salt was dissolved in deionized water and chromato-
graphed on Sephadex LH-20, eluting with deionized water. For
the GABA derivative 1a (LG ) GABA), only a single chro-
matographic peak was detected by absorption at 254 nm. For
the glutamate derivative 1a (LG ) Glu), only the early portion
of the peak was collected, omitting a later eluting shoulder.
The trifluoroacetate salts 1a (LG ) GABA, Glu) were obtained
after lyophilization as hygroscopic, colorless powders in 70 and
68% yield, respectively. To reduce the CF3CO2H content of 1a
(LG ) Glu), 1 equiv of triethylamine was added to a suspension
of the salt in EtOAc, followed by filtration of the solid product;
triethylammonium trifluoroacetate was present in the filtrate,
and negligible 1a (LG ) Glu) was found in the EtOAc. The
4440 J. Org. Chem., Vol. 70, No. 11, 2005

Photochemical Cleavage and Release of Carboxylic Acids
20.3, 20.5, 40.8, 46.4, 50.0, 67.0, 127.3, 128.8, 129.5, 133.3,
195.0, 171.3, 193.1. Anal. Calcd for C17H23NO4: C, 66.86; H,
7.59; N, 4.59. Found: C, 67.11; H, 7.74; N, 4.60.
of ether and 2 g of Florisil (60-100 mesh) were added, followed
by suction filtration. The filtrate was washed with 5% NaHCO3
and saturated NaCl and dried over Na2SO4. After concentra-
tion in vacuo, the residue was purified by MPLC to give 0.75
g (75% yield). The spectral data for the R-keto amide product
were as follows (signals in parentheses due to a second
conformer): 1 H NMR (CDCl3) δ 1.08 (t, J ) 6.9 Hz, 3H), 3.62
(s, 2H), 3.75 (q, J ) 6.9 Hz, 2H), 4.90 (s, 2H) (5.08 (s, 2H)),
7.02 (m, 2H), 7.30 (m, 8H); 13C NMR (CDCl3) δ 12.9, 41.0, 44.5,
67.1, 127.6 (127.4), 128.7 (128.6), 129.7 (129.4), 133.4, 139.4,
163.8, 170.3, 192.5; MS (m/z) 311 (2), 163 (71), 148 (91), 120
(100), 105 (72), 77 (59), 51 (12). Anal. Calcd for C19H19NO4:
C, 70.14; H, 5.89; N, 4.30. Found: C, 69.90; H, 6.01; N, 4.41.
Preparative Photolysis of r-Keto Amide 1a (LG )
PhCO2-) in Aqueous CH3CN. A solution of 1.0 g (3.8 mmol)
of 1a (LG ) PhCO2-) in 25 mL of 50% aqueous acetonitrile in
a quartz tube was flushed with N2 and irradiated through a
Pyrex filter with a water-jacketed Hanovia 450 W medium-
pressure mercury lamp for 3 h. The sample was at room
temperature during photolysis. NMR analysis of an aliquot
showed 94% conversion. The photolysate was extracted three
times with ethyl acetate. To further facilitate removal of
reactant, benzoic acid, and 3a, ethyl acetate was added to the
aqueous phase, and the mixture was frozen in dry ice over-
night. After thawing, the organic phase was removed, and the
addition of ethyl acetate and freezing was repeated two more
times. The final aqueous solution, after separation of ethyl
acetate, was lyophilized to give the diastereomeric hemiacetals
2a. A small amount of oxazolidinone 3a was isolated from the
concentrated organic extracts by MPLC, eluting with 50%
ethyl acetate in hexane.
The spectral data for the major diastereomer of hemiacetals
2a were as follows: 1H NMR (CDCl3) δ 1.11 (t, J ) 7.2 Hz, 3
H), 1.35 (d, J ) 5.5 Hz, 3 H), 1.52 (s, 3 H), 3.11 (dq, J ) 14.0,
7.2 Hz, 1 H), 3.45 (dq, J ) 14.0, 7.2 Hz, 1 H), 5.30 (q, J ) 5.5
Hz, 1 H); 13C NMR (CDCl3) δ 12.7, 20.1, 23.5, 34.6, 83.6, 98.4,
169.4. The spectral data for the minor diastereomer of hemi-
acetals 2a were as follows: 1H NMR (CDCl3) δ 1.11 (t, J ) 7.2
Hz, 3 H), 1.44 (d, J ) 5.5 Hz, 3 H), 1.47 (s, 3 H), 3.09 (dq, J )
14.0, 7.2 Hz, 1 H), 3.51 (dq, J ) 14.0, 7.2 Hz, 1 H), 5.10 (q, J
) 5.5 Hz, 1 H); 13C NMR (CDCl3) δ 12.6, 21.3, 22.8, 34.8, 83.9,
99.0, 169.0. Anal. Calcd for diastereomeric mixture of hemi-
acetals C7H13NO3‚0.26 H2O: C, 50.76; H, 8.35; N, 8.46.
Found: C, 50.76; H, 8.28; N, 8.66.
The spectral data for oxazolidinone 3a were as follows: 1H
NMR (CDCl3) δ 1.20 (t, J ) 7.2 Hz, 3 H), 1.49 (d, J ) 5.4 Hz,
3 H), 3.21 (dq, J ) 14.0, 7.2 Hz, 1 H), 3.69 (dq, J ) 14.0, 7.2
Hz, 1 H), 4.56 (d, J ) 2.4 Hz, 1 H), 4.90 (d, J ) 2.4 Hz, 1 H),
5.44 (q, J ) 5.4 Hz, 1 H); 13C NMR (CDCl3) δ 13.2, 21.0, 35.2,
86.2, 86.9, 150.8, 160.7. The compound was too volatile for
successful elemental analysis.
Preparation of N-BOC-Alanine, 3-(Diisopropylamino)-
2,3-dioxopropyl Ester 1b (LG ) BocAla). To a solution of
3.3 g (18 mmol) of N,N-diisopropyl-2,3-dihydroxypropanamide
and a catalytic amount of DMAP was added 2.6 g (14 mmol)
of 2-(tert-butoxycarbonylamino)propionic acid (N-BOC-ala).
After the carboxylic acid dissolved, 14 mL (1.0 M in CH2Cl2)
of DCC was added under nitrogen over 20 min, and the
reaction was stirred for 24 h at room temperature. The
dicyclohexylurea was removed by suction filtration, and the
filtrate was washed twice with water (10 mL), 5% NaHCO3,
and saturated NaCl, dried over Na2SO4, and concentrated in
vacuo to give 3.4 g of crude N-BOC-alanine 3-(diethylamino)-
2-hydroxy-3-oxopropyl ester 11b (LG ) BocAla) as a colorless
oil. Without further purification, a solution of 3.4 g (9.4 mmol)
of the oil in 30 mL of CH2Cl2 was stirred while adding 2.9 g
(14 mmol) of PCC at room temperature. After 10 h of reaction,
10 mL of ether and 3 g of Florisil (60-100 mesh) were added,
followed by suction filtration. The filtrate was washed with
5% NaHCO3 and saturated NaCl and dried over Na2SO4. After
concentration in vacuo, the residue was purified by MPLC to
give 2.8 g (59% yield) of crystalline product, which was
crystallized from ethyl acetate in hexane to give material with
mp 88-89 °C. The spectral data were as follows: 1H NMR
(CDCl3) δ 1.23 (d, J ) 6.6 Hz, 3H), 1.22 (d, J ) 6.6 Hz, 3H),
1.43 (d, J ) 6.6 Hz, 3H), 1.44 (d, J ) 6.6 Hz, 3H), 1.44 (S, (H),
1.48 (d, J ) 6.6 Hz, 3H), 3.53 (sept, J ) 6.6 Hz, 1H), 3.94 (sept,
J ) 6.6 Hz, 1H), 4.43(m, 1H), 4.97 (d, J ) 17.1 Hz, 3 H), 5.08
(d, J ) 17.1 Hz, 3H); 13C NMR (CDCl3) δ 19.0, 20.4, 21.0, 28.6,
46.6, 49.3, 50.0, 67.0, 80.2, 155.1, 164.8, 173.1, 192.7; MS (m/
z) 358 (M+, 0), 274 (4), 170 (8), 128 (82), 100 (13), 86 (100), 57
(27). Anal. Calcd for C17H30N2O6: C, 56.97; H, 8.44; N, 7.82.
Found: C, 57.07; H, 8.41; N, 7.78.
Preparation of Benzoic Acid, N-Ethyl-N-phenyl-2,3-
dioxopropanamide 1c (LG ) PhCO2-). To a solution of 3.18
g (15.2 mmol) of N-ethyl-N-phenyl-2,3-dihydroxypropanamide
10c in 50 mL of CH2Cl2 were added a catalytic amount DMAP
and 15 mL of (15 mmol) of 1 M DCC in CH2Cl2. While stirring
under N2, a solution of 1.8 g (15 mmol) of benzoic acid in 20
mL of CH2Cl2 was added, followed by 12 h of reaction. The
dicyclohexylurea was filtered, and the filtrate was concentrated
to obtain crude product. Without further purification, the crude
product was dissolved in 40 mL of CH2Cl2, and 4.6 g (22 mmol)
of PCC was added. The reaction was stirred for 6 h, and then
100 mL of ether and 5 g of Florisil (60-100 mesh) were added,
followed by suction filtration. The filtrate was washed with
5% NaHCO3 and saturated NaCl and dried over Na2SO4. After
concentration in vacuo, the residue was purified by MPLC to
give 2.8 g (60% yield), based on starting amide 10c. Crystal-
lization from EtOAc in hexane gave colorless crystals, mp 86-
89 °C. The spectral data for the R-keto amide product were as
Preparative Photolysis of r-Keto Amide 1b (LG )
PhCO2-) in Aqueous CH3CN. A solution of 0.50 g (1.7 mmol)
of 1b (LG ) PhCO2-) in 25 mL of 50% aqueous acetonitrile in
a quartz tube was flushed with N2 and irradiated through a
Pyrex filter with a water-jacketed Hanovia 450 W medium-
pressure mercury lamp for 3 h. The sample was at room
temperature during photolysis. After extraction with ethyl
acetate three times, the combined extracts were washed with
5% NaHCO3 and saturated NaCl and dried over MgSO4, and
the solvent was removed in vacuo to give an oil. The oil was
chromatographed (MPLC), eluting with 25% EtOAc in hexane,
to give starting material 1b (LG ) PhCO2-), which eluted first,
and a colorless crystalline product 3b, mp 47-48 °C. The
spectral data were as follows: 1H NMR (CDCl3) δ 1.50 (d, J )
6.6 Hz, 6 H), 1.53 (s, 6 H), 3.47 (sept, J ) 6.6 Hz, 1H), 4.46 (d,
J ) 2.1 Hz, 1 H), 4.83 (d, J ) 2.1 Hz, 1 H); 13C NMR (CDCl3)
δ 20.7, 27.1, 46.5, 85.0, 95.3, 150.2, 159.7. Anal. Calcd for
oxazolidinone 3b C9H15NO2: C, 63.88; H, 8.93; N, 8.28.
Found: C, 63.60; H, 9.08; N, 8.05.
1
follows: H NMR (CDCl3) δ 1.19 (t, J ) 6.9 Hz, 3H), 3.87 (q,
J ) 6.9 Hz, 2H), 5.15 (S, 2H), 7.43 (m, 2H), 7.48 (m, 5H) 7.60
(m, 1H), 8.05 (m, 2H); 13C NMR (CDCl3) δ 12.9, 44.6, 67.2,
126.69 (127.63), 128.6, 129.6, 129.1, 130.0, 133.6, 139.5, 163.9,
165.4, 192.6; MS (m/z) 311 (M+, 1), 163 (60), 148 (84), 120 (100),
105 (78), 77 (68). Anal. Calcd for C18H17NO4: C, 69.44; H, 5.50;
N, 4.50. Found: C, 69.47; H, 5.62; N, 4.68.
Preparation of Phenylacetic Acid, N-Ethyl-N-phenyl-
2,3-dioxopropanamide 1c (LG ) PhCH2CO2-). To a solu-
tion of 0.64 g (3.0 mmol) of N-ethyl-N-phenyl-2,3-dihydroxy-
propanamide 10c in 20 mL of CH2Cl2 were added a catalytic
amount of DMAP and 2.5 mL (2.5 mmol) of 1 M DCC in CH2-
Cl2. While stirring, a solution of 0.34 g (2.5 mmol) of phenyl-
acetic acid in 5 mL of CH2Cl2 was added under nitrogen. After
12 h of reaction, the dicyclohexylurea was removed by filtra-
tion, and the filtrate was concentrated to obtain the crude
product. Without further purification, the crude product was
dissolved in 20 mL of CH2Cl2, and 0.74 g (3.4 mmol) of PCC
was added. The reaction was stirred for 6 h, and then 100 mL
Preparative of Photolysis of Acetic Acid and Phenyl-
acetic Acid Derivatives of r-Keto Amide 1a (LG )
J. Org. Chem, Vol. 70, No. 11, 2005 4441

Ma et al.
CH3CO2-, PhCH2CO2-) in CD3CN. A solution of 0.30 g (1.1
mmol) of 1a (LG ) PhCH2CO2-) in CD3CN in an NMR tube
was purged with N2 and then irradiated through a Pyrex filter
with a water-jacketed Hanovia 450 W medium-pressure
mercury lamp for 75 min. The temperature was 26 °C during
photolysis. The NMR sample was directly applied to the MPLC
column and eluted with 50% ethyl acetate in hexane to give
unreacted 1a (LG ) PhCH2CO2-), which eluted first, followed
by phenylacetate ester 12a (LG ) PhCH2CO2-) and a minor
General Procedure for Product Quantum Yield De-
terminations. The semimicrooptical bench for quantum yield
determinations is similar to the apparatus described by
Zimmerman.21 Light from a 200 W high-pressure mercury
lamp was passed through an Oriel monochromator, which was
set to 310 nm wavelength, and 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 actinometer.
The photolysate 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 acti-
nometer. Light output was monitored by ferrioxalate actinom-
etry22 using the splitting ratio technique. The samples were
not purged with nitrogen.
General Procedure for Photolyses of Carboxylic Acid
Derivatives of r-Keto Amides 1a,b in D2O Solutions and
Controls on Hydrolytic Stability of 1a,b in the Dark.
Samples of 0.07-0.16 mmol of R-keto amides 1a,b in 1 mL of
D2O, or in 0.25 mM phosphate ion buffer in D2O (pD 5.4), or
50% D2O in CD3CN were contained in NMR tubes. The
temperature throughout the photolyses was 26 °C. Samples
were mounted beside a water-jacketed 450 W medium-pres-
sure mercury lamp equipped with a Pyrex filter. Samples were
also kept in the dark to determine stability at room temper-
ature. The results in Tables 1-3 generally required 0.5-1 h
photolysis times. Yields were determined by 1H NMR spec-
troscopy using DMSO as the standard, except for 1a (LG )
Glu), which used DMF, and 1a (LG ) GABA), which used
glycine.
1
amount of oxazolidinone 3a. The H NMR spectrum 12a (LG
) PhCH2CO2-) showed ca. 10% of a minor C-N conformer,
which exhibited a >CHO2CCH2Ph methine quartet at δ 6.8.
The spectral data for 12a (LG ) PhCH2CO2-) were as follows:
1H NMR (CDCl3) δ 1.11 (t, J ) 7.2 Hz, 3 H), 1.54 (d, J ) 6.0
Hz, 3 H), 2.48 (s, 3 H), 3.19 (dq, J ) 14.0, 7.2 Hz, 1 H), 3.34
(dq, J ) 14.0, 7.2 Hz, 1 H), 3.58 (s, 2 H), 6.63 (q, J ) 6.0 Hz,
1 H), 7.1-7.4 (m, 5 H); 13C NMR (CDCl3) δ 14.8, 19.2, 27.5,
27.5, 35.7, 41.9, 78.8, 127.1, 128.4, 128.8, 132.7, 166.9, 170.3,
197.4. Anal. Calcd for C15H19NO4: C, 64.97; H, 6.91; N, 5.05.
Found: C, 64.97; H, 6.92; N, 5.16.
In the case of 1a (LG ) CH3CO2-), the same mmol amount
of starting material was irradiated, and the mixture was
separated by MPLC to give starting material 1a (LG )
CH3CO2-) as the first eluting peak followed by 12a (LG )
CH3CO2-), which was obtained as an oil. The spectral data
for 12a (LG ) CH3CO2-) were as follows: 1H NMR (CDCl3) δ
1.11 (t, J ) 7.2 Hz, 3 H), 1.55 (d, J ) 6.0 Hz, 3 H), 2.04 (s, 3
H), 2.51(s, 3 H), 3.32 (dq, J ) 14.0, 7.2 Hz, 1 H), 3.48 (dq, J )
14.0, 7.2 Hz, 1 H), 6.32 (q, J ) 6.0 Hz, 1 H); 13C NMR (CDCl3)
δ 14.9, 19.2, 21.7, 27.5, 35.8, 78.4, 167.0, 169.9, 197.4. Anal.
Calcd for C9H15NO4: C, 53.72; H, 7.51; N, 6.96. Found: C,
53.91; H, 7.38; N, 7.23.
Preparative Photolysis of r-Keto Amide 1c (LG )
PhCH2CO2-) in Aqueous CH3CN. A solution of 1c (LG )
PhCH2CO2-) in 25 mL of 30% aqueous CH3CN was photolyzed
for 48 h. The sample was at room temperature during
photolysis. The light brown photolysate was extracted by
EtOAc and concentrated in vacuo. The crude colorless solid
was chromatographed (MPLC), eluting with 25% EtOAc in
hexane to obtain a product, which was crystallized from ether
in EtOAc to obtain diasteromeric hemiacetals 2c, mp 122-
124 °C. The spectral data for the major diastereomer were as
follows: 1H NMR (CDCl3) δ 1.24 (d, J ) 5.4 Hz, 3H), 1.55 (s,
3H), 4.2 (br, 1H), 5.71 (q, J ) 5.4 Hz, 1H), 7.26 (m, 5H); 13C
NMR (CDCl3) δ 20.7, 24.3, 85.6, 99.1, 123.9, 127.0, 129.4, 134.9,
168.9. The spectral data for the minor diastereomer were as
follows: 1H NMR (CDCl3) δ 1.35 (d, J ) 5.4 Hz, 3H), 1.49 (s,
3H), 4.2 (br, 1H), 5.58 (q, J ) 5.4 Hz, 1H), 7.11 (m, 5H); 13C
NMR (CDCl3) δ 22.0, 23.7, 86.3, 99.8, 123.2, 127.0, 129.5, 134.9,
168.5. Anal. Calcd for the mixture of diastereomers C11H13-
NO3: C, 63.74; H, 6.32; N, 6.76. Found: C, 63.57; H, 6.23; N,
6.80.
Acknowledgment. Acknowledgment is made to the
donors of the Petroleum Research Fund, administered
by the American Chemical Society (M.G.S.), and to the
National Science Foundation for a Career Award (R.R.)
for support of this research.
Supporting Information Available: ∆OD vs time plots
-
for bleaching of bromocresol green by 1a (LG ) CH3CO2
,
BocAla, 4-NCC6H4CO2-) and by 1b (CH3CO2-, PhCO2-, Boc-
Ala); 1H and 13C NMR spectral data for 3a and 10c; and
procedures for preparation of 9a,b, 10a-c, and 11b (LG )
PhCO2-). This material is available free of charge via the
Internet at http://pubs.acs.org.
JO050246S
(21) Zimmerman, H. E. Mol. Photochem. 1971, 3, 281-92.
(22) Hatchard, C. G.; Parker, C. A. Proc. R. Soc. London 1956, 235,
518-36.
4442 J. Org. Chem., Vol. 70, No. 11, 2005