VOL. 52, 1964 BIOCHEMISTRY: BUCHANAN, BACHOFEN, AND ARNON 839

* This investigation was supported by research grants from the USPHS and from the National Science Foundation. t Predoctoral trainee on USPHS training grant in Biochemistry, 2T1-GM-53. 1 See Gurd, F. R. N., L. J. Banaszak, E. H. Eylar, and A. J. Veros, footnote in 4 Richards, F. M., Ann. Rev. Biochem., 32, 269 (1963). Also see Banaszak, L. J., and F. R. N. Gurd. J. Biol. Chem., 239, 1836 (1964). 2 Rupley, J., personal communication, in press. 3 Doscher, M. S., and F. M. Richards, J. Biol. Chem., 238, 2399 (1963). 4 Sigler, P. B., and H. C. Skinner, Biochem. Biophys. Res. Commun., 13, 236 (1963). 5 Anson, M. L., J. Gen. Physiol., 20, 663 (1937). 6 Allan, B. J., P. J. Keller, and H. Neurath, Biochemistry, 3, 40 (1964). 7 Ludwig, M. L., I. C. Paul, G. S. Pawley, and W. N. Lipscomb, these PROCEEDINGS, 50, 282 (1963). 8 Moore, S., and W. H. Stein, J. Biol. Chem., 176, 367 (1948). 9 Ibid., 211, 907 (1954). 10 Lumry, R., E. L. Smith, and R. R. Glantz, J. Am. Chem. Soc., 73, 4330 (1951). 11 Sabatini, D. D., K. Bensch, and R. J. Barrnett, J. Cell Biol., 17, 19 (1963). 12 Kendrew, J. C., in Models and Enzyme Structures, Brookhaven Symposia in Biology, No. 15 (1962), p. 216. 13 Bar-Eli, A., and E. Katchalski, Nature, 188, 856 (1960). 14 Cebra, J. J., D. Gival, H. I. Silman, and E. Katchalski, J. Biol. Chem., 236, 1720 (1961).

ROLE OF IN THE REDUCTIVE ASSIMILATION OF C02 AND ACETATE BY EXTRACTS OF THE PHOTOSYNTHETIC BACTERIUM, CHROMA TIUM BY BOB B. BUCHANAN, REINHARD BACHOFEN, AND DANIEL I. ARNON* DEPARTMENT OF CELL PHYSIOLOGY, UNIVERSITY OF CALIFORNIA, BERKELEY Communicated July 27, 1964 A ferredoxin-dependent reductive synthesis of pyruvate from C02, H2, and acetyl- CoA-henceforth referred to as the pyruvate synthase system (eq. 1)-has re- cently been obtained' with cell-free extracts of the obligately anaerobic bacterium,. Clostridium pasteurianum. ferredoxin Acetyl phosphate + CO2 + H2 coA ) Pyruvate + Pi (1) This article presents the first evidence for the ferredoxin-dependent pyruvate synthase system as a mechanism for reductive CO2 assimilation in a photosynthetic organism. A ferredoxin-dependent synthesis of pyruvate, followed by a synthesis of amino acids, was obtained with cell-free extracts of the photosynthetic bacterium, Chromatium, that were supplied with H2, C02, and acetate or its derivatives. The pyruvate synthase system includes a new primary CO2 fixation reaction (as defined by Wood and Stjernholm2) that is consistent with previously reported carbon labeling patterns of amino acids isolated from bacterial cells that were sup- plied with labeled CO2 or acetate. Over a decade ago, Ehrensvard and associates,3 working with the photosynthetic bacterium Rhodospirillum rubrum, and Tomlin- son,4 working with the nonphotosynthetic anaerobe Clostridium kluyveri, obtained labeling iata which suggested the operation of some unknown enzymic mecha- Downloaded by guest on October 1, 2021 840 BIOCHEMISTRY: BUCHANAN, BACHOFEN, AND ARNON PROC. N. A. S.

nism for the condensation of C02 with acetate (or a derivative thereof) to give a C3 compound (e.g., pyruvate) that was then used for the synthesis of amino acids. M\ore recently, similar reactions were indicated by the work of Sadler and Stanier5 and Hoare6 on the photoassimilation of acetate and C02 by intact cells of Chlorobium limicola and R. rubrum, respectively (see also reviews by Wood and Stjernholm,2 Krampitz,7 and Elsden8 9). Prior to the recognition of the pyruvate synthase system, it was generally con- sidered,10' 11 that, as in other photosynthetic or chemosynthetic organisms, the only primary C02 fixation reaction in Chromatium, which leads to the formation of a C3 compound, is the carboxylation of ribulose diphosphate followed by its split to two molecules of phosphoglycerate. The pyruvate synthase system, which we now report, constitutes an alternative mechanism for a synthesis of a C3 compound that leads to the formation of alanine, aspartate, and glutamate. These amino acids and not phosphorylated sugars or phosphoglycerate are the main soluble products of photosynthesis by Chromatium.'0' 1 The pyruvate synthase system provides directly the C3 carbon skeleton for alanine. The C4 and C0 carbon skele- tons needed for aspartate and glutamate are readily supplied by the previously demonstrated10 reactions in Chromatium extracts, i.e., the carboxylation of phos- phoenolpyruvate (or of pyruvate in the presence of ATP) to oxalacetate, and the condensations of acetyl with either pyruvate or oxalacetate. That photosynthetic bacteria contain ferredoxin, on which the operation of the pyruvate synthase system depends, is now well established. Chromatium was the first bacterial species from which the protein now called ferredoxin was isolated (under the name pyridine nucleotide reductase, see Table 4 in ref. 12), before the term ferredoxin was coined and applied by M\ortenson et al."I to an iron-containing protein which they isolated from C. pasteurianum. Crystalline Chromatium ferredoxin (Fig. 1) is similar to crystalline ferredoxin from C. pasteurianum,14 with respect to absorption spectra, molecular weight, iron and "inorganic sulfide" content;"5 however, Chromatium ferredoxin was found" to be about 70 mv more elec- tronegative than Clostridium ferredoxin.14 The present investigation also confirms and extends the earlier evidence1 that, in the presence of hydrogen gas, the sole contribution of radiant energy to the photo- of Chromatium is photosynthetic phosphorylation, i.e., a light-induced formation of ATP. As noted previously,10' 12 this conclusion is of particular interest because Chromatium is one of the few strict anaerobes that are unique in the living world in being also strict phototrophs that can grow only in the light. Unlike, for example, the green alga Chlorella or photosynthetic bacteria of the genus Rhodo- spirillum, Chromatium cannot replace its anaerobic, light-dependent mode of life by an aerobic, heterotrophic metabolism in the dark.16 Methods.-Chromatium, strain D, was grown in the light in the "carbonate" medium of Arnon et al.," harvested by centrifugation, and stored at 4VC. Five gm of cell paste was suspended with the aid of a magnetic stirrer in 10 ml of 0.02 M potassium phosphate buffer, pH 6.5. The cell suspension was subjected to sonic oscillation for 15 min and then centrifuged for 10 min at 12,000 X g in a refrigerated centrifuge. The residue was discarded. Ferredoxin was removed from the cell- free supernatant fluid by passing it through a 2 X 5 cm DEAE-cellulose column, equilibrated with 0.02 M potassium phosphate buffer, pH 6.5. The DEAE-cellulose column was eluted with the same buffer to give a volume of the eluate (henceforth called cell extract) equal to the initial volume of the original supernatant fluid. The protein content of the cell extract was determined Downloaded by guest on October 1, 2021 VOL. 52, 1964 BIOCHEMISTRY: BUCHANAN, BACHOFEN, AND ARNON 841

.~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~0

M~~~~

?~~~~~~~~~~ ~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~ ~......

.g 'a...~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~....

3:s__ ....\..'s:' i~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~Hi::']~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~..a'..a

*~~~ ~.: ~.. .s ~ ~ ~ ......

FIG. 1.-Microphotograph of crystalline Chromatium ferredoxin. Magnification, 785 X. (Bacho- fen, Oda, and Arnon, 1963.)

by the phenol colorimetric method, as modified by Rabinowitz and Pricer, 8 and the extract was diluted with the same phosphate buffer to give a final protein concentration of 25 mg/ml. Twenty- five mg of protein in the cell extract corresponded to approximately 1 mg of bacterial chloro- phyll.'9 The cell extract contained the necessary for the synthesis of pyruvate from CO2, H2, Downloaded by guest on October 1, 2021 842 BIOCHEMISTRY: BUCHANAN, BACHOFEN, AND ARNON PROC. N. A. S.

and acetate (or its derivatives), as well as for the subsequent utilization of the newly formed pyruvate, mainly for the synthesis of amino acids. The extract also contained "chromatophores," i.e., particles containing bacteriochlorophyll that were active in catalyzing photophosphorylation. The used were isolated and purified by the methods of Tagawa and Arnon.14 The ferredoxins were reduced with hydrogen gas by the native Chromatium hydrogenase which was present in the cell extract. The enzyme reactions were carried out in Warburg vessels equilibrated with the desired gas. After 10 min of equilibration, the taps were closed and the vessels were shaken for 60 min prior to adding the cell extract from a side arm. This preincubation was effective in overcoming a marked 30-45-min lag observed when the cell extract was tipped in immediately after equilibra- tion with the gas. Total C1402 fixed was determined after the reaction mixture was acidified with 0.3 ml of 12 N HC1 and centrifuged. Of the clear supernatant fluid, 0.2 ml was added to 0.1 ml of 2,4-dinitro- phenylhydrazine reagent, dried under a heat lamp, and counted with a conventional thin-window Geiger-Mueller counter. Pyruvate was isolated and counted as the 2,4-dinitrophenylhydrazine derivative as previously described.' The 2,4-dinitrophenylhydrazine derivative of pyruvate was identified by paper chromatography and radioautography as previously described.' Amino acids were identified by two-dimensional paper chromatographs, developed inb utanol-acetic acid-water (52:14:35) and water-saturated phenol. Radioautography and cochromatography with the authentic amino acids established the identity of the isolated C'4-labeled glutamate, aspartate, and alanine. Results and Discussion.-The pyruvate synthase system in Chromatium: Table 1 shows that, in the dark, the DEAE-treated extracts of Chromatium, unlike those of TABLE 1 C. pasteurianum,l formed relatively little REDUCTIVE SYNTHESIS OF PYRUVATE FROM pyruvate from acetyl phosphate, C02, C1402, H2, AND ACETYL PHOSPHATE BY and H2 gas in the absence of the carbonyl Chromatium EXTRACTS IN THE DARK C1402 fixed trappig reagent, semicarbazide. In the as pyruvate presence of semicarbazide, formation of Treatment (cpm) Complete 13,900 pyruvate required acetyl phosphate, co- Minus MnCl2 15,750 enzyme A, ferredoxin, and hydrogen gas. semicarbazide 2, 210 coenzyme A 590 No pyruvate was formed when H2 was ferredoxin 150 replaced by argon, with or without the acetyl phosphate 20 a o Complete but H2 replaced by addition of DPNH2 (or TPNH2). argon 130 The requirements for pyruvate synthe- Complete but H2 replaced by . . argon and DPNH2 70 SiS by Chromatium extracts are similar to The complete system contained cell extracts (see those found previously with extracts of Methods) of Chromatium (10 mg protein), 200 jug of C. to by ferredoxin from C. pa8teurianum, and the following pasteurianuml and those found in jsmoles: potassium phosphate buffer, pH 6.5, 200; MnClI, 4; neutralized semicarbazide, 40; Wolfe and O'Kane for the degradation of coenzyme A, 0.5; acetyl phosphate, 50; bicarbonate (1.1 X 106 cpm), 10. Finalcarbon-14volume, pyruvate and the pyruvate-CO2 exchange 3.0 ml. Gas phase, Hi (unless otherwise indi-exrcsoC.1 * 20, 21 cated). The reaction was carried out at 300 for 60 reaction in extracts of C. butyricum. min in the dark. However, under our experimental condi- tions, the pyruvate synthase system was not dependent on the addition of thiamine pyrophosphate or a divalent cation such as manganese (Table 1). The pyruvate synthase system and the synthesis of amino acids: Figure 2 shows the time course of pyruvate formation and of total C1402 assimilation by Chromatium extracts when DPN and the amino group donor, glutamine, were present and when semicarbazide was omitted. Under these conditions, the main products of C1402 assimilation were identified as aspartate, glutamate, and alanine; these amino acids were reported previously to be the principal products of C1402 assimilation by whole cells and cell-free extracts of Chromatium.10 11 Figure 2 also shows that the net Downloaded by guest on October 1, 2021 VOL. 52, 1964 BIOCHEMISTRY: BUCHANAN, BACHOFEN, AND ARNON 843

150 Ferredoxin-dependent C1402 assimilation by extracts of Chromatium FIG. 2.-Ferredoxin-dependent re- Totol ductive assimilation of C1402. The C'4fixed reaction mixture contained cell loot ./ extracts of Chromatium (10 mg protein), 200 ,ug C. pasteurianum ferredoxin, and the following in ;&moles: potassium phosphate buf- fer, pH 6.5, 200; MnCl2, 4; co- enzyme A, 0.1; acetyl phosphate, 50 50; DPN, 2; glutamine, 25; car- bon-14 bicarbonate (1.1 X 106 cpm), 10. Final volume, 3.0 ml. Gas phase, H2. The reaction was CP4-pyruvote carried out at 30° C in the dark. C) 0 30 60 90 120 minutes

synthesis of pyruvate ceased after 15 min, whereas the total assimilation of C1402 proceeded linearly for at least 2 hr. This suggested that the pyruvate synthase system was the principal path for the assimilation of CO2. Figure 3 supports this conclusion by showing that, in short-term experiments, the incorporation of C1402 into pyruvate was greater than the incorporation of C1402 into amino acids. The increase in pyruvate accumulation ceased after a few minutes; thereafter, the con- version of pyruvate into amino acids became the dominant pathway of CO2 as- similation. As shown in Table 2, total C1402 assimilation (mainly into amino acids) differed from pyruvate synthesis in having additional requirements for DPN and MnCl2. TPN was ineffective as a substitute for DPN. The dependence of total C1402

Ferredoxin-dependent synthesis 150 Fendoxin-dependent C0. assimilation of pyruvate and amino acids by by extracts of Chromatium j 60 extracts of Chromatium

140 2 4 6 8 v ~40 C _Amino acids 0 .

20 50 5 CF-Pyruvote Cromlium ferredoxin e-O Closatdhium ferredoxin o-o

minutes 0 FIG. 3.-Ferredoxin-dependent 0Ferreimand50 I100miCrogramis150 200 synthesis(2.2uoe0X ofcro-4of0 pyruvatep)wr6iabntand8haminode.reute10mlaino0rmtuetat.Cytlb 10 acids. Experimental conditions FIG. 4.-Effect of Chromatium as described for Fig. 2, except that and Clostridium ferredoxin on 2 /Amoles of carbon-14 bicarbonate the reductive assimilation Of C'402 (2.2 X 106 cpm) were added. by Chromatium extracts. Crystal- Pyruvate was isolated as described line ferredoxins prepared from C. previously;' amino acids were pasteurianum and Chromatium isolated by a modification of the cells were added as indicated; Dowex-1 colunm chromatography reaction time was 60 min. Other procedure described by Busch, experimental conditions were as Hurlbert, and Potter.2 given for Fig. 2. Downloaded by guest on October 1, 2021 844 BIOCHEMISTRY: BUCHANAN, BACHOFEN, AND ARNON PROC. N. A. S.

TABLE 2 TABLE 3 REDUCTIVE ASSIMILATION OF C1402 AND EQUIVALENCE OF ACETATE PLUS ATP, ACETYL PHOSPHATE BY Chromatium AcETYL-COA, AND ACETYL PHOSPHATE IN EXTRACTS IN THE DARK THE REDUCTIVE ASSIMILATION OF C1402 BY Total C1402 Chromatium EXTRACTS IN THE DARK assimilated Total Treatment (cpm) assimilatedC1402 Complete 72,500 Addition (cpm) Minus MnCl2 35 500 Acetate, ATP 58,500 coenzyme A 16,200 Acetate, ADP 3,000 DPN* 3,600 Acetyl-CoA 58,500 acetyl phosphate 600 Acetyl phosphate 78,500 "t ferredoxin 200 The reaction mixture contained 10 Complete but H2 replacedrsmolesby of acetate or of each of its deriva- argon 400 tives as shown above. 10 Amoles of ATP or ADP was added where indicated. The * TPN in the absence of DPN gave 5,300 cpm. reaction was carried out for 60 min; other The reaction was carried out for 60 min; other experimental conditions were as given experimental conditions were as given for Fig. 2. for Fig. 2.

assimilation on the concentration of ferredoxin is shown in Figure 4. Although the ferredoxins of C. pasteurianum and Chromatium were interchangeable in this system, lower amounts of Chromatium ferredoxin were sufficient to saturate the system. Other experiments on amino acid synthesis by dialyzed extracts of Chromatium cells revealed an additional requirement for an amino group donor such as glut- amine or ammonium salt. In the course of this investigation, Chromatium extracts were found to carry out a ferredoxin-dependent reduction of DPN by hydrogen gas. It is possible, therefore, that the DPNH2 produced in this manner was used in reductive aminations that gave rise to one or more of the amino acids that were formed in our reac- tion mixtures. Acetyl-CoA as the CO2 acceptor: Table 3 shows that synthetic acetyl-CoA was effective in replacing acetyl phosphate as the C1402 acceptor in the pyruvate syn- thase system of Chromatium extracts. We have previously reported' that coenzyme A is required for the pyruvate synthase system of C. pasteurianum. Mortlock, Valentine, and Wolfe22 found that, in extracts of C. butyricum, coenzyme A replaced inorganic phosphate in the "phosphoroclastic" degradation of pyruvate; and Wolfe and O'Kane2 observed that coenzyme A is required for the pyruvate-C02 exchange reaction. We have found similar requirements for coenzyme A, in extracts of Chromatium, for the degradation of pyruvate and the pyruvate-CO2 exchange re- action. These observations on Clostridium species and Chromatium, as well as the evidence presented in Tables 1, 2, and 3, suggest that the effective CO2 acceptor in the pyruvate synthase system is acetyl-CoA. Chromatium extracts appear to form acetyl-CoA either from acetyl phosphate or from acetate and ATP. Acetate plus ADP and inorganic phosphate were found to be ineffective (Table 3). These results indicate that Chromatium extracts con- tain the enzymes phosphotransacetylase,2' which converts acetyl phosphate to acetyl-CoA, and acetokinase,24 which, in the presence of ATP, converts acetate to acetyl phosphate. However, the possibility cannot be excluded that Chromatium extracts contain also another pathway for an ATP-dependent activation of acetate to its coenzyme A thiolester, namely, the acetyl-CoA synthetase system.25 The acetyl-CoA synthetase system has already been demonstrated26 in another photo- synthetic bacterium, Rhodospirillum rubrum. Downloaded by guest on October 1, 2021 VOL. 52, 1964 BIOCHEMISTRY: BUCHANAN, BACHOFEN, AND ARNON 845

TABLE 4 EQUIVALENCE OF LIGHT AND ATP IN THE REDUCTIVE ASSIMILATION OF C1402 AND ACETATE BY Chromatium EXTRACTS

- Light - - Dark--- Total C1402 Total C1402 assimilated assimilated Treatment (cpm) Treatment (cpm) Complete 15,300 Complete 3,900 Minus MnCl2 7,100 Complete, ATP added 17,200 " coenzyme A 4,700 " acetate 4,300 " DPN 400 i ferredoxin 2,300 Complete but H2 replaced by argon 1,500 The reaction was carried out at 200 for 60 min. Each vessel contained 0.5 Amole coenzyme A and 5 J.moles ADP. 10 gmoles ATP was added where indicated. Illumination, 20,000 lux. Other experimental conditions were as given for Fig. 2. Equivalence of light and A TP for the pyruvate synthase system: As already stated, the assimilation of C 1402 via the pyruvate synthase system depended on the avail- ability of a C2 acceptor, i.e., acetyl-CoA or its equivalent. In the presence of co- enzyme A, acetate itself served as a source of acetyl-CoA but only when ATP was supplied to the reaction mixture. Since these experiments were carried out in the dark, ADP and inorganic phosphate were ineffective as substitutes for ATP (Table 3). Table 4 shows that, under illumination and in the presence of coenzyme A, Chromatium extracts were able to use acetate as a precursor of acetyl-CoA, without an exogenous supply of ATP. ATP was formed from ADP and inorganic phosphate via photosynthetic phosphorylation by the chromatophores present in the extract. Without exogenous ATP, C1402 assimilation, in the presence of acetate, became light- dependent, retaining at the same time its requirements for ferredoxin, coenzyme A, H2, DPN, and manganese. The contribution of light was limited to the formation of ATP. This is evidenced by the effect of added ATP in the dark (Table 4) and also by the dual inhibitory effect of an added hexokinase-glucose system on the assimilation of C1402: in the dark when ATP was supplied, or in the light when ATP was formed from ADP and inorganic phosphate (Table 5). Concluding Remarks.-The present results with extracts of Chromatium and those reported earlier1 with extracts TABLE 5 of C. pasteurianum have revealed a new, EFFECT OF THE HEXOKINASE SYSTEM ON ASSIMILATION BY Chromatium ferredoxin-dependent primaryC02 fixation EXTRACTSC'402 IN THE LIGHT AND IN THE DARK reaction, based on the operation of the Total C1402 assimilated pyruvate synthase system in at least some Treatment (cpm) photosynthetic and nonphotosynthetic Dark, control 800 Dark, ATP 17,500 anaerobes. In extracts of C. pasteuria- Dark, ATP, hexokinase 1,400 Dark, ATP, hexokinase, glu- num, relatively large amounts of pyru- cose60600 vate accumulate, but in extracts of Chro- Light 9,200 matium, pyruvate does not accumulate Light, hexokinase 700 unless a trapping agent is supplied. In- Light, hexokinase, glucose 600 The reaction was carried out at 200 for 60 min. stead, Chromatium extracts convert the Each vessel contained 10 jumoles acetate and 5 newly to amino acids, MgCl2; 2 jsmnoles ATP, 0.75 mg crystal- newly formed pyruvate to amino acids, ~~~Aimolesline hexokinase (Sigma Chemical Co., St. Louis, provided that an amino group donor and indicated.Mo.), and 20Otherpmnolesexperimentalglucose wereconditionsadded wherewere for 2 and for DPNDPNarearepresent.present. ~~~~~~~Tableas given4 ("light"Fig.treatments).("dark" treatments) Downloaded by guest on October 1, 2021 846 BIOCHEMISTRY: BUCHANAN, BACHOFEN, AND ARNON PROC. N. A. S.

These results are of particular interest because amino acids are the main soluble products of photosynthesis in Chromatium.10 11 It has been generally accepted that the primary CO2 fixation reaction in the synthesis of amino acids by Chroma- tium must always be the carboxylation of ribulose diphosphate,27 resulting in the formation of 3-phosphoglycerate, but the evidence for this view is no longer com- pelling. The pyruvate synthase system is an alternative pathway for the formation of C3 compounds, leading to amino acid synthesis, that is independent of ribulose 1,5-diphosphate carboxylase, hitherto regarded as the enzyme that is never by- passed in autotrophic assimilation of CO2. Additional evidence that this enzyme has indeed been bypassed in the present experiments was found in the insensitivity of the pyruvate synthase system to cyanide-an inhibitor to which ribulose 1,5- diphosphate carboxylase is especially sensitive.25 A primary fixation of CO2 by way of the pyruvate synthase system makes use of the strong reducing potential of ferredoxin directly, without the mediation of pyridine nucleotides, thereby avoiding a drop in potential of about 100 mv.14 The experiments with Chromatium extracts have shown that, in photosynthetic bacteria, this efficient use of reducing power is accompanied by the contribution of radiant energy for the synthesis of ATP by way of photosynthetic phosphoryla- tion. Abbreviations: Acetyl-CoA, acetyl coenzyme A thiolester; DEAE, diethylaminoethyl; DPN, TPN, di- and triphosphopyridine nucleotide; ATP, adenosine triphosphate; ADP, adenosine diphosphate. * Aided by grants from the USPHS and the Office of Naval Research. 1 Bachofen, R., B. B. Buchanan, and D. I. Arnon, these PROCEEDINGS, 51, 690 (1964). 2 Wood, H. G., and R. L. Stjernholm, in The Bacteria, ed. R. Y. Stanier and I. C. Gunsalus (New York: Academic Press, 1962), vol. 3, p. 41. 3 Cutinelli, C., G. Ehrensvard, L. Reio, E. Saluste, and R. Stjernholm, Arkiv. Kemi, 3, 315 (1951). 4 Tomlinson, N., J. Biol. Chem., 209, 597 (1954). 6 Sadler, N. R., and R. Y. Stanier, these PROCEEDINGS, 46, 1328 (1960). 6 Hoare, D. S., Biochem. J., 87, 284 (1963). 7 Krampitz, L. O., in The Bacteria, ed. R. Y. Stanier and I. C. Gunsalus (New York: Academic Press, 1961), vol. 2, p. 209. .8 Elsden, S. R., in The Bacteria, ed. R. Y. Stanier and I. C. Gunsalus (New York: Academic Press, 1962), vol. 3, p. 1. 9 Elsden, S. R., Federation Proc., 21, 1047 (1962). 10 Losada, M., A. V. Trebst, S. Ogata, and D. I. Arnon, Nature, 186, 753 (1960). " Fuller, R. C., R. M. Smillie, E. C. Sisler, and H. L. Kornberg, J. Biol. Chem., 236, 2140 (1961). 12 Losada, M., F. R. Whatley, and D. I. Arnon, Nature, 190, 606 (1961). 13 Mortenson, L. E., R. C. Valentine, and J. E. Carnahan, Biochem. Biophys. Res. Commun., 7, 448 (1962). 14 Tagawa, K., and D. I. Arnon, Nature, 195, 537 (1962); see also Tagawa, K., and D. I. Arnon, Modern Methods of Plant Analysis, 7, 595 (1964). 16 Bachofen, R., and D. I. Arnon, manuscript in preparation. 16 Van Niel, C. B., Ark. Mikrobiol., 3, 1 (1931); Muller, F. M., Ark. Mikrobiol., 4, 131 (1933). 17 Arnon, D. I., V. S. R. Das, and J. D. Anderson, in "Studies on microalgae and photosyn- thetic bacteria," Plant Cell Physiol., special issue (1963), p. 529. 18 Rabinowitz, J. C., and W. E. Pricer, Jr., J. Biol. Chem., 237, 2898 (1962). 19 Anderson, I. C., and R. C. Fuller, Arch. Biochem. Biophys., 76, 168 (1958). Downloaded by guest on October 1, 2021 VOL. 52, 1964 ASTRONOMY: A. G. WILSON 847

20 Wolfe, R. S., and D. J. O'Kane, J. Biol. Chenm.., 205, 755 (1953). 21 Ibid., 215, 637 (1955). 22 Mortlock, R. P., R. C. Valentine, and R.. S. Wolfe, J. Biol. Chem., 234, 1653 (1959). 23 Stadtman, E. R., J. Biol. Chem., 196, 527, 535 (1952). 24 Lipmann, F., J. Biol. Chem., 155, 55 (1944). 2' Berg, P., J. Biol. Chem., 222, 991 (1956). 26Eisenberg, M. H., Biochim. Biophyp. Acta, 16, 58 (1955). 27 Quayle, J. R., Ann. Rev. Microbiol., 15, 119 (1961). 28Trebst, A. V., M. Losada, and D. I. Arnon, J. Biol. Chem., 235, 840 (1960). 29 Busch, H., R. B. Hurlbert, and V. R. Potter, J. Biol. Chem., 196, 717 (1952).

DISCRETIZED STRUCTURE IN THE DISTRIBUTION OF CLUSTERS OF GALAXIES* BY ALBERT G. WILSON

THE RAND CORPORATION, SANTA MONICA, CALIFORNIA Communicated by T. Y. ThomaS, July 20, 1964 From the basic Einstein equations positing the equivalence of the geometrical and physical tensors, RAB - 1/2RgAB = KTABE Edelen" 2 and Thomas3 have independently predicted discretization in the geo- metrical sizes of galaxies. Theory suggests that the diameters or major axes of the galaxies should be, under certain conditions, proportional to a sequence of eigen numbers of the form [n(n + 1) ]1/2, where n is a positive integer. This theoretically predicted size discretization has been observationally confirmed by Wilson4, 5 for samples of field and cluster galaxies in the case of ellipticals of small eccentricity. The evidence for discretized distributions in sizes of galaxies suggests the pos- sibility that there exist further regularities in the large-scale distributions of matter which have not been suspected. The present paper reports an observed regularity in the mean red shifts of clusters of galaxies which appears unlikely for randomly distributed clusters. There is at present no cosmological theory which predicts such regularities. This is not surprising, however, since most current cosmological models assume that the observed granularities in the distribution of matter- galaxies and clusters of galaxies-may be approximated by a uniform, smooth distribution. No theory which ignores the granularities can be held accountable for structure in the granularities. The epistemological difficulties which arise from smoothing assumptions have been well expressed by Neyman:6 "The contrast be- tween the domain of current observations of individual galaxies and their clusters on the one hand, and the theory dealing with the smoothed-out substratum on the other, is the more striking because every effort to verify empirically the conclusions of the theory must deal with observations of objects whose very existence this theory ignores. Thus, in order to effect a verification it is necessary to adopt a number of ad hoc hypotheses and, as a result, the conclusions are open to question." Furthermore, when observations are designed and interpreted with specific theoretical models, there is a tendency to ignore information which does not bear Downloaded by guest on October 1, 2021