Metalloproteins in the evolution of photosynthesis

57

BioSystems, 14 (1981) 57--80 © Elsevier/North-Holland Scientific Publishers Ltd.

M E T A L L O P R O T E I N S IN THE E V O L U T I O N OF PHOTOSYNTHESIS

R. CAMMACK, K.K. RAO and D.O. HALL

University o f London, King's College, 68, Half Moon Lane, London SE 24 9JF, United Kingdom (Received January 8th, 1981) Certain metalloproteins are common to all photosynthetic electron transfer chains. These include soluble proteins such as ferredoxins and cytochromes of tlle c2 type, and membrane-bound components such as cytochrome b, c~ and the Rieske iron-sulphur protein. The sequence of electron transfer Quinone ~ (cyt b, Fe-S, cyt cl) ~ cyt c~ indicates a common precursor to these systems and to the mitochondrial respiratory chain. In cyanobacteria the cytochrome c~ can be interchanged with the copper protein plastocyanin, and furthermore in chloroplasts of higher plants the latter is used exclusively. The ferredoxins in anaerobic photosynthetic bacteria a r e mostly of the [4Fe-4S] type, probably derived from those of the fermentative bacteria. These could readily be formed in the earliest cells from iron, sulphide and a very simple peptide. In the oxygen-evolving cyanobacteria and the aerobic halobacteria the [2Fe-2S] ferredoxins predominate. The electron transfer chains of the cyanobacteria have been incorporated almost unchanged into the chloroplasts of plants. The electron transfer chains of purple photosynthetic bacteria were probably the precursors of the mitochondrial respiratory chain, as shown by similarities of cytochromes c 2 and succinate dehydrogenase. However a different origin of the eukaryotic cytoplasm is indicated by the presence of the copper/zinc superoxide dismutase.

Photosynthetic microorganisms must have appeared early in the history o f Life on Earth. At first, there would have been sources of chemical energy available, such as fermentable compounds, but in order for life to be sustained on a global scale the organisms must have developed the ability to harness the renewable energy from the sun. Sediments 3.4 × 109 years old have been found to contain stromatolites, which are usually taken as evidence for the existence of photosynthetic microorganisms {Walter et al., 1980). (It is however difficult to determine whether these or other microfo,~sils represent photosynthetic organisms or not). There is currently much discussion a b o u t the conditions on the Earth which prevailed during the early stages of the evolution of life. Comparison with the atmospheres of the other inner planets has suggested that tile atmosphere of the Earth soon after it fo:rmed consisted of oxidized gases, principally carbon dioxide, carbon m o n o x i d e and ni~Lrogen, and that liquid water would be present (Henderson-Sellers et al.,

1980). Hydrogen would be present only in small amounts as it would diffuse into space. This is in contrast to the earlier hypothesis of an atmosphere o f reducing gases such as methane and ammonia (see Berkner and Marshall, 1965). In either case, the geological evidence clearly indicates that molecular oxygen was absent. It seems reasonable to suppose that the first organisms would use the simplest metabolic system consistent with the energy resources available, and that as these resources were depleted, more elaborate systems would evolve. Thus it is thought that the first organisms carried o u t fermentation of the organic c o m p o u n d s that were formed abiogenically, just as the anaerobic bacteria such as C l o s t r i d i u m existing today. As these organic compounds became depleted, organisms developed that could regenerate them b y photosynthesis (Broda, 1975). It is likely that there are no remains o f the earliest photosynthetic organisms, and we can only speculate as to their form and metabolism.

58 There was probably a period of rapid developm e n t in which the most efficient systems survived. The components of these systems that developed are found in the energy metabolism of present-day organisms: the use of chlorophylls as light-harvesting pigments, electron transport associated with the generation o f a transmembrane potential, and the formation of ATP by a vectorial ATPase. By examining the photosynthetic organisms of the present day we can hope to determine what types of metabolism are feasible, and may have existed in the early forms of life. Also by a detailed study of the structure and sequence o f their proteins, it may be possible to determine the course of evolutionary divergence from these ancestral forms.

Types of microbial photosynthesis The photosynthetic process can be summarised as H:X + (acceptor)~h_~t X + (reduced acceptor) + A D P + Pi + ATP

(1) (reduced acceptor) + CO2 + ATP -* [CH20] + A D P + Pi (2) The thermodynamically unfavourable first reaction is driven by a light-dependent reaction involving chlorophyll and utilizes a membrane-bound electron transport system to reduce an acceptor (either ferredoxin or pyridine nucleotide) (Crofts and Wood, 1978). Additional ATP can be formed by cyclic electron transport leading to a transmembrane potential with no net electron transport to an acceptor. The ATP and reduced acceptor are then used in CO: fixation to form carbohydrate. Photosynthetic bacteria can be divided into greens and purples, depending on the t y p e of photosynthetic pigments present (Clayton and Sistrom, 1978). In the sulphur bacteria such as the green Chlorobium or purple

Chromatium the electron d o n o r is an inorganic c o m p o u n d such as sulphide or thiosulphate (Fig. 1). These bacteria are strict anaerobes. In the non-sulphur bacteria such as the green Chloroflexus (Castenholz, 1979) or purple Rhodospirillum the donors are organic c o m p o u n d s such as succinate (Fig. 1). The non-sulphur bacteria can adapt to the presence o f oxygen, when they repress the formation of chlorophyll and develop a cytochrome c oxidase (either c y t o c h r o m e a/a3 or o, or both}. U n d e r these conditions there is a strong similarity to the respiratory chain of mitochondria (Fig. 2). The cyanobacteria (blue-green algae) carry o u t a more complex photosynthetic reaction in which water is the electron donor, being converted to oxygen. Although the reduction of pyridine nucleotide by water using the energy of one quantum o f light is thermodynamically feasible, no photosynthetic organism can achieve it. To carry o u t the electron transfer requires two light-dependent reactions, photosystems II and I, as outlined in Fig. 3. In some ways, P h o t o s y s t e m I is analogous to the reaction centre o f the green photosynthetic b a c t e r i a (Crofts and Wood, 1978}. Both have low-potential iron-sulphur proteins as electron acceptors (Malkin and Bearden, 1978}. This analogy is supported by the observation that certain cyanobacteria such as Oscillatoria limnetica can grow in the presence of sulphide and use it as an electron donor instead of water, suppressing Photosystem II (Padan, 1979). Photosystem II is analogous to the reaction centre of the purple non-sulphur bacteria in its redox potential span (Fig. 1). The system that produces oxygen from water is a unique feature of cyanobacterial and plant photosynthesis, and appears to contain manganese although its structure has n o t y e t been defined. Olson (1978) has suggested that it might have originated in bacteria that used nitrogenous c o m p o u n d s such as hydrazine or nitrite as electron donors. Lumsden and Hall (1975) have n o t e d that superoxide dismutase, an enzyme designed to

59

EomV -06

PL,~RPLEBACTERIA

GREENBACTERIA

( Fe-S?I~.--..,.Fd.~

-Of,

FAO NAO

-02

i.a_.

I F~'s b~ ATP

h~.

F_~bf ~ I Fe-S

11"

hV

lre-sI

UOorMO~

0

\

ADP ~ATP

cyt.c2

PSi0

^, PS"/0 u~ Bchlaorb I Cyfochrome~"AOP [ oxi,:lase ]~ATP

08

,,..J'H2S M~Q~.ICyfc553

Sch[a

02

Fig. 1.

Schematic representations of electron transport chains of purple non~ulphur bacteria and green sulphur bacteria. Not all the k n o w n electron carriers are included.

(ov - -06

• -04

MITOCHONDRIA NAD LF?

0

06

08 Fig. 2.

.Succinafe

' J O ~

02 OZ.

~'-=ATP

"~ATP i

Cyfochrome I oxJCase

(ADP ",, ATP

[ O2

Outline sctieme of the mitochondrial respiratory chain.

alleviate the effects o f oxygen toxicity, is also a manganese-containing protein in some microorganisms, and suggest that the two proteins may have a common origin. All the prokaryotic photosynthetic organisms appear to have been present early in the history of the Earth. The cyanobacteria were probably the principal cause of the oxygen which increased dramatically in the atmosphere about 2.2 billion years ago immediately after the formation of the oxidized banded iron formations (see Broda, 1975). Towe (1978) has argued that oxygen could have been produced by photolysis prior to this. They fix CO2, producing reduced carbon compounds and 02. If these were simply recombined, for example by respiratory bacteria, no net accumulation of oxygen would take place. However, if the recycling of carbon compounds by fermentation, photosynthesis, etc., resulted in the production of some hydrogen, and this was lost from the planet, the result would be the loss of reducing equivalents and the net accumulation o f oxygen could take place:

60

E'o v

CHLOROPLASTS & CYANOBACTERIA /

- -0-6

Z

/

/

~

/

-0"~*

d ~ F AO

"~AOP

/

/ /

-0-2

/ / hV

/

Q"~pQ_

e"

0-2 ~ P C

or

Cytcss3 0-~

b4

A'rP

PHOTOSYSTEM I

- 06

0,

PHOTOSYSTEM 11 Fig. 3. Outline scheme of oxygen-evolving photosynthesis.

Evolutionary studies of metallopmteins Since the fossil record is uninformative as to what types of organism were present in the early stages of evolution, we have to rely on molecular evidence such as sequences of proteins and nucleic acids. Metalloproteins are relatively easy to study as markers of evolutionary change, particularly small molecules such as ferredoxins and cytochromes c which are soluble, and readily extracted and characterised by their optical absorption. They are present in many types of organism, adapted to different metabolic processes. An increasing number of amino acid sequences of metalloproteins are being determined. From these it is possible to deduce how the different metabolic processes are related to each other. In this article we are mainly concerned with the analogies between different types of photosynthetic processes, rather than the

differences between them. For example it is found that certain cytochromes and ironsulphur proteins are found in similar positions in the different electron transfer chains (Figs. 1--3). The detailed structures of these proteins can give information about how the bioenergetic systems of microorganisms, and the chloroplasts and mitochondria are related to each other (Schwartz and Dayhoff, 1978). The closer the similarity of sequences between proteins, the more recently they are likely to have diverged from a common ancestor. The study of evolutionary relationships by comparing amino acid or nucleotide sequences is well established for eukaryotes. In the animal kingdom the evidence of evolutionary divergence derived from sequences such as cytochrome c and haemoglobin agrees well with the fossil record. Evolution of the plants can therefore be investigated in this way with some confidence. However the derivation of

61 the course oil evolution by amino acid sequence and protein structure is more difficult for microorganisms. Firstly, the bacteria are metabolically more diverse, and proteins such as ferredoxins and c y t o c h r o m e c can take different functions in different organisms. Additional sections of amino acid sequence may be added to assist the protein in its interactions with particular enzymic systems. It is possible that proteins could evolve b y deletion as well as by insertion of such s e q u e n c ~ . Secondly, there is considerable scope for gene transfer between the bacteria. It is noticeable that in almost all cases the enzyme system for any particular metabolic process, e.g. protein synthesis, ATP formations, nitrogen fixation, and even the minor pathways used b y specialized organisms, is structurally similar in all microorganisms. This implies that once each system has been developed and optimized, it was incorporated into all organisms using that process; t h e y did n o t have to develop them separately. As an extreme view one ~ could consider a c o m m o n " p o o l " o f genetic material among the prokaryotes, so that each p r o k a r y o t e is a collection of genes, selected to specify the p .articular metabolic processes required for its own environment and way of life. In this view, the idea of an "evolutionary t r e e " o f the microorganisms would clearly be meaningless. Variations in protein structure would be adaptations to different conditions such as different extracellular pH, growth temperature, oxygen tension, etc. Since the evolutionary success of a microorganism depends on the most efficient use of energy resources, selection pressure on the efficiency of enzyme systems is stronger than in higher organisms. One 'therefore might expect there would be less scope for so-called " n e u t r a l " mutations in prol~ein sequence. In practice, it seems that it is n o t necessary to take so pessimistic a view and sequence data can be useful in determining the course of evolution o f photosynthetic organisms. The possibilities of convergent evolution of

proteins and genetic transfer from one organism to another must be taken into account. Also the sequences of a protein even from closely-related species may have undergone a considerable n u m b e r of amino acid substitutions over the long period of time during which these organisms have existed. The test of convergent evolution is to look at those features of a molecule that are n o t directly related to its function. If two ferredoxins, for example, resemble each other in t h e sequences a r o u n d the ironsulphur cluster, this could be convergent evolution because they both have to contain that cluster. However, if t h e y are similar in other regions of the chain it seems likely that they evolved from a c o m m o n ancestor. In fact it is f o u n d both for ferredoxins and cytochrome c that there are several families of proteins which are distantly related, if at all, to each other. Within each family there is a range of different proteins which are clearly related in sequence and threedimensional structure. The classification o f microorganisms by comparison of amino acid sequences should be simplest when comparing closely-related species. It is f o u n d that the similarities between different species do n o t correlate with the present classification of the photosynthetic bacteria, as for example with the sequences of cytochromes c {Ambler et al., 1979). However it must be remembered that the taxonomic criteria at present employed such as morphology and conditions of growth are designed principally as an aid to identification of species. The amino acid sequence data can be combined to obtain a different, self~onsistent classification on evolutionary lines (Dickerson, 1980a).

Iron-sulphur proteins The iron-sulphur proteins have a strong claim to have been among the first electron transfer agents to have evolved on Earth. The

62 ferredoxins and rubredoxins, which are the simplest molecules of this class, now occur abundantly in most types of microorganisms. They are small soluble protein molecules containing iron-sulphur clusters of various types (Fig. 4) (Lovenberg, 1973, 1977; Ran and Hall, 1977; Yoch and Carithers, 1979; Sweeney and Rabinowitz, 1980; Rao and Cammack, 1981). The work of R.H. Holm and his group has shown that compounds identical to the 1Fe, [2Fe-2S] and [4Fe-4S] clusters in the rubredoxins and ferredoxins can readily be formed non-enzymically by mixing an iron salt, a suitable thiol ligand and, where appropriate, sulphide (see Holm, 1977; Averill and Orme-Johnson, 1978}. These compounds are quite stable in the absence of oxygen, so long as excess thiol is present. Perhaps surprisingly the apparently complex [4Fe-4S] clusters are most readily formed by the majority of thiol compounds; the other

dusters require ligands with a particular steric restriction in the configuration of the sulphur atoms. It may be noted that among ferredoxins of the present day, the simplest ones, with the lowest molecular weight, contain [4Fe-4S] clusters. The first compounds synthesized by Holm's group were soluble in organic solvent only, but they showed that it is possible to make watersoluble derivatives with hydrophilic ligands. This has been extended to the synthesis of polypeptides containing several residues of the amino acid cysteine to form a primitive type of ferredoxin (Que et al., 1974; Gunter et al., 1979). When mixed with iron and sulphide, these compounds usually form [4Fe-4S] clusters. However certain short sequences appear to favour [ 2Fe-2S ] clusters (Balasubramanian, A., Ridge, B. and Rydon, H.N., unpublished observations). The functions of the protein in a ferredoxin therefore include protection of the

,¢~

¢~,

Fe

.¢~

Os 2

ill

I c.

Fig. 4. Structures of the one-, two-, three- and four-iron cluster types found in rubredoxin and ferredoxins. From Cammack, 1980. The Fe atom is bonded to cysteine S in rubredoxin (upper left).

63 cluster f r o m o x y g e n (if present) and determination o f the t y p e o f iron-sulphur cluster. This in t u r n will d e t e r m i n e t he range o f r e d o x potential over which the p r o t e i n will operate. T h e synthetic iron-sulphur c o m p o u n d s can undergo o x i d a t i o n and reduction. In water the m i d p o i n t r e d u c t i o n potentials (--400 mV to - - 5 0 0 mV) are comparable with those in ferredoxins from anaerobic bacteria such as C l o s t r i d i u r n or f r o m chloroplasts and algae (Hill et al., 1977). Moreover t he c o m p o u n d s have been f o u n d to substitute f or ferredoxins as electron transfer agents in biological systems. Adams et al. (1977, 1980) showed

TABLE

t hat water-soluble [4Fe-4S] c o m p o u n d s can transfer electrons efficiently t o bacterial hydrogenase. This shows t h a t during t he early stages o f evolution a very simple ferredoxin-like molecule would have been a useful electron transfer agent ( E c k and D a y h o f f , 1966; Hall et al., 1971). T he versatility o f t he iron-sulphur clusters is d e m o n s t r a t e d by t h e highly evolved ferredoxins and rubredoxins of t h e present day (Table 1). These proteins contain 1Fe, [2Fe-2S], [4Fe-4S] or t he r e c e n t l y discovered [3Fe-3S] clusters or various combinations o f these. Their m i dpoi nt r e d o x

1

Rubredoxins and ferredoxins (for Refs. see Cammack, 1979) Protein

1-centre Rubredoxin 2-centre Rubredoxin "Desulforedoxin" 2-Fe ferredoxin (Plant) 2-Fe ferredoxin (blue-green algae) 2-Fe ferredoxin 2-Fe ferredoxin 2-Fe ferredoxin Fe-S protein I Fe-S protein II HiPIP 4-Fe ferredoxin Ferredoxin II Ferredoxin

Typical source

Clostridium pasteurianum Pseudomonas oleovorans Desulphovibrio gigas

Mr )< 1 0

Fe-S -3

mV

6

Fe

-60

19

2Fe 2Fe

--35

7.6

Spinacia oleracea

2x 3.8 10.5

[2Fe-2S]

-420

Spirulina platensis

10.5

[ 2Fe-2S ]

--390

Halobacterium halobium Pseudomonas putida C. pasteurianum A z o to bacter vinelandii A. vinelandii Ch ro matiu m vinosu m Bacillus stearo thermophilus Rhodospiriilum rubrum Thermus thermophilus

15 12.5 25

[2Fe-2S] [2Fe-2S]

21

[ 2Fe-2S]

21 9.5

[2Fe-2S]

-345 -240 -300 -350 -225 + 350 -280 -430

8.5

14.5 9

[2Fe-2S]

[4Fe-4S]=+:3÷ [4Fe.4S]1 ÷:~÷ [4Fe-4S]l*;:* [4Fe-4S ] and

6

2 [4Fe-4S]1÷;2÷

-530 -250 +340 -420 +355 -380 -400

10 18 24

2 [4Fe-4S]l*; 2÷ 314Fe.4S]1÷; 2÷ 414Fe-4S]=*: 3÷

--490 --455 --130

[ 2Fe-2S ] or

Ferredoxin I (Fe-S protein IIi:) Ferredoxin IV

A. vinelandii

14.5

R. rubrum

14

8-Fe ferredoxin

C. pasteurianum Peptococcus aerogenes Chr. vinosum Desulfovibrio gigas

8-Fe ferredoxin Ferredoxin I Ferredoxin II

Em b

- D. g/gas

a Mr = relative molecular mass (mol. wt). b E m = midpoint reduction potential, relative to the standard hydrogen electrode.

[3Fe-3S] [4Fe-4S]~÷;3÷ [3Fe-3S] 7-8 Fe, S

64 1=I

~..~

~

~ o

~-.~

.~. -~ ~

i ,~~ ~ °

¢~

4~

~

~

,.~

.~

s,,.

0

o

9

I=I

0

.~ ~"~

o~

III

¢~

~

ltl

"0

_.="

~

~

~

~

I~ ..,=

;~:

~

~

~'

..

I=I 0

I= III

BI < 4-

o~ 4~ I=

+

0

£ 0

Z~..,

0

0~

~ = ~

65 potentials can vary between - 4 9 0 mV and +350 mV, so that they can operate in a wide range of electron transfer systems. The redox potentials are determined by the configuration of the protein. The influence of the proteiw on redox potential is well known in cytochromes and flavoproteins, where mechanisms such as variation in exposure of the haem or flavin to water, and the formation and breaking of internal hydrogen bonds have been suggested (SteUwagen, 1978; Adman, 1979). In the iron-sulphur proteins, the protein can also influence the type of iron-sulphur cluster that is formed, and in the case of the [4Fe-4S] clusters, the type of redox change that it can undergo. The ferredoxins are used as electron transfer agents in numerous systems of soluble and membrane-bound enzymes (Table 2). Rubredoxins, iron-sulphur proteins with single iron atoms bound to cysteine sulphur and no labile sulphide, occur in fermentative and photos:~nthetic bacteria but their function is at present unknown.

Ferredoxins with [4Fe-4S] clusters The small ferredoxins from anaerobic bacteria such as Clostridium and Desulfovibrio, contain [4Fe-4S] clusters which, as already noted, are similar in chemical properties to the stable [4Fe-4S] chemical analogues. The eight-iron ferredoxins contain two such clusters, with the polypeptide of about 6000 daltons. This forms a compact unit which can transfer two electrons either singly or together. Ol~her ferredoxins contain a single [4Fe-4S] cluster in a polypeptide of 6000 daltons (Desulfovibrio spp.) or 8500 daltons (Bacillus spp.). In the amino acid sequences of these proteins (Fig. 5) the four cysteines which attach to each cluster consist of three closely spaced and one more distant cysteine, as can be seen from the 3-dimensional structure of Peptococcus aerogenes ferredoxin (Fig. 6). It appears that there are

steric restrictions on placing four closelyspaced cysteines around one cluster (Van Rooten and De Coen, 1974}. Because it is of small size and contains mostly the simpler amino acids that are readily synthesized abiogenically, the eight-iron ferredoxin of the type found in C. butyricum seems likely to have been an early form of electron carrier (Eck and Dayhoff, 1966; Hall et al., 1971). The two halves of the ferredoxin molecule are closely homologous and this has been frequently cited as a likely example of gene duplication from a polypeptide of about 28 residues (see Matsubara et al., 1968). It is of interest that when C. acidi-urici ferredoxin apoprotein was enzymically cleaved into two sections of this size and the separate halves were reconstituted, the resulting molecules still appeared to be dimers with two clusters each (Orme-Johnson, 1972). This suggests that the postulated 28-residue ancestral protein was still a dimer. In this way the clusters are wrapped up in the minimum amount of protein in the most efficient way. Von Heijne et al. (1978) have suggested that such a 28-residue polypeptide could have also been a precursor of the rubredoxins and flavodoxins. All of these ferredoxins have relatively low midpoint potentials, near to that of the H*/H2 couple ( - 4 2 0 mV). The reduction process is a change from oxidation level +2 (corresponding to 2Fe3* + 2Fe 2÷) to oxidation level +1 (1Fe 3÷ + 3Fe:*). The latter state gives a characteristic electron paramagnetic resonance (EPR) spectrum around g = 1.96 (Fig. 7). The HiPIPs (high-potential iron,sulphur proteins) of photosynthetic bacteria such as Chromatium vinosum also contain one [4Fe-4S] cluster per molecule and are now recognised as a type of ferredoxin. However they have much higher midpoint redox potentials (+350 mV). They achieve this by operating between oxidation levels +2 and +3 (3Fe~÷ + 1Fe2÷). The +3 state can be recognized by an EPR spectrum with g-values greater than 2 (Fig. 7). If the protein

66 IIIII1~1

~o °~

~o~

~

~

1

e

~

N ~ ~.~

>><>><~<

~o~

>>>>~II>

I ~ ~ 1 >>>>> <<>>> ~

1 ~ ~ 1 ~111 I1~

lli[lll~ fill

~ ~~~

l~I> ~

I ~ I ~ ~ ~111111 IIIIIII IIIIIII

I

I~oo

I

I ~ >

I

IIII

°~ ~ ,.~c

Iltl ~

~

O

Z llIII

,~

o~ ~

"-~

I-.t g~N Z

IIIII1~1 IIIII1>1

~'~

~ . ~

ZZ m "N ' ~

67

Qosfridium pasfeurianum Rubredoxin

Chromafium vinosum HiPIP

O~ l:c O= S O= Cw-.~ Fig. 6. X-ray crystal structures of rubredoxin, HiPIP and eight-iron ferredoxin. Only the a-carbon positions of the amino acid residues and cysteinyl sulphurs are shown. Redrawn from Watenpaugh et al. (1979), Carter et al. (1974) and Adman et al. (1973).

structure is unfolded, the cluster can be reduced to the + 1 state, like the low potential ferredoxins (Cammack, 1973). Thus, by modifying the protein environment, the [4Fe-4S] clusters can be changed from a strong reducing agent to a strong oxidant. The arrangement of cysteine residues in the HiPIPs is quite different from that in the low-potential fe~redoxins (Fig. 6). The amino acid sequences .,how no homology and the patterns o f folding of the protein chains are quite different (Tedro et al., 1979). Thus the two types of protein appear to be quite unrelated, and presumably t h e y evolved the [ 4Fe-4S ] cluster separately. Recently a class of proteins has been discovered in bacteria, which is related to the low-potential ferredoxins but have high-

potential clusters. These are typified by ferredoxin I from Azotobacter vinelandii and ferredoxin from Pseudomonas ovalis. It is likely that ferredoxins I and IV from Rhodospirillum rubrum are similar (Yoch and Carithers, 1979). T h e y contain two clusters per molecule, one of high potential {+350 mV) and one of low potential ( - 4 0 0 mV). The function of these clusters with such disparate potentials is not clear. The former appears to be a [4Fe-4S] ~÷;3÷ cluster as in HiPIP. The latter appears, from spectroscopic and X-ray crystallographic analysis to be a novel type of [3Fe-3S] d u s t e r (Fig. 4) (Emptage et al., 1980; Stout et al., 1980). This type of cluster also gives an EPR signal at g = 2.01 in the oxidized state (Fig. 7b). The amino acid

68 g-value

22 I

20

21 l

I

I

I

19 I--

18

~

I

I

l

Ferret..._

~Desulfowbrio glgas Ferredoxin E

J

Bacillus stearothermophilus Ferredoxin

Chromafiurn

/vinosum

J 0 30

2 I__

I

032

~J

I

I 0 3L,

L

HiPIP

I 036

I

Fig. 7. EPR spectra of ferredoxins with [4Fe-4S] and [3Fe-3S] clusters. Spectra were recorded at about 15 K.

sequence of P. ovalis ferredoxin bears a close resemblance to the eight-iron ferredoxins (Fig. 5). Thus a similar a~rangement of cysteine residues in a protein can a c c o m m o d a t e a [4Fe-4S] 2÷;1÷, [4Fe-4S] 2÷;3÷ or [3Fe-3S] cluster. Presumably it is the other amino acid residues around the cluster that determine which t y p e will be favoured. Moura et al. (1978) have reported an even more striking instance of the effect of protein conformation on the t y p e and redox potential of clusters. The ferredoxin in D. g/gas has two forms. Ferredoxin I is a trimeric unit, with a [4Fe-4S] cluster in each subunit and a midpoint potential o f - 4 5 5 mV while ferredoxin II is a tetrameric unit, with a [3Fe-3S] cluster in each subunit (Huynh et al., 1980), and a midpoint potential of - 1 3 0 mV (Cammack et al., 1977). Both ferredoxins are formed from the same subunit, of which the amino acid sequence has been determined (Fig. 5).

-In this way, t w o forms of the same protein could function at two points in the same electron transfer chain {Moura et al., 1978). Apart from the HiPIPs, all the ferredoxins with [4Fe-4S] and [3Fe-3S] clusters show strong similarities in their amino acid sequences (Fig. 5) particularly in the arrangement of cysteine residues. Presumably they share a c o m m o n ancestor. The ferredoxins from the green and purple photosynthetic bacteria can be considered as being derived from those of fermentative bacteria such as Clostridium, b y insertion of one or t w o extra peptide sequences. Thus the ferredoxins ' f r o m the green photosynthetic bacteria may be considered as derived b y insertion of a " l o o p " between residues 41 and 47, and the ferredoxin from the purple sulphur bacterium C. vinosum b y the further addition of a C-terminal sequence. These additions have the effect of making the redox potential more negative, - 4 9 0 mV, than those in the fermentative bacteria. The photosynthetic bacterial ferredoxins also have an additional function, as electron donors to the 2-oxoacid reductases which are involved in the CO2 fixation cycle (Evans et al., 1966). The ferredoxins from the non-sulphur purple bacteria are o f higher molecular weight (Yoch et al., 1978) b u t none of their sequences are y e t known. The present data would suggest an evolutionary sequence: fermentative bacteria -+ green sulphur bacteria ~ purple sulphur bacteria -* purple non-sulphur bacteria.

F e ~ e d o x i n s with [2Fe-2S] clusters As already noted, the [2Fe-2S] clusters require a more rigid arrangement: of thiol ligands to prevent isomerisation into [4Fe-4S] clusters. Accordingly, the ferredoxins that contain these clusters have larger polypeptides than the simple eight- and four-iron ferredoxins, and generally only contain one cluster per molecule. T h e y give EPR spectra in the reduced state, around g = 1.94 (Fig. 8). [2Fe-2S] proteins are f o u n d in anaerobic

69

22 1

g-v~ltue 2.0

21 I

I

I

I

19

I

I

1.8 I

I

Pseudomonas Ferredoxin

/

~~~oH~obacteri halobium um

/~

II

030

i

I

Ferredoxin

I

I

,

0.32 0.3~, Hagnetic field/T

,

036

L

Fig. 8. EPR spectra of ferredoxins with [2Fe-2S] clusters. Spectra were recorded at about 30 K.

organisms, including Clostrid iu m pasterianu m (Cardenas et d., 1976) and photosynthetic bacteria (Evans et al., 1971) where their function is not known. However, they are much more abundant in the oxygen~evolving cyanobacteria and plants (Table 2), and are involved in other oxygen
(1969) detected the production of superoxide radicals (O~) from autoxidation of C. pasteurianum [4Fe-4S] ferredoxin but could not detect them on oxidation of a number of [2Fe-2S] ferredoxins. Autoxidation of spinach ferredoxin does produce O~, (Misra and Fridovich, 1971) but in smaller amounts. Allen {1975) proposed that reduced spinach ferredoxin can react rapidly with O~, producing the less toxic H202. The amino acid sequences of the [2Fe-2S] ferredoxins (Fig. 9) are dissimilar from those of the [4Fe-4S] type ferredoxins and their patterns of protein folding are different (Fig. 10,6). A possible relationship to the eight-iron ferredoxins has been suggested (Matsubara et al., 1980) but this is difficult to prove, since traces of homology might be the consequence of both types having to accommodate iron-sulphur clu stem. The sequences of [2Fe-2S] ferredoxins from anaerobes would be of great interest in the evolution of this class of proteins, but so far none have been determined. The sequences of ferredoxins from cyanobacteria and plants show a considerable degree of homology (Fig. 9) (Matsubara et al., 1980) but they are dissimilar from those from Ps. putida and mammalian adrenal glands (Yasunobu et al., 1973). This does not necessarily argue against a common ancestry. In the case of the cytochromes c2, it was found that although the amino acid sequences had diverged to such an extent that it was difficult to recognise any homology between the sequences, the three-dimensional folding patterns of the proteins were very similar (Dickerson, 1980b). The two-iron ferredoxins do not crystallize well and so far we only have the detailed structure of one protein, from the cyanobacterium Spirulina platensis (Fig. 10). So far the only evidence for structural homology within the two-iron ferredoxins remains their circular dichroism spectra. Although the [2Fe-2S] cluster is believed to have no inherent chirality, all the [2Fe-2S] ferredoxins show a characteristic C.D. spectrum in

70

z

i,.~ I"~ I.~ ~'.] ~

~..=

I'~ I-'] I'~ I-~ I'~

~

~

..=

<<<<<<<<<<

<<<<~<<<<< <<<><<<<<< =l'~

L

.

.

.

.

.

.

.

.

.

.

.

.

.

.

o .

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

,

. . . . . .

.-

71

\

Fig. 10. X-ray crystal structure of ferredoxin from Spirulina platensis. Redrawn from Fukuyama et al. (1980).

the visible and near ultraviolet regions, with a peak around 4:30 nm (Palmer et al., 1967) (Fig. 11). The [4Fe-4S] by contrast show weak and rather variable CD structure. The mason for this optical asymmetry is unknown, but presumably arises from a particular arrangement of the protein domains around the [2Fe-2S] cluster. The conservation of the CD spectra in different [2Fe-2S] proteins suggests that they all evolved from the same ancestral protein. In the oxygen~volving photosynthetic organisms, the principal function of the [2Fe-2S] ferredoxins is in photosynthetic electron transport from Photosystem I to NADP via a flavoprotein. However, they

can serve as electron donors to a number of other systems (Table 2), and can also regulate the activity of a number of enzymes of the Calvin-Benson cycle of CO2 fixation, through the thioredoxin system (Buchanan et al., 1979). In many organisms, two feFredoxins with homologous but significantly different amino acid sequences have been found. These may have different redox potentials and be adapted to different functions (Hutson et al., 1978). The only other bacterial ferredoxin which shows homology with those of cyanobacteria and plants, is that from Halobacterium halobium. As Fig. 9 shows, the latter protein has additional residues at the N and C termi-

72

.

3OO CircuLar Dichroism.

.

.

.

.

.

.

.

.

•

.

.

.

Wavelength, .

.

.

.

.

.

400 r x I~ I

.

.

.

.

.

.

nm .

.

.

5OO a

.

.

.

.

.

.

.

.

6OO

.

i

Haloba~erium

difficult to understand in terms of the concept of the archaebacteria as a totally different line of descent f r o m the prokaryotes (Magrum et at., 1978).

Iron.sulphur enzymes

I

y.-

~

Spinach

1-,, \

X,./ \ J \,,/

~x

/

~ - /

Pig Adrenal

\

/

~j !

.

.

.

.

.

.

.

.

.

|

.

.

.

.

.

.

.

.

.

I

.

.

.

.

.

.

.

•

.

.

.

.

.

.

.

.

•

Fig. 11. Circular dichroism spectra of [2Fe-2S] ferredoxins. From Caramack (1979).

nals, but the homology with the ferredoxin from the cyanobacteria is remarkable (Hase et al., 1978). Since halobacteria do not carry out any photosynthetic electron transport the presence of large amounts of this protein (Kerscher et al., 1976) implies that it has another important function. Halobacterial ferredoxins have been found to o p e r a t e in 2-oxoacid metabolism (Kerscher et al., 1977) and the reduction of nitrite (Werber et al., 1978). This seems to provide an example of gene transfer from an autotrophic photosynthetic bacterium to halobacteria. The homology with plant-type ferredoxins is

Apart from the simple ferredoxins, ironsulphur clusters are found in m a n y e n z y m e systems, which also contain other groups such as chlorophyll, haem, flavin or molybdenum. Many such conjugated iron-sulphur proteins are known (see Cammack, 1979; Yoch and Carithers, 1979). In most of these proteins the iron-sulphur cluster appears to act as an electron carrier from one group to another. Only in the case of hydrogenase is there evidence for a direct catalytic activity of an iron-sulphur cluster., In a few enzymes, notably aconitase (Ruzicka and Beinert, 1978) and glutamine amidotransferase (Averill et al., 1980), iron-sulphur clusters have been found in proteins with no electron transfer function. In these cases, a regulatory role has been suggested. Iron-sulphur clusters are also found in the membrane-bound electron transfer systems of" photosynthesis and respiration, which lead to the formation of a transmembrane potential and hence ATP. Here a system of the clusters may be required to transfer electrons from one side of the membrane to the other. As emphasized in Figs. 1, 2 and 3, the photosynthetic electron transfer chains of green and purple photosynthetic bacteria, cyanobacteria and plants, as well as the respiratory chain of mitochondria, contain a complex o f an iron,sulphur protein in association with cytochromes b and c,, which accepts electrons from reduced quinones (see Crofts and Wood, 1978). By analogy with mitochondria, the complex is supposed to be associated with the site of inhibition by antimycin A, and a site of coupling with ATP synthesis. The iron-sulphur protein was first isolated from mitochondria by Rieske ( 1976} and it has been detected by EPR spectroscopy

73 in photosynthetic bacterial chromatophores (Prince and Dutton, 1976) and chloroplasts (Malkin and Rearden, 1978). It appears to contain a [2Fe..2S] cluster but is noteworthy for its high redox potential and an EPR signal in the reduced state at g = 1.90. It is also unusual in that its redox potential is pHdependent, indicating that it is involved in hydrogen rather than electron transfer. Succinate dehydrogenase is a membranebound enzyme, with two subunits, an ironsulphur flavoprotein and an iron-sulphur protein. The enzymes of Rhodospirillum rubrum and mammalian mitochondria are remarkably similar (Davis et al., 1977). Indeed the subunits of the two systems can be cros~reacted to form active hybrid enzymes. This reinforces the concept of a similarity between the respiratory chains of purple nonsulphur bacteria and mitochondria. Cytochromes All of these membrane-bound systems contain various cytochromes b, and in each case there is evidence for a cytochrome analogous to mitochondrial cytochrome c,, called cytochrome css5 in photosynthetic bacteria and cytochrome f in chloroplasts and cyanobacteria (Wood, 1980). The electron acceptor of this complex is either a soluble cytochrome c or c:, or plastocyanin in chloroplasts and some cyanobacteria. To extend the analogy between these systems, certain cyanobacteria and algae can use either plastocyanin or a c2-type cytochrome css: depending on conditions of growth (Wood, 1978; Bohner et al., 1980). The soluble c-type cytochromes have been extensively studied. M a n y amino acid sequences and several X-ray crystal structures have been detemained (Dickerson, 1980b) and give one of the best examples of the use of "neutral" mutations to study the ancestry of proteins. When this study was extended further to the cytochromes c2 and other small, soluble

cytochromes of prokaryotes, the sequences were so dissimilar that little homology was discernible, apart from the sequence Cys-X-XCys-His in the haem-binding domain near the N terminus. However a comparison of the protein structures showed a consistent pattern of polypeptide folding (Fig. 12). The large apparent differences could be explained as additional loops or deletions near to the surface of the molecule, away from the haem ring, as with the ferredoxins. When the sequences of the appropriate complementary regions were compared, there was evident homology between them. The differences in structure may partly reflect a change in specificity from one enzyme system to another. Differences in reactivity with nitrite reductase and cytochrome oxidase have been observed (Kamen et al., 1977). Because of their similarity, Dickerson (1980b) has suggested that all the c-type cytochromes of this class, with their diverse and confusing nomenclature, should be called cytochromes c2, particularly since they follow cytochrome c~ in the electron transfer chain.

Copperproteins Electron transfer proteins containing copper are of relatively positive reduction potential. They are absent from anaerobic bacteria such as photosynthetic bacteria, possibly because of the extreme insolubility of copper sulphide (Osterberg, 1974) and presumably they evolved after the atmosphere became oxidizing. As already noted, the copper protein plastocyanin takes the place of cytochrome c5s3 in the oxygen-evolving photosynthetic electron transfer chains of plants and certain cyanobacteria. Plastocyanin belongs to the Type I or "Blue ~ copper proteins (Fee, 1975; MalmstrSm et al., 1975), so called because of their intense colour in the oxidized state. In these redox proteins, the binding site is designed to provide a coordination geometry which is favourable for both valence states of copper,

74 (hlorobium fhl3su[fafophilum

Rhodospirii[um rubrum

~una

Fig. 12. X-ray crystal structures of cytochromes of the c 2 type. Redrawn from Dickerson (1980b).

since Cu(II) favours a square planar configuration while Cu(I) prefers tetrahedral. In plastocyanin the copper is bound to two histidine nitrogens, a cysteine sulphur and a methionine sulphur in a highly distorted tetrahedral site. As a result the Cu(II) form is destabilized relative to the tetragonal Cu 2÷ (aquo) ion by 200 mV, giving the protein a more positive redox potential. The amino acid sequences of a number of plastocyanins have been compared (Boulter

et al., 1977) and they show a higher rate of mutation than the cytochromes c, comparable with that of the [2Fe-2S] ferredoxins. This information can be combined with other sequence data to derive a "natural" classification of the higher plants, and to lzace the evolution of chloroplasts from cyanobacteria (Schwartz and Dayhoff, 1978). Plastocyanin is related to the blue copper protein azurin of non-photosynthetic bacteria such as Pseudornonas and Alcaligenes.

75

Fig. 13. X-ray crystal structures of Populus nigra (poplar) plastocyanin and Pseudomonas aeruginosa azurin. Redrawn from Colman et al. (1978) and Adman et aL (1978).

Although their mnino acid sequences show little homology, statistical analysis indicates that they are related (Ryden and Lundgren, 1979). Most significantly, the ligands to copper in azurin (His 46, Cys 112, His 117 and Met 121) are similarly arranged to those in plastocyanin (His 37, Cys 84, His 87 and Met 92). Both proteins have a similar ~-barrel structure (Fig. 13) (Adman, 1979).

Superoxide dismutases

Superoxide dismutase is a metalloenzyme that catalyzes the conversion of the superoxide [O~] free radical to molecular oxygen and hydrogen peroxide (McCord and Fridovich, 1968L O~ + O~ + 2H ÷ :,~p~o~io~ ' O5 + H 2 0 2 d ism u t ase

The electronic structure of the oxygen molecule is such that its reduction favours the univalent pathway: O5 -* 02 -~ H202 -~ OH-; superoxide is thus the initial product of oxygen reduction. A number of biological reactions such as the autoxidation of reduced ferredoxins, flavins, haem proteins, etc. generate O~. There is strong evidence that O~, orreactive species derived from it, are toxic, and the superoxide dismutase provides a defence mechanism to the cell by scavenging the superoxide formed (Fridovich, 1976). The spontaneous dismutaion of O~ is a rapid process, but superoxide dismutase accelerates it so that the 02 level in the cell is extremely low. Superoxide dismutas'e activity has been detected in all aerobic organisms examined {with one or two exceptions}. The enzyme also occurs in many anaerobes, which may be exposed to O5 occasionally. Three types of superoxide dismutases have

76 been isolated and characterized (see Rao and Cammack, 1981). All have the same, high catalytic rate, but they contain different metals. The enzyme found in the cytosol of eukaryotic cells is Cu/Zn superoxide dismutase. It has a tool. wt of 32 000 and has two identical subunits, each subunit containing one Cu 2+ and one Zn TM. The Cu/Zn enzyme is very seldom found in prokaryotes; its presence has so far been reported only in two bacteria, Photobacterium leiognathi and Paracoccus denitrificans (see Vignais et al., 1980). Superoxide dismutases occurring in prokaryotes contain iron or manganese as the metal ligand. Fe 3÷ superoxide dismutase has been detected in strict anaerobes such as the Clostridia, sulphate reducers and vhotosynthetic bacteria, and in cyanobacteria, eukaryotic algae and protozoa. Superoxide dismutases with M n 3+ as the metal ligand occur in aerobic bacteria, cyanobacteria, algae, protozoa, fungi and in mitochondria (where it is located in the matrix space). The M n enzyme has tool. wts of 40 000 or 80 000 with 1 or 2 atoms of M n per molecule, respectively. The three types of superoxide dismutases can be distinguished by selective inhibition or inactivation. Thus cyanide causes inhibition of Cu/Zn superoxide dismutase activity but does not affect the Fe and M n enzymes. Hydrogen peroxide inactivates both the Cu/Zn and Fe superoxide dismutase but not the Mn-enzyme.

Structural comparisons of superoxide dismutases X-ray diffraction analysis of the Cu/Zn superoxide dismutase from bovine erythrocytes (Richardson et ai., 1977) indicates that the Cu and Zn are linked to the e n z y m e by a c o m m o n ligand, the imidazole ring of His6~. The active site is designed to expose the copper to the Oi and allow the necessary electron and hydrogen transfers, in contrast to other copper proteins in which the metal is buried in the protein, and w h i c h have very low superoxide dismutase activity. The corn-

plete amino acid sequences of Cu/Zn enzymes from bovine erythrocytes (Steinmann et al., 1974) and yeast (Johansen et ah, 1979)have been determined. The two proteins are highly homologous with 55% identity in the sequences--the positions of all the amino acids liganded to the metal are identical in the t w o sequences. The complete amino acid sequences of M n superoxide dismutases from E. coli and Bacillus stearothermophilus are k n o w n (Brock and Walker, 1980); 6(~o of the amino acid residues are invariant in the two sequences. Harris et al. (1980) have made a comparison of the amino terminal sequences of the M n and Fe enzymes from a variety of prokaryotic and eukaryotic species. They found no clear distinction in the sequences of the Fe and M n enzymes. The sequence data suggest a c o m m o n origin for both types of superoxide dismutases. The similarity in the sequences of M n superoxide dismutases from prokaryotes and mitochondria lends support to the endosymbiotic origin of mitochondria. Jackson et al. (1978) detected a cyanideinsensitive dismutase in chloroplasts which they suggested was the M n enzyme. Recently, however, Satin and Bridges (1980) have reported the isolation of the Fe enzyme from the higher plant Brassica campestris. The sequence study also showed no evidence for any homology between the Cu/ Zn superoxide dismutase and the Fe (or Mn) enzyme. The Cu/Zn enzyme found mainly in plants and mammals would have had an independent origin. From a study of the distribution of S O D in various photosynthetic organisms, Asada et al. (1977) propose that the Cu/Zn enzyme developed at first in ferns and mosses. However, Henry and Hall (1977) proposed that certain chlorophycean algae (eukaryote, greens) are related to higher plants since they do possess a Cu/Zn SOD. The superoxide dismutases may have originated to protect against an oxygen atmosphere. Alternatively, they m a y have existed before the oxidizing atmosphere, to remove Oi formed by ultraviolet and ionising

77 radiation (Towe, 1978; Lumsden and Hall, 1975). Since the Fe superoxide dismutases are found in ferrnentative and photosynthetic bacteria, they may have appeared first and transferred to the cyanobacteria and possibly chloroplasts. The Mn superoxide dismutase is found mostly in aerobes and was incorporated into the mitochondria. The Cu/Zn enzyme of the eukaryotic cytoplasm appears to be a different line of development.

References Adams, M.W.W., S.G. Reeves, D.O. Hall, G. Christou, B. Ridge and H.N. Rydon, 1977, Biological activity of synthetic tetranuclear iron-sulphur analogues of the active sites of ferredoxins. Biochem. Biophys. Res. Commun. 79, 1184. Adams, M.W.W., K.K. Rao, D.O. Hall, G. Christou and C.D. Garner, 1980, Biological activity of synthetic molybdenum-iron-sulphur, iron-sulphur and iron-selenium analogues of ferredoxin-type centres. Biochim. Biophys. Acta. 589, 1. Adman, E.T., 1979, A comparison of the structures of electron transfer proteins. Biochim Biophys Acta 549,107. Adman, E.T,, L.C. Sieker and L.H. Jensen, 1973, The structure of a bacterial ferredoxin. J. Biol. Chem. 248, 3987. Adman, E.T., R.E. Stenkamp, L.C. Sieker and L.H. Jensen, 1978, A cyrstallographic model for azurin at 3~k resolution. J. Mol. Biol. 123, 23. Allen, J.F., 1975, A two-step mechanism for the photosynthetic ]'eduction of oxygen by ferredoxin. Biochem. Biophys. Res. Commun. 66, 36. Ambler, R.P., M. Daniel, J. Herrnoso, T.E. Meyer, R. Bartsch and M.D. Kamen, 1979, Cytochrome c~ sequence variation among the recognised species of purple norsulphur-photosynthetic bacteria. Nature 278, 659. Asada, K., S. Kanematsuand K. Uchida, 1977, Superoxide dismutases in photosynthetic organisms: Absence of the cuprozinc enzyme in eukaryotic algae. Arch. Biochem. Biophys. 179,243. Averill, B.A., A. Dwivedi, P. Debrunner, S.J. Vollmer, J.Y. Wong and R.L. Switzer, 1980, Evidence for a tetranuclear iron sulphur centre in glutamine phosphoribosylp yrophosphate amidotransferase from Bacillus subtilis. J. Biol. Chem. 255, 6007. Averill, B.A. and W.H• Orme-Johnson, 1978, Ironsulphur proteins and their synthetic analogs, in: Metal ions in biological systems, Vol. 7, H. Sigel, (ed.) (Marcel Dekker, New York) pp. 127.

Berkner, L.V. and L.C. Marshall, 1965, History of major atmospheric components. Proc. Natl. Acad. Sci. USA 53, 1215. Bohner, H., H. Merkle, P. Kroneck and P. Biiger, 1980, High variability of the electron carrier plastocyanin in mic~'oalgae. Eur. J. Biochem. 105, 603• Boulter, D., B.G. Haslett, D. Peacock, J.A.M. Ramshaw and M.D. Scawen, 1977, Chemistry, function and evolution of plastocyanin, in: International review of biochemistry, Vol 13, D.H. Northcote (ed.) (University Park Press, Baltimore) pp. 1. Brock, C.J. and J.E. Walker, 1980, Superoxide dismutase from Bacillus stearothermoph~us. Complete amino acid sequence of a manganese enzyme. Biochemistry 19, 2873. Broda, E. 1975, The evolution of the bioenergetic processes (Pergamon Press, Oxford). Bruschi, M. 1979, Amino acid sequence of Desulfovibrio gigas ferredoxin: revisions. Biochem. Biophys. Res. Commun. 91,623. Buchanan, B.B., R.A. Wolosiuk and P. Schurmann, 1979, Thioredoxin and enzyme regulation. Trends Biochem. Sci. 4, 93. Cammack, R. 1973, "Super-reduction" of Chromatium high-potential iron-sulphur proteins in the presence of dimethyl sulphoxide. Biochem. Biophys. Res. Commun. 54,548. Cammack, R. 1979, Functional aspects of ironsulphur proteins, in: Metalloproteins: Structure, Function and Clinical Aspects, U. Weser (ed.) (Thieme Verlag, Stuttgart) pp. 162. Cammack, R. 1980, New developments in ironsulphur proteins. Nature 286, 442. Carnmack, R., K.K. Rao, D.O. Hall, J.G. Moura, A.V. Xavier, M. Bruschi, J. Le Gall, A. Deville and J.P. Gayda, 1977b, Spectroscopic studies of the oxidation-reduction properties of three forms of ferredoxin from Desulfovibrio gigas. Biochim. Biophys. Acta 490,311. Cardenas, J., L.E. Mortenson and D.C. York, 1976, • Purification and properties of paramagnetic protein from Clostridium pasteurianum W5. Biochim. Biophys. Acta 434,244. Carter, C.W., J. Kraut, S.T. Freer, N.-G. Xuong, R.A. Alden and R.G. Bartsch, 1974, Two-angstrom crystal structure of oxidized Chromatium high potential iron protein. J. Biol. Chem. 249, 4212. Castenholz, R.W., 1979, Evolution and ecology of thermophilic microorganisms, in: Strategies of microbial life in extreme environments. Dahlen Conference, M. Shilo (ed.), (Dahlen Conference, Berlin) pp. 373--392. Clayton, R.K. and W.R. Sistrom, (ed.), 1978, The Photosynthetic bacteria (Plenum, New York and London).

78 Crofts, A.R. and P.M. Wood, 1978, Photosynthetic electron transport chains of plants and bacteria and their role as proton pumps, in: Current topics of bioenergetics, Vol. 7, D. Rao Sanadi and L.P. Vernon (eds.) (The Academic Press, London) pp. 175--244. Colman, P.M., H.C. Freeman, J.M. Guss, M. Murata, V.A. Norris, J.A.M. Ramshaw and M.P. Vankappa, 1978, X-ray crystal structure analysis of Plastocyanin at 2.7 ~ resolution. Nature 272, 319. Davis, K.A., Y. Hatefi, I.P. Crawford and H. Baltscheffsky, 1977, Purification, molecular properties a ~ l amino acid composition of the subunits of Rhodospirillum rubrum succinate dehydrogenase. Arch Biochem. Biophys. 1 8 0 , 4 5 9 . Dickerson, R.E., 1980a, Evolution and gene transfer in purple photosynthetic bacteria. Nature 283, 210. Dickerson, R.E. 1980b, Cytochrome c and the evolution of energy metabolism. Sci. Am. 242 (3), 136. Eck, R.V. and M.O. Dayhoff, 1966, Evolution of the structure of ferredoxin based on living relics of primitive amino acid sequences. Science 152, 363. Emptage, M.H., T.A. Kent, B.H. Huynh, J. Rawlings, W.H. Orme-Johnson and E. Munck, 1980, On the nature of the iron-sulphur centres in a ferredoxin from Azotobacter vinelandii. J. Biol. Chem. 255, 1793. Estabrook, R.W., K. Suzuki, J.I. Mason, J. Baron, W.E. Taylor, E.R. Simpson, J. Purvis and J. McCarthy, 1973, Adrenodoxin: An iron-sulphur protein of adrenal cortex mitochondria, in: Ironsulphur proteins, Vol. 1. W. Lovenberg (ed.) (Academic Press, New York) pp. 193--224. Evans, M.C.W., B.B. Buchanan and D.I. Arnon, 1966, A new ferredoxin-dependent carbon reduction cycle in a photo-synthetic bacterium. Proc. Natl. Acad. Sci. USA 5 5 , 9 2 8 . Evans, M.C.W., R.V. Smith, A. Telfer and R. Cammack, 1971, Iron-sulphur proteins in photosynthetic microorganisms, in: Proc. 1st Eur. Biophys. Cong. Baden, E. Broda, A. Locker and H. Springer Lederer (eds.) (Vienna Medical Academy) pp. 115--119. Fee, J.A., 1975, Copper proteins - - systems containing the "Blue copper" centre. Struct. Bonding 23, 1. Fridovich, I., 1976, Oxygen radicals, hydrogen peroxide and oxygen toxicity, in: Free radicals in biology, W.A. Pryor (ed.) (Academic Press, New York} pp. 239--277. Fukuyama, K., T. Hase, S. Matsumotos, T. Tsikihara, Y. Katsube, N. Tanaka, M. Kakudo, K. Wada and H. Matsubara, 1980, Structure of S. platensis [2Fe-2S] ferredoxin and evolution of chloroplastt y p e ferredoxins. Nature 2 8 6 , 5 2 2 . Gunter, M., B. Ridge, H.N. R y d o n and R. Sharpe, 1979, Toward a synthesis of Clostridiurn butyri-

cure apoferredoxin: two tetradecapeptides com-

prising half the sequence (Residues 7--20 and 21--34) Bioinorg. Chem. 8 , 3 7 1 . Hall, D.O., R. Cammack and K.K. Rao, 1971, Role for ferredoxins in the origin of life and biological evolution. Nature 233, 136. Harris, J.I., A.D. Auffret, F.D. Northrop and J.E. Walker, 1980, Structural comparisons of superoxide dismutases. Eur. J. Biochem. 106, 297. Hase, T., H. Matsubara and M.C.W. Evans, 1977, Amino acid sequence of Chromatium vinosum ferredoxin: Rev. J. Biochem. 81, 1745. Hase, T., S. Wakabayashi, H. Matsubara, L. Kerscher, D. Oesterhelt, K.K. Rao and D.O. Hall, 1978, Complete amino acid sequence o f Halobacterium halobium ferredoxin containing N-acetyl lysine residue: J. Biochem. 83, 1659. Henderson-Sellers, A., A. Benlow and A.J. Meadows, 1980, The early atmospheres of the terrestrial planets. Q. J. R. Astro. Soc. 21, 74. Henry, L.E.A. and D.O. Hall, 1977, Superoxide dismutase in green algae: An evolutionary survey. Plant and Cell Physiology, Special issue on Photosynthetic Organelles, pp. 377--382. Hill, C.L., J. Renaud, R.H. Holm and L.E. Mortenson, 1977, Synthetic analogues of the active sites of Iron-sulphur proteins. 15. Comparative polarographic potentials of the [Fe4S4(SR),] 2--a Clostridium pasteurianum ferredoxin redox couples. J. Am. Chem. Soc. 99, 2549. Holm, R.H., 1977, Synthetic approaches to the active sites of iron-sulphur-proteins. Acc. Chem. Res. 10, 427. Hutson, G.K., L.J. Rogers, B.G. Haslett, D. Boulter and R. Cammack, 1978, Comparative studies on two ferredoxins from the Cyanobacterium Nostoc Strain MAC. Biochem. J. 172,465. Huynh, B.H., J.J.G. Moura, I. Moura, T.A. Kent, J. LeGall, A.V. Xavier and E. Munck, 1980, Evidence for a three-iron centre in a ferredoxin from Desulfovibrio gigas. J. Biol. Chem. 255, 3242. Jackson, C.', J. Dench, A.L. Moore, B. Halliwell, C.J. F o y e r and D.O. Hall, 1978, Subcellular localisation and identification of superoxide dismutase in the leaves o f higher plants. Eur. J. Biochem. 91, 399. Johansen, J.T., C. Overballe-Petersen, B. Martin, V. Hasemann and I. Svendsen, 1979, The complete amino acid sequence of copper zinc superoxide dismutase from Saccharomyces cerevisiae. Carlsberg. Res. Commun. 4 4 , 2 0 1 . Kamen, M.D., B.J. Errede and T.E. Meyer, 1977, Comparative studies of cytochromes c, in: Evolution of protein molecules, H. Matsubara and T. Yamanaka (eds.) (Japan Scientific Societies Press. T o k y o ) pp. 373--385.

79 Kerscher, L. and D. Oesterhelt, 1977, Ferredoxin is the coenzyme of -keto oxidoreductases in Halobacterium halobium. FEBS Lett. 83, 197. Kerscher, L., D. Oesterhelt, R. Cammack and D.O. Hall 1976, A new plant-type ferredoxin from halobacteria. Eur. J. Biochem. 71,101. Lovenberg, W., (ed.) 1973, 1977, Iron-sulphur Proteins, Vol. 1, 2, 1973, Vol. 3, 1977 (Academic Press, New York). Lumsden, J. and D.O. Hall, 1975, Superoxide dismutase in photosynthetic organisms provides an evolutionary hypothesis. Nature 257,670. Magrum, L.J., K.R. Luehrsen and C.R. Woese, 1978, Are extreme halophiles actually "Bacteria"? J. Mol. Evol. 11, 1.. Malkin, R. and A.J. Bearden, 1978, Membrane-bound iron-sulphur in photosynthetic systems. Biochem. Biophys. Acta ~;05, 147. MalmstrSm, B.G.. L.-E. Andreasson and B. Reinhammer, 1975, Copper containing oxidases and superoxide dismutase, in: The enzymes, Vol. 12, part B, P.D. Boyer (ed.) (Academic Press, New York) pp. 507--579. Matsubara, H., T. Hase, S. Wakahayashi and K. Wada, 1980, Structure and evolution of chloroplast- and bacterial-type of ferredoxins, in: Evolution of protein structure and function, Vol. 21, D.S. Sigman and M.A.B. Brazier (eds.), U.C.L.A. Forum in Medical Science (Academic Press, New York) p. 245. Matsuhara, H., T.H. Jukes and C.R. Cantor, 1968, Structural and! evolutionary relationships of ferredoxins, in: Structure, function and. evolution of proteins, l~rookhaven symposium Biology No. 21, pp. 201-216. McCord, J.M. and I. Fridovich, 1968, The reduction of Cytochrome c by milk xanthine oxidase. J. Biol. Chem. 243, 5453. Misra, H.P. and I. Fridovich, 1971, The generation of superoxide radical during the autoxidation of ferredoxins. J. E;iol. Chem. 246, 6886. Moura, J.J.G., A.V. Xavier, E.C. Hatchikian and J. Le Gall, 1978, Structural control of the redox potentials and of the physiological activity by oligomerization of ferredoxin. FEBS Lett. 89, 177. Olson, J.M., 1978, Pre cambrian evolution of photosynthetic and respiratory organisms, in: Evolutionary biology, Vol. 11, M.K. Hecht, W.C. Steere and B. Wallace (eds.) (Plenum Publishing Corporation, New York) pp. ~L--37. Orme Johnson, W.H., 1972, Tryptic cleavage of Clostridium acidi urici apoferredoxin and reconstitution of the separated fragments. Biochem. Soc. Trans. 1, 30. Orme~Johnson, W.H. and H. Beinert, 1979, On the formation of the superoxide anion radial during

the reaction of reduced iron-sulphur proteins with oxygen. Biochem. Biophys. Res. Commun. 36, 905. Osterberg, R., 1974, Origins of metal ions in biology. Nature 249,382. Padan, E., 1979, Facultative anoxygenic photosynthesis in cyanobacteria. Annu. Rev. Plant Physiol. 30, 27. Palmer, G., H. Britzinger and R.W. Estabrook, 1967, Spectroscopic studies on spinach ferredoxin and adrenodoxin. Biochemistry 6, 1658. Prince, R.C. and P.L. Dutton, 1976, Further studies on the Rieske iron-sulphur centre in mitochondrial and photosynthetic systems: A pK on the oxidised form. FEBS Lett. 65, 117. Que, L., J.R. Anglin, M.A. Bobrick, A. Davison and R.H. Holm, 1974, Synthetic Analogs of the active sites of iron-sulphur proteins IX. Formation and some electronic reactivity properties of Fe4S4 Gly cyl-cysteinyl-glycyl oligopeptide complexes obtained by ligand substitution reactions. J. Am. Chem. Soc. 96, 6042. Rao, K.K. and R. Cammack, 1981, Ferredoxin and superoxide dismutase on the evolution of microorganisms, in: Molecular and cellular aspects of microbial evolution. Society of General Microbiology Symposium, Edinburgh (Cambridge University Press). Rao, K.K. and D.O. Hall, 1977, Chemistry and evolution of ferredoxins and hydrogenases, in: The evolution of metalloenzymes, metalloproteins and related materials, G.J. Leigh (ed.) (London Symposium Press) pp. 39--64. Richardson, J.S., K.A. Thomas, B.H. Rubin and D.C. Richardson, 1975, Crystal structure of bovine Cu, Zn superoxide dismutase at 3 ~ resolution: chain tracing and metal ligands. Proc. Natl. Acad. Sci. USA 72, 1349. Rieske, J.S., 1976, Composition structure and function of Complex III of the respiratory chain. Biochim. Biophys. Acta. 456,195. Ruzicka, F.J. and H. Beinert, 1978, The soluble high potential type iron-sulphur protein from mitochondria is aconitase. J. Biol. Chem. 253, 2514. Ryden, L. and J.-O. Lundgren, 1979, On the evolution of blue proteins. Biochimie 61,781. Salin, M.L. and S.M. Bridges, 1980, Isolation and characterization of an iron-containing superoxide dismutase from a eucaryote, Brassicia campestris Arch. Biochem. Biophys. 201, 369. Schwartz, M. and M.O. Dayhoff, 1978, Origins of prokaryotes, eukaryotes, mitochondria and chloroplasts. Science 199,395. Steinman, H.M., V.R. Naik, J.L. Abernethy and R.L. Hill, 1974, Bovine erythrocyte superoxide dismutase. J. Biol. Chem. 249, 7326. Stellwagen, E. 1978, Haem exposure as the determi-

80 nant of oxidation reduction in haem proteins. Nature 275, 73. Stout, C.D., D. Ghosh, V. Pattabhi and A.H. Robbins, 1980, Iron-sulphur clusters in Azotobacter ferredoxin at 2.5 ~ resolution. J. Biol. Chem. 255, 1797. Sweeney, V. and J.C. Rabinowitz, 1980, Proteins containing 4Fe-4S clusters: An overview. Annu. Rev. Biochem. 49, 1139. Tedro, S.M., T.E. Meyer and M.D. Kamen, 1979, Primary structure of a high potential, four-iron sulphur ferredoxin from the photosynthetic bacterium Rhodospirillum tenue. J. Biol. Chem. 254, 1495. Towe, K.M., 1978, Early Precambrian oxygen: a case against photosynthesis. Nature 274, 657. Van Rooten, M. and J.L. De Coen, 1974, Etude conformationelle d'une ferredoxine bacterielle. Arch. Int, Physiol. Biochem. 82,793. Vignais, P.M., M.F. Henry, A. Terech and J. Chabert, 1980, Production of superoxide anion and superoxide dismutase in Paracoccus denitrificans, in: Chemical and biochemical aspects of superoxide and superoxide dismutases, J.V. Bannister and H.A.O. Hill (eds.) (Elsevier/North-Holland, Amsterdam) pp. 154--159. Von Heijne, G., C. Blomberg and H. Baltscheffsky,

1978, Early evolution of cellular electron transport: Molecular models for the ferredoxin rubredoxin-flavodoxin region. Origins Life 9, 27. Walter, M.R., R. Buick and J.S.R. Dunlop, 1980, Stromatolites 3400--3500 17 yr old from the North Pole area, Western Australia. Nature 284, 443. Watenpaugh, K.D., L.C. Sieker and L.H. Jensen, 1969, The Structure of rubredoxin at 1.2 A resolution K. Mol. Biol. 131,509. Werber, M.M. and M. Mevarech, 1978, Induction of a dissimilatory reduction pathway of nitrate in Halobacterium of the Dead Sea. Arch. Biochem. Biophys. 186, 60. Wood, P.M., 1978,~Interchangeable copper and iron proteins in algal photosynthesis. Studies on plastocyanin and cytochrome c-552 in Chlamydomonas Eur. J. Biochem. 87, 9. Wood, P.M., 1980, Do photosynthetic bacteria contain cytochrome c? Bioehem, J. 189, 385. Yasunobu, K.T. and M. Tanaka, 1973, The types, distribution in nature structure-function and evolutionary data of the iron-sulphur proteins, in: Iron-sulfur proteins, Vol. 2 , W. Lovenberg (ed.) (Academic Press, New York) pp. 27--130. Yoch, D.C. and R.P. Carithers, 1979, Bacterial ironsulphur proteins. M icrobiol. Rev. 43, 384.