Photorelease of carboxylic acids from 1-acyl-7-nitroindolines in aqueous solution: Rapid and efficient photorelease of L-glutamate [4]

Full text of DOI:10.1021/ja990931e

J. Am. Chem. Soc. 1999, 121, 6503-6504
6503
Scheme 1a
Photorelease of Carboxylic Acids from
1-Acyl-7-nitroindolines in Aqueous Solution: Rapid
and Efficient Photorelease of L-Glutamate1
George Papageorgiou,† David C. Ogden,† Andreas Barth,‡ and
John E. T. Corrie*,†
National Institute for Medical Research, The Ridgeway
Mill Hill, London NW7 1AA, UK
Institut fu¨r Biophysik, Johann Wolfgang Goethe UniVersita¨t
D-60590 Frankfurt, Germany
ReceiVed March 22, 1999
ReVised Manuscript ReceiVed April 30, 1999
Photorelease of biologically active compounds from photo-
cleavable (caged) precursors is a useful tool to study biological
processes2 but rapid, efficient release of neuroactive amino acids
has been elusive. We and others3 have approached the problem
with various photolabile protecting groups. Among the better rea-
gents are p-hydroxyphenacyl and 2,2′-dinitrobenzhydryl esters of
Givens4a and Hess,4b respectively. A recent report5 describes pho-
torelease from 7-hydroxycoumarin-4-ylmethyl carbamates, but
these are rate-limited by decarboxylation of the carbamate salt (k ≈
150 s-1, pH 7, 21 °C).6 We now describe stable 1-acyl-7-nitro-
indolines that rapidly and efficiently photorelease carboxylates,
including L-glutamate, in neutral aqueous solution. Related rea-
gents undergo clean photolysis in dioxane-CH2Cl2 with ∼1%
water to yield a carboxylic acid and nitroindoline.7a Photosol-
volysis by the water was shown, but no reaction mechanism was
given.7 We found the reaction takes a different course in aqueous
solution.
a Reagents: (a) Tl(NO3)3-MeOH-HClO4; (b) HCl-H2O-MeOH; (c)
glutaric anhydride; (d) L-HO2C(CH2)2CH(NHBoc)CO2But-EDC-DMAP;
(e) BH3-THF; (f) Et2NP(OBut)2-1H-tetrazole, then MCPBA; (g)
NaNO3-Tfa.
than 8, confirming a deleterious effect of the bromo substituent.
Clean photolysis of 8 was shown by an isosbestic point at 365
nm in spectra of a solution photolyzed for increasing times (e65%
conversion). The photoproduct (λmax 412 nm) of the protecting
In compounds 8-10 a 5-substituent ensured nitration at C-7.
Previous work7 used 5-bromo compounds, but a heavy atom might
lower the photolysis efficiency. We used a CH2CO2Me group that
was also expected to enhance aqueous solubility. Indoline 3 was
prepared by Tl(NO3)3 oxidation8 of 1,5-diacetylindoline 1 and
acidic methanolysis of 2 (Scheme 1). Acylation and further
transformation of the introduced acyl group as required (4 f 6)
followed by nitration9 concurrently removed tert-butyl protecting
groups (in 6 and 7) to give 8-10. The route avoids a difficult
acylation of 7-nitroindolines,7b especially with sensitive side chains
as in glutamic acid. 5-Bromo compound 11 was prepared by a
related route from 1-acetyl-5-bromoindoline.
group was not the 7-nitroindoline 12 (λmax 450 nm) but the
7-nitrosoindole 1310 (Scheme 2). Reaction with a thiol abolished
the 412 nm chromophore, presumably by converting the nitroso
to the corresponding hydroxylamine.12,13 Quantitative amino acid
analysis of part-photolyzed solutions of 10 ((dithiothreitol to react
with released 13) and HPLC assay of starting material consump-
tion showed glutamate release at 1:1 stoichiometry. Photolysis
of 8 in CH2Cl2-dioxane-H2O (2:3:0.05) cleanly gave nitroin-
doline 12, confirming previous data.7a Photolysis in the organic
solvent was at least 2-fold more efficient than in water.
Formation of different products implies a changed mechanism.
Different photoreactivity of nitroaryl compounds in aqueous or
organic solvents has been explained by formation of a highly
polarized π,π* triplet state in water and an n,π* state in organic
solvents.14 We performed some mechanistic studies, but full details
of the reaction mechanism will require additional work. First, flash
photolysis of 9 (that has no carboxylate group to interfere with
measurements on the photolytically released carboxylate) coupled
with FTIR difference spectroscopy13 showed antisymmetric stretch
of the released carboxylate at 1553 cm-1 in normal or [18O]water
(97% isotopic abundance). Thus, the oxygen atom introduced to
form the carboxylate is not from solvent. Second, we measured
product formation kinetics, specifically of the released proton and
carboxylate. Proton release was studied by flash photolysis (320
nm laser, 1 µs pulse, 20 °C) coupled with time-resolved absorption
spectroscopy (615 nm) of a pH 7 solution containing 8 and a pH
indicator (bromothymol blue), as described for other examples.15
Photolysis of 8 and 11 in neutral aqueous solution (without
excluding O2) showed 11 was converted ∼2.5-fold less efficiently
† National Institute for Medical Research.
‡ Johann Wolfgang Goethe Universita¨t.
(1) Patent Appl. GB 99/06192.1, filed 18 March 1999 covers this work.
(2) Corrie, J. E. T.; Trentham, D. R. In Bioorganic Photochemistry;
Morrison, H., Ed.; Wiley: New York, 1993; Vol. 2, pp 243-305.
(3) See refs 3-7 of Papageorgiou, G.; Corrie, J. E. T. Tetrahedron 1999,
55, 237 for a comprehensive list.
(4) (a) Givens, R. S.; Jung, A.; Park, C. H.; Weber, J.; Bartlett, W. J. Am.
Chem. Soc. 1997, 119, 8369. (b) Gee, K. R.; Niu, L.; Schaper, K.; Jayaraman,
V.; Hess, G. P. Biochemistry 1999, 38, 3140.
(5) Furuta, T.; Wang, S. S.; Dantzker, J. L.; Dore, T. M.; Bybee, W. J.;
Callaway, E. M.; Denk, W.; Tsien, R. Y. Proc. Natl. Acad. Sci. U.S.A. 1999,
96, 1193.
(6) Papageorgiou, G.; Corrie, J. E. T. Tetrahedron 1997, 53, 3917.
(7) (a) Amit, B.; Ben-Efraim, D. A.; Patchornik, A. J. Am. Chem. Soc.
1976, 98, 843. (b) Pass, S.; Amit, B.; Patchornik, A. J. Am. Chem. Soc. 1981,
103, 7674.
(8) McKillop, A.; Swann, B. P.; Taylor, E. C. J. Am. Chem. Soc. 1971,
93, 4920.
(9) Mortensen, M. B.; Kamounah, F. S.; Christensen, J. B. Org. Prep. Proc.
Int. 1996, 28, 123.
(10) 7-Nitrosoindole itself has been described from an unrelated photolysis
of a 7-nitroindole derivative.11
(14) (a) Wan, P.; Yates, K. Can. J. Chem. 1986, 64, 2076. (b) Wan, P.;
Muralidharan, S. J. Am. Chem. Soc. 1988, 110, 4336.
(15) (a) Walker, J. W.; Reid, G. P.; McCray, J. A.; Trentham, D. R. J. Am.
Chem. Soc. 1988, 110, 7170. (b) Corrie, J. E. T.; De Santis, A.; Katayama,
Y.; Khodakhah, K.; Messenger, J. B.; Ogden, D. C.; Trentham, D. R. J.
Physiol. 1993, 465, 1.
(11) Kotera, M.; Bourdat, A. G.; Defrancq, E.; Lhomme, J. J. Am. Chem.
Soc. 1998, 120, 11810.
(12) Zuman, P.; Shah, B. Chem. ReV. 1994, 94, 1621 and references therein.
(13) Barth, A.; Corrie, J. E. T.; Gradwell, M. J.; Maeda, Y.; Ma¨ntele, W.;
Meier, T.; Trentham, D. R. J. Am. Chem. Soc. 1997, 119, 4149.
10.1021/ja990931e CCC: $18.00 © 1999 American Chemical Society
Published on Web 06/24/1999

6504 J. Am. Chem. Soc., Vol. 121, No. 27, 1999
Communications to the Editor
The absorbance change of the indicator revealed biphasic acidi-
fication of the solution. The major signal (∼85%) was complete
within the laser pulse (k > 105 s-1), and the minor component
had k ≈ 9 s-1. Time-resolved IR measurements13 at 1557 cm-1
upon flash photolysis of 9 (351 nm laser, 9 ns pulse, spectral
resolution 25 cm-1, 25 °C, pH 7) to monitor carboxylate formation
showed “instantaneous” decrease in absorption from the nearby
disappearing nitro absorption (1539 cm-1), followed by biphasic
absorption increase with major (∼80%) and minor components
at k ≈ 2700 and ∼10 s-1, respectively. The latter process may
correspond with the slow phase of proton release. An overlapping
band on the high-frequency side of the 1553 cm-1 band makes
complete interpretation uncertain, but the demonstrated release
of carboxylates and initial biological tests (see below) suggest
the faster process principally represents carboxylate release (t1/2
≈ 0.26 ms). Future work will investigate if the slower process
also forms carboxylate. Scheme 2 shows a provisional mechanism
to accommodate most of the data. Oxygen transfer from the nitro
to the acyl group has been shown for photolysis of related 1-acyl-
8-nitrotetrahydroquinolines in various organic solvents, but the
byproduct of the heterocycle was not characterized.16 A feature
of the scheme is that H+ loss precedes release of the carboxylate,
as required by the kinetic data. The mechanism does not consider
the slower, minor processes seen for the H+ and carboxylate
signals. Further experiments are needed to substantiate the
proposal, account for the minor signals, and understand the
different photochemistry in organic solvent.
Figure 1. Current response of a cultured rat cerebellar granule neuron
to photolytic release of L-glutamate from 10. The arrow marks the start
of the 1-ms flash from a xenon arc flash lamp21,22 through a Schott UG11
filter and focused to a 200 µm diameter spot at the preparation.23
L-Glutamate precursor 10 was tested in primary cultures of rat
cerebellar granule neurons for its pharmacological properties and
ability to activate glutamate ion channels upon photolysis. Cells
were voltage-clamped at -65 mV with whole-cell patch clamp,19,20
and glutamate sensitivity was established by responses to 20-ms
pulses of L-glutamate, applied at 0.25 Hz by iontophoresis from
a 1-µm tipped glass pipet. Replacing the bath solution with one
containing 10 (1 mM) caused no activation of current nor
diminished the peak response to iontophoretic pulses of L-
glutamate. Furthermore, when phosphate 9 was used instead of
the glutamate precursor 10 ((iontophoretic application of L-
glutamate), it showed no effect before, during, or after photolysis.
The control experiments during and after photolysis confirm that
byproduct 13 has no adverse effect on glutamate receptors or
resting membrane properties of granule neurons. Flash photolysis
of 10 resulted in a fast rise of inward current (Figure 1) with t1/2
0.74 ms (SD 0.13 ms, n ) 8), followed by a decline comprising
fast (mean t1/2 2.9 ms) and slow time courses. The latter likely
arises from glutamate diffusion away from the photolyzed region.
The time course of initial decline may represent desensitization
of non-NMDA glutamate channels. During the slow decline there
was an increase in low-frequency noise of the glutamate-activated
current, consistent with gating of glutamate-activated ion channels.
Our data indicate that 10 is inactive as an agonist or antagonist
at glutamate ion channels of both NMDA- and non-NMDA-
activated types, and that release of L-glutamate by flash photolysis
of 10 is on the same time scale as the light pulse.
To conclude, electrophysiological characterization indicates that
10 is a promising reagent for rapid photorelease of L-glutamate.
The photochemistry is markedly solvent-dependent, with forma-
tion of the nitrosoindole 13 as the principal byproduct in aqueous
solution. Future detailed mechanistic studies will define more
precisely the time course of the photorelease. We are extending
the range of amino acids derivatized with the nitroindoline.
Acknowledgment. We thank Dr. M. J. Gradwell for help with
structural assignment of 13 and the MRC Biomedical NMR Centre for
facilities, Dr. V. R. N. Munasinghe for recording NMR spectra, Drs. S.A.
Howell and K. Welham for low- and high-resolution MS, respectively,
and Dr. P. Wan for comments on the manuscript. The research was
supported by the Medical Research Council under the NAHH Initiative.
Scheme 2. Possible Photolysis Mechanism for
1-Acyl-7-nitroindolines in Aqueous Solution
As well as release rates of bioeffector species upon flash irradi-
ation, important properties of caged compounds include stability
in aqueous solution and extent of photolysis by a single flash.
The amide bond of 1-acylnitroindolines was stable at pH 7; at
pH 12, 30 °C, half-times for amide hydrolysis of 9 and 10 were
29 and 6 h, respectively.17 Faster hydrolysis of the glutamate
derivative 10 indicated that its free amino group competes with
external base but hydrolysis can be disregarded during purifica-
tion, storage, and use of these compounds near neutral pH. The
product quantum yield (Qp) for 10 was determined by laser flash
irradiation (347 nm) of a solution containing 10 and 1-(2-
nitrophenyl)ethyl phosphate.18 Conversions by a single flash (∼90
mJ) were 9.7 and 23%, respectively (reverse-phase HPLC),
leading to Qp ) 0.043 by comparison with the known value18
(0.54) for the caged phosphate and after correction for the different
extinction coefficients.
Supporting Information Available: Synthetic details, hydrolysis and
photolysis studies (PDF). This material is available free of charge via
the Internet at http://pubs.acs.org.
JA990931E
(16) Amit, B.; Ben-Efraim, D. A.; Patchornik, A. J. Chem. Soc., Perkin
Trans. 1 1976, 57.
(17) Hydrolytic stability is of the amide bond. In all of these compounds,
the ester group in the side chain hydrolyzed slowly at pH 7, but in separate
experiments 10 and the salt of its corresponding free acid (isolated as a minor
byproduct) were found to photorelease glutamate with equal efficiency. Ester
hydrolysis is thus unimportant. The stabilities quoted are from experiments
on the free acids of 9 and 10 and were determined from reverse-phase HPLC
measurements of disappearance of starting compounds. The free acids avoided
analytical complications from fast cleavage of the ester group.
(19) All electrophysiological experiments were in Mg2+-free mammalian
Ringer’s solution containing 100 µM glycine.
(20) Hamill, O. P.; Marty, A.; Neher, E.; Sakmann, B.; Sigworth, F. L.
Pflu¨gers ArchiV 1981, 391, 85.
(21) Rapp, G.; Gu¨th, K. Pflu¨gers ArchiV 1985, 411, 200.
(22) Full lamp output converted 40% of caged ATP in a 1-ms pulse. From
relative Qp values and extinction coefficients, ∼15% conversion of the 1 mM
solution of 10 is expected.
(18) Kaplan, J. H.; Forbush, B.; Hoffman, J. F. Biochemistry 1978, 17, 1929.
(23) Khodakhah, K.; Ogden, D. J. Physiol. 1995, 487, 343.