VOL. 46, 1960 BIOCHEMISTRY: PULLMAN, SPANJAARD, AND BERTHIER 1011

that the cause of yield decline is due to a loss in soil fertility and not due to weedcompetition.

The author is indebted to the many individuals who assisted during the courseof the study, and particularly to Professors G. E. Hutchinson and E. S. Deevey ofYale University for their criticisms, suggestions, and facilities.

* The field work of this study was partiallypupported by the Henry and Grace E. DohertyCharitable Foundation, Inc.

'Emerson, R. A., Cornell Univ. Unpub. ms. (n.d).2 Hester, J. A., Jr., Ph.D. Thesis, Univ. California, Los Angeles (1954).3Kempton, J. H., Carnegie Inst. Wash. Rept Gov't Mexico 12th year Chichen ltza Project

and Allied Investigations (1935).4Steggerda, M., CarnepieInst. Wash. Pub. 531 (1941).5Sapper, K., Peterm. Mitth. 127, 1 (1899).6 Simmons, C. S., J. M. T~rano, and H. Pinto, Ministerio de Agricultura, Servicio Cooperativo

Interamericano de Agricultura, Guatemala, Unpub. ms. (n.d).7Walkley, A., and I. A. Black, Soil Sci., 37, 29 (1934).8 Jackson, M. L., Soil "Chemical Analysis," Prentice-Hall (1958).9 Truog, E., J. Am. Soc. Agron, 22, 874 (1930).

10 Kolterman, D. W., Ph.D. Thesis, Univ. Wisconsin (1952)."Reed, J. F., and R. W. Commings, Soil Sci., 59, 103 (1945).12 Stevens, W. L., Biometrika, 35, 346 (1948).

FEATURES OF THE ELECTRONIC STRUCTURE OF THEIRON-PORPHYRIN COMPLEXES WITH SPECIAL REFERENCE TO THE

OXIDO-REDUCTIVE PROPERTIES OF CYTOCHROMES*

BY BERNARD PULLMAN, CLAUDE SPANJAARD, AND GASTON BERTHIER

UNIVERSITk DE PARIS, INSTITUT DE BIOLOGIE PHYSICO-CHIMIQUE, PARIS

Communicated by Albert Szent-Gyorgyi, May 25, 1960

The method of molecular orbitals of quantum chemistry' has been used for thestudy of the electronic structure of the iron-porphyrin complexes of biologicalimportance. The hemoproteins which carry these complexes may be classifiedinto three principal groups: (a) the oxygen carrying hemoproteins, (b) the cyto-chromes, which are electron-transfer agents, and (c) several important enzymessuch as catalase or peroxidase.From the physico-chemical point of view the iron-porphyrin complexes may be

considered as being of two fundamental types2: (1) the so-called "ionic complexes"characterized by the presence of a large number of unpaired electrons (4 in ferrouscomplexes and 5 in ferric complexes), the outstanding examples of which are hemo-globin and myoglobin and (2) the so-called "covalent complexes" which either haveno unpaired electrons (e.g. oxyhemoglobin and ferrocytochrome c), or have onesuch electron (e.g. ferricytochrome c).3 In modern ligand field theory thesetwo types of complexes correspond to electron configurations for the metal ions inoctahedral coordination which are respectively of high and low spin.4 Effectivelyin this theory the five 3d levels of the transition metals are split in the field of theligand (a cubic force field) into two kinds:

1012 BIOCHEMISTRY: PULLMAN, SPANJAARD, AND BERTHIER PROC. N. A. S.

(a) those which point in the direction of maximum intensity of the field and whichmay form a bonds. These are the d,2 and d.2-y2 orbitals (de orbitals);

(b) those which bisect these directions and which may form or bonds. These arethe dxy dyze and dxz orbitals (df orbitals).The electron configurations for the Fe++ and Fe+++ ions in octahedral coordina-

tion which correspond to the two types of possible complexes are shown in Table1. The total spins are obtained by applying Hundt's rules of maximum multi-plicity.

TABLE 1ELECTRON CONFIGURATIONS OF THE CATION IN IRON-PORPHYRIN COMPLEXES

ElectronIon configuration Total spin Type

Fe++ (de)4 (dy)2 2 High spin-essentially ionic(d0)6 0 Low spin-essentially covalent

Fe + + + (de) I (dy)2 21/2 High spin-essentially ionic(dc-)5 l/2 Low spin-essentially covalent

As will be indicated in the description of the method of calculation, the cationmay act, in its interaction with the ligand, as an electron-acceptor in u- bond for-mation or as an electron-donor in wr bond formation. The distinction betweenthe two types of complexes corresponds largely to the participation of the 3delectrons to the binding in the covalent complexes and to the absence of such aparticipation in ionic complexes.

The Method.-The method employed is the molecular orbital method in its L.C.A.O. approxi-mation, overlap included, the fundamental aim of the work being the determination of the molec-ular orbitals which may establish themselves between the electronic orbitals of the porphyrinring and those of the metal cation. For reasons of simplicity most of our calculations have beencarried out as far as the porphyrins are concerned with an unsubstituted porphin ring. It seemsnevertheless highly probable that they are well representative of the general nature of the princi-pal types of electronic interactions which may manifest themselves between the metal cation andthe different types of substituted porphyrins corresponding to the compounds considered in thispaper. Most of the substituents fixed in biochemicals on the periphery of the porphin ring aresaturated groups but, in fact, even the direct conjugation of, say, two vinyl groups with the por-phin ring (as in protoporphyrin) may bring about only small modifications in the electronic proper-ties of the ring. This has been effectively established in the case of metal free porphyrins4a andmost probably holds also for the iron porphyrin complexes.As a preliminary condition for the establishment of molecular orbitals common to the cation-

porphyrin complex, there are the symmetry requirements. The porphin molecule is a planarconjugated skeleton belonging to the symmetry group D4O. Its 0- and ir molecular orbitals belongconsequently to the different irreducible representations of the group which are Ao,, A,,,, A2,, A2,,B1i, Blu, B2,, B2,, El,, and E1u.5

Building of the molecular orbitals of the complex: (1). vr-orbitals:-The Xr orbitals of the porphincorrespond to the Alu, A,,, Big, B2,, and El,, representations. The r orbitals of the iron cationwhich may combine with those of the porphin must belong to the same representations. Theones which obey this condition are 3dxy and 3d4, (Elg) and 4pz (A2,,). Consequently:

(a) In ionic complexes, whether ferrous or ferric, in which the d orbitals of iron do not par-ticipate in the binding, the only orbital of the iron which combines with those of the porphyrinis its 4p, orbital, which is vacant in the isolated metal. In the ferrous complexes the iron con-serves one 3d lone pair and four unpaired 3d electrons, in the ferric complexes it conserves five un-paired electrons.

(b) In covalent ferrous complexes, the iron contributes four 3d electrons to the Xr bindin'Z ofthe complex. These four d electrons come from the 3dx, and 3dy, orbitals. Effectively, themeasurement of magnetic susceptibilities or electron resonance6 of a number of metal complexes of

VOL. 46, 1960 BIOCHEMISTRY: PULLMAN, SPANJAARD, AND BERTHIER 1013

porphyrins and phthalocyanines have indicated, as the most probable arrangement of the 3d or-bitals in the metals, the following one:

3d.2 - 7/23d,2

3dZ, - 3dyz3d,7/

In the ferrous covalent complexes, the six available 3d electrons will occupy in pairs the threelowest orbitals: 3d,7/, 3dz, and 3d7/2. The only ones which have the proper symmetry and whichparticipate in the 7r bonding of the system are the four electrons occupying the 3dr, and the 3dyzorbitals. In the ferric covalent complexes, the cation contributes only three electrons to the 7rbinding. These come from the same 3dz, and 3d7,, orbitals which in this case contain only threeelectrons.7The 7r system of porphin involves 24 molecular orbitals and 26 electrons; that of the ionic fer-

rous or ferric complexes a system of 25 molecular orbitals and 26 electrons; that of the covalentferrous complexes a system of 27 molecular orbitals and 30 electrons; that of the covalent ferriccomplexes and system of 27 molecular orbitals and 29 electrons.

(2) a orbitals: The calculations on a binding were limited to the interaction between the cationand the four central nitrogens of the porphyrin. The orbitals of the iron which may combinewith those of the nitrogens are: the 3d,2 and the 4s orbitals (A,,), the 3d,2 - y2 orbital (B10), the4p, and 4p, orbitals (El.). In the ionic complexes, whether ferrous or ferric, the 3d orbitals re-main untouched and it is only the 4s, 4p,, and 4pv, vacant orbitals of the cation which combinewith the a orbitals of the nitrogens. The a system in this case is thus constituted only by theeight electrons supplied by the four nitrogens. It leads to seven 0f molecular orbitals. In thecovalent complexes, whether ferrous or ferric, the 3d,2 and 3d,2 - 7/2 orbitals combine also with thea orbitals of the nitrogens. But here again the iron does not contribute any electrons to thebinding. The a system in this case is thus constituted again by the eight electrons supplied bythe four nitrogens, but it leads to nine a molecular orbitals.

The parameters: The details on the evaluation of the appropriate parameters are presentedelsewhere.8 We shall indicate here merely the adopted values.The Coulomb integrals of the atoms being of the general form a1 = ac + 5Siyc-c (where ac

and yc-- c are, respectively, the Coulomb integrals of the carbon atom and the exchange integralsof carbon-carbon bonds, overlap included), the values adopted for the coefficient 5i of the variousheteroatoms are:

0.7 for the 7r orbitals of the porphyrin nitrogens1.9 for the hybridized a orbitals of the porphyrin nitrogens-0.35 for the 3d orbitals of Fe ++-0.30 for the 3d orbitals of Fe+++-3 for the 4p orbitals of Fe++ and Fe+++-1.8 for the 4s orbitals of Fe++ and Fe...

The exchange integrals have been evaluated by:

a2 + a3eij = ij-S, 2

+ aand,8ij = KSij 2

with K = 2 for wr bonds and K = 1.67 for af bonds (following Wolfsberg and Helmholz9), the over-lap integrals Si, being either calculated or obtained through interpolations in existing tables.Two more remarks concerning the calculations may be useful.(1) We have limited our study to the examination of the electronic interactions between the

cation and the porphyrin. The calculations do not include the effect of further coordinations withconjugated or other groups jwhich in the hemoproteins may be present above and below theplane of the iron-porphyrin complex.

(2) We have already presented in a previous publication results of calculations on the distribu-

1014 BIOCHEMISTRY: PULLMAN, SPANJAARD, AND BERTHIER PROC. N. A. S.

tion of energy levels and electronic charges in the porphinl'. The data presented in this paperfor the same molecule differ slightly from the previous ones. This is due to the slightly differentapproximations used in the two calculations. The general aspect of the results is, of course, verysimilar in the two publications.

Results and Discussion.-(1) Distribution of energy levels: Table 2 representsschematically the energies of the molecular orbitals in the porphin ring and in thedifferent types of iron-porphyrin complexes. As is well known, these energies areof the general form Ej = a + miy where a is the Coulomb integral and My the reso-nance integral, overlap included, of the molecular orbital method. Generally,positive values of mi correspond to occupied (bonding) molecular orbitals, negativevalues of m1 to empty (antibonding) orbitals. The smallest positive value of mcorresponds then to the highest occupied molecular orbital and the smallest negativevalue of m1 to the lowest empty molecular orbital.

TABLE 2ENERGIES OF MOLECULAR ORBITALS AND THEIR TYPES

Ionic ferro- or Covalent CovalentPorphin ferriporphyrins ferroporphyrins ferriporphyrins

-5.721 Of -5.720 a

-5.051 X -5.051 r -5.051 r-4.503 a -4.321 a -4.321 f-4.321 Of -4.321 a -4.321 f-4.321 a -3.985 X -3.985 IT

-3.913 X -3.913 X -3.985 X -3.985 r-3.878 X -3.878 r -3.913 7 -3.913 X

m- 3.878 X -3.878 X -3.152 a -3.129 aEmpty molecular - 3.830 7 -2.856 X -2.856 X -2.856 rorbitals -2.677 X -2.677 7r -2.677 7 -2.677X

-2.235 X -2.235 r -2.545 X -2.543 X-2.235 X -2.235 X -2.545 X -2.543 X-1.918 Xr -1.918 7 -1.918 7 -1.918 X-0.728 r -0.728 X -0.806 X -0.781 X-0.252 X -0.252 X -0.806 X -0.781 X-0.252 X -0.252 X -0.728 7 -0.728 X

................-0.682 a -0.645 a... .... .. ... ... ..- 0.085 Xr -0.075*X.....-0.085 X -0.075* r

+0.309 7 +0.331 X +0.331 7 +0.331 X+0.537 7 +0.537 X +0.537 X +0.537 X+0.770 X +0.770 X +0.771 X +0.771 X+0.770 X +0.770 X +0.782 X +0.782 X+0.771 7 +0.771 X +0.870 X +0.874 X+0.773 X +0.782 v +0.870 F +0.874 F

. ~~+0.7735 X +0.782 Xr +0.870 X +0.874 F

Occupied molecular +0.935 X +0.935 7 +0.970 X +0.973 Forbitals ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~+0.9357r +0.935 w +0.970 F +0.973orbitals +1.159 7r +1.159 n +1.159 w +1.159w

+1.493 X +1.493 7X +1.493 X +1.493 X+1.572 7 +1.572 F +1.617 X +1.620 F+1.572 Fr +1.572 F +1.617 71 +1.620 Fr+1.621 71 +1.621 F +1.621 F +1.621 F

+1.900 a +1.961 a +1.961 aI ~~~~~+1.961 a' +1.961 a' +1.961 or

+1.961 a +2.087 a +2.094 or

+1.963 a +2.241 a +2.256 f

* Two degenerate 7 orbitals containing 3 electrons.

We shall be essentially interested in this discussion in the values of the energiesof the 7r orbitals and in particular in those of the highest occupied and lowest empty ofthese orbitals. These two values are effectively of particular importance, being

VOL. 46, 1960 BIOCHEMISTRY: PULLMAN, SPANJAARD, AND BERTHIER 1015

related, respectively, to the ionization potential or the electron-donor capacity andthe electroaffinity or electron-acceptor capacity of the compounds."The examination of the data of Table 2 indicates that as far as the values of the

energies of these two essential orbitals are concerned there is a clear-cut distinction be-tween the ionic and the covalent complexes.Thus in porphin itself the mi coefficients of both the highest filled and the lowest

empty molecular orbitals have relatively small absolute values. As already indi-cated in a previous publication by Pullman and Perault'0 this signifies that porphinshould be a relatively good ir electron-donor and 7r electron-acceptor. It can be seenthat essentially the same situation prevails in the ionic ferro- and ferriporphyrins.In fact, the energy of the lowest empty 7r molecular orbital is even practically thesame in these complexes and in porphin; that of the highest filled 7 molecularorbital is only slightly lowered. Consequently the ionic ferro- and ferriporphyrinsshould be both relatively good 7r electron-donors or acceptors. These r electrons areessentially those of the porphin skeleton. The 3d electrons of iron do not partici-pate in the binding. In fact, it is these 3d electrons, whose orbitals lie above thebinding 7r orbitals that will be essentially involved in an oxido-reduction processconcerning these compounds.

The situation is entirely different in the covalent complexes. In these complexesthe 7 orbitals involve both the pz electrons of the porphin and the d electrons of the ironand the energies of the highest filled and lowest empty 7r molecular orbitals are dis-tributed in a manner completely different from that of the ionic complexes. Thus inthe covalent ferroporphyrins the highest filled orbital, which is doubly degeneratedand carries four electrons, is very high lying. In fact, the sign of its coefficientm, is even that which is generally associated with molecular orbitals which may onlybe occupied in the excited states of molecules. This situation has already beenencountered previously in a few other biochemicals, e.g., reduced riboflavinleucomethylene blue, chlorpromazine etc... 12 It means that the compound willpossess particularly strong electron-donor properties. The lowest empty r molecularorbital has been raised very appreciably, so that the compound should not haveany marked electron-acceptor properties.

In connection with the electron-donor properties of the covalent ferroporphyrinsit may also be particularly useful to remark that although it is difficult to determineby theoretical calculation the energy of the iron 3drV lone pair with respect to theenergies of the molecular 7r orbitals, experimental electron resonance studies seemto indicate that the 3d2, orbital is more stable than the highest 7 molecular orbi-tal.7Y 1 Consequently, the oxidation of this type of complexes must involve effec-tively the departure of an electron from the highest filled or molecular orbitalwhich, as stated and as will be seen in more detail in the next part of this paper,contains electrons originating from the 7r orbitals of the porphin and the 3drz and3dyz orbitals of the iron.

In the covalent ferriporphyrins the situation is at first sight quite similar tothat just discussed for the covalent ferroporphyrins: very high lying doubtly de-generated highest filled 7r molecular orbital and high lying, too, lowest empty 7rmolecular orbital. But there is one essential difference between these two cases.In the covalent ferroporphyrins the two degenerate highest filled 7 molecularorbitals were both entirely filled and contained, each of them, two electrons. In

1016 BIOCHEMISTRY: PULLMAN, SPANJAARD, AND BERTHIER PROC. N. A. S.

the covalent ferriporphyrins there are only three electrons to occupy the two de-generated highest filled orbitals. One of these orbitals is thus only half-filled.Being able to accept one more electron it must consequently be considered also as thelowest empty orbital. The coefficient of this orbital being very small in absolute value,this signifies that the compound should possess particularly pronounced electron-acceptor properties. One must, of course, bear in mind that, for the same reason,the compound should also be able to manifest pronounced electron-donor properties.Nevertheless, it may be remarked that the departure of one more electron from thedoubly degenerated highest filled orbital of a covalent ferriporphyrin would lead toa highly reactive and consequently probably unstable biradical.The results just discussed on the distribution of the energies of the ir-molecular

orbitals in the covalent complexes seem to be particularly useful for the inter-pretation of the biological function of the cytochromes. In fact, the cytochromesprobably have a number of highly specialized functions, but the most importantgroup of them seems to be involved in the oxidation chain in the respiration of agreat many tissues. In particular, they represent the essential electron carriersbetween the reduced flavoproteins and molecular oxygen. They thus participatein a series of reactions in which the complexes shuttle between the ferrous and ferricstates. The main cytochrome chain, which is located in the mitochondria, com-prises cyt. b, c, a, and a3 The only member of this chain which has been separatedfrom the mitochondria and completely purified is cytochrome c.14 The calculationspresented in Table 2, although representative to a large extent of the general featuresof the distribution of energy levels in the cytochromes, refer in particular to thiscytochrome. Effectively the substituents which are present, in cytochrome c,on the periphery of the porphyrin ring are either saturated groups or if they containmobile electrons these are far away from the ring and separated from it by CH2groups, so that these substituents do not exert any important influence on thecloud of the ir electrons of the complex.

B. Pullman and A. Pullman have shown in a previous study'2 that the mecha-nism of the functioning of the respiratory coenzymes, diphospho- or triphospho-pyridine nucleotides and flavoproteins, may be related to the energies of the highestfilled and lowest empty 7r molecular orbitals of the oxidized and reduced forms ofthese compounds. It now becomes obvious that the mode of action of the cyto-chromes probably involves a mechanism which is very similar to that proposed for thecoenzymes. Thus the existence of a very high-lying, in fact even antibonding,highest filled molecular orbital in ferrocytochromes (covalent ferroporphyrin) sig-nifies that these molecules must possess extremely pronounced electron-donorproperties. Similarly, the existence of a very low-lying half-empty molecularorbital in ferricytochromes (covalent ferriporphyrin) signifies that this form of thesemolecules must possess extremely pronounced electron-acceptor properties. Thus,just as in the case of the respiratory coenzymes, the oxido-reduction of the cyto-chromes is associated with an organization of the molecular orbitals involved in thedeparture or arrival of an electron which is of such a nature that in each case thereis a particularly high-flying filled orbital associated with the reduced form and aparticularly low-lying empty orbital associated with the oxidized form. These twoforms will consequently have a natural tendency of, respectively, giving off oraccepting an electron. It may be observed that this situation does not involve,

VOL. 46, 1960 BIOCHEMISTRY: PULLMAN, SPANJAARD, AND BERTHIER 1017

as it does in the respiratory coenzymes, any drastic re-distribution of the energiesof the two involved molecular orbitals.The existence of oxidation states higher than Fe3+ does not appear to have

been observed in the cytochromes. The calculated energy of the highest occupied7r molecular orbital in ferricytochrome indicates the possibility of formation ofsuch states. Following the authoritative opinion of Slater,"5 it would not be sur-prising if such states were formed. 16 However, it must be remarked that a covalentiron-porphyrin in the oxidation state of Fe4+ would involve two unpaired electronsat the two degenerate highest occupied 7r molecular orbitals and would thus be ahighly reactive and consequently unstable biradical.As quoted before, the situation described here and which applies directly to

cytochrome c is representative of the whole family of cytochromes. It is true thatin cytochromes b and a there are complementary r-electron-possessing substituentsdirectly conjugated with the porphyrin ring (two vinyl substituents in cytochromeb and one vinyl and one formyl substituent in cytochrome a), but it seems plausibleto admit that the presence of these substituents can produce only very small modi-fications in the electronic structure of the unsubstituted iron-porphyrins. Underthese conditions it seems probable that the differences observed in the biochemicalbehavior of the different cytochromes, in particular their autoxidability or non-autoxidability, are related to other structural factors which unfortunately had to beneglected in this study too, and which may be the nature of the two complementarygroups coordinated with the metal above and below the plane of the iron-porphyrincomplex and the interactions between the complex and the protein portion of thehemoprotein. All these interactions exert a very strong influence on the effectiveelectronegativity of the metal cation, this electronegativity being also very sensi-tive to the modifications of the bond distances between the cation and the co-ordinating atoms. (See, e.g., ref. 3.) Such modifications of electronegativityare able to exert appreciable changes in the distribution of electronic indices.

(2) Distribution of the electronic charges:-The examination of the distributionof the electronic charges brings complementary information on the differences ofstructure between the two types of iron-porphyrin complexes and also, in partic-ular, on the nature of the interaction between the porphyrin and the iron electronsin the cytochromes.The distribution of the mobile electronic charges in the different types of com-

plexes studied is illustrated in Figure 1. It may be observed that:(a) In the ionic complexes (whether ferrous or ferric) the iron atom gains a

very small fraction of the 7r electron system of the porphyrin.(b) In the covalent complexes the iron gives away a large fraction of its w

(d) electrons to the porphyrin ring. The fraction so transferred represents 4e -1.442e = 2.558e in the ferrous complexes and 3e - 1.325e = 1.675e in the ferriccomplexes. As might have been expected3' 4the trivalent cation acts as a weakerde electron-donor than the divalent cation.

This distribution of the 7r electrons needs to be completed by the distribution ofthe ar electrons in the central part of the molecules, this distribution being asso-ciated with strong charge displacements. This is illustrated in Figure 2. In alltypes of complexes the displacement of the a charges is in favor of the iron cation,which gains variable quantities of these a charges, donated by the nitrogens of the

1018 BIOCHEMISTRY: PULLMAN, SPANJAARD, AND BERTHIER PROC. N. A. S.

1,040 1,039

0,976 0,9770934A0,9~~043 09395

N 1,525 N 6

NH HN N -Fe-N0,065

N N

PORPHIN IONIC COMPLEXES(FERROUS AND FERRIC)

1,151 1,112

1071156 11d11096N1,488 N1'7

N-Fe -N N. N1,442 -Fe-N

N N

COVALENT FERROUS COVALENT FERRICCOMPLEXES COMPLEXES

FIG. 1.-Distribution of ir electrons.

N1'86 N 1'7 N1'729

*- Fe-N 1N-Fe-N N-Fe-N

NN N

IONIC COMPLEXES COVALENT FERROUS COVALENT FERRICCOMPLEXES COMPLEXES

FIG. 2.-Distribution of the a electrons in the central part of the complexes.

porphyrin ring. This gain is much greater in the covalent complexes than in theionic ones, and slightly greater in the covalent ferriporphyrins than in the covalentferroporphyrins.The superposition of the charge distributions of Figures 1 and 2 leads to the

distribution of total charges. The cation thus bears a total electronic charge of0.520e in the ionic complexes, of 2.472e in the covalent ferrous complexes and of2.410e in the covalent ferric complexes. Its formal charges, with respect to FeO,are consequently of 1.480e in ionic ferroporphyrins, of 2.480e in ionic ferriporphyrins,

VoL. 46, 1960 BIOCHEMISTRY: PULLMAN, SPANJAARD, AND BERTHIER 1019

of 3.528e in covalent ferroporphyrins and of 3.590e in covalent ferriporphyrins.It is obvious that because of its higher formal positive charge the ferrous cation of acytochrome will be appreciably more electronegative than the ferrous cation ofhemoglobin.

It may also be observed that the concentrations of electronic charges on theperipheral carbon atoms of the covalent complexes are much greater than those ofthe ionic complexes.Another aspect of the electronic distribution which it may be useful to indicate,

particularly in relation tothe discussion on the 01037 0,037oxido-reduction of cyto- 0os,0 054 003610051chromes, concerns the | N004 N0003form of the molecular or-bitals involved in this N- Fe N N-Fe-Nmechanism. Figure 3 1°0,93 10,205presents the distribution N Nof an electron on the high-est filled orbital of the co-valent complexes. It isseen that the electron is (a) (b)

FIG. 3.-Distribution of an electron on the highest filled orbitalgreatly spread over the of covalent (a) ferroporphyrins, (b) ferriporphyrins.whole molecular skeletonand although the largest local concentration of it is found on the Fe cation, this partrepresents only 2/lo of the electron, the remaining 8/lo being distributed essentiallyover the 16 C atoms of the porphyrin.

* Sponsored by the U.S. Public Health Service, Grant 3073 (National Cancer Institute).1 For the general description of the method see, e.g., B. Pullman and A. Pullman, Les theories

electroniques de la chimie organique (Paris: Masson, 1952).2 See, e.g., J. Wyman, Jr., Advances in Protein Chemistry, 4, 407 (1948), or S. Granick in Chemi-

cal Pathways of Metabolism, ed. D. M. Greenberg (New York: Academic Press, 1954), vol. 1I,p. 287.

a Williams, R. J. P., Chem. Rev., 56, 299 (1956).4 Williams, R. J. P., in The Enzymes, ed. P. D. Boyer, H. Lardy, and K. Myrbick (New York:

Academic Press, 1959), vol. 1, p. 391.4a Pullman, B., and Pullman A., Results of Quantum Mechanical Calculations of the Electronic

Structure of Biochemicals (Paris: Institut de Biologie Physico-Chimique, 1960), Vol. 1.5See, e.g., H. Eyring, J. Walter, and G. E. Kimball, Quantum Chemistry (New York: J. Wiley

and Sons, 1944), Appendix V1I.6 Griffith, J. S., Disc. Far. Soc., 26, 81 (1958); Gibson, J. F., D. J. E. Ingram, and D. Schonland,

Disc. Far. Soc., 26, 72 (1958).7 Griffith, J. S., Nature, 180, 31 (1957).8 Spanjaard, C., Dipl6me d'gtudes superieures, Facult6 des Sciences de Paris (1960).9 Wolfsberg, M., aid L. Helmhofz, J. Chem. Phys., 20, 837 (1952).10 Pullman, B., and A. M. Perault, these PROCEEDINGS, 45, 1476 (1959).11 Pullman, B., and A. Pullman, these PROCEEDINGS, 44, 1197 (1958).12 Pullman, B.. and A. Pullman, these PROCEEDINGS, 45, 136 (1959); Biochim. Biophys. Acta,

35, 535 (1959); G. Karreman, 1. Isenberg, and A. Szent-Gy6rgyi, Science, 130, 1191 (1959).'3 Gibson, J. F., and D. J. E. Ingram, Nature, 180, 29 (1957).14Ehrenberg, A., and H. Theorell, Nature, 176, 158 (1955)."5Slater, E. C., in Biochemical Society Symposia, ed. E. M. Crook (Cambridge: University

Press, 1958), no. 15, "Metals and Enzyme Activity," p. 76.

1020 BIOCHEMISTRY: S. B. WEISS PROC. N. A. S.

16Such higher oxidation states of iron have been observed in peroxidase and ferrimyoglobins:Theorell, H., A. Ehrenberg, and B. Chance, Arch. Biochem. Biophys., 37, 237 (1952); George, P.,and D. H. Irvine, J. Chem. Soc., 3142 (1954); George, P., and D. H. Irvine, Biochem. J., 60, 596(1955); George, P., and J. S. Griffith in The Enzymes, ed. P. D. Boyer, H. Lardy, and K. Myrback(New York: Adademic Press, 1959), vol. I, p. 347.

ENZYMATIC INCORPORATION OF RIBONUCLEOSIDETRIPHOSPHA TES INTO THE INTERPOLYNUCLEOTIDE LINKAGES OF

RIBONUCLEIC ACID*

BY SAMUEL B. WEISS

ARGONNE CANCER RESEARCH HOSPITAL, AND THE DEPARTMENT OF BIOCHEMISTRY,THE UNIVERSITY OF CHICAGO

Communicated by Charles Huggins, June 20, 1960

Two separate reactions have been described for the enzymatic polymerizationof nucleotides. In 1955, Grunberg-Manago and Ochoal reported that an enzymeisolated from the microorganism Azotobacter vinelandii catalyzes the synthesisof highly polymerized ribonucleotides from 5'-ribonucleoside diphosphates withthe release of inorganic phosphate. Because of its similarity to the action ofphosphorylase on polysaccharides, i.e., its reversible phosphorolysis, the enzymewas called polynucleotide phosphorylase. Since polynucleotide phosphorylaseacts on single as well as mixtures of ribonucleoside diphosphates, to form singleor mixed polymers, its role as a general mechanism for the biosynthesis of specificribonucleic acids has been debated.2 This enzyme is widely distributed in micro-organisms and its presence in higher forms has been briefly reported by Hilmoeand Heppel.3DNA4 polymerase was first described in 1956 by Kornberg, Lehman, and Simms5

in extracts from E. coli. Further investigations by this group6 clearly showed thatthis enzyme catalyzes a net synthesis of deoxyribonucleic acid which is dependenton the presence of all four deoxynucleoside triphosphates, as well as magnesiumions and "primer" DNA, and occurs with the elimination of inorganic pyrophos-phate. A mammalian system which catalyzes a reaction similar to DNA poly-merase has been reported.7

Terminal addition of one or more ribonucleotides to RNA molecules is anotherreaction observed by a number of investigators.8-13 This reaction is specific forthe ribonucleoside triphosphate and results in the release of inorganic pyrophos-phate. However, no extensive polymerization occurs.

In view of the importance of specific ribonucleic acid molecules as carriers forindividual amino acids in protein synthesis,14' 1 and the widely held view thatRNA may provide templates for the synthesis of specific proteins, the presentwork was undertaken. In the course of this investigation, a new type of reactionfor the incorporation of ribonucleotides has been found. This paper will describethe enzymatic incorporation of ribonucleoside triphosphates into the interpoly-nucleotide linkages of RNA, by an enzyme system of mammalian origin, which