Proc. Natl. Acad. Sci. USAVol. 77, No. 12, pp. 7237-7241, December 1980Biophysics

A new approach to time-resolved studies of ATP-requiring biologicalsystems: Laser flash photolysis of caged ATP

(spectrophotometry/actomyosin ATPase/transient kinetics)

JAMES A. MCCRAY*, LEO HERBETTEt, TORU KIHARA*, AND DAVID R. TRENTHAMtBiophysics Group, Department of Physics and Atmospheric Sciences, Drexel University, Philadelphia, Pennsylvania 19104; and tDepartment of Biochemistry

and Biophysics, University of Pennsylvania, Philadelphia, Pennsylvania 19104

Communicated by Mildred Cohn, September 15, 1980

ABSTRACT 2-Nitrobenzyl derivatives have been used forseveral years as photolabile protecting groups in synthetic or-ganic chemistry. Recently, P3-1(2-nitro)phenylethyladenosine5'-triphosphate "caged ATP," was synthesized and its photolysiswas shown to generate ATP in situ. This and related reactionshave great potential for structural and kinetic studies of bothintact and soluble biological systems and it is thus importantto define the kinetic characteristics of the photolytic reaction.Caged ATP (2.5 mM) was photolyzed at 347 nm by a single 30-nsec pulse from a frequency-doubled ruby laser of 25 mJ energyto generate 500 .M ATP. The kinetics of the overall reactionwere determined by monitoring the kinetics of ATP-induceddissociation of actomyosin, a reaction of known kinetic char-acteristics. Release of 500 IAM ATP was found to be controlledby a process having a rate constant of 2.2 X log0H+J sec-1 at220C at pH 5.8-9.5, which corresponds to 220 sec- at pH 7. Thisprocess is believed to be the breakdown of an aci-nitro com-pound, which was identified on the basis of its spectral prop-erties and the photochromicity of related 2-nitrobenzyl com-pounds.

Photodissociation of carbon monoxide and dioxygen from he-moglobin (1, 2) is an important tool for analysis of hemoglobinfunctions. It would be valuable if a similar approach could bedeveloped to encompass biological reactions in general. Inprinciple, it should be possible to generate photochemicallyreactants such as metabolites from biologically inert photoly-zable compounds. This would allow us to study intact macro-molecular systems in which limitations arise because of thedifficulty of introducing reactants to their sites of reaction ina time shorter than the characteristic time of the reaction underinvestigation.

First steps toward this goal of photochemical deprotectionhave been reviewed by Knowles (3). These include the gener-ation of inorganic phosphate from 3-nitrophenyl phosphate byHavinga et al. (4) in 1956 and the work of Patchornik et al. (5)and Engels and Schlaeger (6), who described the synthesis ofphotolabile amino acid, peptide, and cyclic AMP precursorsand suggested possible applications of their methods. Alterna-tively, inhibitors can be photogenerated, as has been shown byLester et al. (7) in their studies of ion-channel blockademechanisms. These ideas were taken further by Kaplan et al.(8), who synthesized a y-phosphate ester of ATP, P3-1-(2-nitro)phenylethyladenosine 5'-triphosphate (I), which theycalled caged ATP. They showed that caged ATP was inerttoward the sodium pump, could be incorporated into vesiclesand, on continuous irradiation (on the time scale of seconds),formed ATP.

The publication costs of this article were defrayed in part by page

charge payment. This article must therefore be hereby marked "ad-vertisement" in accordance with 18 U. S. C. §1734 solely to indicatethis fact.

7237

The photolysis of I results in the formation of 2-nitrosoace-tophenone (5) and a proton, depending on the ionization stateof ATP (Eq. 1).

CHs NHC

QNO° ° P.O O WN(N)7 . (Ca + H+ ATP

NO3 - NO [1

I H

In the work described here, the amount and rate of ATPgeneration after laser flash photolysis of caged ATP was mea-sured to determine the potential of caged ATP as an ATP pre-cursor for use in kinetic studies of intact biological systems. Itis shown that the energy per pulse from a frequency doubledruby laser generates ATP in sufficient concentration for bio-logical studies. By using a well characterized biological reac-tion-the ATP-induced dissociation of actomyosin-the overallrate constant of the dark reactions that lead to the liberation ofATP on photon absorption was determined. Spectroscopicanalysis of the photolysis gives some insight into the mechanismof the conversion of.caged ATP to ATP.

MATERIALS AND METHODSActomyosin was prepared by mixing purified F actin andmyosin, both of which were isolated from rabbit skeletal muscle(9, 10).Caged ATP was synthesized by condensing 1-(2-nitro)-

phenylethyl phosphate and ADP morpholidate as described byKaplan et al. (8). 1-(2-Nitro)phenylethyl phosphate was syn-thesized from 1-(2-nitro)phenylethanol and 2-cyanoethylphosphate by the method of Tener (11). The 2-cyanoethylgroup was removed from the phosphate diester intermediateby treatment with 1 M NaOH at 100'C for 10 min, and theproduct was partitioned between water and chloroform. ThepH of the aqueous phase was adjusted to 2 by addition of Dowex50 H+, and the resulting filtrate was adjusted to pH 7.5 by ad-dition of aqueous Ba(OH)2. Some barium phosphate precipi-tated and was filtered off. Two volumes of ethanol were added,and t~he barium salt of 1-(2-nitro)phenylethyl phosphate pre-cipitated over 2 days at 5YC. The white solid was washed withethanol and ether and dried. The product (25% yield) wascharacterized as anhydrous barium 1-(2-nitro)phenylethylphosphate, Mr 382, by the UV spectrum (Xmax = 265 nm; e =4240 M-1 cm-') of the free acid, which had been titrated to pH7.5 by addition of NaOH (2.0 mol/mol). The free acid wasobtained by dissolving a weighed amount of the barium salt inwater in the presence of Dowex 50 H+. The relatively low yield

Dow

nloa

ded

by g

uest

on

Nov

embe

r 17

, 202

0

Proc. Natl. Acad. Sci. USA 77 (1980)

of the barium salt may be due to the formation of 2-nitrostyreneduring phosphate diester hydrolysis.

Assay Methods. Caged ATP was analyzed by using high-performance liquid chromatography on an NH2-Bondapakcolumn (Waters Associates) and elutitig with 0.35 M(NH4)2HP04 adjusted to pH 4.0 with HC1. In addition, sampleswere exposed for 2 min to 330-nm light in a fluorimeter andanalyzed on the same column for conversion to ATP. The col-umn eluate was usually analyzed by its absorption at 260 nm,although absorption at 300 nm was occasionally used to, dis-tinguish caged ATP from other nucleotides (8). In addition, anATP meter (Johnson Foundation, University of Pennsylvania)was used to detect ATP formation quantitatively in the 10-1000pmol range by the luciferin/luciferase assay.The stoichiometry of proton release on photolysis of caged

ATP was studied spectrophotometically by using pH indicators(phenol red and chlorophenol red at pH values above and below7, respectively). The buffering capacity of the solutions wasmeasured by adding standard aliquots of NaOH or HC1 andmeasuring the changes of the pH indicators. In all the pH ex-periments, the solutions were adjusted to the appropriate pHvalue, and this was rechecked immediately after the experi-ment. In no case, did the pH deviate by more than 0.1 unit fromthe initial value.

Stopped-flow studies of ATP-induced actomyosin dissociationwere done essentially as described by Finlayson et al. (12).

Laser Photolysis. An air-cooled Brewster angle ruby laserwas combined with a 2 cm X 1 cm2 potassium dihydrogenphosphate (KDP) frequency doubler, cut for first-index-matching conditions and angle tuned, to produce as many347-nm photons as possible. The ruby laser was passively Qswitched with cryptocyanine (1,1'-diethyl-4,4'-carbocyanineiodide) (Eastman dye A 10220) to produce a 1.3-J Q-switchedpulse that was then frequency doubled to give up to 40 mJ at347 nm in 30 nsec. The ruby 694-nm pulse that passed throughthe KDP crystal was blocked by a blue filter (Corning glass no.5031) that allowed the 347-nm pulse to pass with some atten-uation. A Wratten no. 29 filter was placed in front of the crystalto block the laser pump light. The energy of the resulting laserpulse was measured by using a Scientech 362 disc calorimeterenergy meter. If the KDP crystal was deliberately misalignedso that index matching could not be achieved, the laser energyfell to 5% of its maximum value, indicating that it was mostly347-nm light. The laser photolysis experiments were of twotypes. For those in which the analysis of ATP was performedsubsequent to the laser pulse (i.e., the luciferin-luciferase orhigh-performance liquid chromatography assays), the laserpulse impinged on a 0.2 cm2 X 0.2 cm custom-made quartz cellwhose dimensions were optimized for ATP yield. For directspectrophotometry after the laser pulse, the sample was con-tained in a quartz cell in which the measuring light passedthrough the sample orthogonal to the laser beam. The lightsource of the spectrophotometer was a tungsten iodide lamp.The measuring light was passed through a monochromator andoptical fibers and was detected by an EMI 9592B photomulti-plier with gelatin Wratten 2B filters at the front surface. Theresulting voltage traces were photographed on a Tektronix 7904oscilloscope.

RESULTSTo determine the suitability of the frequency-doubled rubylaser for the photolysis studies, the conversion of caged ATP toATP per incident photon was measured at different wave-lengths. The peak of the action spectrum occurred at 315 nm.At 347 nm, the conversion to ATP was 50% of the maximumvalue per incident photon. The action spectrum is shifted about

50 nm to the red of the nitrobenzyl absorption peak: 1-(2-nitro)phenylethyl phosphate has Xmax at 265 nm. However,caged ATP is fluorescent, and the fluorescent excitation spec-trum matched that of the photolysis action spectrum.The amount of ATP produced on laser photolysis was linearly

related to the laser energy (Fig. 1). Typically, a laser pulse of25 mj generated 500 ,uM from 2.5 mM caged ATP. The con-version to ATP decreased at concentrations of caged ATPgreater than 2.5 mM due to the relatively high absorption ofthe solution [which depends on the geometry of the cell and theextinction coefficient of the caged ATP (which is 660 M-1 cm-'at 347 nm)]. There was quantitative agreement between theATP generated per photon from the laser and that from thecontinuous-light output of a 200-W mercury arc lamp. Theyield of ATP was independent of pH at pH 6-8.Having shown that sufficient light is generated by the laser

for significant ATP release, we next investigated the subsequentdark reactions that result in the release of ATP. We used threeprobes of these reactions; first, we investigated the possibilitythat spectrally observable intermediates were formed; second,we monitored the release of protons that occurs during thecourse of the reaction and, third, we monitored ATP releasethrough its reaction with actomyosin.When photolysis of caged ATP was monitored in the near

UV, a spectral change was observed (Fig. 2A); the spectrum ofthe maximum absorption change is shown in Fig. 2B. Themagnitude of the absorption increase was proportional to theconcentration of caged ATP and to the energy of the 347-nmlaser pulse. As a solution of caged ATP was progressively pho-tolysed down to 20% of its initial concentration, the magnitudeof the absorption change decreased proportionately, suggestingthat the product (2-nitrosoacetophenone) did not generate asignal at this wavelength.The increase in absorption occurred within the time resolu-

tion of the spectrophotometer (5 ,usec). The time course of thetransient decay had a single-exponential form and a rate con-stant (X) that was [H+] dependent (Fig. 3); it follows from thesedata that X = 2.2 X 109 [H+]sec'1. X was not altered by theaddition of 1 mM EDTA or 5 mM MgCl2 at pH 7, showing theabsence of a divalent metal ion effect on this reaction.The overall stoichiometry of proton release accompanying

caged ATP photolysis was as predicted from Eq. 1. Thus, at pH

30

H 20 -

0

0

10

0 1 0 20 30Energy, mJ

FIG. 1. Dependence of conversion of caged ATP (2.5 mM) to ATPat pH 7 and 220C on energy output at 347 nm of a frequency-doubledruby laser pulse.

7238 Biophysics: McCray et al.

Dow

nloa

ded

by g

uest

on

Nov

embe

r 17

, 202

0

Proc. Natl. Acad. Sci. USA 77 (1980) 7239

A

1()0-

LI

a:

Ir i I -I

fasI .a svr

ALi s II

*50 1fl5(('

B0

0

0

380 400 420A, nm

440 460 480

FIG. 2. (A) Spectral change observed at 406 nm after laser flashphotolysis of 2.5 mM caged ATP in 0.1 M Tris at pH 8.1 and 220C.The spectrophotometer was set to 100%o transmission before the laserpulse, as shown at the left-hand edge of the photograph. (B) Nor-malized partial spectrum of the transient species. The spectrum was

not extended to shorter wavelengths because of spectral interferencefrom the laser pulse, with necessitated protecting the photomultiplierwhich below 380 nm.

7 (or at pH 5.7 in the presence of MgCl2), net proton release wasstoichiometric with the amount of ATP formed. However, atpH 5.7, ATP3- was the predominant species formed in thepresence of EDTA and less than 0.2 protons were released perATP molecule generated. Proton release occurred within thetime resolution of the spectrophotometer, and there was sub-sequent proton uptake of relatively small amplitude.

ATP causes the dissociation of actomyosin in a reaction thatis a first-order dependent on ATP concentration and is thereforean excellent probe to monitor the release of ATP from its pre-cursor into the medium. Dissociation of actomyosin is accom-panied by an approximately linear decrease in turbidity, whichresults in an increase in transmission of the solution (12). Theturbidity changes after flash photolysis of caged ATP in thepresence of actomyosin at various pH values are shown in Fig.4. In the absence of caged ATP, no transmission change occursat any pH (e.g., the upper trace of Fig. 4). At pH 6.3, there isno lag in actomyosin dissociation after the laser pulse. However,a lag phase is present at pH 7.2 and becomes more pronouncedat pH 8.1. The presence of the lag phase at the higher pH valuescan be correlated with the decrease in X of the pH-dependenttransient decay. The variation in amplitude in the three tracesis probably due to some inhomogeneity and hence variation inturbidity of the viscous actomyosin solutions. [Transient andsteady-state kinetic studies of the acto-subfragment 1 ATPasehave shown that ATP does not necessarily induce completedissociation of the proteins, which also could give rise to variable

102..-) -

IO .;)(-bME NO

OrONIE()m.....

102.- -

a)

10' X0

6.5 7.5 8.5 9.5pH

FIG. 3. Dependence of X on pH under the conditions describedin Fig. 2. Tris (0.1 M) neutralized with HCl.

102.Eu...

Laser

flash

FIG. 4. Kinetics of actomyosin dissociation after flash photolysisof caged ATP. Reaction solution contained 16 ;M actomyosin (16yMmyosin heads and F actin at 1.15 mg/ml), 1.9mM caged ATP, 5mMMgCl2, 0.6 M KCl, and 20 mM 2(N-morpholino)ethane sulfonic acid(pH 6.3), or 20mM imidazole (pH 7.2), or 20mM Tris (pH 8.1). In theupper trace, caged ATP was omitted from the solution.

1

.0

$n.1.0

a)

-)

nt

Biophysics: McCray et al.

Dow

nloa

ded

by g

uest

on

Nov

embe

r 17

, 202

0

Proc. Natl. Acad. Sci. USA 77 (1980)

FIG. 5. Kinetic simulation of 16 yM actomyosin dissociation ongeneration of 75 ,4M ATP from caged ATP photolysis at 220C. In thesimulation, two steps control the rate of actomyosin dissociation asdescribed in the text (compare with Fig. 4).

signal amplitude (13, 14). However, this is unlikely to be im-portant at the high ionic strength used in these experiments. ]

It was shown that caged ATP did not interfere with the re-action of ATP with actomyosin. For example, when 3 mMcaged ATP was incubated with actomyosin and then mixedwith ATP in a stopped-flow apparatus, the kinetics of theATP-induced dissociation were unaffected by the presence ofcaged ATP.To simulate the kinetics of ATP release into the medium, it

was necessary to determine the amount of ATP formed. CagedATP was photolysed under conditions described in the legendto Fig. 4 except for the absence of actomyosin; 75 AM ATP wasgenerated at each pH value. The second-order rate constantassociated with ATP dissociation of actomyosin in 0.6 M KCIwas measured in the pH 6-9 range. The rate constant, whichwas independent of pH (to within 20%), was 3.8 X 105 M-1sec1.

It is important to know whether or not proteins are damagedby 347-nm light. Such analysis has been done principally onintact vesicles containing sarcoplasmic reticulum ATPase(unpublished results) on which neutron and x-ray diffractionstudies were being carried out. At a light intensity more than10 times that occurring in a single laser pulse, there was nodetectable effect on either the structure or the ability of thesystem to pump calcium.

DISCUSSIONExperiments in which the energy output from a laser is corre-lated with the amount of ATP formation show that sufficientATP is generated in a single pulse for analysis of many inter-esting biological systems. Thus, this technique can be used forstructure and function studies of such diverse ATP-requiringsystems as the calcium pump, nitrogenase, and muscle con-traction, as well as for rapid photogeneration of other bio-chemical metabolites. Engels and Schlaeger (6) and Kaplan etal. (8) have shown that caged nucleotides can be introducedacross cellular and other membrane boundaries. Caged-nu-cleotides appear to bind weakly, if at all, to nucleotide sites onproteins (6, 8). We can detect no binding of caged ATP to sar-coplasmic reticulum calcium pump ATPase or to the actomy-osin ATPase system. Caged ATP was neither an inhibitor nora substrate of nitrogen fixation catalyzed by the nitrogenase ofClostridium pasteurianum (unpublished data).The photochemical properties of caged ATP are advanta-

geous from the biochemical viewpoint in that 347-nm light doesnot appear to damage proteins, which are generally essentiallytransparent at that wavelength. Furthermore, caged ATP isconvenient to work with in that it is stable in water at neutralpH and is not detectably photolyzed over a few minutes insubdued daylight.

Characterizing photolytic intermediates and associatedproton changes is important because this allows the study ofreaction mechanism. The photochemistry of 2-nitrobenzylcompounds has been reviewed by Morrison (15). Many of thesecompounds exhibit photochromicity; i.e., transient color for-mation on photon absorption, followed by regeneration of thestarting material (16). Weinstein et al. (17) have analysed thephotochromicity of 2-(2-nitrobenzyl)pyridine by using aspectroscopic technique. They detected a transient interme-diate, which they considered to be an aci-nitro compound,followed by reversion to starting material. On this basis, a ru-dimentary mechanism can be proposed for the photolysis ofcaged ATP (eq. 2) in which the conjugate base II of an aci-nitrocompound is formed. The evidence that II is involved in ATPformation (rather than in a side reaction) comes from the ac-tomyosin kinetics. The breakdown of II has the rate constantX = 2.2 X 109 [H+]sec-1. The second step of Eq. 2 is thereforeacid catalysed, even though no overall proton uptake will occurduring this step (unless conditions are such that ATP3- isformed).

NH2

caged ATP - M±. _ I... C 0 W

NtO0- IIO- O' 0 y0- HO ON

II

[2]

\) C + ATP'NO

The identification of II as the compound whose breakdowndetermines the rate of ATP formation leads to suggestions ofcaged ATP analogues that would photolyse more rapidly. Forexample, the work of Weinstein et al. (17) suggests that sub-stitution of electron-donating groups in the 4-position of thenitrobenzyl group would accelerate the decay of the transientspecies shown in Fig. 2.A critical test of whether the intermediate formulated as an

aci-nitro compound in Eq. 2 is involved in ATP formation wasprovided by the actomyosin dissociation.experiments. Theexpected time course of actomyosin dissociation was simulatedon the basis that the transient species is an intermediate and thatits breakdown at 2.2 X 109 [H+]sec-1 (i.e., 220 sec-1 at pH 7)is followed by ATP-induced actomyosin dissociation with a rateconstant of 3.8 X 105 M-1 sec-'. The simulation (Fig. 5) cor-relates well with the observed time course of actomyosin dis-sociation (see Fig. 4). We therefore conclude that the inter-mediate is involved in ATP formation and that its breakdownis rate limiting in the overall process. It follows that the rate ofATP formation occurs in an exponential process having a rateconstant 2.2 X 109 [H+]sec-I at 22°C. The time resolution ofthe technique is faster at low pH; at pH 6 for example, 50% ofthe ATP will be released within 0.3 msec.

Photolysis of 2-nitrobenzyl derivatives of biologically im-portant reactants is a promising technique for structural andkinetic studies of organized biological systems for which con-ventional mixing and rapid-reaction techniques are inappro-priate. The small (a 30-,ul optical cell was used here) and, ifnecessary sealed, reaction chamber used makes it a useful ap-proach for soluble systems when economy of biological materialis important or anaerobic conditions are required.

We thank Dr. A. Weber for helpful discussion and Ms. Fanny Itshakfor excellent technical assistance. This research was supported byNational Institutes of Health Grants HL 18708 and HL 15835 (to thePennsylvania Muscle Institute), National Science Foundation GrantPCM 76-82578, the Whitehall Foundation, and the Muscular Dys-trophy Association of America.

7240 Biophysics: McCray et al.

Dow

nloa

ded

by g

uest

on

Nov

embe

r 17

, 202

0

Biophysics: McCray et al.

1. Gibson, Q. H. (1959) Biochem. J. 71, 293-303.2. McCray, J. A. (1972) Biochem. Biophys. Res. Commun. 47,

187-193.3. Knowles, J. R. (1972) Acc. Chem. Res. 5, 155-160.4. Havinga, E., De Jongh, R. 0. & Dorst, W. (1956) Rec. Trav.

Chim. 75, 378-83.5. Patchornik, A., Amit, B. & Woodward, R. B. (1970) J. Am. Chem.

Soc. 92,6333-6335.6. Engels, J. & Schlaeger, E.-J. (1977) J. Med. Chem. 20,907-911.7. Lester, H. A., Krouse, M. E., Nass, M. M., Wassermann, N. H. &

Erlanger, B. F. (1979) Nature (London) 280,509-510.8. Kaplan, J. H., Forbush, B. & Hoffman, J. H. (1978) Biochemistry

17, 1929-1935.9. Lehrer, S. S. & Kerwar, G. (1972) Biochemistry 11, 1211-

1217.

Proc. Nati. Acad. Sci. USA 77 (1980) 7241

10. Weeds, A. G. & Hartley, B. S. (1968) Biochem. J. 107, 531-548.

11. Tener, G. M. (1961) J. Am. Chem. Soc. 83, 159-168.12. Finlayson, B., Lymn, R. W. & Taylor, E. W. (1969) Biochemistry

8,811-819.13. Stein, L. A., Schwarts, R. P., Chock, P. B. & Eisenberg, E. (1979)

Biochemistry 18, 3895-3909.14. Wagner, P. D. & Weeds, A. G. (1979) Biochemistry 18,2260-

2266.15. Morrison, H. A. (1969) in The Chemistry of the Nitro and Nitroso

Groups, ed. Feuer, H. (John Wiley & Sons, New York), Part 1,pp. 165-213.

16. Dessauer, R. & Paris, J. (1963) Adv. Photochem. 1, 275-321.17. Weinstein, J., Bluhm, A. L. & Sousa, J. A. (1966) J. Org. Chem.

31, 1983-1985.

Dow

nloa

ded

by g

uest

on

Nov

embe

r 17

, 202

0