Organic Cofactors Participated More Frequently Than Transition Metals in Redox Reactions of Primitive Proteins Hong-Fang Ji, Lei Chen, and Hong-Yu Zhang*

http://www.paper.edu.cn Hypothesis

Organic cofactors participated more frequently than transition metals in redox reactions of primitive proteins Hong-Fang Ji, Lei Chen, and Hong-Yu Zhang*

Summary cobalt). Since transition metals were highly available in Protein redox reactions are one of the most basic and primordial world and NAD, NADP, FMN, FAD have been important biochemical actions. As amino acids are weak (1) redox mediators, most protein redox functions are under- considered vestiges of a previous RNA world, both kinds of taken by protein cofactors, which include organic ligands redox cofactors were likely to be accessible in the pre-protein and transition metal ions. Since both kinds of redox world (indeed, some of the organic redox cofactors can be cofactors were available in the pre-protein RNA world, it is synthesized by non-protein enzymatic catalysis(2,3)). Thus, challenging to explore which one was more involved in there is the intriguing question—which kinds of cofactors were redox processes of primitive proteins? In this paper, using an examination of the redox cofactor usage of more involved in redox processes of primitive proteins? putative ancient proteins, we infer that organic ligands Ten years ago, Daniel and Danson argued that, since the participated more frequently than transition metals in origin and early evolution of life occurred at high temperatures redox reactions of primitive proteins, at least as protein (>958C) and the organic redox cofactors are unstable at these cofactors. This is further supported by the relative temperatures, in the most primitive organisms, transition abundance of amino acids in the primordial world. metals (e.g. iron) were more likely than organic coenzymes Supplementary material for this article can be found on (4) the BioEssays website (http://www.mrw.interscience. (e.g. NAD/P(H)) to carry out the redox functions of proteins. wiley.com/suppmat/0265-9247/suppmat/index.html). However, to our knowledge whether the common ancestor BioEssays 30:766–771, 2008. ß 2008 Wiley Periodicals, of extant life liked it hot or not is still in debate.(5–7) More Inc. importantly, through analyzing the cofactor usage of redox proteins (with Enzyme Classification of E.C. 1._._._) derived Introduction from a hyperthermophile Pyrococcus furiosus (which is Protein redox reactions, consisting of a matched set (oxidation located at the root of the evolutionary tree, with an Optimal and reduction), have been recognized as one of the most basic Growth Temperature of 1008C), we found that most (82.4%) of and important biochemical actions in all organisms, because the redox cofactors are organic compounds (Table 1). Taken they are associated with energy production, material meta- together, we think that Daniel and Danson’s inference is not bolism and other key biological processes. As amino acids are conclusive and which kind of cofactors participated more weak redox mediators, most protein redox functions are frequently in redox reactions of primitive proteins deserves undertaken by protein cofactors, which include organic ligands further investigation. As this issue deals with the very early (e.g. nicotinamide-adenine-dinucleotide (NAD), nicotinamide- stage of life, there is no supporting fossil evidence. However, adenine-dinucleotide phosphate (NADP), flavin mononucleo- thanks to the rapid progress in genomics and bioinformatics, tide (FMN) and flavin-adenine dinucleotide (FAD)) and some putative ancient protein sets have been established from transition metals (e.g. copper, iron, manganese, nickel and which we can obtain some meaningful clues to this intriguing question. Shandong Provincial Research Center for Bioinformatic Engineering and Technique, Center for Advanced Study, Shandong University of Redox cofactor usage in putative Technology, Zibo 255049, P. R. China. Funding agencies: This study was supported by the National Basic ancient proteins Research Program of China (grant 2003CB114400) and the National Through a large-scale phylogenomic analysis on 174 pro- Natural Science Foundation of China (grant 30570383 and 30700113). teomes, Caetano-Anolle´s and co-workers established a *Correspondence to: Hong-Yu Zhang, Center for Advanced Study, chronology for proteins.(8) The most ancient proteins (with Shandong University of Technology, Zibo 255049, P. R. China. Ancestry Value of 0 0.05) consist of 163 members, from E-mail: [email protected] DOI 10.1002/bies.20788 which 39 redox proteins (with Enzyme Classification of E.C. Published online in Wiley InterScience (www.interscience.wiley.com). 1._._._) can be identified (Table 2). According to the catalytic reactions presented in the comments of SWISS-PROT/

766 BioEssays 30.8 BioEssays 30:766–771, ß 2008 Wiley Periodicals, Inc.

转载中国科技论文在线 http://www.paper.edu.cn Hypothesis

Table 1. Redox proteins contained in hyperthermophile Pyrococcus furiosus

Redox proteins (EC number)a Entry namea Cofactorb

Shikimate dehydrogenase (1.1.1.25) AROE_PYRFU NADP,NADPH Ketol-acid reductoisomerase (1.1.1.86) ILVC_PYRFU NADP,NADPH Histidinol dehydrogenase (1.1.1.23) HISX_PYRFU NAD, NADH Glyoxylate reductase (1.1.1.26) GYAR_PYRFU NAD, NADH Probable L-threonine 3-dehydrogenase (1.1.1.103) TDH_PYRFU NAD, NADH Inosine-50-monophosphate dehydrogenase (1.1.1.205) IMDH_PYRFU NAD, NADH Coenzyme A disulﬁde reductase (1.8.1.14) CDR_PYRFU NAD(P), NAD(P)H Glycerol-1-phosphate dehydrogenase [NAD(P)] (1.1.1.261) EGSA_PYRFU NAD(P), NAD(P)H Glyceraldehyde-3-phosphate dehydrogenase (1.2.1.59) G3P_PYRFU NAD(P), NAD(P)H Glutamate dehydrogenase (1.4.1.3) DHE3_PYRFU NAD(P), NAD(P)H Pyruvate/ketoisovalerate oxidoreductases common subunit gamma (1.2.7.7) PORC_PYRFU CoA, acetyl-CoA L-aspartate oxidase (1.4.3.16) NADB_PYRFU FADc Dihydroorotate dehydrogenase (1.3.3.1) PYRD_PYRFU FMNc Probable glycine dehydrogenase [decarboxylating] subunit 2 (1.4.4.2) GCSPB_PYRFU Pyridoxal phosphatec Superoxide reductase (1.15.1.2) SOR_PYRFU Ironc Tungsten-containing aldehyde ferredoxin oxidoreductase (1.2.7.5) AOR_PYRFU Fe-Sc Hydrogenase (1.12.2.1) Q59667_PYRFU Fe2þ,Ni2þc Sarcosine oxidase subunit alpha (1.5.3.1) Q8U025_PYRFU —d Probable peroxiredoxin (1.11.1.15) TDXH_PYRFU —d Prephenate dehydrogenase (1.3.1.12) Q8U096_PYRFU —d Glucose 1-dehydrogenase (1.1.1.47) Q8TZX2_PYRFU —d Pyruvate synthase subunit porA (1.2.7.1) PORA_PYRFU —d Phosphoglycerate dehydrogenase (1.1.1.95) Q8U3T5_PYRFU —d Ribonucleoside-diphosphate reductase (1.17.4.1) P95484_PYRFU —d

aData from SWISS-PROT/TrEMBL.9 bAccording to the catalytic reactions presented in the comments of SWISS-PROT/TrEMBL.9 cAccording to the cofactor information presented in the comments of SWISS-PROT/TrEMBL.9 dCofactor information is absent in the comments of SWISS-PROT/TrEMBL.9

TrEMBL,(9) the redox cofactors were pinpointed for 34 redox agree well with a recent discovery based on the pattern analysis proteins (Table 2). It can be found that most (94.1%) of the of ligand-protein mapping that NAD and NADP are among the proteins depend on organic cofactors rather than metallic earliest cofactors bound to proteins and the predicted most counterparts, in which NAD/P(H) are most widely shared ancient host proteins of these ligands pertain to c.2 fold.(12) (96.9%). Recently, Prachumwat and Li provided another set of early Elucidation of redox cofactor preference in proteins by classifying the yeast proteome into five age primitive proteins groups.(10) The oldest age group includes yeast proteins that It is of obvious interest to elucidate why transition metals can be traced back to eubacterial genomes, which consists of were less popular than organic cofactors to undertake the 1806 members (according to Saccharomyces Genome Data- redox functions of primitive proteins. Thirty years ago, White base (SGD), Oct 18, 2006),(11) from which 100 redox proteins proposed that coenzyme-binding sites of contemporary (with Enzyme Classification of E.C. 1._._._) can be extracted. enzymes may be relics of the earliest proteins,(1) which According to the catalytic reaction annotations provided by inspired us to speculate that the clues to elucidating the redox SWISS-PROT/TrEMBL,(9) the redox cofactors were deter- cofactor preference in primitive proteins may reside in the mined for 95 proteins, in which organic cofactors (82) are cofactor-binding residues of contemporary redox proteins. dominant and NAD/P(H) (67) are most popular (See Table S1 Recently, based on an extensive analysis of metal-binding in Supplementary Material). sites of metalloproteins, Kasampalidis et al. indicated that the The predominance of organic cofactors in two relatively copper and iron-coordinating residues are much more con- independently established putative ancient protein datasets served than those of other positions during evolution and the strongly suggests that organic cofactors were more involved metal-binding sites mainly consist of reducing amino acids, i.e. than metallic counterparts in the redox functions of primitive His, Cys and Met.(13) This conclusion is further supported by proteins and NAD/P(H) were the most widely used redox our analyses. Through identifying the metal-binding residues cofactors. In addition, as shown in Table 2, most (64.3%) of redox metalloproteins (with Enzyme Classification of E.C. of NAD/P(H)-containing redox proteins belong to NAD(P)- 1._._._, containing copper,iron, manganese, nickel, cobalt and binding Rossmann-fold domain fold (c.2). All of these findings selected from PDB at a non-redundant level), we found that

BioEssays 30.8 767 中国科技论文在线 http://www.paper.edu.cn Hypothesis

Table 2. Putative ancient redox proteins derived from MANET database

Putative ancient redox proteins (EC number)a Folda Cofactorc

UDP-glucose 6-dehydrogenase (1.1.1.22) c.2.1.6 NADþ, NADH aldehyde reductase (1.1.1.21) c.1.7.1 NAD(P)þ, NAD(P)H alcohol dehydrogenase (NADPþ) (1.1.1.2) c.1.7.1 NADPþ, NADPH aspartate-semialdehyde dehydrogenase (1.2.1.11) c.2.1.3 NADPþ, NADPH homoserine dehydrogenas (1.1.1.3) c.2.1.3 NADPþ, NADPH dTDP-4-dehydrorhamnose reductase (1.1.1.133) c.2.1.2 NADPþ, NADPH biliverdin reductase (1.3.1.24) c.2.1.3 NAD(P)þ, NAD(P)H MP dehydrogenase (1.1.1.205) c.1.5.1 NADþ, NADH 3-hydroxyacyl-CoA dehydrogenase (1.1.1.35) c.2.1.6 NADþ, NADH 1,6-dihydroxycyclohexa-2,4-diene-1-carboxylate dehydrogenase (1.3.1.25) c.2.1 NADþ, NADH alcohol dehydrogenase (1.1.1.1) c.2.1.1 NADþ, NADH dihydropyrimidine dehydrogenase (NADPþ) (1.3.1.2) c.1.4.1 NADPþ, NADPH phosphoglycerate dehydrogenase (1.1.1.95) d.58.18.1 NADþ, NADH glycerate dehydrogenase (1.1.1.29) c.2.1.4 NADþ, NADH dihydrouracil dehydrogenase (NADþ) (1.3.1.1) —b NADPþ, NADPH hydroxypyruvate reductase (1.1.1.81) —b NAD(P)þ, NAD(P)H benzoate 4-monooxygenase (1.14.13.12) —b NAD(P)þ, NAD(P)H L-threonine 3-dehydrogenase (1.1.1.103) —b NADþ, NADH NADH dehydrogenase (quinone) (1.6.99.5) —b NADþ, NADH NADH dehydrogenase (ubiquinone) (1.6.5.3) —b NADþ, NADH precorrin-3B synthase (1.14.13.83) —b NADþ, NADH precorrin-2 dehydrogenase (1.3.1.76) —b NADþ, NADH protochlorophyllide reductase (1.3.1.33) —b NADPþ, NADPH xanthine dehydrogenase (1.17.1.4) —b NADþ, NADH 3-hydroxybutyryl-CoA dehydrogenase (1.1.1.157) —b NADPþ, NADPH precorrin-6A reductase (1.3.1.54) —b NADPþ, NADPH magnesium-protoporphyrin IX monomethyl ester (oxidative) cyclase (1.14.13.81) —b NADPþ, NADPH glutamyl-tRNA reductase (1.2.1.70) —b NADPþ, NADPH GMP reductase (1.7.1.7) —b NADPþ, NADPH sulﬁte reductase (NADPH) (1.8.1.2) —b NADPþ, NADPH cinnamyl-alcohol dehydrogenase (1.1.1.195) —b NADPþ, NADPH dihydroorotate oxidase (1.3.3.1) —b FMNd coproporphyrinogen dehydrogenase (1.3.99.22) —b Fe-Sd sulﬁte reductase (ferredoxin) (1.8.7.1) —b Fe-Sd Benzoyl-CoA reductase (1.3.99.15) —b —e dihydroorotate dehydrogenase (1.3.99.11) c.1.4.1 —e adenylyl-sulfate reductase (1.8.99.2) —b —e glucoside 3-dehydrogenase (1.1.99.13) —b —e cinnamoyl-CoA reductase (1.2.1.44) c.2.1 —e

aData from Molecular Ancestry Network (MANET) (http://www.manet.uiuc.edu/).8 bFold information can not be determined by Molecular Ancestry Network (MANET) (http://www.manet.uiuc.edu/).8 cAccording to the catalytic reactions presented in the comments of SWISS-PROT/TrEMBL.9 dAccording to the cofactor information presented in the comments of SWISS-PROT/TrEMBL.9 eCofactor information is absent in the comments of SWISS-PROT/TrEMBL.9

His, Cys and Met, especially His, indeed play a major role in In addition, the average Shannon’s information theoretic fastening transition metals (Table3), which can be understood, entropies(13,15) of cofactor-binding and non-binding positions at least in part, in terms of their skills of modulating redox indicate that the organic cofactor-binding residues are also potentials of transition metal ions. In comparison, through much more conserved than others during evolution (Fig. 1). scrutinizing the binding sites of redox proteins derived from The conservation and preference of cofactor-binding Annotated Database of Druggable Binding Sites from the PDB residues in contemporary redox proteins imply that the (sc-PDB)(14) (with Enzyme Classification of E.C. 1._._._ and ancestral metallic redox proteins depended largely on reduc- also selected at a non-redundant level), we found that the most ing amino acids to bind cofactors, while the ancestral non- popular residues to bind organic redox cofactors (i.e. NAD, metallic redox proteins mainly relied on neutral or near-neutral NADP,FMN and FAD) are neutral or near-neutral amino acids, amino acids to fasten redox cofactors. According to the yields such as Gly (12% 16%), Ala (10%), Ser (12% 14%) and of prebiotically synthesized amino acids(16–19) and the evolu- Thr (10% 14%), and reducing amino acids, e.g. His (0.4% tionary theory of genetic code,(20–23) it is reasonable to infer 0.8%), Cys (2%) and Met (1%), are very rare (Table 3). that, in the primordial world, neutral or near-neutral amino

768 BioEssays 30.8 中国科技论文在线 http://www.paper.edu.cn Hypothesis

Table 3. Amino acid compositions for binding sites of redox proteins containing different cofactors

Copper Iron Manganese Nickel Cobalt NAD NADP FMN FAD Amino acida (%)b (%)c (%)d (%)e (%)f (%)g (%)h (%)i (%)j

Gly 1.2 1.6 0.0 0.0 0.0 15.1 15.7 13.6 12.7 Ala 0.5 0.8 0.0 0.0 0.0 10.3 11.4 7.8 9.7 Ser 0.7 2.4 0.0 0.0 0.0 12.7 12.8 13.2 14.0 Asp 2.7 7.0 11.9 15.8 12.5 5.1 5.9 2.8 3.3 Glu 1.0 4.7 32.6 14.3 32.3 1.8 1.5 1.8 1.8 Val 0.0 0.3 0.0 0.0 0.0 3.1 2.8 4.3 4.7 Leu 0.7 0.5 0.0 0.0 0.0 1.5 0.8 1.5 1.3 Ile 0.0 0.0 0.0 0.0 0.0 1.4 0.6 0.5 0.5 Pro 0.2 0.0 0.0 0.0 0.0 1.2 0.8 0.8 0.3 Thr 0.0 0.8 0.0 0.0 0.0 10.0 13.2 12.6 12.2 Phe 0.0 0.4 0.0 0.0 0.0 1.5 1.3 0.8 1.1 Tyr 1.7 7.1 3.3 0.0 0.0 3.0 3.2 1.5 2.8 Arg 0.0 0.4 0.0 0.0 0.0 12.9 7.8 15.8 14.0 His 41.3 48.4 45.6 38.2 49.0 0.7 0.4 0.8 0.7 Trp 2.6 0.0 0.0 0.0 0.0 0.8 3.2 1.8 1.5 Asn 0.0 0.3 6.6 2.1 0.0 4.7 4.4 6.6 7.4 Gln 0.7 1.1 0.0 4.0 6.2 3.2 1.9 3.8 3.4 Lys 0.0 0.9 0.0 0.0 0.0 7.2 9.4 7.2 5.3 Cys 24.9 19.8 0.0 25.6 0.0 2.6 2.3 2.3 2.3 Met 21.8 3.5 0.0 0.0 0.0 1.2 0.6 0.5 1.0

aArranged in the order suggested by coevolution theory of genetic code.(21) bAverage values for 300 copper-containing proteins (Table S2). cAverage values for 443 iron-containing proteins (Table S2). dAverage values for 39 manganese-containing proteins (Table S2). eAverage values for 35 nickel-containing proteins (Table S2). fAverage values for 18 cobalt-containing proteins (Table S2). gAverage values for 142 NAD-containing proteins (Table S2). hAverage values for 73 NADP-containing proteins (Table S2). iAverage values for 124 FMN-containing proteins (Table S2). jAverage values for 150 FAD-containing proteins (Table S2).

acids (e.g. Gly, Ala, Ser and Thr) were much more abundant proteins was likely to be limited by the low abundance of than reducing amino acids (e.g. His, Cys and Met).(24) Thus, reducing amino acids. the preference of organic redox cofactors to metallic counterparts in primitive proteins can be understood, at least in part, in Conclusions terms of the amino acid abundance in the primordial world. It has been widely recognized that transition metals made That is to say, the usage of metallic cofactors in primitive redox great contributions to the origin of life. For instance, iron played

Figure 1. (Overleaf ) Composition and conservation of organic cofactor-binding residues in some non-metallic redox proteins. A: Schematic representation of NAD and binding residues in alcohol dehydrogenase (PDB entry: 1HSO) (a1), FMN and binding residues in flavodoxin (PDB entry: 1AHN) (a2) and FAD and binding residues in glutathione reductase (PDB entry: 3GRS) (a3). To clearly identify the organic cofactor-binding residues, only the residues forming hydrogen bonds with the cofactors were considered. During the screening process, first, ‘‘builder’’ module in Insight II software (Accelrys, Inc., San Diego, CA) was used to modify the protein and ligand properties, such as hybridization and bond order; then ‘‘measure’’ tool was used to calculate and display the hydrogen bonds. B: Multiple sequence alignments of alcohol dehydrogenases derived from 25 species (b1), flavodoxin from 16 species (b2), and glutathione reductase from 19 species (b3). Symbols above the alignments indicate sequence conservation: (*) 100% conserved identities; (:) highly conserved identities. To quantitatively measure the conservation of each position, Shannon’s information theoretic entropy was calculated by the equation: s(i) ¼S p(k) ln(p(k)), where p(k) is the probability that the ith position in the multiple-aligned sequences is occupied by a residue of class k.(15) The lower the entropy at a position is, the more conserved the residue at the position. The average entropy for NAD-binding residues is 0.59 0.19, significantly lower than that for other residues, 1.18 0.90 (p < 108); the average entropy for FMN-binding residues is 0.76 0.29, significantly lower than that for other residues, 1.66 0.79 (p < 10-5); the average entropy for FAD-binding residues is 0.07 0.02, significantly lower than that for other residues, 1.48 0.74 (p < 109). The present analyses indicate that the organic redox cofactor-binding positions are much more conserved than others during evolution. As NADP is very similar to NAD in structures, their binding sites are very similar to each other and thus the data for NADP are not presented.

BioEssays 30.8 769 中国科技论文在线 http://www.paper.edu.cn Hypothesis

Figure 1.

770 BioEssays 30.8 中国科技论文在线 http://www.paper.edu.cn Hypothesis

a redox catalytic role in the synthesis of biologically significant 8. Kim HS, Mittenthal JE, Caetano-Anolle´s G. 2006. MANET: tracing molecules in the pre-protein world with the help of inorganic evolution of protein architecture in metabolic networks. BMC Bioinform (25–28) 7:351–363. ligands (e.g. sulfides). However, it remains unknown 9. Bairoch A, Boeckmann B. 1994. The SWISS-PROT protein sequence whether transition metals still played an important role in the data bank: current status. Nucleic Acids Res 22:3578–3580. redox functions of primitive proteins. Through examining 10. Prachumwat A, Li WH. 2006. Protein function, connectivity, and duplicability in yeast. Mol Biol Evol 23:30–39. the cofactor usage of putative ancient redox proteins, we 11. Cherry JM, Ball C, Weng S, Juvik G, Schmidt R, et al. 1997. Genetic and found that the redox functions of primitive proteins depended physical maps of Saccharomyces cerevisiae. Nature 387:67–73. more on organic cofactors than transition metals, which can 12. Ji H-F, Kong D-X, Shen L, Chen L-L, Ma B-G, et al. 2007. Distribution patterns of small-molecule ligands in the protein universe and implica- be further explained in terms of the relative amino acid tions for origin of life and drug discovery. Genome Biol 8:R176. abundance in the primordial world. Although the present 13. Kasampalidis IN, Pitas I, Lyroudia K. 2007. Conservation of metal- hypothesis sounds reasonable and is self-consistent, it is still coordinating residues. Proteins 68:123–130. 14. Kellenberger E, Muller P, Schalon C, Bret G, Foata N, et al. 2006. sc- essential to test it by experiments. However, it is rather difficult PDB: an annotated database of druggable binding sites from the protein to directly (prebiotically) simulate the birth of redox proteins data bank. J Chem Inf Model 46:717–727. from scratch. Thus, we suggest that one may use indirect 15. Guharoy M, Chakrabarti P. 2005. Conservation and relative importance of residues across protein-protein interfaces. Proc Natl Acad Sci USA methods to examine the present hypothesis. 102:15447–15452. The scenario revealed in this study implies that the primitive 16. Weber AL, Miller SL. 1981. Reasons for the occurrence of the twenty redox proteins mainly consisted of neutral or near-neutral coded protein amino acids. J Mol Evol 17:273–284. 17. Kobayashi K, Tsuchiya M, Oshima T, Yanagawa H. 1990. Abiotic amino acids and organic cofactors. Hence, our hypothesis will synthesis of amino acids and imidazole by proton irradiation of be supported if one could rationally design and constitute a simulated primitive earth atmospheres. Orig Life Evol Biosph 20:99– protein by neutral or near-neutral amino acids and redox 109. 18. Kobayashi K, Kaneko T, Saito T, Oshima T. 1998. Amino acid formation in organic cofactors. In addition, the neutral or near-neutral gas mixtures by high energy particle irradiation. Orig Life Evol Biosph amino acids-based redox proteins should also be produced by 28:155–165. in vitro selection methodology,(29,30) using organic redox 19. Plankensteiner K, Reiner H, Rode BM. 2006. Amino acids on the rampant primordial Earth: Electric discharges and the hot salty ocean. Mol Divers cofactors as selectors. The former method can be regarded 10:3–7. as a top-down strategy, while the latter is a bottom-up tactic. It 20. Wong JT. 1975. A coevolution theory of the genetic code. Proc Natl Acad will be exciting to see the birth of such kind of proteins, no Sci USA 72:1909–1912. 21. Wong JT. 2005. Coevolution theory of the genetic code at age thirty. matter which strategy is selected. BioEssays 27:416–425. 22. Trifonov EN. 2004. The triplet code from first principles. J Biomol Struct Acknowledgments Dyn 22:1–11. 23. Trifonov EN, Gabdank I, Barash D, Sobolevsky Y. 2006. Primordia vita. We are grateful to Prof. Antonio Lazcano for his helpful deconvolution from modern sequences. Orig Life Evol Biosph 36:559– suggestions. 565. 24. Zhang H-Y. 2007. Exploring the evolution of standard amino-acid alphabet: when genomics meets thermodynamics. Biochem Biophys References Res Commun 359:403–405. 1. White HB. 1976. Coenzymes as fossils of an earlier metabolic state. J Mol 25. Wa¨chtersha¨user G. 1992. Groundworks for an evolutionary biochemistry: Evol 7:101–104. The iron-sulphur world. Prog Biophys Mol Biol 58:85–201. 2. Cleaves HJ, Miller SL. 2001. The nicotinamide biosynthetic pathway is a 26. Cody GD, Boctor NZ, Filley TR, Hazen RM, Scott JH, et al. 2000. by-product of the RNA world. J Mol Evol 52:73–77. Primordial carbonylated iron-sulfur compounds and the synthesis of 3. Huang F, Bugg C, Yarus M. 2000. RNA-catalyzed CoA, NAD, and FAD pyruvate. Science 289:1337–1340. synthesis from phosphopantetheine, NMN, and FMN. Biochemistry 27. Rees DC, Howard JB. 2003. The interface between the biological 39:15548–15555. and inorganic worlds: iron-sulfur metalloclusters. Science 300:929– 4. Daniel RM, Danson MJ. 1995. Did primitive microorganisms use nonhem 930. iron proteins in place of NAD/P? J Mol Evol 40:559–563. 28. Russell MJ, Martin W. 2004. The rocky roots of the acetyl-CoA pathway. 5. Galtier N, Tourasse N, Gouy M. 1999. A nonhyperthermophilic common Trends Biochem Sci 29:358–363. ancestor to extant life forms. Science 283:220–221. 29. Roberts RW, Szostak JW. 1997. RNA-peptide fusions for the in vitro 6. Bada JL, Lazcano A. 2002. Some like it hot, but not the first selection of peptides and proteins. Proc Natl Acad Sci USA 94:12297– biomolecules. Science 296:1982–1983. 12302. 7. Becerra A, Delaye L, Lazcano A, Orgel LE. 2007. Protein disulfide 30. Liu R, Barrick JE, Szostak JW, Robert RW. 2000. Optimized synthesis oxidoreductases and the evolution of thermophily: was the last common of RNA-protein fusions for in vitro protein selection. Methods Enzymol ancestor a heat-loving microbe? J Mol Evol 65:296–303. 318:268–293.

BioEssays 30.8 771