~NALYTIcAI, moCmmfIsmY 39, 170-180 (1971)

Measurement of Simultaneous Synthesis of Inorganic

Pyrophosphate and Adenosine Triphosphate

RICHARD J. GUILLORYl AND R. R. FISHER2

Department of Biochemistry and Molecular Biology, Cornell University, Ithaca, New York l&50

Received June 24, 1970

Inorganic pyrophosphate (PPi) and higher polyphosphates were first recognized in biological systems as constituents of yeast, and other fungi (1). Subsequent work showed PPi to be involved in the biosynthesis and transformation of a wide variety of biological substances (Z-4,9). Re- cently PPi in addition to adenosine triphosphate (ATP) was found to be a product of photophosphorylation in chromatophores of Rhodo- spirillum rubrum (5,6).

Our interest in the mechanism of energy conservation in R. rubrum (10,13) required that we be able to estimate PPi in the presence of a five- to ten-fold higher concentration of ATP. The usual method for PPi analysis, which involves hydrolysis of PPi followed by the deter- mination of inorganic orthophosphate, is inadequate for this purpose because of the concomitant hydrolysis of the pyrophosphate bonds of ATP. In addition, we required a system which would rapidly convert PPi to a stable product not susceptible to a highly active pyrophos- phatase enzyme present in the 11. rubrum membrane (10,23).

In submitochondrial particles, maximum rates of aerobic ATP synthesis from ADP and Pi can be measured only when the terminal phosphate of ,4TP is trapped in a metabolically stable form. The classical trapping system used for mitochondrial oxidative phosphorylation consists of the enzyme hexokinase together with its substrate glucose. In the presence of sufficient hexokinase the rate of formation of stable G-6-P3 from ATP

1 Established Investigator of the American Heart Association. ‘Recipient of a Predoctoral Fellowship (5FOL GM 3503) from the U.S.P.H.S. a Abbreviations used and not defined in the text: Bchl, bacteriochlorophyll; “Pi.

radioactive inorganic orthophosphate ; G-6-P, glucose g-phosphate; G-l-P, glucose l-phosphate.

170

PYROPHOSPHATE AND ATP ASSAY 171

competes effectively with the mitochondrial adenosine triphosphatase. A somewhat similar trapping system, illustrated by equations 1 and 2,

has been devised for the rapid conversion of PP, to a stable product not susceptible to pyrophosphatase activity. The G-6-P formed by the action of adenosine diphosphate glucose pyrophosphorylase (ADPG- ppase) which is activated by pyruvate (11,12) and hexokinase is mea- sured enzymically by oxidation with glucose-6-phosphate dehydrogenase (eg. 3). G-l-P formation is measured by coupling with phosphogluco- mutase and glucose-g-phosphate dehydrogenase (eqs. 3 and 4).

ADPG pyrophosphorylase

PPi + ADPH c ’ XTP + glucose l-phosphate

hexokinase ATP + glucose - ADP + glurose B-phosphate

glucose B-phosphate + SADP+ glucose-&phosphate dehydrogenase

* KADPH + H+

(1)

(2)

+ glucordactone B-phosphate (3)

phosphoglucomutnse

glucose l-phosphate . ’ glucose-6-phosphate (4)

Because of the low metabolic level of PPi synthesis encountered in R. rubrum, it was necessary to develop a sensitive assay for the products of the ADPG-ppase trap, utilizing the fluorescent property of reduced triphosphopyridine nucleotides.

MATERIALS AND METHODS

Glucose-6-phosphate dehydrogenase (n-glucose-6-phosphate: NADP oxido-reductase) (EC 1 .l .1.49) , hexokinase ( ATP : n-hexose-6-phospho- transferase) (EC 2.7.1.1)) and phosphoglucomutase (ar-D-glucose-1,6- diphosphate: a-n-glucose-l-phosphate phosphotransferace) (EC 2.7.5.1) were obtained from Boehringer Mannheim Corporation and adenosine-5’ diphosphoglucose (ADPG) ) adenosine-5’diphosphate (ADP) , adenosine 5’-triphosphate (ATP), or-n-glucose l-phosphate, n-glucose 6-phosphate, triphosphopyridine nucleotide (NADP’) , ,&diphosphopyridine nucleo- tide, reduced form (NADH) , protamine sulfate (Grade 1)) glycylglycine, and oligomycin from Sigma Chemical Company. All other reagents were analytical-grade material obtained from commercial sources. Standard solutions of ATP and PPi were prepared by weighing just prior to their use.

Chromatophores were prepared from photosynthetically grown R. rubrum cells (Van Niel Strain-l) (15) by grinding with sand (16). Bacteriochlorophyll was assayed according to the method of Clayton (17). Particles from R. rubrum cells resolved with respect to adenosine triphosphatase activity were prepared as previously described (10).

172 GTJILLORY AND #FISHER

Preparation of ADPG Pyrophosphorylase

ADPG pyrophosphorylase was purified by the method developed by Furlong (11) and Furlong and Preiss (12) up to the (NH,) &!?O, frac- tionation step following the protamine precipitation.

R. rubrum (5% inoculum) was grown anaerobically in the light for 72 hr (15) and then harvested by centrifugation for 5 min at 40009. The sedimented cells from 10 liters of growth medium were suspended in 200 ml of distilled water and kept frozen until ready to be used.

The thawed cells were washed with 4 liters of cold distilled water followed by 2 liters of 0.01 1M Tris buffer, pH 8.0. The cells (225 gm wet weight) were suspended in 5.5 ml of 0.05 M Tris buffer, pH S.O/O.Ol M MgCl,/O.B m&f dithiothreitol, pH 7.5, per gram of cells (wet weight). The cell suspension was t.hen sonicated using the large probe of a Branson Sonifier (Branson Instruments Inc., Danbury, Connecticut) at its maximum power output. A single batch (1,235 ml) was sonicated for 5 min at 4”, allowed to cool for 5 min, and then sonicated again for 5 min. The sonic extract was centrifuged for 15 min at 4000g and the red supernatant gently decanted from t,he sedimented residue.

The supernatant was brought to 65” in a 70” hot water bath and maintained at 65” for 5 min while the solution was constantly stirred. The heated solution was cooled immediately in an ice-water bath (15 min\ and then centrifuged for 15 min at 16,300g. Following centrifuga- tion, the decanted supernatant (1225 ml) can be kept in the frozen state for use at a later time.

Protamine sulfate (1.6 ml of a 1% freshly prepared solution) was added dropwise t,o each 100 ml of the supernatant at 4” with constant stirring. The protamine sulfate treated suspension was maintained at 4” for 10 min and then centrifuged for 10 min at 17,000g. The addition of protamine sulfate (1.6 ml per 100 ml initial supernatant) was repeated and the precipitate collected at each step until the ADPG-ppase activity was present in the precipitate.

A few drops of 0.3 M phosphate buffer, pH 7.3, was added to each pellet, and the precipitate suspended by means of a stirring rod. The paste was then qxtracted with 10 ml of the same buffer per 100 ml of initial supernatant (120 ml). After 15 min the extract was centrifuged for 15 min at 17,OOOg and the supernatant carefully removed and assayed for activity.

Enzymic Assay for ADPG-ppase

The assay of pyrophosphorylase activity was carried out at room temperature using a Zeiss PMQ-2 spectrophotometer.

PYROPHOSPHATE AND ATP ASSAY 173

The enzyme sample (0.01 to 0.1 ml) was added to the following mixture in a total final volume of 1.0 ml: 30 mM Tris buffer, pH 7.4; 25 mM sodium pyruvate ; 10 mM glucose; 1.5 mM NADPH ; 10 mM MgCl,; 2 mM sodium pyrophosphate, pH 7.4; 1 mM ADPG; 150 pg hexokinase; and 10 pg glucose-6-phosphate dehydrogenase. The reaction was initiated by the addition of 2 mM PPi. The reduction of NADP* was followed by recording the optical density at 340 nm at 1 min inter- vals up to 5 min. The NADPH formed per minute was evaluated using 6.2 n-n?&’ cm-l as the extinction coefficient (18).

A unit of activity is defined as that quantity of enzyme bringing about the formation of 1 pmole of NADPH per minute in the standard assay and the specific activity as pmoles NADPH formed per minute per milligram of protein. The specific activities of various preparations ranged from 4 to 8. Protein was estimated by the biuret procedure (19).

Evaluation of ATP Synthesis (Photophosphorylation)

Photophosphorylation of ADP was carried out at 30” with light illumination at 7.5 X lo4 erg cm-’ set-I. The reaction mixture in 15 mm test tubes contained in a total volume of 3 ml: 30 mM glycylglycine buffer, pH 7.8; 3.3 n&! [32Pi]KH2P04 (3 x 10G cpm/pmole) ; 3 mM MgCl, ; 3 m.iV ADP ; 50 mlM glucose ; 0.4 mg hexokinase ; 16 mM ascor- bate; and chromatophores (50 pg BChl). The reaction was terminated at 5 min by immersion of the thin-walled incubation tube in a boiling water bath for 5 min. The heated tubes were cooled in an ice bath for 15 min and centrifuged for 5 min in an International clinical centrifuge (model CL, International Equipment Co., Needham Heights, Mass.).

An aliquot (0.1 ml) of the protein-free supernatant was added to 5 ml of 1% ammonium molybdate in 0.096 N perchloric acid and the ATP synthesis evaluated according to the procedure of Lindberg and Ernster (20). In addition, ,aliquots of the prot’ein-free supernatant were taken directly for fluorometric analysis.

Evaluation of PPi Synthesis

Photopyrophosphorylation was carried out at 30” in 15 mm t,est tubes with light illumination at 7.5 X lo4 erg cm-2 set-‘. The reaction mixture contained in a total volume of 1.5 ml: 70 mM glycylglycine buffer, pH 7.8; 18 mM pyruvate; 15 mM [““Pi]KH,PO, (3 X lo6 cpm/pmole) ; 5.7 mM MgCl, ; 0.4 mg hexokinase ; 50 mM glucose ; 70 mM ascorbate; 20 pg oligomycin ; 0.70 mM ADPG ; 1 to 2 units of ADPG-ppase; and chromatophores (50 pug BChl). The reaction was terminated at 10 min by immersion of the thin-walled incubation tubes into a boiling water bath for 5 min. Following cooling in an ice bath and centrifugation,

174 GUILLORY AND aFISHER

aliquots of the supernatant were taken directly for fluorometric analysis or for extraction with isobutanol (20) as described for the evaluation of ATP synthesis.

Fluorometric Measurements

Fluorometric measurements were made with an Eppendorf photometer equipped with a fluorometer attachment (Netheler and Hinz GmbH, Hamburg-Wellings biittel) . A primary filter isolating wavelengths from 313 to 366 nm and a secondary (barrier) filter from 400 to 3000 nm were used. The intensities of the fluorescence changes were recorded on a 10 in. linear recorder (Beckman Instruments, Inc., Fullerton, Cali- fornia) .

Prior to the fluorometric measurements a stock solution of reduced pyridine nucleotide (NADH) was prepared and used for the stand- ardization of the Eppendorf fluorescence photometer. The concentration of the stock solution of NADH (2.5 X lo-” M) in 0.2 M glycylglycine, pH 7.4, was accurately determined by its ,absorption at 340 nm with a Zeiss model PMQ-2 spectrophotometer. Appropriate dilutions of the stock solution were made in 0.2 M glycylglycine buffer and the degree of fluorescence of the diluted samples evaluated with the Eppendorf. In all cases, the fluorometric instrument was set initially at zero fluores- cence with 0.2 M glycylglycine, pH 7.4. The majority of our experiments required that the multiple-step switch of the instrument be set at 7. However, this can be varied depending upon the particular sensitivity required and the level of background fluorescence encountered. The calculated concentration of the NADH solution was then plotted against the measured increment in fluorescence.

The magnitude of pyridine nucleotide reduction during the analysis of samples containing the products of the PPi trapping system was evaluated using this standard curve.

Analysis of Products from the PPi Trapping System

Aliquots of the supernatants obtained after heat denaturation were used for the fluorometric analysis of glucose 6-phosphate and glucose l- phosphate. The syntehsis of pyrophosphate in the presence of the ADPG- ppase-hexokinase couple results in an equivalent formation of glucose 6-phosphate and glucose l-phosphate. An increment in glucose 6-phos- phate in excess of glucose l-phosphate was taken to represent ATP synthesis.

An aliquot of the heat-treated supernatant (0.005 to 0.10 ml) repre- senting 5 to 15 mymoles of PPi was added to a quartz curvet containing in a final volume of 2 ml; 100 mM glycylglycine buffer, pH 7.6; 5 mM

PYROPHOSPHATE AND ATP ASSAY 175

MgCl,; and 25 pug glucose-6-phosphate dehydrogenase. The reaction was initiated by the addition of 0.25 mM NADP and the fluorometric change was followed (chart speed 0.5 in./min) for 5 min. The subsequent addition of 50 pg of hexokinase together with 20 mM glucose sometimes results in an additional 4% increase in fluorescence and is assumed to represent the presence of a small amount of adenosine triphosphate which cont’aminates the enzyme system, If the heat-treated supernatants arc maintained at 0 to 4” for periods of 6 hr prior to analysis, the fluorescence increment upon the addition of hexokinase and glucose is increased to about 10% of the initial fluorescence change observed with added NADP. We assume that this change represents residue adenylate kinase activity which was not completely destroyed by the heat treat- ment. Under normal assay conditions this activity does not manifest itself. Upon stabilization of the fluorescence change the addition of 50 ,ug of phosphoglucomutase resulted in an additional increase in fluores- cence representing the reaction of glucose l-phosphate.

The increments in fluorescence were evaluated on the basis of standard curves for NADH fluorescence and the quantity of pyrophosphate deter- mined from the concentration of glucose 6-phosphate and glucose l- phosphate assayed. Under controlled conditions this procedure permits the simultaneous determinations of PPi (as glucose l-phosphate) and ATP (as the difference between glucose 6-phosphate and glucose l-phos- phate). The only limiting factor for such an analysis is the quantitative difference between the rate of ATP synthesis as compared with that of PPi synthesis. In R. rubrum the latter represents 10 to 20% of the former (21,24) .

EXPERIMENTAL

Figure 1 illustrates the type of response obtained in the fluorometric enzymic analysis of a solution containing the products of the enzymic pyrophosphate trapping system (upper trace). For this assay 20 meoles of PPi was incubated for 15 min at 25” in the medium described for the evaluation of PPi synthesis and in the absence of Rhodospidlum rubrum chromatophores. The reaction was terminated by heating in a boiling water bath for 10 min. After cooling to ice temperature, the solution was centrifuged. An aliquot (0.05 ml) of the protein-free supernatant representing about 0.7 mpmole of pyrophosphate was added to a quartz cuvet and analyzed for the products of the PPi trapping system. There is an exact stoichiometry for glucose 6-phosphate and glucose l-phos- phate assayed under the above conditions. The lower trace demonstrates the response of a system obtained by omitting PPi from the medium described for the evaluation of PPi synthesis.

176 GUILLORY AND PISHER

60--

Y 207- PHOSPHOGLUCOMUTASE F HEXOKINASE a w’ E

lo- ,.1

2 MINUTES

k ’

NADP

TIME (MINUTES)

FIG. 1. Enzymic-fluorometric analysis of solution containing products of pyrophos- phate trapping system. For conditions see text.

Table 1 represents experiments conducted in the same manner as in Fig. 1 except that quantities of ATP (13-26 mpmoles) were added to- gether with pyrophosphate (5-10 mpmoles) and that oligomycin was omitted from the reaction mixture. From the balance of the phosphoryl- ated sugars assayed, it is obvious that the pyrophosphate trap allows one

TABLE 1

Analysis of Pyrophosphate in the Presence of Adenosine Triphosphate

Expt.

I

II

Reagent added (mpmoles)

ATP PPi

0 5.0

13.0 0

13.0 5.0

0 10.0 26.0 0

26.0 10.0

G-6-P (mpmoles) G-l-P (mpmoles)

Found Theory Found Theory

5.7 5.0 5 0 5.0

13.8 13.0 0.5 0

19.1 18.0 4.6 5.0

11.4 10.0 10.0 10.0

27.6 26.0 1.1 0

36.9 36.0 10.3 10.0

See text for assay procedures.

PYROPHOSPHATE AND ATP ASSAY 177

to accurately determine pyrophosphate quantitatively in the presence of ATP. The difference, glucose 6dphosphate less the quantity of glucose l-phosphate assayed, represents the level of ATP, and the measured concentration of glucose l-phosphate represents the molar concent’ration of pyrophosphate added to the system.

Extraction of R. rubrum chromatophores with 1.9M lithium chloride results in the loss of adenosine triphosphatase activity and the ability of the psrticles to photophosphorylate ADP (10). Table 2 summarizes

TABLE 2 Comparison of Fluorometric and Radiophosphate Determinations of PPi Synthesized

by Chromatophores of R. rubrum Treat,ed with 1.9 M Lithium Chloride (mpmoles organic phosphate ester formed)

Assay t,ime (min) G-6-P

Fluorometric

G-I-P G-6-P + G-l-P Radiophosphate

5 55 71 126 112 10 106 111 217 198 15 175 183 358 310

Photopyrophosphorylation was carried out at 30” with light illumination at 7.5 X lo* erg crnm2 set-I. The reaction mixture contained in a total volume of 1.5 ml: 70 mM glycylglycine buffer, pH 7.8; 19 mM pyruvate; 15 mM [3*P2]KH2P04 (3 X lo6 cpm/ *moles); 5.7 mM MgClz; 0.4 mg hexokinase; 50 mM glucose; 70 mM ascorbate; 20 rg oligomycin; 0.70 mM ADPG; 2 units ADPG-ppase; and chromatophores (50 pg Bchl).

See text for assay procedures.

an experiment demonstrating the time course of pyrophosphate synthesis in such particles and compares the enzymic fluorometric analysis with that using radiophosphate. It is evident that in this chromatophore prep- aration PPi synthesis is still functional in the absence of photophos- phorylation of ADP and that the enzymic fluorometric and radiolabel- ing procedures are in good agreement.

Baltscheffsky and von Stedingk (5) have shown that oligomycin is an effective inhibitor of light-induced ATP synthesis in chromatophores of R. rubrum. Table 3 shows that the illumination of chromatophores in the absence of oligomycin (line 1) and t.he presence of the PP, trapping system results in the formation of substantial amounts of G-6-P in ex- cess of G-l-P. Since this synthesis is not found during incubation in the absence of the trapping system (line Z), it most probably represents photophosphorylation of ADP which is formed as a result of the trapping of PPi (eqs. 1 and 2). Increasing concentrations of oligomycin bring about a reduction in the amount of ATP (i.e., G-6-P) synthesis with only a slight change in the quantity of PPi assayed. At 40 to 60 pg of

178 GUILLORY AND FISHER

TABLE 3 Effect of Oligomycin on Pyrophosphate Synthesis in Chromatophores of R rubrum

(rmoles substance formed per 50 pg BChl per 5 min)

Incubation conditions G-6-P G-l-P

G-6-P less

G-l-P

Ratio G-6-P/ G-l-P

1. Complete 3.0 0.25 2.75 12.0 2. Less ADPG, ADPG-ppase 0.021 0.021 0 - 3. plus oligomycin 20 pg 0.640 0.340 0.300 1.88 4. plus oligomycin 40 pg 0.370 0.355 0.015 1.04 5 plus oligomycin 60 pg 0.285 0.290 0.005 0.98

Experimental procedure as described in the text except that oligomycin was omitted from the complete assay system and added as indicated.

oligomycin the G-6-P to G-l-P was approximately 1, indicating that only PPi was being synthesized.

DISCUSSION

The recognition of a large variety of pyrophosphorylase reactions in- dicates that there are indeed many biological sources of pyrophosphate. The recent finding of a photopyrophosphorylation reaction in R. rubrum chromatophore confirms early suggestions (4,22) that pathways might exist which would involve pyrophosphorolysis of inorganic phosphate as well as phosphorolysis of ADP.

The assay procedure described in this report permits the quantitative evaluation of pyrophosphate in a sensitive, specific, and unambiguous manner. In addition, correlation of the G-6-P and G-l-P concentrations, the products of the enzymic trapping system, allows one to distinguish ATP from PPi synthesis. Assays using 39Pi preclude this distinction. The pyrophosphate trap also permits the measurement of a low rate of pyro- phosphate synthesis unemcumbered by pyrophosphatase activity.

While this assay procedure has been developed primarily for photo- pyrophosphorylation in R. rubrum the trapping system should have general applicability. During the progress of this work, Johnson et al. (14) reported the use of UDP glucose pyrophosphorylase (in combina- tion with phosphoglucomutase and glucose-6-phosphate dehydrogenase) as an assay for DNA dependent RNA polymerase.

SUMMARY

An adenosine diphosphoglucose pyrophosphorylase trapping ‘system is described which permits the unambiguous, sensitive, and quantitative evaluation of pyrophosphate under conditions in which this substance

PYROPHOSPHATE AND ATP ASSAY 179

is unstable. The system can be used for the evaluation of the rates of PPi and ATP formation when both synthetic reactions are occurring simultaneously. While up to now the method has been applied primarily to photopyrophosphorylation and photophosphorylation in chromato- phores of R. &rum it has, nevertheless, general applicability for bio- energetic systems.

ACKNOWLEDGMENTS

The technical assistance of Mrs. Patricia B&son and Miss Kira Targosh in the growing and harvesting of the Rhodospirillum rubrum cells and in the preparation of ADPG-ppase is greatly appreciated. We wish as well to express our appreciation to Dr. Clement Furlong, Jr., for supplying us with his purification procedure for ADPG-ppase prior to its publication.

This work was carried out under a Grant-in-Aid of the American Heart Associa- tion and supported in part by the “Southern Tier Heart Association of New York” and by grant AM 1075801 the U. S. Public Health S8ervice, National Institute of Arthritis and Metabolic Diseases.

REFERENCES

1. SCHMIDT, G., in “Phosphorus Metabolism” (W. D. McElroy and B. Glass, eds.), Vo. I. p. 443. Johns Hopkins Press, Baltimore, 1961.

2. CORI, C. F., in “A Symposium on Respiratory Enzymes,” p. 175. University of Wisconsin Press, Madison, 1942.

3. CORI, G. T., OCHOA, S., SLEIN, M. W., AND CORI, C. F., Biochim. Biophus. Acta 7, 304 (1951).

4. KORNBERG, A., in “Advances in Enzymology” (F. F. Nord, ed.), Vol. 18, p. 191. Interscience, New York, 1957.

5. BALTSCHEFFSKY, H., AND VO’N STEDINGK, L. V., B&hem. Biophys. Res. Commun. 22, 722 (1966).

6. BALTSCHEFFSKY, H., VON STEDINGK, L. V., HELDT, H. W., AND KLINGENBERG, M.. Science 153, 1120 (1966).

7. BALTSCHEFFSKY, H., Acta Chem. &and. 21, 1973 (1967). 5. LIPMANN, F., in “Advances in Enzymology” (F. F. Nord, ed.), Vol. I, p. 1.

Interscience. New York, 1941. 9. KEISTER, D. L., AND YIKE, N. J., Biochem. Biophys. Res. Commun. 24, 519

(1966). 10. FISHER, R. R., AND GUILTAIRY, R. J., FEBS Letters 3, 27 (1969). 11. FURLONG, C. E., Dissertation, Graduate Division, University of California, Dmavis.

196% 12. FURLONG, C. E., AND PREISS, J., J. Biol. Chem. 244, 2539 (1969). 13. FISHER, R. R., AND GUILMRY, R. J., Abstracts, Sixth Meeting, Federation of

European Biochemical Societies, Madrid, April 1969. 14. JOHNSON, J. C., SHANOFF, M., BASS, S. T., BOEZI, J. A., AND HANSEN, R. G.,

Anal. Biochem. 26, 137 (1968). 15. ORMEROD, J. G., ORMEROD, K. D., AND GEST, H.. Arch. Biochem. Biophys. 94,

449 (1961). 16. BALTSCHEFFSKT, H., Biochim. Biophys. Acta 40, 1 (1960). 17. CLAYTON, R. Ii., in “Bacterial Photosynthesis” (R. Gest, A. San Pietro, and

L. P. Vernon, eds.), p. 495. Antioch Press, Yellow Springs, Ohio, 1963.

180 GUILLORY AND cFISHER

18. BERGMEYER, H.-U., ed., “Methods of Enzymatic Analysis,” p. 1030. Academic Press, New York, 1963.

19. JACOBS, E., JACOB, M., SANADI, D. R., AND BRADLEY, L. B., J. Biol. Chem. 223, 147 (1956).

20. LINDBERG, O., AND ERNSTER, L., in “Methods in Biochemical Analysis” (D. Glick, ed.), Vol. 3, p. 1. Interscience, New York, 1956.

21. GUILMRY, R. J., FISHER, R. R., AND LOWE, L. Y., in preparation. 22. LINDBEBG, O., AND ERNSTER, L., Ezpptl. Cell Res. 3, 209 (1952). 23. BALTSCHEFFSKY. M., BALTSCHEFFSKY, H., AND VON STEDINGK, L. V., Brookhaven

Symp. Biol. 19, 246 (1966). 24. HORIO, T., NISIRIKAWA, K., HORIUTI, Y., AND KAKUNO, T.. in “Comparative Bio-

chemistry and Biophysics of Photosynthesis” (K. Shibata, A. Takamiya, A. T. Jagendorf, and R. C. Fuller, eds.), p. 408. University Park Press, Baltimore, 1968.