Tài liệu Báo cáo khoa học: Open questions in ferredoxin-NADP+ reductase catalytic mechanism ppt

REVIEW ARTICLE Open questions in ferredoxin-NADP + reductase catalytic mechanism Ne ´ stor Carrillo and Eduardo A. Ceccarelli Molecular Biology Division, Instituto de Biologı ´ a Molecular y Celular de Rosario (IBR), Facultad de Ciencias Bioquı ´ micas y Farmace ´ uticas, Universidad Nacional de Rosario, Argentina Ferredoxin (flavodoxin)-NADP(H) reductases (FNR) are ubiquitous flavoenzymes that deliver NADPH or low potential one-electron donors (ferredoxin, flavodoxin) to redox-based metabolisms in plastids, mitochondria and bacteria. The plant-type reductase is also the basic prototype for one of the major families of flavin-containing electron transferases that display common functional and structural properties. Many aspects of FNR biochemistry have been extensively characterized in recent years using a combination of site-directed mutagenesis, steady-state and transient kine- tic experiments, spectroscopy and X-ray crystallography. Despite these considerable advances, various key features in the enzymology of these important reductases remain yet to be explained in molecular terms. This article reviews the current status of these open questions. Measurements of electron transfer rates and binding equilibria indicate that NADP(H) and ferredoxin interactions with FNR result in a reciprocal decrease of affinity, and that this induced-fit step is a mandatory requisite for catalytic turnover. However, the expected conformational movements are not apparent in the reported atomic structures of these flavoenzymes in the free state or in complex with their substrates. The overall reaction catalysed by FNR is freely reversible, but the pathways leading to NADP + or ferredoxin reduction proceed through entirely different kinetic mechanisms. Also, the reductases isolated from various sources undergo inactivating dena- turation on exposure to NADPH and other electron donors that reduce the FAD prosthetic group, a phenomenon that might have profound consequences for FNR function in vivo. The mechanisms underlying this reductive inhibition are so far unknown. Finally, we provide here a rationale to interpret FNR evolution in terms of catalytic efficiency. Using the formalism of the Albery–Knowles theory, we identified which parameter(s) have to be modified to make these reductases even more proficient under a variety of conditions, natural or artificial. Flavoenzymes with FNR activity catalyse a number of reactions with potential importance for biotechnological processes, so that modifi- cation of their catalytic competence is relevant on both scientific and technical grounds. Keywords: ferredoxin-NADP(H) reductase; flavoproteins; oxidoreductases; ferredoxin; flavodoxin; catalytic mechanism; X-ray crystallography; steady-state kinetics; transient kinetics; enzyme evolution. Portrait of a reductase Ferredoxin-NADP(H) reductases (FNR, EC 1.18.1.2) con- stitute a family of hydrophilic, monomeric enzymes that contain noncovalently bound FAD as a prosthetic group. The first FNR was isolated from pea thylakoids in the mid- 1950s [1]. Shortly thereafter, Shin and Arnon [2] showed that the physiological role of the chloroplast reductase was to catalyse the final step of photosynthetic electron trans- port, namely, the electron transfer from the iron-sulphur protein ferredoxin (Fd), reduced by photosystem I, to NADP + (Eqn 1). This reaction provides the NADPH necessary for CO 2 assimilation in plants and cyanobacteria. 2FdðFe 2þ ÞþNADP þ þ H þ !  2FdðFe 3þ Þ þ NADPH ð1Þ Equation 1 reflects one of the most conspicuous features of FNR, its ability to split electrons between obligatory one- and two-electron carriers, which is a direct consequence of the biochemical properties of its prosthetic group. FAD and other flavins (Fl) can exist in three different redox states: oxidized, one-electron reduced (semiquinone) radical and fully reduced hydroquinone, containing, respectively, 18, 19 and 20 electrons in a p orbital system constructed from 16p atomic orbitals [3]. The isoalloxazine ring also has 7r lone electron pairs available for protonation, providing ample opportunity for tautomers. When free in solution, flavin semiquinone radicals disproportionate to the oxidized and reduced forms (Eqn 2), but when buried in a protein they are much less prone to dismutation. This results in a considerable stabilization of the semiquinone, which in turn allows flavoproteins, in general, to readily engage in mono- and Correspondence to N. Carrillo, IBR, Facultad de Ciencias Bioquı ´ micas y Farmace ´ uticas, Universidad Nacional de Rosario, Suipacha 531 (S2002LRK) Rosario, Argentina. Fax: + 54 341 439 0465, Tel.: + 54 341 435 0661, E-mail: carrill@arnet.com.ar Abbreviations:cytc,cytochromec; Fl, flavin; Fd, ferredoxin; FNR, ferredoxin-NADP(H) reductase; Fld, flavodoxin; ox, oxidized; red, reduced; sq, semiquinone. Enzyme: Ferredoxin-NADP(H) reductases (FNR, EC 1.18.1.2). (Received 13 January 2003, revised 4 March 2003, accepted 13 March 2003) Eur. J. Biochem. 270, 1900–1915 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03566.x bielectronic exchange reactions. This disposition is favoured in FNR by the small differences existing in the redox potentials (E m ) for the various one- and two-electron transfer processes, that is )331, )314 and )323 mV for the oxidized/semiquinone couple, the semiquinone/hydro- quinone pair and the two-electron FAD reduction, respect- ively. H 2 Fl Æ þ H 2 Fl Æ !  HFl ox þ H 2 Fl À red þ H þ ð2Þ FNR functions are by no means confined to photosyn- thesis. Following the isolation of the chloroplast reductase, flavoproteins with FNR activity were found in photo- trophic and heterotrophic bacteria, animal and yeast mitochondria, and apicoplasts of obligate intracellular parasites [4–10]. Studies in a number of species showed that FNR operates as a general electronic switch at the bifurcation steps of many different electron transfer pathways. In heterotrophic organisms and tissues, inclu- ding roots and other nonphotosynthetic plant organs, the reaction represented by Eqn (1) is driven toward Fd reduction, mobilizing low-potential electrons for a diversity of metabolisms. They include steroid hydroxylation, nitrate reduction, nitrogen and hydrogen fixation, anaerobic pyruvate assimilation, desaturation of fatty acids and synthesis of amino acids and deoxyribonucleotides (Table 1, reviewed in [11]). Van Thor et al. [12] have proposed that this ÔbackwardÕ reaction might also occur in photosynthetic cells of cyanobacteria, being a rate-limiting step of cyclic electron flow around photosystem I. Some bacteria and algae possess another flavoprotein, FMN-containing flavodoxin (Fld), that is able to effi- ciently replace Fd as the electron partner of FNR in different metabolic routes, including photosynthesis [8]. In cyanobacteria, Fld expression is induced under condi- tions of iron deficit, when the [2Fe)2S] cluster of Fd cannot be assembled [8]. In other prokaryotes, flavodoxins are synthesized constitutively, or induced by oxidants [13]. Both Fld and FNR participate in detoxification of reactive oxygen species in aerobic and facultative bacteria [13,14]. FNR displays a strong preference for NADP(H) and is a very poor NAD(H) oxidoreductase. In contrast, the site of electron donation from reduced flavin appears to be open to a remarkable variety of adventitious oxidants of very different structure and properties. The reaction is largely irreversible and led the original FNR discoverers to label it as a thylakoid-bound ÔNADPH diaphoraseÕ [1]. NADPH þ H þ þ nA ox ! NADP þ þ nA red ð3Þ A ox and A red represent the oxidized and reduced forms of the electron partner, and the term n equals one or two depending on whether the oxidant behaves as a two- or a one-electron carrier, respectively. The list of acceptors comprises ferricyanide and other complexed transition metals, substituted phenols, nitroderivatives, tetrazolium salts, NAD + (transhydrogenase activity), viologens, qui- nones and cytochromes (reviewed in [15]). By contrast, the oxidase activity of FNR is very low [15,16], suggesting that there might be restrictions to the formation of the caged radical pair and/or the covalent (C4a)-flavin hydroperoxide intermediates required for efficient oxygen reduction [17], although other mechanisms cannot be ruled out. The oxidase reaction is enhanced several-fold by many FNR acceptors, including one-electron reduced Fd or Fld, viologens, nitroderivatives and quinones, that can readily engage in oxygen-dependent redox cycling leading to superoxide formation [16,18]. Diaphorase activity is prob- ably devoid of physiological meaning in most cases, but it has paid an enormous service to the understanding of FNR function and catalytic mechanism. In addition, some of these artificial reactions might have technological relevance for bioremediation and the pharmaceutical industry [18,19]. This brief account illustrates the plasticity of FNR as a catalyst and its ubiquity among living organisms. Studies on the enzymology of this reductase began in the late 1960s, employing both steady-state and rapid kinetic measure- ments, and a comprehensive model describing the various steps of reaction 1 was formulated by Batie and Kamin in 1984. Since then, an impressive amount of information on FNR structure, function and biogenesis has been gathered, providing clues to understand holoenzyme assembly, sub- strate binding and catalysis in molecular terms. However, several key features of FNR enzymology remain obscure, and could not be accounted for, or reconciled with the available structural-functional data. The aim of this article is to review these open questions in FNR biochemistry, as a tool and conceptual framework for future research. It is worth noting that a thorough understanding of FNR catalytic mechanism has become increasingly important after the recognition of this reductase as the prototype for a Table 1. Functions associated with ferredoxin-NADP(H) reductase in different organisms. Organism or tissue Function Related partner or metabolism References Chloroplasts NADP + reduction Photosynthetic electron transport [2] cyanobacterial vegetative cells [8] Nonphotosynthetic plastids Fd reduction Nitrogen assimilation [72,73] Cyanobacterial heterocysts Fd or Fld reduction Dinitrogen fixation [8] Mitochondria Fd reduction Steroid hydroxilation [58] Fatty acid desaturation E. coli Fld reduction Activation of anaerobic enzymes [7] E. coli NADPH oxidation Protection against oxidative stress [14] Clostridia Fd reduction Hydrogen fixation [4] Methanogenic bacteria Fd reduction Methane assimilation [5] Apicomplexan parasites Fd reduction Fatty acid desaturation [10] Ó FEBS 2003 Ferredoxin-NADP(H) reductase catalytic mechanism (Eur. J. Biochem. 270) 1901 large family of flavin-containing enzymes displaying similar structures and reaction mechanisms [20–26]. The two-domain structure of ferredoxin- NADP(H) reductases Three-dimensional models of the oxidized and fully reduced forms of spinach FNR, refined to 1.7 A ˚ resolution, were first reported by Karplus and coworkers [20–23]. The flavoprotein molecule is made up of two structural domains, each containing approximately 150 amino acids (Fig. 1A). The carboxyl terminal region includes most of the residues involved in NADP(H) binding, whereas the large cleft between the two domains accommodates the FAD group. The isoalloxazine ring system is tightly bound to the amino terminal domain through hydrogen bonds and van der Waals contacts [23]. It stacks between the aromatic side chains of two tyrosine residues, represented in spinach FNR by Tyr96 on the si-face and Tyr314 (the carboxyl terminus) on the re-face (Fig. 1B). The phenol ring of Tyr314 and the flavin group are coplanar in such a way as to maximize p-orbital overlap [20]. A large portion of the isoalloxazine moiety is shielded from the bulk solution by the side chain of the carboxyl terminal tyrosine, but the edge of the dimethyl benzyl ring that participates in electron transfer remains exposed to solvent in the native holoenzyme [23]. Crystal structures have also been resolved for other FNR proteins, including those present in pea, paprika and maize chloroplasts [27–29], corn root plastids [30], cyanobacteria [31], Escherichia coli [32] and Azotobacter vinelandii [33]. Despite ample variations in amino acid sequences, the chain topologies of all these proteins are highly conserved, with most differences occurring at the loops between the invariant secondary structure elements [20–23,27–33]. In the chloroplastic and cyanobacterial reductases, as in most flavoproteins, the FAD molecule binds in an extended form, with the 2¢-P-AMP moiety wrapped up by a short sheet- loop-sheet motif of the apoprotein [20–23,27–31]. In contrast, the adenosine in the E. coli FAD bends back from the diphosphate so that the nitrogen at position 7 of the adenine group forms a hydrogen bond to nitrogen 1 of the isoalloxazine. This interaction is further stabilized by stacking of the two terminal aromatic side chains; the phenol ring of Tyr247 with the flavin, as in plant FNR, and the indole of Trp248 with the adenine [32]. The FAD cofactor of the A. vinelandii FNR is also twisted and stabilized in a similar way, but the environment of the prosthetic group presents some unique features in this reductase. The most important difference is the absence of the aromatic interaction on the re-face of the isoalloxazine that is typical of plant, cyanobacterial and E. coli FNR proteins [33]. Instead, the tyrosine position in the Azotobacter enzyme is occupied by an alanine (Ala254). Despite these major modifications in the active site region, the FMN halves of FAD display very similar conformations when the various homologous proteins are superimposed [20–23,27–33]. Several amino acid residues important for the structural integrity of the native holoenzyme, for electron transfer, and for FAD, NADP(H), ferredoxin and flavodoxin binding, have been identified using a combination of chemical modification experiments, site-directed mutagenesis and X-ray diffraction studies. This very interesting aspect of FNR biochemistry will only be addressed here in relation to the catalytic mechanism of the enzyme. Further information on these topics can be found in a number of articles and reviews [23,31,34–57]. Finally, it is noteworthy that not all flavoenzymes displaying FNR activity belong to the FNR class. The adrenodoxin reductases found in animal and yeast mito- chondria and their bacterial homologues represent a curious case. They are hydrophilic, monomeric proteins made up of FAD and NADP(H) domains that can freely exchange electron partners (ferredoxin, flavodoxin, adrenodoxin) with plant-type FNR [58]. Many features of NADP(H) docking and catalysis are also similar, although reaction geometry is different [58]. However, mitochondrial reduc- tases are unrelated in sequence to their chloroplast coun- terparts, and the structural data indicate that they actually belong to the disulphide oxidoreductase family of flavopro- teins, whose prototype is glutathione reductase [6,58]. The plant-type and mitochondrial-type FNR progenies thus represent two different and independent origins, followed by a remarkable case of convergent evolution to yield proteins with essentially the same enzymatic properties. The reaction pathway Batie and Kamin [59,60] formulated the first detailed pathway for the FNR-mediated electron transfer between NADP(H) and ferredoxin, using data from binding equili- bria, steady-state kinetics and rapid mixing experiments (Fig. 2). The overall reaction was interpreted as an ordered two-substrate process, with NADP + binding first. Under these assumptions, the kinetics were shown to be consistent with the formation of ternary complexes as intermediates of the catalytic mechanism (Fig. 2, steps 2 and 5). Substrate- binding parameters and rate constants were determined for the complete pathway mediated by both plant and bacterial reductases, and for several individual steps (Table 2). Turn- over numbers in the range of 200–600 s )1 have been reported for the spinach and Anabaena enzymes, whereas E. coli FNRismuchlessactive(Table2).Go ´ mez-Moreno and coworkers have also studied the electron transfer to and from flavodoxin [46,47 and references therein]. The reverse reaction, that is the electron transport from NADPH to Fd (or Fld), is routinely measured in vitro through a coupled assay, using cytochrome c (cyt c) as final electron acceptor (Eqns 4 and 5). NADPH þ 2FdðFe 3þ Þ !  NADP þ þ H þ þ 2FdðFe 2þ Þð4Þ Fd ðFe 2þ Þþcyt cðFe 3þ Þ!FdðFe 3þ Þþcyt cðFe 2þ Þð5Þ In the following sections, we will analyse the various steps involved in FNR catalysis, with emphasis in unresolved or controversial questions. Most of the discussion will be based on data obtained with the plant and cyanobacterial flavoenzymes. The reductases present in E. coli and other bacterial species are less well characterized, but it is already clear that they display a number of differences with respect to their plant-type counterparts. These distinct features will also be addressed when pertinent. 1902 N. Carrillo and E. A. Ceccarelli (Eur. J. Biochem. 270) Ó FEBS 2003 Fig. 1. The Ca polypeptide backbone and the active site region of plant-type ferredoxin–NADP(H) reductase. (A) FNR is a two-domain flavoprotein. The computer graphic is based on X-ray diffraction data for the spinach enzyme [23], with the FAD binding domain shown in blue, the NADP(H) binding domain in pink, and the FAD prosthetic group in yellow. (B) Detailed view of the isoalloxazine ring system in FNR, displaying relevant interactions with active site amino acid residues. The phenol ring of the carboxyl terminal tyrosine 314 shields the re-side of the flavin from solvent. The figure was drawn using SWISS - PDBVIEWER 3.7 and rendered with POV - RAY . Ó FEBS 2003 Ferredoxin-NADP(H) reductase catalytic mechanism (Eur. J. Biochem. 270) 1903 NADP(H) binding (Fig. 2, steps 1 and 9) Even though no chemistry is expected to occur between the oxidized nicotinamide and the oxidized flavin, the role of NADP + as leading substrate during FNR turnover is supported by a number of kinetic measurements. In the spinach enzyme, electron transfer from reduced Fd to FNR is too slow (k obs ¼ 40–80 s )1 ) to account for steady-state rates of nucleotide reduction (k obs >500s )1 ). NADP + binding greatly accelerates this reaction (Table 2), indicating that the presence of the coenzyme at the active site is a prerequisite for catalysis [60]. Possible mechanisms involved in this activation are discussed in the next section. Formation of the binary complex has been measured in vitro by a variety of techniques, most conspicuously differential spectroscopy. Binding isotherms fit to simple hyperbolic functions in all cases. Dissociation constants increased with the ionic strength of the medium (I)andthe Fd concentration, with K d ¼ 6–20 l M at I % 100 m M for plant and Anabaena reductases [43,44,50,59,61]. Karplus and coworkers [20,23] have proposed that stacking of the nicotinamide ring onto the re-face of the isoalloxazine moiety during enzyme turnover requires displacement of the aromatic side chain of the carboxyl terminal tyrosine. This thermodynamically unfavoured process results in a decrease of the binding affinity for NADP(H) relative to those of the 2¢-P-AMP and 2¢-P-ADP-ribose analogues [61]. A similar movement has been postulated for the penultimate tyrosine residue of E. coli FNR, although in this case the carboxyl terminal tryptophan interacting with the adenosine moiety of FAD might also undergo a conformational change upon substrate binding [32,62–64]. Complex formation is then interpreted as a two-step binding of the nucleotide to a bipartite site (Fig. 3). The first step involves a strong interaction of FNR with the adenosine part of NADP(H), followed by isomerization leading to nicotinamide docking and, eventually, hydride transfer (Eqn 6). FNR þ NADPðHÞ !  FNR NADPðHÞ !  FNR NADPðHÞð6Þ Where FNRÆNADP(H) and FNRÆNADP(H) represent complexes with the coenzyme bound through the adeno- sine, or the adenosine and nicotinamide portions, respect- ively. The second step in Eqn 6 is energetically costly and weakens the entire interaction to a remarkable extent. Measurements of complex formation in vitro ledtothe amazing conclusion that under saturating conditions less than 20% of the nicotinamide is placed in contact with the flavin, and is therefore available for hydride transfer [50]. The model of Fig. 3 explains why crystals of FNR with bound analogues could be obtained with the wild-type enzyme, whereas those involving NADP(H) complexes were only possible with engineered FNR proteins in which the carboxyl terminal tyrosine had been replaced by nonaromatic residues such as serine [27]. In these FNR mutants, rearrangement of the phenol group is no longer required, and both pockets of the binding site would be readily accessible. Figure 4 shows amino acid residues involved in recogni- tion of the adenosine-ribose and the NMN portions of the dinucleotide, as identified through structural and mutage- nesis studies [23,27,31,51]. They include residues displaying charge interactions with the specific 2¢-phosphate group presumably responsible for discrimination against NAD(H). The increase in coenzyme affinity caused by substitutions of the carboxyl terminal tyrosine was so dramatic that the resulting FNR mutants were able to avidly incorporate NADP + during biosynthesis in E. coli and keep it bound through the purification and crystallization steps [27]. Incidentally, once the carboxyl terminal residue is replaced, the electrostatic interactions at the 2¢-phosphate group of NADP + are no longer sufficient to discriminate between the coenzymes, and FNR becomes an efficient NAD(H) oxidoreductase [50]. The reversible NADPH release depicted in step 9 also represents the initial event of the reverse reaction, NADPH- ferredoxin reductase, and will be addressed in further detail in a forthcoming chapter, when discussing that FNR activity. Fig. 2. The electron transfer mechanism of ferredoxin-NADP(H) reductase. The various steps of the catalytic pathway were initially proposed by Batie and Kamin [60] on the basis of kinetic and binding experiments on the spinach FNR. Oxidized forms are white, one- electron reduced forms are light grey and two-electron reduced forms are dark grey. 1904 N. Carrillo and E. A. Ceccarelli (Eur. J. Biochem. 270) Ó FEBS 2003 Electron transfer from reduced ferredoxin to ferredoxin-NADP(H) reductase (Fig. 2, steps 2–4) A recent article by Hurley et al. [57] has provided a very comprehensive and updated review of experiments descri- bing the interaction and electron transfer between Anabaena FNR and Fd. Accordingly, the present section gives only a concise account of these data; the reader is referred to the above mentioned article for a more detailed description of the two processes. Conversion of oxidized FNR to the semiquinone form (FNR sq ) by reduced Fd (or Fld) is too fast to be measured by rapid mixing techniques [44,60,65]. However, the kinetics Fig. 3. Schematic representation of the bipartite NADP(H)-binding mode to ferredoxin–NADP(H) reductase. The model is based on the properties of a tyrosine-to-serine site-directed mutant of pea FNR bound to NADP(H) [27,50]. Sites A and N represent the adenosine- and the nicotinamide- binding regions, respectively, in the active site of FNR. Table 2. Kinetic and binding parameters for various activities and interactions of ferredoxin-NADP(H) reductase. Binding and kinetic parameters were averaged from experiments carried out at I £ 100 m M . (Original sources cited in the text.) When dispersion among reported values exceeded 50%, the interval between extreme data is indicated. ND, not determined. Reaction FNR source K m or K d (l M ) k obs or k cat (s )1 ) Spinach leaves Anabaena E. coli Spinach leaves Anabaena E. coli Binding FNR ox – NADP + 15 6 ND – – – FNR ox –Fd ox (Fld ox ) <1 4 (3) a 0.5 (2) a –– – Electron transfer Fd red fi FNR fi NADP + 10 b ND ND 600 >200 ND 1 b Fd red (Fld red ) fi FNR ox ND 10 ND >600 c 60 c 6200 (>600) c,a 250 (>600) c,a 8 (25) a ND (8) a FNR red fi NADP + ND ND 500 >600 ND NADPH fi FNR fi Fd ox fi cyt c 10 b 6 b 4 b 250 225 20 1 b 15 b 2 b NADPH fi FNR fi Fld ox fi cyt c ND ND 33 b 4 b 7 b ND 23–80 5 NADPH fi FNR ox <2 ND <5 >600 d 200 d >600 d >140 d 22 FNR red fi Fd ox (Fld ox ) ND ND <5 (<5) a ND >600 (3) a 2 (0.01) a NADPH fi FNR fi K 3 Fe(CN) 6 30 b 100 b 23 b 170 b 10 b 24 b 550 225–520 27 a Values in parentheses are those obtained with Fld. b The upper and lower numerals indicate K m estimates for electron donors (Fd red , NADPH) and acceptors (NADP + ,Fd ox , Fld ox ,K 3 Fe(CN) 6 ), respectively. c The upper and lower values reported provide rates of transfer for the first and second electron, respectively. d The upper and lower numerals show rates of formation of charge–transfer complexes and rates of hydride transfer, respectively. Ó FEBS 2003 Ferredoxin-NADP(H) reductase catalytic mechanism (Eur. J. Biochem. 270) 1905 of this reaction could be resolved for the Anabaena reductase by using laser flash photolysis, yielding a k obs of about 6000 s )1 and a K d of 1.7 l M for the transient FNR ox – Fd red complex at I ¼ 100 m M [40,44,57]. Taking into account the many contacts that are required to establish this protein–protein interaction, it is somehow surprising that FNR can efficiently accommodate Fd or Fld, two proteins that differ completely in their primary, secondary and tertiary structures. Other similarities must exist to account for their functional equivalence, in spite of the lack of homology. It is interesting to note, in this context, that the molecular association of FNR with its electron partners is steered by electrostatic interactions [37,40,47,56,66,67]. Using the Hodgkin index as a similarity measure, Ullmann et al. [68] showed that Fd and Fld could be completely overlapped on the basis of their surface electrostatic potentials. The active sites and prosthetic groups of both proteins, rather than their centres of mass, coincided in the alignment [68]. Transient associations of the electron carriers in different oxidation states are generally not amenable to structural studies, but binary complexes of oxidized FNR and Fd could be resolved by X-ray crystallography for both the Anabaena and maize couples [29,69]. The resulting struc- tures provided insightful data to complement chemical cross-linking and mutagenesis studies, and helped to model flavodoxin docking [46]. Fd binds to a concave region of the FAD domain of maize FNR (Fig. 5A), burying an accessible area of % 800 A ˚ 2 in each partner, which repre- sents about 5% and 15% of the total surface areas of FNR and Fd, respectively [29]. The FAD and [2Fe)2S] redox centres are sufficiently close (6.0 A ˚ ) for direct electron transfer through the space between the two prosthetic groups (Fig. 5A–C). Distribution of surface charge and calculations of the molecular dipole moments confirm the relevance of complementary patches of basic and acidic residues in FNR and Fd, respectively [66,67]. These polar groups play a major role in determining the relative orientation of the two electron carriers in the initial nonproductive complex [40,66]. Attainment of the func- tional conformations competent for electron exchange requires further, fine adjustments, stabilized by a combina- tion of well-defined hydrogen bonds, salt bridges, van der Waals interactions and hydrophobic packing forces origin- ating from the dehydration of the protein–protein interface [40,47,48,52,57]. When the FNR molecules of the corn and cyanobacterial complexes are superimposed, the Fd part- ners appear rotated by an angle of 96°, indicating that many protein–protein interactions are different in the two systems [29,69]. Hurley et al. [57] have therefore proposed that, as in the case of Fld binding, the crucial parameters selected during evolution might be proximity of the prosthetic groups in a nonpolar environment to facilitate direct electron transfer. Fig. 4. The nucleotide-binding site of ferredoxin–NADP(H) reductase. View of NADP(H) bound to the pea FNR-Y308S mutant reveals the intimate interactions made by both the 2¢-P-AMP portion of the ligand and the nicotinamide. NADP(H) is depicted in blue, FAD in yellow, amino acids in grey. Dashed lines mark interactions of £ 3.5 A ˚ thatmayengageinhydrogenbonds.ResiduesarelabelledasinpeaFNR(numbersinparentheses are those of spinach FNR). The figure shows two water molecules that display hydrogen interactions with both NADP(H) and the protein backbone. Modified from Deng et al. [27], drawn using SWISS - PDBVIEWER 3.7 and rendered with POV - RAY . 1906 N. Carrillo and E. A. Ceccarelli (Eur. J. Biochem. 270) Ó FEBS 2003 Full reduction of ferredoxin-NADP(H) reductase (Fig. 2, steps 5–7) When spinach FNR ox is mixed with excess Fd red in a stopped flow system, all the flavoprotein is converted into the semiquinone form in the dead time of the instrument [60]. Transfer of the second electron, however, is too slow to be compatible with steady-state catalysis, as already indica- ted. The latter process actually involves various steps, which are the dissociation of Fd ox (Fig. 2, step 4), binding of Fd red (step 5) and flavin reduction (step 6). The reaction is strongly inhibited by Fd ox and stimulated by NADP + , Fig. 5. Structure of the ferredoxin–NADP(H) reductase–ferredoxin complex. (A) View of the maize leaf bipartite FNR–Fd complex with the ribbon diagram of Fd coloured in red, the FAD-binding domain of FNR in blue and the NADP(H) binding domain in pink. (B) Hypothetical tripartite NADP(H)–FNR–Fd complex. View of the superposition of the maize leaf bipartite FNR–Fd complex and the pea FNR–NADP(H) complex. Pea and maize FNR polypeptides were superimposed by the least square fitting of the isoalloxazine ring of FAD. The two domains of maize FNR are shown in light blue, Fd in red, FAD in yellow and the NADP(H) from the pea FNR–NADP(H) crystals in deep blue. White arrows indicate the [2Fe)2S] cluster (green), the isoalloxazine ring of FAD and the nicotinamide ring of NADP(H). (C) Superimposed view of the active site structure of the maize FNR complexed with Fd (in red), the free maize FNR (in light blue) and the pea FNR complexed with NADP(H) (in grey). The pyridine nucleotide is represented in blue. Note that in the three superimposed structures E312 (306) undergoes significant movements upon complex formation. To facilitate the observation, E306 from pea FNR was omitted in the main figure and included in the inset. The models were based on the detailed structures of the bipartite complexes reported by Kurisu et al. [29] and Deng et al. [27]. The figure was drawn using SWISS - PDBVIEWER 3.7 and rendered with POV - RAY . Ó FEBS 2003 Ferredoxin-NADP(H) reductase catalytic mechanism (Eur. J. Biochem. 270) 1907 indicating that step 4 is the rate-limiting step, and that NADP + facilitates Fd ox release, allowing the entire reaction to proceed at a rapid pace through steps 4 and 8 [60]. The sequence of events depicted in Fig. 2 agrees well with other experimental observations. A ternary complex between the two oxidized proteins and NADP + is readily formed in vitro as measured by differential spectroscopy. The affinity of FNR for Fd ox decreased > 10-fold on addition of NADP + ,andvice versa, indicating a strong case of negative cooperativity for binding [59]. Essentially the same results were obtained when using NADPH [59]. In this case, dissociation of the two enzyme–product complexes, rather than formation of the FNR–substrate species, are the most demanding steps that limit turnover. NADP + reduction and product release (Fig. 2, steps 8 and 9) Full FNR reduction renders a two-electron reduced Fd ox – FNR red –NADP + complex that requires hydride transfer to the nucleotide and dissociation to complete the catalytic cycle. The sequence of steps proposed in Fig. 2 complies with a canonical compulsory order mechanism, although alternative pathways could be envisaged. For instance, it is conceivable that Fd ox could dissociate from the FNR red – NADP + complex prior to nucleotide reduction. Rapid mixing experiments provided the main empirical support for the proposed reaction order, showing that in a mixture of the three components, the oxidation of FNR red byNADP + was faster than electron transfer from Fd red to the reductase [60]. Even in the absence of Fd, NADP + reduction by FNR red proceeds at % 500 s )1 , a rate compatible with catalysis [42,60]. Open question: the molecular bases of cooperativity As indicated previously, interactions of ferredoxin and NADP + with FNR exhibit reciprocal negative cooperati- vity, which is translated, paradoxically, into positive cooper- ativity at the kinetic level [60,61]. It was expected therefore that complex formation should lead to modifications in the structure of the active site of the flavoprotein. Conforma- tional movements resulting from nucleotide binding, how- ever, appear to be largely restricted to displacement of the carboxyl terminal tyrosine, as judged by the crystal structures of wild-type and mutant FNR in complex with NADP(H) and analogues [23,27,31]. We speculate that motion of the phenol ring of tyrosine is responsible for the decrease in FNR affinity for Fd, but the actual position of this residue when pushed away by the entering nicotinamide is unknown. In a tyrosine-to-tryptophan mutant of pea FNR that allows for % 40% of nicotinamide occupancy, the displaced indole ring failed to adopt a single ordered position [27]. Interaction of FNR and Fd, on the contrary, does lead to structural changes in the two electron carriers relative to the conformations of the free proteins [29]. On complex formation, the NADP(H) domain is displaced slightly as a single unit, and the side chain of Glu312 (numbering of spinach FNR) moves to hydrogen bonding distance of the hydroxyl group of Ser96 [29]. These two residues are highly conserved among reductases of different origins, and their charge, size and polarity are crucial to optimize the active site geometry for electron and hydride transfer [38,43, 45,49]. The protein–protein interaction also affects the microenvironments of the two prosthetic groups. The redox potentials (E m ) of Fd and FNR were shifted by )25 mV and +20 mV, respectively, facilitating electron transfer in the photosynthetic direction, namely, from Fd red to FNR ox [70]. It is not clear how these observed or putative structural changes in the active site region of FNR correlate with the induced-fit mechanism deduced from kinetic measurements. Dorowski et al. [28] have proposed that Fd binding might favour displacement of the carboxyl terminal tyrosine by nestling the phenol group into a hydrophobic pocket of the iron–sulphur protein. These authors advanced further and challenged the model of Fig. 2 by suggesting that Fd is the leader substrate that assists in NADP + binding. Although such a mechanism would be at odds with the reported decrease in NADP + affinity upon Fd attachment, the two models can be reconciled by assuming that the dinucleotide binds first in a nonproductive manner through its 2¢-P-AMP portion. Fd may then interact with the tyrosyl residue, favouring nicotinamide docking and establishing a loosely bound complex compatible with turnover [28]. It is clear that the carboxyl terminal tyrosine plays a pivotal role during FNR catalysis, but further research will be required to understand its actual function and importance. The backward reaction is a mechanistic puzzle Ferredoxin (or flavodoxin) reduction is the most widely distributed function of FNR-type proteins. Table 1 pro- vides a summary of metabolic routes that require such an activity from either FNR or adrenodoxin reductase. In cyanobacteria, a single FNR species functions as NADP + reductase in vegetative cells and as Fd reductase in heterocysts [8], whereas two distinct isoforms fulfil these roles in chloroplasts and nonphotosynthetic plastids of vascular plants. In the latter case, tissue specificity is determined at the transcriptional level by cis-acting regula- tory elements [71,72]. Interestingly, the redox potentials of these reductases and those of their corresponding ferredox- ins have been tuned by evolution to favour the physiological direction of electron transport [30]. However, the four proteins can be readily exchanged in vitro when assayed in a variety of reactions, indicating that the major force driving NADP + or Fd reduction in vivo would be the availability of substrates [30,73]. NADPH binding to oxidized FNR leads to rapid hydride exchange between the nucleotide and the reductase, result- ing in a succession of charge-transfer complexes involving flavin and nicotinamide (Eqn 7, species in brackets), whose formation can be followed by the appearance of long wavelength absorbance signals [61]. FNR ox þ NADPH !  FNR ox NADPH½ !  FNR red NADP þ ÂÃ ð7Þ The presence of various molecular species complicates the quantitative estimation of binding equilibria. Batie and Kamin [61] obtained an upper limit of about 2 l M for the K d of the spinach FNR ox ÆNADPH complex, indicating that 1908 N. Carrillo and E. A. Ceccarelli (Eur. J. Biochem. 270) Ó FEBS 2003 nucleotide binding to FNR ox is tighter in the reduced state. Complex formation is very rapid (k obs > 500 s )1 ), followed by slower hydride transfer to the flavin at 200 s )1 (Table 2). Electron transfer from FNR red to Fd ox istoofasttobe followed by stopped-flow techniques [43,44,52]. All the previous reactions proceed at velocities that are compatible with the steady-state rate of Fd reduction (k obs ¼ 200– 250 s )1 ), as measured by the cyt c reductase assay [30,34,35,43,44,52,57]. The reversible nature of the various steps involved in FNR-mediated NADP + reduction suggested that electron transfer from NADPH to Fd should proceed by a reversed version of the ordered pathway of Fig. 2. The forward and backward reactions are shown, in Cleland’s notation, in Fig. 6A,B, respectively. Assuming that v ¼ k )6 [FNR ox ] [NADPH] ) k 6 [FNR ox ÆNADPH + FNR red ÆNADP + ], then in the absence of products the velocity equation for Fd reduction (Fig. 6B) will be: Where K m , K d and V m have their conventional meanings and K m ¢ (Fd) represents the sum of the K m values for the successive interactions of the two molecules of Fd ox with FNR red and FNR sq (Fig. 6B). Equation 8 predicts that double reciprocal plots of v against the concentrations of any of the two substrates should yield straight lines intersecting in the fourth quadrant, as it occurs with the forward reaction. There is no simple way to measure Fd (or Fld) reduction, because the reduced acceptor is reoxidized by dissolved oxygen with k cat % 40 s )1 [16]. Therefore, kinetic evaluation of the Fd reductase activity requires measure- ments under strict anaerobiosis, and experiments of this kind have not yet been carried out with the plant-type flavopro- teins. As indicated before, cytochrome c is usually employed as a final acceptor that competes favourably with dioxygen for the spare electron of reduced Fd (Eqns 4 and 5). In vivo, the physiological acceptor enzymes (such as thioredoxin reductase, nitrate reductase or dihydroascorbate reductase) would play a similar role, preventing accumulation of the reduced ÔdoxinsÕ that, under the oxygen tensions prevailing in aerobic cells, could otherwise facilitate the propagation of superoxide radicals and other toxic oxygen derivatives. Wan and Jarrett [63] have measured the anaerobic oxidation of NADPH by an E. coli system made up of FNR and either ferredoxin or any of the two bacterial flavodoxins. This reductase species is remarkably slow, turning over at 0.15 s )1 for Fd and about 0.004 s )1 for the flavodoxins [63]. The rates of individual electron transfer steps are consistent with this slow pace of catalysis ([62,63], summarized in Table 2). Moreover, E. coli FNR mediates direct reduction of cytochrome c at % 5s )1 ,withthisrate being enhanced only about twofold by the addition of saturating flavodoxin [62]. Similar low activities have been obtained with the A. vinelandii [74] and Rhodobacter capsulatus (C. Bittel, N. Carrillo and N. Cortez, IBR, Rosario, Argentina, unpublished observations) reductases. The collected results indicate that these bacterial FNR forms display distinct catalytic features, and their comparison with the plant-type enzymes needs to be considered with caution. Surprisingly, when spinach Fd reduction was measured in vitro by the cyt c assay, parallel lines were obtained in 1/v vs. 1/[NADPH] plots, suggesting a two-step transfer Ôping- pongÕ mechanism without formation of a ternary complex [75]. The diaphorase activity (with various electron partners) also conforms to a double-displacement mechanism [75,76], although in this case the electronic route between the flavin Fig. 6. The forward and reverse reactions catalysed by ferredoxin-NADP(H) reductase. NADP + reduction (A) follows the compul- sory ordered pathway of Fig. 2, whereas two alternative mechanisms are proposed for electron transfer in the reverse direction: ordered (B), or two-step transfer (C). E, FNR ox ;F,FNR sq ;G,FNR red ;A,NADP + P, NADPH; B, Fd red ;Q,Fd ox . m ¼ NADPH½Fd ox ½V m K dðNADPHÞ K mðNADPHÞ þ K 0 mðFdÞ NADPH½þK mðNADPHÞ Fd ox ½þ½NADPH½Fd ox  ð8Þ Ó FEBS 2003 Ferredoxin-NADP(H) reductase catalytic mechanism (Eur. J. Biochem. 270) 1909 [...]... nez-Julvez, M., Medina, M & Gomez-Moreno, C (1999) Ferredoxin-NADP+ reductase uses the same site for the interaction with ferredoxin and avodoxin J Biol Inorg Chem 4, 568578 47 Hurley, J.K., Hazzard, J.T., Mart nez-Julvez, M., Medina, M., Gomez-Moreno, C & Tollin, G (1999) Electrostatic forces involved in orienting Anabaena ferredoxin during binding to Anabaena ferredoxin: NADP+ reductase: site-specic... Ceccarelli, E.A (1993) Probing the role of the carboxyl terminal region of plant ferredoxin-NADP+ reductase by site-directed mutagenesis and deletion analysis J Biol Chem 268, 1926719273 37 Aliverti, A., Corrado, M.E & Zanetti, G (1994) Involvement of lysine-88 of spinach ferredoxin-NADP+ reductase in the interaction with ferredoxin FEBS Lett 343, 247250 38 Aliverti, A., Bruns, C.M., Pandini, V.E., Karplus,... ferredoxin- NADP+ reductase Biochim Biophys Acta 1297, 200206 42 Arakaki, A.K., Ceccarelli, E.A & Carrillo, N (1997) Plant-type ferredoxin-NADP+ reductases: a basal structural framework and a multiplicity of functions FASEB J 11, 133140 43 Medina, M., Mart nez-Julvez, M., Hurley, J.K., Tollin, G & Gomez-Moreno, C (1998) Involvement of Glutamic 301 in the catalytic mechanism of ferredoxin-NADP+ reductase. .. glutamic acid 139 of Anabaena ferredoxin-NADP+ reductase in the interaction with substrates Eur J Biochem 269, 49384947 Faro, M., Hurley, J.K., Medina, M., Tollin, G & Gomez-Moreno, C (2002) Flavin photochemistry in the analysis of electron transfer reactions: role of charged and hydrophobic residues at the carboxyl terminus of ferredoxin-NADP+ reductase in the interaction with its substrates Bioelectrochemistry... T., Zanetti, G., Herrmann, R.G & ă Curti, B (1991) Probing the role of Lysine 116 and Lysine 224 in the spinach ferredoxin-NADP+ reductase by site-directed mutagenesis J Biol Chem 266, 1776017763 35 Aliverti, A., Piubelli, L., Zanetti, G., Lubberstedt, T., Herrmann, ă R.G & Curti, B (1993) The role of cysteine residues of spinach ferredoxin-NADP+ reductase as assessed by site-directed mutagenesis Biochemistry... Sanz-Aparicio, J., Tollin, G., Gomez-Moreno, C & Medina, M (1998) Role of Arg100 and Arg264 from Anabaena ferredoxinNADP+ reductase for optimal NADP+ binding and electron transfer Biochemistry 37, 1768017691 45 Aliverti, A., Deng, Z., Ravasi, D., Piubelli, L., Karplus, P.A & Zanetti, G (1998) Probing the function of the invariant glutamyl residue 312 in spinach ferredoxin-NADP+ reductase J Biol Chem... for an extended family of avoprotein reductases Comparison of phthalate dioxygenase reductase with ferredoxin reductase and ferredoxin Protein Sci 2, 21122133 22 Karplus, P.A & Bruns, C.M (1994) Structurefunction relations for ferredoxin reductase J Bioenerg Biomembr 26, 8999 23 Bruns, C.M & Karplus, P.A (1995) Rened crystal structure of spinach ferredoxin-NADP+ oxidoreductase at 1.7 A resolution: oxidized,... (1995) Involvement of serine 96 in the catalytic mechanism of ferredoxin-NADP+ reductase: structure-function relationship as studied by site-directed mutagenesis and X-ray crystallography Biochemistry 34, 83718379 39 Calcaterra, N.B., Pico, G.A., Orellano, E.G., Ottado, J., Carrillo, N & Ceccarelli, E.A (1995) Contribution of the FAD binding site residue tyrosine 308 to the stability of pea ferredoxin-NADP+. .. cDNA sequence of adrenodoxin reductase Identication of NADP+-binding sites in oxidoreductases Eur J Biochem 180, 479484 7 Bianchi, V., Haggard-Ljungquist, E., Pontis, E & Reichard, P (1995) Interruption of the ferredoxin (avodoxin) NADP+ oxidoreductase gene of Escherichia coli does not aect anaerobic growth but increases sensitivity to paraquat J Bacteriol 177, 45284531 8 Razquin, P., Fillat, M.F., Schmitz,... structures of the 60 kDa avoprotein monomer of the sulte reductase indicate a disordered avodoxin-like module J Mol Biol 299, 199212 27 Deng, Z., Aliverti, A., Zanetti, G., Arakaki, A.K., Ottado, J., Orellano, E.G., Calcaterra, N.B., Ceccarelli, E.A., Carrillo, N & Karplus, P.A (1999) A productive NADP+ binding mode of ferredoxin-NADP+ reductase revealed by protein engineering and crystallographic studies . FAD binding domain shown in blue, the NADP(H) binding domain in pink, and the FAD prosthetic group in yellow. (B) Detailed view of the isoalloxazine ring. lvez, M., Medina, M., Go ´ mez-Moreno, C. & Tollin, G. (1999) Electrostatic forces involved in orienting Anabaena ferredoxin during binding to Anabaena