Photochemistry and thermal decarboxylation of α-phosphoryloxy-p- nitrophenylacetates

Article abstract of DOI:10.1111/j.1751-1097.2009.00576.x

α-Carboxy-4-nitrobenzyl phosphate 4 and its derived monomethyl phosphate ester 5 were synthesized and purified by anion-exchange chromatography. A gradient of LiCl was necessary for elution of the anion-exchange column to avoid unexpected thermal decarbox

Full text of DOI:10.1111/j.1751-1097.2009.00576.x

Photochemistry and Photobiology, 2009, 85: 1089–1096
Photochemistry and Thermal Decarboxylation of a-Phosphoryloxy-
p-nitrophenylacetates
John E. T. Corrie1, V. Ranjit N. Munasinghe1, Maria Rudbeck2 and Andreas Barth*2
1MRC National Institute for Medical Research, London, UK
2Department of Biochemistry and Biophysics, Arrhenius Laboratories for Natural Sciences,
Stockholm University, Stockholm, Sweden
Received 12 February 2009, accepted 29 March 2009, DOI: 10.1111 ⁄ j.1751-1097.2009.00576.x
ABSTRACT
a-Carboxy-4-nitrobenzyl phosphate
4
and its derived
monomethyl phosphate ester 5 were synthesized and purified
by anion-exchange chromatography. A gradient of LiCl was
necessary for elution of the anion-exchange column to avoid
unexpected thermal decarboxylation that occurred during
vacuum evaporation when the volatile triethylammonium bicar-
bonate buffer was used. Photolysis of each compound was
accompanied by decarboxylation, and 4 released inorganic
phosphate with near-100% stoichiometry. Time-resolved infra-
red spectroscopy of the photolysis reaction, coupled with
density functional theory calculations of vibrational frequen-
cies, enabled us to infer a mechanism for the photolytic
pathway, although there was some evidence for a second
pathway also being operative. In contrast to the results for 4,
photolysis of 5 appeared to release little or no monomethyl
phosphate.
Scheme 1. Dual pathways in photolysis of CNB-type compounds such
as 1.
occur, as photodecarboxylation of 4-nitrophenylacetate has
been reported (2) to be more efficient than for its 2-nitro
isomer, and decarboxylation was also found (3) to be the
initial step in photolysis of 4-nitromandelate 6 and its
acetate ester 7.
INTRODUCTION
We have recently described (1) aspects of the photochemistry
of a-carboxy-2-nitrobenzyl (CNB) compounds such as 1,
analogs of which have been used for rapid photorelease of
bioeffector species in time-resolved studies of biological
systems (for a bibliography of applications of these com-
pounds in biology, see Corrie et al. [1]). The photoreactivity of
1 and related compounds was shown to involve two separate
pathways, of which one proceeded by the normal 2-nitrobenzyl
photofragmentation, i.e. via the aci-nitro species 2 to release
the product (monomethyl phosphate in the case of 1) and
by-product 2-nitrosophenylglyoxylate 3. A second, photode-
carboxylative pathway of 1 did not result in significant release
of monomethyl phosphate. These processes are summarized in
Scheme 1.
Unlike the ortho-isomers 1 and 9, we found that 5 and, to a
lesser extent, its precursor 4 were unexpectedly susceptible to
thermal decarboxylation under mild conditions. The present
paper describes procedures that avoided this complication and
enabled us to obtain pure samples of both compounds. Studies
of the photochemistry of 4 and 5 are reported, using time-
resolved infrared spectroscopy and supported by a range of
additional techniques, that included density functional theory
(DFT) calculations to assist with assignments of particular
bands in the infrared spectra.
In the course of the work described above, we were
interested in investigating the photochemistry of the related
para-nitro compounds 4 and 5. Our motivation was princi-
pally to see whether the latter compounds, which cannot
undergo the conventional 2-nitrobenzyl-type photofragmen-
tation, nevertheless exhibited any useful photochemistry. We
anticipated that substantial photodecarboxylation would
*Corresponding author email: andreas.barth@dbb.su.se (Andreas Barth)
ꢀ 2009 The Authors. JournalCompilation. The AmericanSocietyofPhotobiology 0031-8655/09
1089

1090 John E. T. Corrie et al.
kept overnight at room temperature and analyzed as above by HPLC,
which showed that the earlier eluted peak had almost totally
disappeared, while the slower peak was enhanced, consistent with
hydrolysis of the carboxylate ester 11. The solution was diluted to
70 mL, adjusted to pH 7.2 and quantified by UV spectroscopy
(583 lmol). The solution was diluted to 1 mS cm)1 conductivity and
desalted by anion-exchange chromatography, using a linear gradient
formed from 10 and 140 mM LiCl (each 1000 mL). Fractions contain-
ing the product were analyzed as above by HPLC (single peak, tR
1.8 min) and quantified by UV spectroscopy (452 lmol). The com-
bined fractions were concentrated under vacuum to ꢀ9.5 mL (bath
temperature <30ꢁC), diluted with EtOH-acetone (3:7 vol ⁄ vol; 95 mL)
and allowed to stand at 4ꢁC overnight. Further processing, as
described above for 4, gave pure 5 (197 lmol; 2.7 mL, 73 mM); 1H
NMR: d 8.27 (m, 2H, 3,5-H), 7.71 (m, 2H, 2,6-H), 5.44 (d, JH,P 8.7,
1H, CH), 3.42 (d, JH,P 10.9, 3H, Me); LRMS (negative ion ES): calcd
for (C9H8NO8P + H)): 290; found: 290.
MATERIALS AND METHODS
General. 1H NMR spectra were recorded on a Varian Unityplus 500
instrument for solutions in D2O with an acetone internal reference.
J values are given in Hz. Analytical anion-exchange chromatography
was performed on a Whatman Partisphere SAX column (Cat. No.
4621-0505) with mobile phase flow rates of 1.5 mL min)1 and
detection at 254 nm. Preparative anion-exchange chromatography
was performed on
a column of DEAE-cellulose (2.5 · 40 cm).
Triethylammonium bicarbonate (TEAB) buffer was prepared by
bubbling CO2 into an ice-cold 1 M solution of redistilled triethylamine
in water until the pH stabilized at ꢀ7.4. When TEAB buffer was used,
pooled column fractions were evaporated at ꢀ1 mmHg and freed from
residual triethylamine by repeated evaporation from methanol. Lith-
ium chloride gradients for elution of preparative anion-exchange
columns were prepared from aqueous solutions of LiCl, without
adjustment of the ambient pH. Preparative reverse-phase chromatog-
raphy was performed on a 7.8 · 300 mm column packed with Waters
C18 Bulk Packing Material (Cat. No. 20594) at a flow rate of
2 mL min)1. Mobile phases for HPLC are specified at relevant
positions in the text. Aqueous solutions of 4 and 5 were quantified
using the spectrum of 4-nitromandelic acid in aqueous solution at pH
4-Nitrobenzyl methyl phosphate, 13. An aqueous solution of
4-nitrobenzyl phosphate 12 (see General) as its Et3NH+ salt
(0.33 mmol in 16.5 mL) was adjusted to pH 4.5 with 1 M HCl, cooled
in ice and stirred vigorously with an ethereal solution of diazomethane
(ꢀ0.3 M, total volume 40 mL), which was added in two aliquots. After
30 min, anion-exchange HPLC (mobile phase 20 mM Na phosphate,
pH 6.0-MeOH 100:13 [vol ⁄ vol]) of the aqueous phase indicated >90%
conversion of the starting material (tR 1.6 and 5.9 min for 13 and 12,
respectively). The aqueous layer was washed with Et2O and the pH
adjusted to 5.5. The solution was pumped onto the preparative reverse-
phase HPLC column, previously equilibrated in 10 mM sodium phos-
phate, pH 5.5, and the column was eluted with the same buffer for
ꢀ1.5 h. The mobile phase was changed to water for 1.5 h and then to
H2O-MeCN (3:2 vol ⁄ vol) for 1.5 h to elute the product. Fractions
containing 13 were concentrated under vacuum to remove MeCN,
adjusted to pH 7.2 and applied to the preparative DEAE-cellulose
column, that was eluted with a linear gradient formed from 10 and
150 mM TEAB (each 1000 mL). Fractions containing the product were
)1
7, kmax = 285 nm (e = 10 100 M cm)1) as the reference chromo-
phore. 4-Nitrobenzyl phosphate 12 was prepared from 4-nitrobenzyl
alcohol via its di-t-butyl ester and subsequent trifluoroacetic acid
treatment, as described for synthesis of other phosphates (1). Its 1H
NMR spectrum was comparable to published data (4,5). 1-(2-
Nitrophenyl)ethyl phosphate 18 was available from previous work (6).
a-Carboxy-4-nitrobenzyl phosphate, 4. A solution of ethyl a-diazo-4-
nitrophenylacetate 10 (0.235 g, 1 mmol) in anhydrous dioxan (20 mL)
was treated with concentrated phosphoric acid (2.0 mL) and the
solution was kept in the dark at room temperature for 4.5 days. The
solution was diluted with water (140 mL), adjusted to pH 5.5 with 1 M
NaOH and extracted with ether (3 · 75 mL). The aqueous solution
was lyophilized to about half its volume and loaded onto the
preparative reverse-phase HPLC column that had been equilibrated
in 50 mM sodium acetate, pH 5.5 and the column was eluted with the
same buffer for ca 1.5 h. The mobile phase was changed to water for
2 h, then to a mixture of water-acetonitrile (9:1 vol ⁄ vol). The fractions
eluted in the latter solvent were analyzed by anion-exchange HPLC
(mobile phase 100 mM Na phosphate, pH 6-MeOH [100:13 vol ⁄ vol])
and those containing the ester 11 (tR 2.6 min) were concentrated under
reduced pressure to remove acetonitrile and adjusted to pH 13 with
1 M KOH. The solution was left overnight at room temperature and
adjusted to pH 7.2 by dropwise addition of 1 M HCl to the rapidly
stirred solution. The acid was added carefully to avoid local regions of
low pH that could have promoted decarboxylation. Analysis of the
solution by anion-exchange HPLC as above showed a single peak
corresponding to 4 (tR 4.7 min). The solution was diluted with water to
1 mS cm)1 conductivity and desalted by anion-exchange chromatog-
raphy using a linear gradient formed from 10 and 200 mM LiCl (each
1000 mL). Fractions containing the product were analyzed by HPLC
(as above), quantified by UV spectroscopy (369 lmol) and concen-
trated under vacuum to a volume of ꢀ8 mL, keeping the bath
temperature below 30ꢁC. This solution was diluted with EtOH-acetone
(3:7 vol ⁄ vol; 80 mL) and allowed to stand at 4ꢁC overnight. The fine
precipitate was collected by centrifugation (1000 g) and the pellet was
washed by resuspension in ice-cold EtOH-acetone (3:7 vol ⁄ vol; 65 mL)
and further centrifugation. After decanting the supernatant, the pellet
was dried under vacuum and dissolved in water (7.0 mL). The
concentration of 4, determined by UV spectroscopy, was 51.2 mM
(36% yield); 1H NMR: d 8.25 (m, 2H, 3,5-H), 7.72 (m, 2H, 2,6-H), 5.44
(d, JH,P 9.0, 1H, CH); LRMS (negative ion ES): calcd for
(C8H5NO8P + H + Li)): 282; found: 282.
)1
quantified by UV analysis at 280 nm (e = 10100 M cm)1). The
solution of 13 (227 lmol) was evaporated in vacuo, re-evaporated three
times from methanol and the residue was dissolved in water (20 mL).
The pH was adjusted to 7.0 and the solution was freeze-dried, then
redissolved in a small volume of water and converted to the sodium salt
(Dowex 50, Na form). The material was concentrated and stored at
)20ꢁC (3 mL, 75 mM); 1H NMR: d 8.25–8.27 (m, 2H, 3,5-H), 7.64–7.66
(m, 2H, 2,6-H), 5.04 (d, 2H, JH,P 7.9, CH2), 3.55 (d, 3H, JH,P 10.8, OMe);
LRMS (negative ion ES): calcd for C8H9NO6P): 246; found: 246.
Time-resolved UV–visible measurements. Absorption transients at
406 nm were triggered by a 30 ns, 308 nm pulse from a LambdaPhysik
LPX105 XeCl excimer laser directed into
a quartz cell with
10 · 10 mm path lengths, that was located in a thermostatted (20ꢁC)
sample holder positioned in a Luzchem LFP-111 transient recorder
(Luzchem Research, Quebec, Canada). Solutions contained the rele-
vant compound (0.5 mM) in 25 mM MOPS (pH 7.0) with 150 mM
NaCl. Data were collected using the Luzchem software and analyzed
in Microsoft ExcelTM. Typically, data from four repeat traces were
averaged before analysis.
Infrared spectroscopic measurements. Infrared difference spectra of
photolysis were recorded as described previously (1), in buffer made by
adjusting a 200 mM aqueous solution of MOPS with 3 M NaOH to pH
7. Samples in [18O]water were prepared by evaporation of the test
compound and buffer (both in H2O solution) on one of the CaF2
IR cell windows and reconstituting the dried spots in isotopic water.
Quantitative and isotopic measurements of inorganic phosphate. A
solution of 4 (0.91 mM) in 20 mM PIPES buffer, pH 7.0 was irradiated
for 45 s (Rayonet RPR-100 photochemical reactor; 16 · 350 nm
lamps) and the irradiated solution and nonirradiated control were
analyzed by anion-exchange HPLC as described above. The extent of
photolysis, determined by comparison of peak heights in the control
and irradiated solutions, was 52%. The control and irradiated
solutions were also assayed for inorganic phosphate, using the
fluorogenic assay based on the coumarin-labeled phosphate binding
protein described elsewhere (7,8). The measured concentrations of
inorganic phosphate in the control and irradiated samples were 0.02
and 0.46 mM, respectively, i.e. net phosphate in the irradiated sample
was 0.44 mM. As the concentration of 4 converted by photolysis was
0.47 mM, this corresponds to an inorganic phosphate yield of ꢀ94%.
a-Carboxy-4-nitrobenzyl methyl phosphate, 5. An aqueous solution
of 4 (619 lmol; 24 mL) was adjusted to pH 4.5 with 1 M HCl, cooled in
ice and stirred vigorously with an ethereal solution of diazomethane
(42.5 mL, ꢀ400 mM). After 30 min the diazomethane was completely
consumed, and anion-exchange HPLC analysis of the aqueous layer
(mobile phase 50 mM sodium phosphate, pH 6-MeOH [100:13
vol ⁄ vol]) gave two peaks, tR 1.2 and 1.8 min. The starting material 4
had tR 5.4 min in this mobile phase. The ether layer was separated and
washed with water (2 · 5 mL) and the aqueous layers were combined
(ꢀ35 mL) and adjusted to pH 13 with 1 M KOH. The solution was

Photochemistry and Photobiology, 2009, 85 1091
For measurement of the isotopic composition of phosphate ion
released from 4 when photolyzed in [18O]water, samples were prepared
as for the time-resolved IR spectra. Thus aliquots of the stock solution
of 4 (1 lL; 94 mM) and 200 mM MOPS, pH 7 (1 lL) were dried onto a
CaF2 window under a stream of nitrogen and rehydrated with
contaminated with a minor amount (ꢀ5%) of a species with a
lower ionic charge. This was shown by comparison (HPLC
and 1H NMR) with an authentic sample (4,5) to be the product
of thermal decarboxylation, namely 4-nitrobenzyl phosphate
12. We subsequently found that formation of this product
could be entirely avoided by elution of the preparative anion-
exchange column with a gradient of lithium chloride, followed
by precipitation of the lithium salt of 4 from the aqueous
eluate with acetone-ethanol. This procedure, formerly devel-
oped (16) for purification of nucleotides on strong anion-
exchangers at low pH, gave almost quantitative recovery of 4
as its lithium salt from the column eluate.
[
18O]water (1 lL; isotopic enrichment 99%). The cell was closed with
a second CaF2 window and assembled into the spectrometer. It was
exposed to two flashes of the photolysis lamp and the sample was
recovered from the disassembled cell windows by addition of water
(15 lL per window). The experiment was repeated three times and the
combined solutions were used for mass spectrometric analysis, using
methods previously described (9). A control sample was treated
identically, except that it was not irradiated. The measured 16O:18
O
ratio was 99.6:0.4, i.e. at natural abundance within the error of the
measurement.
Computation of vibrational frequencies: All calculations were per-
formed using the GAUSSIAN 03 program (10). The geometries were
optimized and the vibrational frequencies computed using hybrid DFT
with the B3LYP functional (11,12) and with the 6-31++G(d,p) basis
set. Solvent effects were taken into account using the conductor-like
polarizable continuum model (CPCM) with the default values for
water (dielectric constant = 78.39). No scaling factors were used on
the vibrational frequencies.
Treatment of a weakly acidic aqueous solution of 4 with
ethereal diazomethane, followed by mild alkaline hydrolysis,
gave the monomethyl phosphate ester 5. The carboxylate and
phosphate groups were both methylated by diazomethane
but the carboxylate ester could then be selectively hydrolyzed,
as previously found for the corresponding o-isomer (1). For 5,
we again investigated purification by anion-exchange
chromatography with a TEAB gradient. In this case, evapo-
ration of the volatile buffer resulted in much more extensive
(>70%) decarboxylation, to give 13, but the lithium chloride
elution procedure described above successfully gave pure 5.
The identity of 13 was confirmed by synthesis of an authentic
sample (see Materials and Methods).
We emphasize that thermal decarboxylation was not
observed during preparation of the corresponding o-nitro
compounds 1 and 9, which were purified by anion-exchange
using TEAB gradients and subsequent vacuum evaporation
(1). For the p-compounds, decarboxylation did not occur
either in highly alkaline solution (pH 13 during the carboxylic
ester hydrolysis in preparation of 5) or in anhydrous trifluo-
roacetic acid (data not shown). Incubations of 4 or 5 in 0.1 M
HCl at 40ꢁC had almost no effect on 4 for at least 24 h, but
resulted in near-complete decarboxylation of 5 within 5 h (rate
constant ꢀ0.8 h)1). The decarboxylation was ꢀ5-fold slower
when 5 was incubated at pH 3 and 40ꢁC. We surmise that
decarboxylation took place during the changes in pH that
occur upon vacuum evaporation of TEAB, where an initial
excursion above pH 9 is followed by a return to mildly acidic
pH (17). The extent of this final acidity alone seems unlikely to
be sufficient to promote the extensive decarboxylation of 5
observed during the TEAB evaporation and further factors are
probably involved, with a specific effect of the triethylammo-
nium ion being a likely contributor. The para-location of the
nitro group must be important and the extensive decarboxyl-
ation of 5, in contrast to the complete lack of this reactivity for
its o-isomer 1, can to an extent be rationalized on the basis of
diminished electronic overlap of the nitro group in the
o-isomer because of steric crowding between the adjacent
substituents. This would result in less effective resonance
stabilization of the carbanion that develops during decarbox-
ylation. In other work, we have described evidence for related
out-of-plane twisting in o-substituted nitroaryl anion radicals
(18). The marked effect of phosphate methylation on the
relative sensitivity to decarboxylation of 4 and 5 implies that
RESULTS AND DISCUSSION
Synthesis and thermal decarboxylation of compounds 4 and 5
We have described elsewhere (3) the synthesis of 4-nitroman-
delic acid 6, together with its O-acetyl derivative 7 and t-butyl
ester 8. In principle, synthesis of 4 and 5 could be achieved from
8 using simple modifications of methods used for synthesis of
the ortho-isomers such as 1 and its precursor 9 (1), but an
alternative starting material provided a more straightforward
synthesis of 4. As shown in Scheme 2, the diazo compound 10,
available in one step from ethyl 4-nitrophenylacetate (13), was
stirred with an excess of concentrated phosphoric acid in
dioxan solution at room temperature to give the product 11.
Preparative reverse-phase HPLC to remove the excess inor-
ganic phosphate and alkaline hydrolysis of the ethyl ester then
gave the phosphate 4.
Scheme 2. Preferred synthetic route to compound 4. Reagents: (a)
conc. H3PO4-dioxan; (b) aq. KOH.
During procedures to isolate 4 in pure form, we encountered
a surprising propensity for thermal decarboxylation, which
was solved as described below. When preparative anion-
exchange column chromatography, eluted with a gradient of
TEAB, was used as we had done (1) for the ortho-nitro
compounds 1 and 9, fractions from the column were analyzed
by HPLC and contained only 4, as would be expected. With
the usual methods for this form of ion-exchange chromatog-
raphy (1,14,15), the volatile buffer salt was removed by
vacuum evaporation of the aqueous solution below 30ꢁC,
followed by several vacuum evaporations from methanol, to
obtain 4 in salt-free form. However, HPLC analysis after the
evaporation process showed that the phosphate 4 had become

1092 John E. T. Corrie et al.
the charge state of the phosphate is also important. We note
that an idiosyncratic thermal decarboxylation of o- and
p-nitrophenylacetic acid derivatives has also been observed
(19) under mildly basic conditions (K2CO3 in warm DMF).
Different mechanisms evidently operate for the latter reaction
and that reported here, as one is promoted by base while the
process we observed occurs in acidic solution.
quantum yields, but nevertheless imply that the quantum
yields of 4 and 5 are of similar magnitude to that of 9 (reported
QP 0.41 [20]).
We next recorded time-resolved IR difference spectra of the
photolysis of these two compounds, also at pH 7 and 20ꢁC. In
each case the most intense peak in the spectra was at
2343 cm)1, arising from CO2 that was formed by photode-
carboxylation, as seen previously for the ortho-CNB com-
pounds (1). For both 4 and 5, formation of CO2 was complete
within the time frame of the first spectrum recorded after the
light flash (0–60 ms) and the band slowly decayed, as described
previously (1), because of hydration of the evolved CO2 (21).
There was no evidence of a second, slow phase of CO2 release
that had been observed previously (1) for 1.
These spectroscopic data, combined with previous work (3)
on the related compounds 6 and 7, led us to a proposed
reaction pathway for photolysis of 4, as shown in Scheme 3. It
is convenient to present the Scheme at this point in order to
more easily discuss the data that support it. Notably, it
incorporates formation of an aci-nitro intermediate 14 as a
consequence of photodecarboxylation rather than by abstrac-
tion of a benzylic proton by the excited nitro group, which is
the conventional route for photolysis of 2-nitrobenzyl com-
pounds (see Scheme 1). The latter pathway is not available in
the 4-nitro series as the benzylic proton is not accessible to the
nitro group.
Spectroscopic studies of the photolysis of 4 and 5
We first recorded transient absorption data at 406 nm and
20ꢁC for 4 and 5, following laser flash irradiation at 308 nm, to
determine whether any species absorbing in this region was
formed. Absorption at ‡400 nm after flash irradiation of
2-nitrobenzyl and related compounds is characteristic of the
formation of an aci-nitro intermediate, such as 2 (Scheme 1).
At pH 7, 4 gave a transient that was fitted by a single
exponential decay at 0.36 s)1 (Fig. 1). A small rapid phase at
ꢀ300 s)1 (relative amplitude ꢀ10% of the main transient; see
inset of Fig. 1) was observed in recordings made on a faster
time scale (data not shown). The kinetics of the main phase of
the decay appear to be similar to those of bands in the IR
spectrum of photolysis that are compatible with the proposed
aci-nitro intermediate 14 (see below). For 5, the main phase
was at ꢀ0.07 s)1 and, as observed for 4, there was a small
faster phase (relative amplitude ꢀ15% of the main transient,
data not shown), in this case at ꢀ40 s)1. The origins of these
faster phases for 4 and 5 remain unclear. The photolysis
efficiencies of each compound, assessed using broad band-
width light centered at 350 nm (Rayonet Photochemical
Reactor) were similar, and slightly higher than for the related
2-nitro compound 9. For example, for identical concentrations
of separate solutions of 4 and 9 irradiated for the same length
of time, conversions of starting materials were 33% and 52%,
respectively. Because of the different chromophores of the two
compounds, these relative conversions do not relate directly to
Scheme 3. Proposed mechanism for photolytic reactions of 4.
Aside from the above-mentioned band for photoreleased
CO2 (data not shown), the photolysis spectrum for 4 recorded
in the period immediately after the light flash (Fig. 2, solid
line) showed a number of well-resolved, strong features that
included positive bands at 1575, 1477 and 1410 cm)1. The
latter two bands appear at similar positions for related 2-nitro
compounds (1), as discussed in more detail below, and are due
to the vibrations of the aci-nitro intermediate. The assignment
of the 1575 cm)1 band is less straightforward. It appears at an
identical position to that reported (1) for the conjugated
carboxylate present in the intermediate 2 (Scheme 1), for
which the band was positively identified by an isotope-induced
shift of its 13C-carboxylate isotopomer. A priori, this band
Figure 1. Transient 406 nm absorbance following 308 nm laser flash
photolysis of an aqueous solution of 4 (0.5 mM) in 25 mM MOPS,
150 mM NaCl at pH 7, 20ꢁC. The red line superimposed on the
experimental points is
a single exponential fit to the data
(kobs = 0.36 s)1). The inset is of a trace recorded on a faster time
scale to show the small 300 s)1 transient at the same wavelength. The
red line is a double exponential fit to the data.

Photochemistry and Photobiology, 2009, 85 1093
strong mode calculated at 1588 cm)1 was localized at the
carboxylate group, while an equally intense band at 1555 cm)1
had a major contribution from stretch of the exocyclic C–C
bond (for the full data set, see Table S2). Downshift of the
latter mode from its value in 14 is consistent with the presence
of the conjugated carboxylate substituent. Thus the calculated
spectrum of 17 is predicted to have two strong bands within a
30 cm)1 interval in the 1500 to 1600 cm)1 region. This is not
observed and thus the presence of 17 is not supported by the
comparison between calculated and experimental spectra.
Figure 2. Infrared difference spectra for photolysis of 4 (94 mM; H2O
solution in 200 mM MOPS, pH 7.0, 20ꢁC), averaged over the time
interval 2–647 ms (—), and at 80 s (ÆÆÆÆÆ) after the photolysis flash. A
strong CO2 band at 2343 cm)1 (Dabsorbanceꢀ30 · 10)3) was present
in the initial spectrum and decayed to ꢀ10% of its initial intensity.
This band is outside the frequency range shown in the figure.
The positive band at 1477 cm)1 was significantly upshifted
from the positions of a similarly shaped band in the spectra of
the aci-nitro intermediates formed from the related 2-nitro
compound 1 (1460 cm)1) and from 2-nitrobenzyl methyl
phosphate (1459 cm)1) (1). We assign this band to the mode
calculated at 1501 cm)1, for which the main contribution was
stretching of the carbon–nitrogen bond of the nitronate group.
The strong band at 1410 cm)1 also seems to be upshifted from
its position at 1373 cm)1 for 2 (1) and from 1398 cm)1 for
2-nitrobenzyl methyl phosphate (see Supporting Information of
Corrie et al. [1], Fig. S2). We assign this band to a delocalized
mode calculated at 1419 cm)1, with contributions from rocking
vibrations of several C–H bonds as well as from C=N
stretching (Table S1). Delocalized vibrations have previously
been inferred for several modes in the aci-nitro intermediate
formed from 1-(2-nitrophenyl)ethyl-caged ATP (27,28).
Negative bands in Fig. 2 represent vibrational modes of the
starting compound 4 that is consumed by photolysis. Those
at 1594 cm)1 (mas of the disappearing carboxylate group),
1521 cm)1 (mas of the disappearing nitro group) and 1351 cm)1
(combined ms of the disappearing carboxylate and nitro
groups) can be confidently assigned (1,27), while the negative
band at 1086 cm)1 is likely to arise from a phosphate vibration
present in the starting material.
could therefore be consistent with the presence of a carbox-
ylate group in the photogenerated intermediate(s) produced by
photoirradiation of 4. However, for the spectrum shown in
Fig. 2, neither the intensity nor frequency of this band was
affected upon photolysis in D2O solvent (data not shown). The
mas carboxylate vibration normally experiences an upshift of
ꢀ10 cm)1 (except for a-aminocarboxylates) (22–24) and an
ꢀ2-fold increase in intensity (23,24) upon a change from H2O
to D2O solvent.
To explore whether the 1575 cm)1 band necessarily indi-
cates a carboxylated aci-nitro intermediate, DFT calculations
(25) in continuum water were performed for the postulated
decarboxylated aci-nitro intermediate 14. The calculations
revealed a number of vibrational modes with strong absorp-
tion coefficients, and these calculated bands are listed in the
Supporting Information (Table S1). DFT-calculated frequen-
cies typically overestimate the observed values (26) and it is
reasonable to assign the strongly absorbing vibrations with
calculated frequencies at 1605, 1501 and 1419 cm)1 to the
observed bands at 1575, 1477 and 1410 cm)1, respectively. The
calculated spacing between these three vibrations is nearly
identical to that observed experimentally. In fact, the observed
band at 1575 cm)1 is distorted by overlap with the adjacent
negative band at 1594 cm)1 so its true maximum would
be slightly upshifted, bringing it nearer to the calculated
frequency.
The calculations showed that the normal mode assigned to
the 1575 cm)1 band was dominated by a stretching vibration
of the exocyclic C–C double bond that carries the phosphate
substituent (Table S1). We conclude that the calculated
spectrum of 14 provides a good description of the experimental
spectrum in the region above 1400 cm)1 which is of interest
here, and assignment of the 1575 cm)1 band to a vibration of a
retained carboxylate group in the aci-nitro intermediate is not
supported either by the experimental evidence or calculations
of vibrational modes. For confirmation, we also made DFT
calculations of vibrational frequencies in the analog of 14 in
which the carboxylate group was retained, i.e. the species 17. A
As shown in Fig. 2, most of the bands in the initial spectrum
decayed in the transition to the final spectrum (mean rate
constant [ SD] 0.13
0.02 s)1, based on single-exponential
fits to the integrated intensities of the decaying bands at 1575 and
1410 cm)1, and the developing bands at 1694 and 1070 cm)1).
This value is slower than that observed for the 406 nm transient
(0.36 s)1) but similar discrepancies in kinetic data from the very
different experimental conditions of the FTIR and UV–visible
measurements have been seen in other work (27).
A notable feature of the final spectrum (dotted line in
Fig. 2) was the appearance of two medium intensity bands at
1070 and 994 cm)1, which corresponded to the inorganic
phosphate ion. The same two bands were present in the
photolysis spectrum of 1-(2-nitrophenyl)ethyl phosphate 18
(1071 and 993 cm)1; see Fig. S1) and correspond to published
data (29) for the phosphate ion (as HPO42), the predominant
species at pH 7). It is possible that a small proportion of the
phosphate ion is already released in the time frame of the

1094 John E. T. Corrie et al.
initial spectrum (full line in Fig. 2), where a weak positive band
at 994 cm)1 and a shoulder near 1070 cm)1 on the negative
band at 1066 cm)1 can be seen. Analytical measurement (see
Materials and Methods) showed that the photolysis yield of
inorganic phosphate was 94%, i.e. almost stoichiometric.
We can now review the above data and consider the extent
to which they support the reaction mechanism shown in
Scheme 3. The data show rapid photodecarboxylation (much
faster than the aci-nitro decay process), essentially quantitative
release of inorganic phosphate, evidence for two pathways
(shown by the IR spectra recorded in normal and [18O]water),
and a nitro group that disappears as a consequence of the light
flash and is not reformed during decay of the intermediate
species (i.e. the intensity of the negative nitro band in the
difference spectra does not change significantly between the
first and last spectra). As discussed above, it is possible that
there is a phase of phosphate release faster than could be
resolved in our IR spectra (i.e. faster than ꢀ10 ms, and
possibly related to the fast phase of the 406 nm transient
shown in Fig. 1), but most of the intensity of the phosphate
bands at 1070 and 994 cm)1 develops on the same slow time
scale over which other bands in the spectra appear or decay.
The pathway shown in Scheme 3 requires photolytic decar-
boxylation, as reported previously (3) for the related com-
pounds 6 and 7, to form the aci-nitro species 14. Subsequent
hydration gives the hemiacetal 15, which evidently breaks
down as rapidly as it is formed to give inorganic phosphate
and 4-nitrosobenzaldehyde 16. The latter compound was
reported (3) to absorb at 1690 cm)1 (KBr disk), i.e. very close
to the observed 1694 cm)1 band. Upon photolysis in isotopic
water, the mechanism implies incorporation of 18O into the
carbonyl group of 16 and its carbonyl band would be
downshifted, resulting in the band that appears near
1660 cm)1. A downshift of ꢀ30 cm)1 between 16O and 18O
isotopomers is typical of an isolated carbonyl group (30). The
incorporation of solvent water into the organic product is
consistent with the lack of isotopic oxygen enrichment in the
released phosphate ion.
The formation of nitroso product 16 is also consistent with
the band at 1507 cm)1 in the final spectrum (N=O stretch)
and disappearance of the nitro band of the starting material.
This contrasts with results of the earlier photolysis study (3) of
4-nitromandelic acid 6 and its acetate 7 in water, where the
products were mostly nitro compounds, and indeed where
both photolysis products from 7 also retained the acetate
group of the starting compound (under anaerobic conditions).
Under aerobic conditions, photolysis of 7 gave 4-nitrobenzal-
dehyde and 4-nitrobenzoate. The conditions in our experi-
ments are essentially anaerobic as the IR samples are
contained in a closed cell, so only the oxygen that is present
in solution when the sample is prepared can be available for
interaction with photogenerated species. The concentration of
oxygen in water exposed to atmospheric gas is only ꢀ0.3 mM
(31) and a single flash in the IR spectrometer typically converts
20–30% of the starting compound (i.e. ꢀ18–28 mM from the
starting 94 mM concentration in the present experiments), so it
is unlikely that oxygen can play a significant role in the
reaction under our conditions.
The high yield of phosphate from photolysis of 4 is of
interest as the presence of a strong CO2 band implies a
substantial extent of photodecarboxylation. This result is in
direct contrast to our previous study (1) of ‘‘conventional’’
CNB-caged compounds (see Scheme 1), where it was firmly
concluded that product release only took place in the portion
of the reaction flux that did not involve photodecarboxylation.
The different structure of the p-compounds described here
must in any case imply a different pathway for product release
than that of Scheme 1 and our proposal, which is compatible
with the available data, is shown in Scheme 3.
This mechanism implies that an oxygen atom from solvent
water becomes incorporated into the by-product 16 via the
conjugate addition process shown, although an alternative
would be nucleophilic displacement by water at the phospho-
rus center. Photolysis in [18O]water solution was carried out to
distinguish these possibilities. Spectra of the photochemically
generated intermediate(s) (full line in Fig. 2) were identical
when recorded in either normal or [18O]water (data not
shown). This was also true for the final photolysis spectra
below 1650 cm)1 (see Fig. S2). Notably, there was no isotope-
induced shift of the bands at 1070 and 994 cm)1 that
correspond to inorganic phosphate, which implies that the
isotope is not incorporated into the released phosphate ion.
This result was confirmed by mass spectrometric analysis and
therefore the phosphate liberated by photolysis cannot arise by
nucleophilic displacement at the phosphate group by solvent
water.
In contrast, the final photolysis spectra in normal and
[18O]water showed some unexpected differences in the region
between 1650 and 1700 cm)1, which indicated that the reaction
course must be more complex than the pathway of Scheme 3
alone. In this region, the band at 1694 cm)1 in normal water
was reduced in [18O]water to ꢀ60% of its intensity and slightly
shifted to 1692 cm)1, while a new band appeared at 1660 cm)1
.
Note that this region is overlaid by the antisymmetric
vibration of bicarbonate ion that develops on a similar time
scale, so relative intensities of bands in this region may be
unreliable. The residual intensity at 1692 cm)1 and new band
at 1660 cm)1 suggest that two species are formed, only one of
which incorporates oxygen from the solvent. The result cannot
be explained by incomplete isotopic enrichment of the
[18O]water, as a control experiment (see Fig. S3) verified the
isotopic composition as being essentially 100%. The relative
intensity of the band at 1692 cm)1 was maintained at five-fold
lower concentration of 4, suggesting that it does not originate
from intermolecular reactions between two molecules of 14, or
between 14 and 4, although such processes have been proposed
for related compounds (3). It was also observed with a similar
amplitude in different buffer solutions (MES, pH 5.5 and 6,
and MOPS, pH 6.5).
Thus, our data are consistent with the mechanism of
Scheme 3 being a substantial route to the released phosphate
ion product, but the results in [18O]water suggest that a second
pathway is also operative. One possibility could have been that
the bicarbonate ion formed by hydration of photoreleased CO2
could perform nucleophilic attack on the aci-nitro species 14 in
competition with the water molecule that is shown in
Scheme 3. When first formed, the bicarbonate ion would

Photochemistry and Photobiology, 2009, 85 1095
contain one 18O and two 16O atoms and could provide a means
to generate the nonlabeled by-product 16. This option is
however ruled out by the kinetics: we have previously reported
the hydration rate of CO2 as 0.055 s)1 under the same
experimental conditions as the present experiments (32). Thus
decay of the aci-nitro species is at least six-fold faster than the
formation of bicarbonate. Furthermore, there was no evidence
for biphasic decay of the CO2 band, i.e. the normal hydration
equilibrium of CO2 was not perturbed by a pathway that
removed the bicarbonate ion from solution.
unearthed for these compounds, as photolysis of 5 appears to
release almost no monomethyl phosphate ion.
The other item of mechanistic and synthetic interest is the
unexpected thermal decarboxylation observed for compounds 4
and 5 during evaporation of TEAB buffers, and its avoidance by
using lithium chloride gradients for elution of these compounds
from preparative ion exchange columns. Clearly an effect either
of the TEAB buffer or the changes in pH that occur during its
evaporation have a role in this process.
Acknowledgements—We are grateful to Chris Toseland and Dr. Martin
Webb for the mass spectrometric measurements, to Dr. David
Trentham for help with recording the 406 nm transient and to
Dr. Margareta Blomberg (Stockholm University) for advice on the
DFT calculations. We thank the MRC Biomedical NMR Centre for
access to facilities, the Knut och Alice Wallenbergs Stiftelse for
There are hints of other reaction possibilities within the
spectra of the photochemically generated intermediate (solid
line in Fig. 1). For example, the possible fast phase of phosphate
ion release that is inferred from the IR spectra might be
associated with the small amplitude fast phase noted in the
decay of the 406 nm transient. In addition, an unexplained
shoulder at ꢀ1468 cm)1 on the low-wavenumber side of the
1477 cm)1 band in the IR spectra of 4 (also present for 5) may be
associated with an intermediate other than 14. Further specu-
lation on other possible pathways does not seem appropriate.
Difference spectra of photolysis were also recorded for 5,
but the evolution of these spectra was markedly different from
those discussed above for 4. Thus, spectra recorded for 5
immediately after the light flash (see Fig. S4) showed a strong
positive CO2 band at 2343 cm)1 and a number of sharp
positive bands, very similar to those in the initial spectrum of 4
in the region above ꢀ1300 cm)1. However, all these bands
decayed very slowly (several minutes), with little evidence for
release of the monomethyl phosphate anion. A weak envelope
developed very slowly at ꢀ1100 cm)1, consistent with a small
amount of this product being formed (see data in fig. 4 of
Corrie et al. [1] for a much clearer band at this frequency,
arising from photolysis of the 2-nitro isomer 1). A notable
feature of the spectra in Fig. S4 was the negative band at
1523 cm)1 that decayed from an intense band in the spectrum
recorded immediately after photolysis to a much weaker band
in the spectrum recorded 12 min after the flash. This band is
from the antisymmetric stretch of the nitro group and the
recovery from its postflash intensity level indicates that the
nitro group is largely reformed during decay of the photo-
chemically generated intermediate, which is in line with results
in the earlier study (3) of compounds 6 and 7.
funding the infrared spectrometer, Vetenskapsradet for supporting
˚
M.R. and the EPSRC National Mass Spectrometry Centre for
negative ion mass spectrometry. This work was supported in part by
a Royal Society ESEP grant.
SUPPORTING INFORMATION
Additional Supporting Information may be found in the online
version of this article:
Figure S1. Infrared difference spectrum for photolysis of
1-(2-nitrophenyl)ethyl phosphate 18, recorded at 20ꢁC for an
H2O solution of the compound (130 mM) in 200 mM MOPS,
pH 7.0.
Figure S2. Infrared difference spectra for photolysis of
a-carboxy-4-nitrobenzyl phosphate, 4, recorded at 20ꢁC for a
solution of the compound (94 mM) in 200 mM MOPS, pH 7.0,
with solvent of either normal water (—) or [18O]water (ÆÆÆÆÆ).
Figure S3. Portion of the IR difference spectrum for
photolysis of 4 (94 mM) in 200 mM MES, pH 6.0, 20ꢁC in
[18O]water as solvent.
Figure S4. Infrared difference spectra for photolysis of
a-carboxy-4-nitrobenzyl methyl phosphate, 5, recorded at
20ꢁC for an H2O solution of the compound (82 mM) in
200 mM MOPS, pH 7.0.
Table S1. DFT-calculated frequencies, intensities and
assignments for normal modes of the intermediate species 14.
Data are shown for the range 1800)935 cm)1
Table S2. DFT-calculated frequencies, intensities and
.
CONCLUSION
assignments for normal modes of the putative intermediate
species 17. Data are shown for the range 1800)940 cm)1
Please note: Wiley-Blackwell are not responsible for the
content or functionality of any supporting information sup-
plied by the authors. Any queries (other than missing material)
should be directed to the corresponding author for the article.
.
This work arose in the course of a related study of a-carboxy-
2-nitrobenzyl photochemistry (1). Overall, it emphasizes the
rather complex photochemistry of the a-carboxy-4-nitrobenzyl
system, where products are evidently very influenced by the
nature of the group attached to the benzyl carbon. Perhaps the
most surprising aspect is the essentially quantitative release of
inorganic phosphate from 4, although this release is on a very
REFERENCES
slow time scale (t ꢀ5 s). It is conceivable that this could
½
1. Corrie, J. E. T., V. R. N. Munasinghe, D. R. Trentham and
A. Barth (2008) Studies of decarboxylation in photolysis of
a-carboxy-2-nitrobenzyl (CNB) caged compounds. Photochem.
Photobiol. Sci. 7, 84–97.
2. Margerum, J. D. and C. T. Petrusis (1969) The photodecarboxyla-
tion of nitrophenylacetate ions. J. Am. Chem. Soc. 91, 2467–2472.
3. Morrison, J., P. Wan, J. E. T. Corrie and V. R. N. Munasinghe
(2003) Chemistry of photogenerated a-hydroxy-p-nitrobenzyl
carbanions in aqueous solution: Protonation vs disproportion-
ation. Can. J. Chem. 81, 586–597.
enable triggering of consecutive reactions induced by two
different substances. A single light flash applied to a mixture of
two caged compounds with different release kinetics would
predominantly liberate the faster one first, followed later by
the slower. The near-stoichiometric release of phosphate ion
from 4 contrasts with significantly sub-stoichiometric product
release from the related a-carboxy-2-nitrobenzyl system (1).
However, there are evidently further complexities still to be

1096 John E. T. Corrie et al.
4. Kluba, M. and A. Zwierzak (1974) Organophosphorus esters. 6.
Phosphorylation of bifunctional organic halides. Rocz. Chem. 48,
1603–1608.
5. Kaplan, J. H., B. Forbush and J. F. Hoffman (1978) Rapid
photolytic release of adenosine 5¢-triphosphate from a protected
analogue: Utilization by the Na:K pump of human red blood cell
ghosts. Biochemistry 17, 1929–1935.
and deoxyribonucleoside 5¢-triphosphates. J. Am. Chem. Soc. 80,
1141–1145.
17. Oiwa, K., J. F. Eccleston, M. Anson, M. Kikumoto, C. T. Davis,
G. P. Reid, M. A. Ferenczi, J. E. T. Corrie, A. Yamada, H.
Nakayama and D. R. Trentham (2000) Comparative single-mol-
ecule and ensemble myosin enzymology: Sulfoindocyanine ATP
and ADP derivatives. Biophys. J. 78, 3048–3071.
6. Corrie, J. E. T., G. P. Reid, D. R. Trentham, M. B. Hursthouse
and M. A. Mazid (1992) Synthesis and absolute stereochemistry of
the two diastereoisomers of P3-1-(2-nitrophenyl)ethyl adenosine
triphosphate (‘caged’ ATP). J. Chem. Soc., Perkin Trans. 1 1015–
1019.
18. Corrie, J. E. T., B. C. Gilbert, V. R. N. Munasinghe and A. C.
Whitwood (2000) EPR studies of the structure of transient
radicals formed in photolytic reactions of some 2-nitrobenzyl
compounds. Characterisation of aryl alkoxy nitroxides and
nitroaromatic radical-anions in the photolysis of caged ATP
and related compounds. J. Chem. Soc., Perkin Trans. 2 2483–2491.
19. Bull, D. J., M. J. Fray, M. C. Mackenny and K. A. Malloy (1996)
Facile decarboxylation of nitrobenzeneacetic acids as a convenient
route to alkylnitrobenzenes. Synlett 647–648.
7. Brune, M., J. L. Hunter, J. E. T. Corrie and M. R. Webb
(1994) The direct, real-time measurement of rapid inorganic
phosphate release using
a novel fluorescent probe and its
application to actomyosin subfragment 1 ATPase. Biochemistry
33, 8262–8271.
20. Walker, J. W., Z. Lu and R. L. Moss (1992) Effects of Ca2+ on the
kinetics of phosphate release in skeletal muscle. J. Biol. Chem. 267,
2459–2466.
8. Brune, M., J. L. Hunter, S. A. Howell, S. R. Martin, T. L. Hazlett,
J. E. T. Corrie and M. R. Webb (1998) The mechanism of the
coumarin-labeled phosphate binding protein, as a probe of inor-
ganic phosphate in the submicromolar range. Biochemistry 37,
10370–10380.
9. Phillips, R. A., J. L. Hunter, J. F. Eccleston and M. R. Webb
(2003) The mechanism of ras GTPase activation by neurofibro-
min. Biochemistry 42, 3956–3965.
10. Frisch, M. J., G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
M. A. Robb, J. R. Cheeseman, J. A. Montgomery Jr, T. Vreven,
K. N. Kudin, J. C. Burant, J. M. Millam, S. S. Iyengar, J. Tomasi,
V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N. Rega,
G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K. Toyota,
R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda,
O. Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P. Hratchian,
J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts,
R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli,
J. W. Ochterski, P. Y. Ayala, K. Morokuma, G. A. Voth,
P. Salvador, J. J. Dannenberg, V. G. Zakrzewski, S. Dapprich,
A. D. Daniels, M. C. Strain, O. Farkas, D. K. Malick, A. D. Rabuck,
K. Raghavachari, J. B. Foresman, J. V. Ortiz, Q. Cui, A. G. Baboul,
S. Clifford, J. Cioslowski, B. B. Stefanov, G. Liu, A. Liashenko,
P. Piskorz, I. Komaromi, R. L. Martin, D. J. Fox, T. Keith,
M. A. Al-Laham, C. Y. Peng, A. Nanayakkara, M. Challacombe,
P. M. W. Gill, B. Johnson, W. Chen, M. W. Wong, C. Gonzalez and
J. A. Pople (2003) Gaussian 03, Revision B.03. Gaussian, Inc.,
Wallingford, CT.
21. Ho, C. and J. M. Sturtevant (1963) The kinetics of hydration of
carbon dioxide at 25ꢁ. J. Biol. Chem. 238, 3499–3501.
22. Tackett, J. E. (1989) FT-IR characterization of metal acetates in
aqueous solution. Appl. Spectrosc. 43, 483–489.
23. Venyaminov, S. Y. and N. N. Kalnin (1990) Quantitative IR
spectroscopy of peptide compounds in water (H2O) solutions. 1.
Spectral parameters of amino acid residue absorption bands.
Biopolymers 30, 1243–1257.
24. Chirgadze, Y. N., O. V. Fedorov and N. P. Trushina (1975)
Estimation of amino acid residue side-chain absorption in the
infrared spectra of protein solutions in heavy water. Biopolymers
14, 679–694.
25. Meier, R. J. (2007) Calculating the vibrational spectra of mole-
cules: An introduction for experimentalists with contemporary
examples. Vib. Spectrosc. 43, 26–37.
26. Hehre, W. J., L. Radom, P. von, R. Schleyer and J. A. Pople
(1986) Ab Initio Molecular Orbital Theory. Wiley, New York.
27. Barth, A., J. E. T. Corrie, M. J. Gradwell, Y. Maeda, W. Mantele,
T. Meier and D. R. Trentham (1997) Time-resolved infrared
spectroscopy of intermediates and products from photolysis of
1-(2-nitrophenyl)ethyl phosphates: Reaction of the 2-nitrosoace-
tophenone by-product with thiols. J. Am. Chem. Soc. 119, 4149–
4159.
28. Barth, A., K. Hauser, W. Mantele, J. E. T. Corrie and
D. R. Trentham (1995) The photochemical release of ATP from
‘‘caged ATP’’ studied by time-resolved infrared spectroscopy.
J. Am. Chem. Soc. 117, 10311–10316.
11. Becke, A. D. (1993) Density-functional thermochemistry 3. The
role of exact exchange. J. Chem. Phys. 98, 5648–5652.
12. Lee, C. T., W. T. Yang and R. G. Parr (1988) Development of the
Colle-Salvetti correlation-energy formula into a functional of the
electron density. Phys. Rev. B 37, 785–789.
29. Chapman, A. C. and L. E. Thirlwell (1964) Spectra of phosphate
compounds. I. The infra-red spectra of orthophosphates. Spec-
trochim. Acta 20, 937–947.
13. Regitz, M. (1965) Eine neue Synthese fur a-Diazo-carbon-
ylverbindungenen. Chem. Ber. 98, 1210–1224.
14. Porath, J. (1955) Use of triethylammonium buffers in ion-
exchange chromatography and electrophoresis. Nature 175, 478.
15. Moffatt, J. G. (1964) A general synthesis of nucleoside 5¢-tri-
phosphates. Can. J. Chem. 42, 599–604.
30. Bryn, M. and M. Calvin (1966) Oxygen-18 exchange reactions of
aldehydes and ketones. J. Am. Chem. Soc. 88, 1916–1922.
31. Murov, S. L., I. Carmichael and G. L. Hug (1993) Handbook of
Photochemistry, 2nd edn. Marcel Dekker, New York.
32. Papageorgiou, G., A. Barth and J. E. T. Corrie (2005) Flash
photolytic release of alcohols from photolabile carbamates or
carbonates is rate-limited by decarboxylation of the photoproduct.
Photochem. Photobiol. Sci. 4, 216–220.
16. Smith, M. and H. G. Khorana (1958) Nucleoside polyphosphates.
VI. An improved and general method for the synthesis of ribo-