Caged CO2 for the Direct Observation of CO2-Consuming Reactions

Article abstract of DOI:10.1002/cbic.201200659

CO₂-consuming reactions, in particular carboxylations, play important roles in technical processes and in nature. Their kinetic behavior and the reaction mechanisms of carboxylating enzymes are difficult to study because CO₂ is inconvenient to handle as a gas, exists in equilibrium with bicarbonate in aqueous solution, and typically yields products that show no significant spectroscopic differences from the reactants in the UV/Vis range. Here we demonstrate the utility of 3-nitrophenylacetic acid and related compounds (caged CO₂) in conjunction with infrared spectroscopy as widely applicable tools for the investigation of such reactions, permitting convenient measurement of the kinetics of CO₂ consumption. The use of isotopically labeled caged CO₂ provides a tool for the assignment of infrared absorption bands, thus aiding insight into reaction intermediates and mechanisms.

Full text of DOI:10.1002/cbic.201200659

CHEMBIOCHEM
FULL PAPERS
DOI: 10.1002/cbic.201200659
Caged CO2 for the Direct Observation of CO2-Consuming
Reactions
Katharina Lommel,[a] Gabriela Schꢀfer,[a] Konstantin Grenader,[b] Christoph Ruland,[a]
Andreas Terfort,[b] Werner Mꢀntele,[a] and Georg Wille*[a]
CO2-consuming reactions, in particular carboxylations, play im-
portant roles in technical processes and in nature. Their kinetic
behavior and the reaction mechanisms of carboxylating en-
zymes are difficult to study because CO2 is inconvenient to
handle as a gas, exists in equilibrium with bicarbonate in aque-
ous solution, and typically yields products that show no signifi-
cant spectroscopic differences from the reactants in the UV/Vis
range. Here we demonstrate the utility of 3-nitrophenylacetic
acid and related compounds (caged CO2) in conjunction with
infrared spectroscopy as widely applicable tools for the investi-
gation of such reactions, permitting convenient measurement
of the kinetics of CO2 consumption. The use of isotopically la-
beled caged CO2 provides a tool for the assignment of infrared
absorption bands, thus aiding insight into reaction intermedi-
ates and mechanisms.
Introduction
CO2-consuming reactions are very common and important
both in the chemical industry and in nature. Moreover, techni-
cal Carbon Capture and Storage (CCS) has been discussed as
one way to reduce atmospheric CO2,[1] but the biological path-
way is even more important: ribulose-1,5-bisphosphate-carbox-
ylase/oxygenase (RuBisCO), the CO2-fixing enzyme in photo-
synthesis, is the most abundant enzyme on earth, providing
the major pathway for removing CO2 from the atmosphere.[2]
Many other carboxylating enzymes play crucial roles in the me-
tabolism of all organisms;[3] examples include the biotin-de-
pendent enzymes involved in fatty acid synthesis (acetyl-coen-
zyme A carboxylase) and gluconeogenesis (pyruvate carboxy-
lase), the carbamoylphosphate synthase of the urea cycle, and
the CO2-fixing phosphoenolpyruvate carboxylase (PEPC).
centration jumps required for the observation of pre-steady-
state kinetics or enzyme-bound intermediates are especially
difficult to generate. Bicarbonate, which exists in a pH-depen-
dent equilibrium with CO2 in aqueous solution, can be used as
a substitute, although the two species clearly differ chemically,
and enzymes usually accept only one of the two as substrate.
We show here a way to tackle these problems by combining
IR spectroscopy with the use of photodecarboxylating com-
pounds.
IR spectroscopy, however, is ideally suited for the observa-
tion of CO2: in aqueous solution the molecule exhibits a very
strong IR absorbance band (absorption coefficient e=1.5ꢁ
106 cm2 molÀ1) at 2343 cmÀ1 [4]
; that is, in a spectral region with-
out overlap from signals of other, specifically biochemically rel-
evant, compounds. Furthermore, the absorbance band of CO2
in aqueous solution lacks the typical substructure from rota-
tional transitions that would be visible in the gas phase, so dis-
solved and atmospheric CO2 can easily be distinguished. IR
spectroscopy has been used previously in an activity assay for
a decarboxylating enzyme[5] and also for the observation of an
enzyme-bound CO2 species in the active site of carbonic anhy-
drase.[6]
Despite their importance and ubiquity, CO2-utilizing enzymes
are generally difficult to study. UV–visible spectroscopy, a stan-
dard tool in enzyme kinetics, often fails because the reactants
(CO2 or bicarbonate and an organic compound) and the prod-
ucts (the carboxylated compound) rarely exhibit distinct spec-
tral features in the UV–visible region. As a result, activity assays
described in the literature are often discontinuous (employing
radioactive 14CO2, for example), or use helper enzymes in cou-
pled reactions, precluding the observation of fast processes.
Furthermore, CO2—being a gas under “biochemical” condi-
tions—is tedious to handle. The rapid and well-defined con-
The second component of our approach—photoactivatable
or caged compounds—are established tools for the rapid gen-
eration of concentration jumps by flash photolysis and subse-
quent observation of chemical reactions.[7–13] Many of these re-
agents release CO2 in addition to their intended effector mole-
cule, either as a result of their mechanism of effector release
(e.g., in the amine-yielding photolysis of carbamates[14,15]) or in
undesired side-reactions that limit the cage’s efficiency.[16,17]
Nitrophenylacetic acids and their anions have been shown
to decarboxylate quickly and efficiently upon irradiation with
UV light,[18] and their good solubilities and stabilities have
made them attractive candidates for use as caged CO2 com-
[a] K. Lommel,+ Dr. G. Schꢀfer,+ C. Ruland, Prof. Dr. W. Mꢀntele, Dr. G. Wille
Institut fꢁr Biophysik, Johann Wolfgang Goethe-Universitꢀt Frankfurt
Max-von-Laue-Str. 1, 60438 Frankfurt am Main (Germany)
E-mail: wille@biophysik.uni-frankfurt.de
[b] K. Grenader, Prof. Dr. A. Terfort
Institut fꢁr Anorganische und Analytische Chemie
Johann Wolfgang Goethe-Universitꢀt Frankfurt
Max-von-Laue-Str. 7, 60438 Frankfurt am Main (Germany)
[+] These authors contributed equally to this work.
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2013, 14, 372 – 380 372

CHEMBIOCHEM
FULL PAPERS
www.chembiochem.org
Scheme 1. Caged CO2 compounds investigated in this study.
pounds. Their mechanisms of photolysis have recently been
studied in exquisite detail, both theoretically and experimental-
ly.[19,20] In particular, 3-nitrophenylacetic acid (mNPAA,
Scheme 1) shows promise as caged CO2. It is a small molecule
that decarboxylates very rapidly and efficiently (tꢀ200 ps,
quantum yield 0.63), releasing just CO2 and one other, inert
side product (3-nitrotoluene). The very fast reaction is a prereq-
uisite for time-resolved IR measurements because any intramo-
lecular reactions of the caged compound will be finished
before the first spectrum is acquired after photolysis. The sim-
plicity of the reaction also helps in the IR band assignment.
To facilitate the identification of CO2-associated bands in the
IR spectra, we also synthesized the isotopically labeled deriva-
tives 2-(3-nitrophenyl)(1-13C)acetic acid (13C-mNPAA) and 2-(3-
nitrophenyl)(18O2)acetic acid (18O-mNPAA). Furthermore, 2,2’-(5-
nitro-1,3-phenylene)diacetic acid (NPDAA) was prepared, with
the aim of generating two moles of CO2 per mole of cage,
thereby reducing the amount of the poorly soluble side prod-
uct.
Figure 1. Difference FTIR spectra of the four compounds. The concentration
of each compound is 20 mm in 100 mm sodium phosphate buffer (pH 7.3),
295 K. Spectra were normalized to equal integral absorbance of the CO2
peak near 2300 cmÀ1 to account for small differences in optical path length
or flash intensities. c: “5 s after photolysis” minus “before photolysis”; ····:
“120 s after photolysis” minus “before photolysis”. The spectra are stacked
on the absorbance scale to avoid overlap. Arrows indicate the symmetric
and antisymmetric ÀNO2 stretching bands.
case by subtracting the IR spectrum before the laser flash from
the spectrum obtained 5 s after light exposure.
mNPAA: The prominent peak at 2343 cmÀ1 results from the
antisymmetric stretching vibration of CO2 physically dissolved
in water. The rotational structure that would be visible in the
spectrum of gaseous CO2 is absent. One set of negative bands
is associated with the loss of carboxylate, as expected (nas =
1565 cmÀ1, ns =1381 cmÀ1). The other two negative bands are
most likely due to a small reduction in the absorption coeffi-
cient of the aromatic nitro group upon decarboxylation (nas =
1537 cmÀ1, ns =1355 cmÀ1). Under the experimental conditions,
a peak height at 2343 cmÀ1 of 3 mOD corresponds to a CO2
concentration of 4 mm: that is, effective photolysis of 20% of
the total cage in the sample. This agrees with the observation
that a second flash on the same sample produced a very simi-
lar “after-minus-before” difference spectrum (not shown) with
an integrated CO2 peak area of about 75% of the first one. The
photolytic efficiency might be somewhat underestimated due
to the fact that some CO2 will have reacted with water before
the first IR spectrum is obtained (see below).
Results and Discussion
Synthesis of derivatives
The synthesis of the 3-nitrophenylacetic acids started from 1-
(chloromethyl)-3-nitrobenzene via the corresponding nitriles
(obtained with KCN or K13CN), which were then hydrolyzed
with H2O or H218O. The diacid NPDAA was prepared in a similar
fashion by starting from 5-nitroisophthalic acid via the alcohol
and the corresponding bromomethyl derivatives. Mass spectra
1
and H and 13C NMR spectra confirmed the structures of the
products. The mass spectral analysis of 18O-mNPAA showed
that 18O-labeling was not complete, most likely due to expo-
sure of the substance to acidic aqueous conditions during the
final extraction. It is known that in aqueous solutions carboxyl-
ic acids exchange carbonyl oxygen atoms with the solvent.[21]
NPDAA: The photoinduced difference spectrum of the diacid
is very similar to that of mNPAA. For equal peak areas of CO2,
the peaks associated with the nitro group (arrows in Figure 1)
are smaller by about 40%. The effect indicates that under the
experimental conditions used here [flash duration (ꢀ20 ns)>
photochemical relaxation (tꢀ200 ps in mNPAA)], two mole-
cules of CO2 are indeed released from at least a proportion of
the cage molecules. From a mechanistic standpoint, it would
be intriguing to look at the decarboxylation of this diacid on
the ultrafast time scale and with theoretical approaches. In par-
ticular, it would be interesting to see whether both carboxy-
lates can be cleaved simultaneously from the excited state.
This, however, would require ultrafast spectroscopic techniques
Photolysis of caged CO2 compounds
Solutions of all compounds were photolyzed by use of a UV
light pulse (308 nm, ꢀ20 ns, 220 mJ pulse energy) from a XeCl
excimer laser. Although the absorbance of mNPAA at 308 nm
is only ꢀ25% of that at the maximum at 270 nm,[19] the use of
this wavelength avoids excessive absorption by proteins in the
sample. Figure 1 shows the difference spectra obtained in each
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2013, 14, 372 – 380 373

CHEMBIOCHEM
FULL PAPERS
www.chembiochem.org
and much shorter laser pulses, and was not the focus of the
present study.
18O-mNPAA: Two effects are immediately apparent. Hetero-
geneous labeling of the cage compound, as indicated by mass
spectrometry, yields both C18O2 (2309 cmÀ1) and C16O18O
(2327 cmÀ1) in a ratio of ꢀ80:20. C16O2 is initially almost com-
pletely absent. Also, the corresponding carboxylate peaks are
downshifted, so that one of them coincides with the nitro
group peak at 1358 cmÀ1, whereas the other carboxylate peak
at 1557 cmÀ1 almost coincides with the second nitro group
peak at 1540 cmÀ1.
13C-mNPAA: 13CO2 exhibits an isotope effect on the frequency
of the antisymmetric CO2 stretching vibration that is even
larger than for C18O2: the band is shifted towards 2278 cmÀ1
.
The corresponding carboxylate peaks are also further down-
shifted and now coincide with those of the nitro group
(1358 cmÀ1, 1529 cmÀ1).
Figure 2. Reaction behavior of photolytically released CO2 (t=0 s) under dif-
ferent conditions (100 mm Na-HEPES buffer, 295 K), monitored at 2343 cmÀ1
Except for the measurement with uninhibited carbonic anhydrase, curves
were scaled to match with that of mNPAA at pH 7.5 without additives, to
.
Under the experimental conditions, in particular at pH 7.3,
the released CO2 will react with water to form bicarbonate via
carbonic acid. The end point of the reaction after 120 s is also
account for small variations in optical path length and pulse intensity in the
different samples.
shown in Figure 1. Bicarbonate causes very broad bands with
À
maxima around 1650 cmÀ1 (1600 cmÀ1 for H13CO3
) and
À
1360 cmÀ1 (1330 cmÀ1 for H13CO3 ).
Hydration of CO2
After its photolytic release from the caged compounds, CO2
will re-equilibrate in water, forming bicarbonate via carbonic
acid. At pH 7.5 as used here, the main pathway is the reaction
between CO2 and water, not hydroxide, so only the equilibrium
concentrations, and not the equilibration rate, will depend on
pH. The new equilibrium concentration of CO2 at pH 6.5 will
thus be higher than that at pH 7.5, but the equilibrium will be
attained at the same rate.
This can be clearly seen in Figure 2. Exponential fits of the
decay curves in all cases yield an observed pseudo first-order
rate constant of k=0.077 sÀ1 (t=13 s). The literature gives a
value of 0.043 sÀ1 (258C),[22] which is in quite good agreement
with the number we found, considering that no efforts were
made to correct for the contribution of the reverse reaction or
a likely catalytic effect of the buffer.
Figure 3. Time course of isotope exchange of C18O2 after photolytic release
from 18O-mNPAA: 50 mm 18O-mNPAA in 100 mm phosphate buffer, pH 5.5,
293 K was flashed after baseline acquisition (at 51.2 s).
In many organisms hydration of CO2 is greatly accelerated
by the action of carbonic anhydrase.[23] This effect can be clear-
ly seen in Figure 2: upon addition of 200 mgmLÀ1 of carbonic
anhydrase (c=7 mm) to the sample, the equilibrium is reached
before the first spectrum could be obtained, but the final equi-
librium level of CO2 is the same as before. When acetazol-
amide, an inhibitor of carbonic anhydrase,[23] was also added
(20 mm), the effect of the enzyme was suppressed, as would be
expected.
mixed C18O16O intermediate, finally resulting in close to 100%
C16O2.
Reactions between CO2 and amines
Reactions between CO2 and amines are of great technical im-
portance because they represent the major method by which
CO2 is selectively removed from exhaust gases. From the re-
sulting carbamates, CO2 can be recovered in a concentrated
form and processed further.[1] The formation of carbamates is
also important in enzyme catalysis in, for example, the forma-
tion of the functioning active sites in RuBisCO,[24] urease,[25] and
structurally related enzymes, the reaction mechanisms of
biotin-dependent enzymes,[26] or the regulation of hemoglobin
A demonstration for the reaction between CO2 and water is
given in Figure 3. 18O-mNPAA was photolyzed at acidic pH
(100 mm phosphate buffer pH 5.5 in H216O), so only a small
fraction of the released CO2 would have been converted to bi-
carbonate in the final equilibrium. Nevertheless, continuing re-
action with water leads to the disappearance of C18O2 via the
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2013, 14, 372 – 380 374

CHEMBIOCHEM
FULL PAPERS
www.chembiochem.org
activity.[27] We therefore investigated the reaction behavior of
photolytically released CO2 with a primary, a secondary, and a
tertiary amine: ethanolamine (EA), N-ethyl-aminoethanol (EAE),
and N,N-diethyl-aminoethanol (DEAE). The primary and secon-
dary amines should yield carbamates under the chosen reac-
tion conditions, whereas the tertiary amine, lacking the ability
to deprotonate at the nitrogen atom after nucleophilic attack
on CO2, should not. Figure 4 shows the results. The photode-
carboxylation of mNPAA is unaffected even in the presence of
500 mm amine, as can be seen from the prominent negative
carboxylate bands and also the CO2 peak in the case of DEAE.
spectral region between 1520 and 1420 cmÀ1. This is a clear
difference from the photolysis of the cage alone, which gives
no signal in this spectral region. Most of these bands shift
upon isotopic replacement of 12CO2 with 13CO2, and must
therefore involve vibrational contributions from the CO2-de-
rived carbon atom.
In the case of the tertiary amine the CO2 peaks are clearly
visible, unlike in the cases of the primary and secondary
amines. These peaks also decay, albeit slowly, with a rate con-
stant of ꢀ0.14 sÀ1, due to the reaction between CO2, water,
and hydroxide under the basic conditions (pH 9.2) in the
sample. However, they are accompanied not by distinct posi-
tive product bands between 1520 and 1420 cmÀ1, but rather
only by the very broad bands of bicarbonate around 1650 and
1360 cmÀ1 in a result very similar to those seen in the spectra
obtained from the cage alone (Figure 1).
Binding of CO2 to RuBisCO
One application for the caged CO2 is the investigation of car-
boxylating enzymes. We first investigated the basic process of
CO2 binding to the active site of RuBisCO. Even though CO2 is
the carboxylating reagent of ribulose-1,5-bisphosphate during
turnover, one extra molecule of CO2 is first required for the
generation of a fully functional active site. It carboxylates a con-
served lysine residue to form a carbamate, which then acts as
a binding ligand for a magnesium ion that is required for activ-
ity.[24,30–32] We set out to observe this process of initial CO2-
binding directly, and investigated the reaction behavior of
photolytically released CO2 in the presence of two different Ru-
BisCO species and one control protein—bovine serum albumin
(BSA). The last of these should not show any specific reactivity
towards CO2.
Figure 4. Reactions between amines and photolytically released CO2. Spectra
were obtained during the 5 s immediately following photolysis (500 mm
amine, 200 mm sodium borate/HCl buffer, pH 9.2, 275 K). c: mNPAA, ····:
13C-mNPAA. The spectra are stacked on the absorbance scale to avoid over-
lap.
For RuBisCO from Spinacea oleracea (spinach, SoRuBisCO)
the purification procedure we used yields an enzyme that has
only 1% of the activity of the fully activated enzyme, due to
the lack of carbamate in the active site.[33] In fact, the standard
coupled activity assay of SoRuBisCO shows a distinct lag phase
in the NADH consumption, indicating that the bicarbonate in
the activity assay first has to activate the enzyme before sub-
strate turnover can start. Recombinant RuBisCO from Rhodo-
spirillum rubrum (RrRuBisCO), on the other hand, produced
linear NADH consumption plots without any lag phase in the
same assay, meaning either that the activation in this case is
so fast that it cannot be observed during the activity assay, or
that the enzyme has its carbamate already in place due to
more favorable binding equilibrium constants. These two pos-
sibilities cannot be distinguished easily by this assay because
the activating agent, CO2, is also the substrate of the subse-
quent reaction.
Both the primary and the secondary amine react with the re-
leased CO2 on a timescale that is too fast for our experimental
setup. Even at a maximum time resolution of 50 ms no trace
of any CO2 peak could be detected (not shown). If we conser-
vatively estimate that in the noisier individual spectra obtaina-
ble after 50 ms we would be able to detect a CO2 peak if it
were more than 5% of the originally released ꢀ4 mm CO2, we
obtain a lower limit for the bimolecular rate constant of carba-
mate formation of 0.12 m3 molÀ1 sÀ1. This is in agreement with
literature data for the reaction between ethanolamine and CO2
in aqueous solution—1.5 m3 molÀ1 sÀ1 (277 K).[28]
Precise assignment of the product peaks in the region be-
tween 1800–1200 cmÀ1 to their corresponding molecular vibra-
tions is very challenging. Even though the formation of carba-
mate seems fairly simple, there are accompanying side reac-
tions that all contribute to the IR difference signal. These in-
clude formation of bicarbonate and carbonate, amine protona-
tion, buffer response, and the formation of hydrogen bonds
with the solvent. Furthermore, the hydroxy groups of the
amino alcohols chosen here were shown—at least in a theoreti-
cal study—to form intramolecular hydrogen bonds with the
carbamates.[29] At the very least, we can conclude that positive
product peaks appear especially are particularly evident in the
In the IR measurements with photolytically released CO2,
both BSA and RrRuBisCO yielded results very similar to those
seen with the cage alone: the initially high IR absorbance of
photolytically released CO2 slowly decreased until equilibrium
was reached (Figure 5). This implies that the RrRuBisCO must
indeed already have a fully functioning active site with a carba-
mate in place, as indicated by the absence of a lag phase in
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2013, 14, 372 – 380 375

CHEMBIOCHEM
FULL PAPERS
www.chembiochem.org
Figure 6. Stacked infrared spectra of the substrate (a: phosphoenolpyru-
vate) and the products (····: phosphate and c: oxaloacetate) of TspPEPC
(100 mm in aqueous solution, pH 8.5, water background subtracted). Arrow
positions indicate the wavenumbers used for kinetic observations.
Figure 5. Reaction behavior of CO2 in the presence of different proteins
5 s after photolytic CO2 release (20 mm mNPAA, 100 mm HEPES, pH 7.5,
280 K, protein concentrations: c: BSA 150 mgmLÀ1, a: RrRuBisCO
220 mgmLÀ1, 4.1 mm active sites; ····: SoRuBisCO 150 mgmLÀ1, 2.2 mm active
sites). The spectra are stacked on the absorbance scale to avoid overlap.
for each reaction partner (Figure 7). In the same assay with
0.01 mgmLÀ1 carbonic anhydrase, the reaction is complete
within the dead time of the instrument, due to the high
TspPEPC concentration (not shown). This shows that TspPEPC
uses bicarbonate rather than CO2 as its substrate, which is in
accordance with the literature.[35] Without carbonic anhydrase,
the observed reaction rates are limited by the hydration of
CO2.
the activity assay. The behavior would then be no different
from that in the case of BSA.
In contrast, in the sample with SoRuBisCO the re-equilibra-
tion was too fast to be observed with the experimental setup.
The resulting IR spectrum shows only the broad bicarbonate
bands around 1650 and 1360 cmÀ1. Unexpectedly, no carba-
mate-associated bands between 1520 and 1420 cmÀ1 are visi-
ble above noise level (compare Figure 4). The reason for this is
unclear. Contamination of the preparation with carbonic anhy-
drase is unlikely because the results are the same in the pres-
ence of 50 mm of the carbonic anhydrase inhibitor acetazol-
amide (not shown).[34] A possible explanation would be an in-
trinsic carbonic anhydrase-like activity of SoRuBisCO, which, to
the best of our knowledge, has not been described before. We
note, though, that the protein concentrations typically em-
ployed in the study of SoRuBisCO are much lower than those
we used here (2.2 mm of active sites), which would make the
effect easy to overlook. Definite confirmation of a carbonic-
anhydrase-like activity of SoRuBisCO requires further investiga-
tion.
Reaction between CO2 and phosphoenolpyruvate
carboxylase
To demonstrate the possibility of observing enzymatic turnover
with the caged CO2, we chose phosphoenolpyruvate carboxy-
lase (PEPC). This enzyme catalyzes the primary CO2 fixation in
the photosynthetic pathways of C4 plants and Crassulaceae
and in certain microorganisms provides oxaloacetate for the
citric acid cycle through an anaplerotic reaction.[35] For our pur-
poses, it has the advantage that substrates and products are
conveniently distinguishable by characteristic infrared spectral
features (Figure 6).
Figure 7. Time courses for four components in a reaction mixture containing
0.1m HEPES/HCl, pH 8.5, 10 mm MgCl2, 20 mm potassium phosphoenolpyru-
vate, 0.2 mm acetyl-coenzyme A, and 14 mgmLÀ1 TspPEPC. Dotted lines:
single exponential fits.
Conclusions
In this study we have demonstrated that photodecarboxylating
reagents (“caged CO2”) are indeed useful tools for the investi-
gation of reactions involving carbon dioxide. They allow us to
obtain shifts of the CO2 concentration in solution on the nano-
We recombinantly expressed PEPC from Thermus sp.
(TspPEPC) in E. coli.[36] The photolytic release of CO2 starts the
reaction, which can be monitored at wavelengths characteristic
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2013, 14, 372 – 380 376

CHEMBIOCHEM
FULL PAPERS
www.chembiochem.org
rapid-scan mode with an acquisition rate of 180 kHz and a spectral
resolution of 4 cmÀ1. The Fourier transform was carried out with
Blackman–Harris three-term apodization and a zero-filling factor of
two.
second timescale. Because, immediately after photolysis, the
system is far from the slow CO2/bicarbonate equilibrium, much
higher CO2 concentrations can be reached over a wide pH
range than by the more usual application of bicarbonate. If, on
the other hand, the photolytically triggered, fast release of bi-
carbonate is desired, carbonic anhydrase can be added to the
sample to accelerate the hydration of the CO2.
Static FTIR spectroscopy: Static FTIR absorption spectra of
TspPEPC substrates and products were recorded with an Alpha
spectrometer (Bruker Optik, Ettlingen), fitted with an in-house built
ATR unit,[37] with water as background. Spectra were calculated
from the averages of 128 scans with 2 cmÀ1 spectral resolution by
Fourier transformation with the inbuilt parameters.
Once CO2 has been released, the course of its reactions can
be easily monitored by IR spectroscopy. Its antisymmetric
stretching vibration yields an intense signal at 2343 cmÀ1: that
is, in a part of the IR spectrum where only very few other com-
pounds contribute. IR bands associated with reaction products
can also be monitored, although here the precise assignment
to species and vibrational modes is more difficult. As a tool for
this assignment, the isotopically labeled variants of mNPAA
presented here should be very useful. Quantum chemical ap-
proaches can also be considered for calculating normal modes
and guiding IR band assignment.
Protein purification: Bovine serum albumin and carbonic anhy-
drase from bovine erythrocytes were purchased from Sigma–Al-
drich Chemie GmbH (Munich) and used without further purifica-
tion.
SoRuBisCO: RuBisCO from S. oleracea was purified from fresh spi-
nach obtained from the local supermarket with a modified proce-
dure based on that described by Salvucci et al.[33] Instead of
a Mono-Q column for anion exchange chromatography, we used
a Q-Sepharose Fast Flow column (GE Healthcare, Munich). No acti-
vation through addition of bicarbonate was carried out after the
purification, and buffer exchange with Amicon centrifugation devi-
ces was also carried out without added bicarbonate.
The very fast photolysis of the compounds that we used as
caged CO2 should permit the study of reactions on the nano-
second timescale. With current FTIR technology we are able to
observe the kinetics of reactions with time constants of
ꢀ20 ms, permitting the observation of CO2 hydration or
enzyme-catalyzed product formation if the enzyme concentra-
tion is sufficiently low. Here the caged CO2 can be applied in
a universal activity assay for carboxylating enzymes.
RrRuBisCO: RuBisCO from R. rubrum was produced by heterologous
expression in E. coli and by a modified procedure based on that by
Somerville.[38] The gene for the protein was ordered in a proprietary
plasmid conferring ampicillin resistance from Mr. Gene GmbH (Re-
gensburg) based on the amino acid sequence of UniProt entry
Q2RRP5. The gene was under the control of a T7 promoter, and its
codon usage had been optimized for expression in E. coli. Compe-
tent cells of E. coli strain BL21(DE3) were transformed with this
plasmid and grown in terrific broth (TB) up to an OD600 of ꢀ0.8. Ex-
pression of the RrRuBisCO gene was induced by addition of isopro-
pyl b-d-1-thiogalactopyranoside (IPTG, final concentration 0.5 mm)
and incubation at 378C overnight. Harvested cells were disrupted
by sonication or in a bead mill, and particulate matter was re-
moved from the lysate by ultracentrifugation (140000g, 45 min).
The supernatant was removed, and protamine sulfate was added
to a final concentration of 0.3% (w/v) from a 3% (w/v) stock solu-
tion. The precipitate was removed by ultracentrifugation
(100000g, 15 min). The supernatant was brought to 508C for
10 min in a water bath and then cooled on ice. The precipitate
was again removed by ultracentrifugation (100000g, 15 min), the
supernatant was applied to a Q-Sepharose fast flow column (GE
Healthcare) equilibrated with buffer [Tris·HCl (pH 8, 50 mm), MgCl2
(50 mm)], dithiothreitol (DTT, 1 mm), and the protein was eluted
with a rising NaCl gradient in the same buffer (0–400 mm NaCl,
three column volumes). RrRuBisCO-containing fractions (as judged
by an SDS-PAGE gel) were pooled, concentrated to a final volume
of 3 mL, applied to a Sephacryl S75 column (GE Healthcare,
Munich) equilibrated with buffer [HEPES (pH 7.5, 20 mm), NaCl
(100 mm)], and eluted with the same buffer. Again, the RrRuBisCO-
containing fractions were pooled and diluted to a final conductivi-
ty of <1 mScmÀ1. The protein was then applied to a Source Q
column (GE Healthcare, Munich) equilibrated with buffer [HEPES
(pH 7.5, 20 mm)] and eluted with a rising NaCl gradient in the
same buffer (0–300 mm NaCl, five column volumes). RrRuBisCO-
containing fractions were pooled, concentrated, and frozen at
À708C until use.
Further technical developments should open up new possi-
bilities for the direct observation of even faster reactions. In
particular, the advent of new tunable quantum cascade lasers
as intense IR light sources should allow the observation of very
fast reaction kinetics at specific and variable wavelengths.
Our approach could even be applied to the direct spectro-
scopic observation of enzyme-bound reaction intermediates.
For that purpose, the enzyme concentration must be very high
to yield a sufficiently strong IR signal, and substrate consump-
tion would therefore normally be very rapid. There are still sev-
eral possible means to observe otherwise short-lived species:
the reaction can be carried out at a lower temperature where
turnover is slow. In this case, enzymes from thermophilic or-
ganisms can be used advantageously. Reaction intermediates
can also be trapped kinetically, for example, if a second sub-
strate that is required for turnover is omitted from the sample.
Finally, genetically engineered enzyme variants that are
blocked at certain reactions steps could allow the accumula-
tion of reaction intermediates.
Experimental Section
Flash photolysis and rapid-scan FTIR spectroscopy: Rapid-scan
FTIR spectra were recorded with a modified IFS 66 spectrometer
(Bruker Optik, Ettlingen) and HgCdTe detector. Samples were pre-
pared in CaF2 cuvettes with a nominal optical path length of 5 mm
and mounted in a temperature-controlled sample holder. Photoly-
sis was induced with a XeCl excimer laser pulse at 308 nm focused
on the sample (tꢀ20 ns, Q=220 mJ, Model RD-EXC-200, Radiant
Dyes Laser & Accessories GmbH, Wermelskirchen). The pulse was
triggered by use of the spectrometer software OPUS and synchron-
ized with IR spectrum acquisition. IR spectra were obtained in
TspPEPC: TspPEPC from T. sp. was produced by heterologous ex-
pression in E. coli. The gene for the protein was ordered in a propri-
etary plasmid conferring ampicillin resistance from Mr. Gene, based
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2013, 14, 372 – 380 377

CHEMBIOCHEM
FULL PAPERS
www.chembiochem.org
on the amino acid sequence of UniProt entry P51060. The gene
was under the control of a T7 promoter, and its codon usage had
been optimized for expression in E. coli. Competent cells of E. coli
strain BL21(DE3) were transformed with this plasmid and grown
overnight in medium (500 mL). Cell were harvested, washed in M9
medium, and used to inoculate 10 L of M9 medium. Cells were
grown up to an OD600 of ꢀ0.8. Expression of the TspPEPC gene
was induced by addition of IPTG (final concentration 0.2 mm) and
incubation at room temperature overnight. Harvested cells were
resuspended in buffer [Tris·HCl (pH 8, 50 mm), DTT (1 mm), ethyl-
enediaminetetraacetic acid (EDTA, 1 mm)] and disrupted by sonica-
tion. Particulate matter was removed from the lysate by ultracentri-
fugation (140000 g, 45 min). The supernatant was removed and
(NH4)2SO4 (40%, w/v) was added. The precipitate was collected by
centrifugation. The pellet was dissolved in buffer [Tris·HCl (pH 8,
50 mm), DTT (1 mm), EDTA (1 mm)], dialyzed against this buffer (2ꢁ
2 L) overnight, and centrifuged again to remove insoluble protein.
The supernatant was applied to a Q-Sepharose fast flow column
(GE Healthcare) equilibrated with buffer [Tris·HCl (pH 8, 50 mm),
DTT (1 mm), EDTA (1 mm)], and the protein was eluted with
a rising NaCl gradient in the same buffer (0–500 mm NaCl, three
column volumes). TspPEPC-containing fractions (as judged by an
SDS-PAGE gel) were pooled, concentrated to a final volume of
3 mL, applied to a Sephacryl S100 column (GE Healthcare) equili-
brated with buffer [Tris·HCl (pH 8, 50 mm), NaCl (150 mm)], and
eluted with the same buffer. TspPEPC-containing fractions were
pooled, concentrated, and frozen at À708C until use.
(Fisons), respectively. IR spectra were recorded with a Nicolet 6700
FTIR spectrometer. Melting points were determined with a “Dr. Tot-
toli apparatus” (Buchi Corporation) without correction.
2-(3-Nitrophenyl)[1-13C]acetonitrile (1): K13CN (0.42 g, 6.24 mmol) was
added to a solution of 1-(chloromethyl)-3-nitrobenzene (1.00 g,
5.85 mmol) and [18]crown-6 (0.15 g, 0.57 mmol) in acetonitrile
(10 cm3), and the bluish suspension was stirred for 24 h at 258C.
Methylene dichloride (25 cm3) was then added, and the now
brownish suspension was filtered. The filtrate was washed with
water (2ꢁ25 cm3), dried (MgSO4), and concentrated under reduced
pressure to afford a brownish oil, which was used in the following
step without further purification.
2-(3-Nitrophenyl)[1-13C]acetic acid (2, 13C-mNPAA, Scheme 1): A solu-
tion of crude 1 in half-concentrated HCl (20 cm3) was heated to
1008C for 18 h. After the system had cooled down, ethyl acetate
(20 cm3) was added, and the mixture was basified with aqueous
NaOH. The resulting combined aqueous extracts were acidified
with aqueous HCl and extracted with methylene dichloride. The
combined organic layers were dried over MgSO4 and concentrated
to afford product 2 (0.46 g, 45%) as a slightly brownish solid.
1H NMR (CDCl3, 250 MHz): d=8.15–8.18 (m, 2H), 7.62–7.65 (m, 1H),
7.49–7.56 (m, 1H), 3.79 ppm (d, J=8.0 Hz, 1H); 13C NMR (CDCl3,
250 MHz): d=176.29, 135.80, 135.15, 129.73, 124.67, 122.72, 116.65,
40.44 ppm (d, J=55.7 Hz).
2-(3-Nitrophenyl)acetonitrile (3): Potassium cyanide (3.8 g,
57.1 mmol) was added to a solution of 1-(chloromethyl)-3-nitro-
benzene (10.0 g, 58.5 mmol) and [18]crown-6 (1.5 g, 5.7 mmol) in
acetonitrile (100 cm3), and the greenish suspension was stirred for
18 h at 258C. Methylene dichloride (50 cm3) was then added, and
the now brownish suspension was filtered. The filtrate was washed
with water (2ꢁ50 cm3), dried (MgSO4), and concentrated under
reduced pressure to yield a brownish oil, which was fractionally
distilled in vacuo to afford product 3 (4.4 g, 48%) as a yellowish oil,
containing [18]crown-6 (6%). 1H NMR (CDCl3, 250 MHz): d=8.21–
8.23 (m, 2H), 7.71–7.74 (m, 1H), 7.57–7.64 (m, 1H), 3.89 ppm (s,
2H).
For use in the FTIR measurements, protein samples were buffer-ex-
changed and concentrated with Amicon centrifugation devices
and the buffer specific for the measurements, which also contained
the caged compound.
RuBisCO activity assays: RuBisCO activity was tested with a modi-
fied method based on that of Norton et al. in a Hitachi U-2000
spectrophotometer.[39] To avoid primary amines that might interfere
with carbamate formation, we used 4-(2-hydroxyethyl)piperazine-1-
ethanesulfonic acid (HEPES, pKa 7.5) as a buffer instead of tris(hy-
droxymethyl)aminomethane (Tris).
2-(3-Nitrophenyl)[18O]acetic acid (4, 18O-mNPAA, Scheme 1): A solu-
tion of 3 (0.30 g) in H218O (1 cm3) in a screw-cap vial was exposed
to a stream of gaseous HCl for 2 h until the solution solidified. The
vial was sealed and heated for 24 h to 1008C. After the system had
cooled down, ethyl acetate was added, and the mixture was basi-
fied with aqueous NaOH. The resulting combined aqueous extracts
were acidified with aqueous HCl and extracted with methylene
dichloride. The combined organic layers were concentrated under
reduced pressure to afford 4 (0.4 g, 114%) as a slightly brownish
solid still containing some water. 1H NMR (CDCl3, 250 MHz): d=
8.16–8.18 (m, 2H), 7.62–7.65 (m, 1H), 7.49–7.56 (m, 1H), 3.79 ppm
(d, J=8.0 Hz, 1H); MS (ESÀ): m/z (%): 135.4 (100), 181.6 (6.8)
[MÀH]À 18O/16O, 183.7 (17.6), [MÀH]À 18O/18O.
TspPEPC activity assay: TspPEPC activity was tested in a modified
coupled assay with malate dehydrogenase[40] in the following mix-
ture: Tris acetate (pH 8.5, 0.1m), KHCO3 (10 mm), MgSO4 (10 mm),
potassium phosphoenolpyruvate (2 mm), acetyl-coenzyme A
(0.2 mm) as an allosteric activator, NADH (0.1 mm), malate dehydro-
genase (5 IU), TspPEPC (0.1–0.2 mgmLÀ1). The oxidation of NADH
to NAD was followed with a Hitachi U-2000 spectrophotometer. To
avoid primary amines that might interfere with carbamate forma-
tion, we used HEPES (pKa 7.5) as a buffer instead of Tris for the IR
measurements.
Syntheses
General: Unless otherwise noted, all reagents were obtained com-
mercially and used without further purification. Solvents for chro-
matography were technical grade and distilled prior to use. Sol-
vents for reactions were purchased as reagent grade and distilled
prior to use. Analytical TLC was performed on Macherey–Nagel
Polygram SIL G/UV254 precoated plastic sheets for TLC, and visuali-
zation was achieved by irradiation with UV light. Column chroma-
tography was performed with silica gel (Merck 60, particle size
0.040–0.063 mm). Solvent mixture ratios are understood as
(5-Nitro-1,3-phenylene)dimethanol (5): A solution of 5-nitroisophthal-
ic acid (4.3 g, 20.4 mmol) in tetrahydrofuran (25 cm3) was cooled to
08C, and borane-tetrahydrofuran (1n, 100 cm3, 100 mmol) was
added dropwise over 1 h. The mixture was allowed to warm up
slowly to 258C and stirred for 36 h. Methanol (20 cm3) was added
slowly, and the mixture was filtered and concentrated. The residue
was dissolved in ethyl acetate (30 cm3), washed with water (2ꢁ
15 cm3), dried (MgSO4), filtered, and concentrated to afford the
known compound 5 (3.6 g,98%) as a yellow solid, which was used
in the following step without further purification. 1H NMR (CDCl3,
250 MHz): d=8.15 (s, 2H), 7.72 (s, 1H), 4.82 ppm (s, 4H) was in line
with the literature.[41]
1
volume/volume. H NMR and 13C NMR spectra were recorded with
Bruker AM 250/AV 300 instruments in CDCl3/[D6]DMSO. ESI-MS/
MALDI-TOF mass spectra were obtained with a VG Platform II in-
strument with a Quadrupol Analyzer and a VG Tofspec instrument
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2013, 14, 372 – 380 378

CHEMBIOCHEM
FULL PAPERS
www.chembiochem.org
1,3-Bis(bromomethyl)-5-nitrobenzene (6): Phosphorus tribromide
(1.9 cm3, 20 mmol) was added dropwise to a solution of crude 5 in
benzene (60 cm3). The mixture was stirred for 30 min at 258C and
then heated to reflux for 2.5 h. After cooling down, the mixture
was poured onto ice and extracted with diethyl ether (2ꢁ15 cm3).
The combined organic extracts were then washed with water
(20 cm3) and brine (20 cm3), dried (MgSO4), filtered, and concentrat-
ed under reduced pressure to afford the product 6 (5.8 g, 94%) as
a slightly yellowish solid, which was used in the following step
1
without further purification. H NMR (CDCl3, 250 MHz): d=8.19 (s,
2H), 7.75 (s, 1H), 4.52 ppm (s, 4H) was in line with the literature.[42]
2,2’-(5-Nitro-1,3-phenylene)diacetonitrile (7): [18]Crown-6 (0.18 g,
0.67 mmol), KCN (0.49 g, 7.36 mmol), water (3.2 cm3), and a small
amount of KI were added at room temperature to a solution of 6
(1.03 g, 3.35 mmol) in acetonitrile (40 mL). After having been
stirred for 72 h, the reaction mixture was worked up with water
and ethyl acetate. The combined organic extracts were washed
with brine, dried (MgSO4), and concentrated under reduced pres-
sure. Column chromatography (SiO2) of the residue with hexane/
ethyl acetate as the eluent provided 7 (0.47 g, 59%) as a yellowish
solid. M.p. 110–1128C; 1H NMR (CDCl3, 250 MHz): d=8.22 (s, 2H),
7.72 (s, 1H), 3.92 ppm (s, 4H); 13C NMR (CDCl3, 250 MHz): d=
147.42, 133.47, 133.37, 123.11, 116.16, 23.60 ppm; IR (ATR): n˜ =
3073.95, 3041.51, 2954.50, 2925.33, 2324.25, 2257.48, 1782.36,
1745.31, 1623.88, 1532.37, 1464.76, 1440.71, 1409.91, 1346.95,
1321.31, 1308.39, 1238.25, 1213.66, 1157.27, 1102.96, 1013.54,
996.43, 951.90, 927.58, 898.32, 853.15, 784.55, 741.64, 664.19,
656.29 cmÀ1; MS (ESÀ): m/z: 199.9 [MÀH]À; elemental analysis
calcd (%) for C10H7N3O2: C 59.70, H 3.51, N 20.89; found: C 59.76, H
3.42, N 21.12.
Figure 8. Single-crystal structure of 7. Displacement ellipsoids are drawn at
50% probability level. Nitrile group bond lengths amount to 1.1392(16) ꢃ
(dN2···C8) and 1.1394(16) ꢃ (dN3···C10), in line with the literature.[45]
2,2’-(5-Nitro-1,3-phenylene)diacetic acid (8, NPDAA, Scheme 1): Half-
concentrated HCl (20 cm3) was added to 7 (0.45 g, 2.24 mmol), and
the solution was heated to reflux for 16 h. After the system had
cooled down, ethyl acetate was added, and the mixture was basi-
fied with aqueous NaOH. The resulting combined aqueous extracts
were acidified with aqueous HCl and extracted with ethyl acetate.
The combined organic layers were dried (MgSO4) and concentrated
under reduced pressure to afford 8 (0.52 g, 97%) as a slightly
yellowish solid. M.p. decomposition at 203–2048C; 1H NMR
([D6]DMSO, 300 MHz): d=12.51 (brs, 1H), 8.06 (s, 2H), 7.62 (s, 1H),
3.77 ppm (s, 4H); 13C NMR ([D6]DMSO, 300 MHz): d=171.99,
147.58, 137.53, 137.08, 122.65, 39.64 ppm; IR (ATR): n˜ =2897.43,
1691.10, 1536.90, 1444.82, 1415.72, 1346.62, 1300.76, 1275.75,
1254.05, 1226.63, 1184.04, 1147.21, 1100.20, 982.91, 915.81, 875.43,
813.14, 765.03, 744.59, 723.08, 668.75 cmÀ1; MS (MALDI-TOF): m/z:
262.836 [M+Na]+, 278.969 [M+K]+.
Crystals of this compound were obtained as colorless blocks by
slow evaporation of a CDCl3 solution. Crystal data for 7 were ob-
tained with a STOE IPDS-II diffractometer and graphite-monochro-
mated MoKa radiation. The structures were solved by direct meth-
ods by use of the program SHELXS[43] and refined with full-matrix,
least-squares on F2 with use of the program SHELXL-97.[44]
CCDC 850844 contains the supplementary crystallographic data for
this paper. These data can be obtained free of charge from the
Cambridge Crystallographic Data Centre via http://www.ccdc.ca-
m.ac.uk/data_request/cif.
Acknowledgements
The authors thank Hans-Werner Mꢁller for excellent technical as-
sistance, Fabian Dornuf for help with measurements, Prof.
Alexander Heckel for advice, and the Adolf-Messer-Foundation
and the “Fçrderfonds” of the Goethe University Frankfurt for fi-
nancial support.
¯
Crystal data for 7: Triclinic, C10H7N3O2, space group P1 with a=
8.0125(10) ꢃ, b=8.0705(10) ꢃ, c=8.8205(11) ꢃ, V=461.43(10) ꢃ3,
Z=2, 1calcd =1.448 mgmÀ13, m=0.105 mmÀ1, and F(000)=208. Crys-
tal size: 0.37ꢁ0.36ꢁ0.34 mm3. Independent reflections: 1717 with
R
int =0.0303. The structure was solved by direct methods (SHELXS-
Keywords: caged CO2 · carboxylation · enzyme catalysis · IR
spectroscopy · photolysis
97) and refined by use of full-matrix, least-squares difference Fouri-
er techniques. The final agreement factors are R1 =0.0333 and
wR2 =0.0936 [I>2s(I)].
[1] Working Group III of the IPCC in IPCC Special Report on Carbon Dioxide
Capture and Storage (Eds.: B. Metz, O. Davidson, H. de Coninck, M. Loos,
L. Meyer), Cambridge University Press, Cambridge, 2005.
[2] R. J. Ellis, Trends Biochem. Sci. 1979, 4, 241–244.
[3] S. M. Glueck, S. Gꢄmꢄs, W. M. F. Fabian, K. Faber, Chem. Soc. Rev. 2010,
39, 313–328.
[4] M. Falk, A. G. Miller, Vib. Spectrosc. 1992, 4, 105–108.
[5] M. Muthusamy, M. R. Burrell, R. N. F. Thorneley, S. Bornemann, Biochem-
istry 2006, 45, 10667–10673.
[6] M. E. Riepe, J. H. Wang, J. Biol. Chem. 1968, 243, 2779–2787.
[7] A. P. Pelliccioli, J. Wirz, Photochem. Photobiol. Sci. 2002, 1, 441–458.
[8] A. Barth in Dynamic Studies in Biology—Phototriggers, Photoswitches and
Caged Biomolecules (Eds.: M. Goeldner, R. Givens), Wiley-VCH, Weinheim,
2005, pp. 369–399.
As can be seen in Figure 8, the molecule is approximately planar.
The planarity is disturbed by a slight twist of the CH2CN substitu-
ents: C(4)-C(3)-C(7)-C(8)=171.69(10)8 and C(4)-C(5)-C(9)-C(10)=
176.16(10)8. The unit cell consists of two molecules interconvertible
through an inversion center. The molecules form layers consisting
of staggered chain-like structures along [010]: the distance be-
tween the molecules in a chain-like structure amounts to 2.629 ꢃ
(dO2···H4), and the distance between chain-like structures to 9.810 ꢃ
(dN1···N1’). The layers are oriented parallel to (101) and alternate like-
¯
wise through an inversion center accordant to the P1 space group.
The layer distance amounts to 3.619 ꢃ.
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2013, 14, 372 – 380 379

CHEMBIOCHEM
FULL PAPERS
www.chembiochem.org
[9] G. C. R. Ellis-Davies, Nat. Methods 2007, 4, 619–628.
[10] G. Mayer, A. Heckel, Angew. Chem. 2006, 118, 5020–5042; Angew. Chem.
Int. Ed. 2006, 45, 4900–4921.
[11] H.-M. Lee, D. R. Larson, D. S. Lawrence, ACS Chem. Biol. 2009, 4, 409–
427.
[12] C. Brieke, F. Rohrbach, A. Gottschalk, G. Mayer, A. Heckel, Angew. Chem.
2012, 124, 8572–8604; Angew. Chem. Int. Ed. 2012, 51, 8446–8476.
[13] W. Mꢀntele, Trends Biochem. Sci. 1993, 18, 197–202.
[14] G. Papageorgiou, J. E. T. Corrie, Tetrahedron 1997, 53, 3917–3932.
[15] G. Papageorgiou, A. Barth, J. E. T. Corrie, Photochem. Photobiol. Sci.
2005, 4, 216–220.
[16] A. Barth, S. R. Martin, J. E. T. Corrie, Photochem. Photobiol. Sci. 2006, 5,
107–115.
[17] J. E. T. Corrie, V. R. N. Munasinghe, D. R. Trentham, A. Barth, Photochem.
Photobiol. Sci. 2008, 7, 84–97.
[18] J. D. Margerum, C. T. Petrusis, J. Am. Chem. Soc. 1969, 91, 2467–2472.
[19] K. Neumann, M.-K. Verhoefen, J.-M. Mewes, A. Dreuw, J. Wachtveitl,
Phys. Chem. Chem. Phys. 2011, 13, 17367–17377.
[20] J.-M. Mewes, K. Neumann, M.-K. Verhoefen, G. Wille, J. Wachtveitl, A.
Dreuw, ChemPhysChem 2011, 12, 2077–2080.
[31] G. H. Lorimer, M. R. Badger, T. J. Andrews, Biochemistry 1976, 15, 529–
536.
[32] I. Andersson, S. Knight, G. Schneider, Y. Lindqvist, T. Lundqvist, C.-I. Brꢀn-
dꢅn, G. H. Lorimer, Nature 1989, 337, 229–234.
[33] M. E. Salvucci, A. R. Portis, W. L. Ogren, Anal. Biochem. 1986, 153, 97–
101.
[34] M. Kandel, A. G. Gornall, D. L. Cybulsky, S. I. Kandel, J. Biol. Chem. 1978,
253, 679–685.
[35] K. Izui, H. Matsumura, T. Furumoto, Y. Kai, Annu. Rev. Plant. Biol. 2004,
55, 69–84.
[36] T. Nakamura, I. Yoshioka, M. Takahashi, H. Toh, K. Izui, J. Biochem. 1995,
118, 319–324.
[37] A. Roth, F. Dornuf, O. Klein, D. Schneditz, H. Hafner-Gießauf, W. Mꢀntele,
Anal. Bioanal. Chem. 2012, 403, 391–399.
[38] C. R. Somerville, S. C. Somerville, Mol. Gen. Genet. 1984, 193, 214–219.
[39] I. L. Norton, M. H. Welch, F. C. Hartman, J. Biol. Chem. 1975, 250, 8062–
8068.
[40] K. Terada, T. Murata, K. Izui, J. Biochem. 1991, 109, 49–54.
[41] H. A. Staab, L. Schanne, C. Krieger, V. Taglieber, Chem. Ber. 1985, 118,
1204–1229.
[21] R. L. Redington, J. Phys. Chem. 1976, 80, 229–235.
[22] B. H. Gibbons, J. T. Edsall, J. Biol. Chem. 1963, 238, 3502–3507.
[23] V. M. Krishnamurthy, G. K. Kaufman, A. R. Urbach, I. Gitlin, K. L. Gudiksen,
D. B. Weibel, G. M. Whitesides, Chem. Rev. 2008, 108, 946–1051.
[24] W. W. Cleland, T. J. Andrews, S. Gutteridge, F. C. Hartman, G. H. Lorimer,
Chem. Rev. 1998, 98, 549–562.
[42] S. Ghorai, A. Bhattacharjya, A. Basak, A. Mitra, R. T. Williamson, J. Org.
Chem. 2003, 68, 617–620.
[43] G. M. Sheldrick, Acta Crystallogr. Sect. A Found Crystallogr. 1990, 46,
467–473.
[44] G. M. Sheldrick, Acta Crystallogr. Sect. A Found. Crystallogr. 2008, 64,
112–122.
[25] E. Jabri, M. Carr, R. Hausinger, P. Karplus, Science 1995, 268, 998–1004.
[26] P. V. Attwood, J. C. Wallace, Acc. Chem. Res. 2002, 35, 113–120.
[27] G. Gros, H. S. Rollema, R. E. Forster, J. Biol. Chem. 1981, 256, 5471–5480.
[28] G. F. Versteeg, L. A. J. Van Dijck, W. P. M. Van Swaaij, Chem. Eng.
Commun. 1996, 144, 113–158.
[45] J.-Y. Le Questel, M. Berthelot, C. Laurence, J. Phys. Org. Chem. 2000, 13,
347–358.
Received: October 15, 2012
[29] E. F. da Silva, H. F. Svendsen, Ind. Eng. Chem. Res. 2006, 45, 2497–2504.
[30] W. A. Laing, J. T. Christeller, Biochem. J. 1976, 159, 563–570.
Published online on January 16, 2013
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2013, 14, 372 – 380 380