HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Alagaratnam, S. Articles by Canters, G. W. Search for Related Content PubMed PubMed Citation Articles by Alagaratnam, S. Articles by Canters, G. W. Social Bookmarking                   What's this?

A crystallographic study of Cys69Ala flavodoxin II from Azotobacter vinelandii: Structural determinants of redox potential

¹ Leiden Institute of Chemistry, Leiden University, 2300 RA Leiden, The Netherlands
² Laboratory of Biophysical Chemistry, University of Groningen, 9747 AG Groningen, The Netherlands
³ Department of Biochemistry and Molecular Biology, University of Parma, Italy
⁴ Laboratory of Biochemistry, Department of Agrotechnology and Food Sciences, Wageningen University, 6703 HA Wageningen, The Netherlands

Reprint requests to: Gerard W. Canters, Leiden Institute of Chemistry, P.O. Box 9502, 2300 RA Leiden, The Netherlands; e-mail: canters{at}chem.leidenuniv.nl; fax: +31-71-527-4349.

(RECEIVED May 10, 2005; FINAL REVISION June 22, 2005; ACCEPTED June 23, 2005)

	Abstract

TOP Abstract Introduction Results and Discussion Materials and methods References

 
Flavodoxin II from Azotobacter vinelandii is a "long-chain" flavodoxin and has one of the lowest E₁ midpoint potentials found within the flavodoxin family. To better understand the relationship between structural features and redox potentials, the oxidized form of the C69A mutant of this flavodoxin was crystallized and its three-dimensional structure determined to a resolution of 2.25 Å by molecular replacement. Its overall fold is similar to that of other flavodoxins, with a central five-stranded parallel -sheet flanked on either side by -helices. An eight-residue insertion, compared with other long-chain flavodoxins, forms a short 3₁₀ helix preceding the start of the ₃ helix. The flavin mononucleotide (FMN) cofactor is flanked by a leucine on its re face instead of the more conserved tryptophan, resulting in a more solvent-accessible FMN binding site and stabilization of the hydroquinone (hq) state. In particular the absence of a hydrogen bond to the N5 atom of the oxidized FMN was identified, which destabilizes the ox form, as well as an exceptionally large patch of acidic residues in the vicinity of the FMN N1 atom, which destabilizes the hq form. It is also argued that the presence of a Gly at position 58 in the sequence stabilizes the semiquinone (sq) form, as a result, raising the E₂ value in particular.

Keywords: flavodoxin; FMN; hydrogen bonding; polarity; redox potentials

Abbreviations: FMN, flavin mononucleotide • RMS, root mean square • ox, oxidized • sq, semiquinone • hq, hydroquinone.

Article and publication are at http://www.proteinscience.org/cgi/doi/10.1110/ps.051582605.

	Introduction

TOP Abstract Introduction Results and Discussion Materials and methods References

 
The redox potential of an electron transfer protein is of prime importance in relation to its function, where the correlation between protein structure and redox potential helps explain how nature has adapted proteins to their specific functions. In the past, this has been studied for a range of flavodoxins, small (14–23 kDa) acidic / proteins that contain a single noncovalently bound flavin mononucleotide (FMN) cofactor, from different organisms. In vivo they act as remarkably versatile low potential one-electron donors in a range of reactions (Mayhew and Tollin 1992) in mainly prokaryotic organisms, including obligate and facultative anaerobes, microaerophiles, and photosynthetic cyanobacteria, as well as in both red and green eukaryotic algae. In the photosynthetic bacteria Anabaena, for example, it replaces ferredoxin as an electron shuttle from photosystem I to ferredoxin-NADP⁺ reductase under iron-deficient conditions (Sykes and Rogers 1984).

Flavodoxin II from Azotobacter vinelandii ATCC 478, the subject of this article, has been implicated in electron transfer to nitrogenase, where its synthesis was coinduced with nitrogenase upon introduction of A. vinelandii cells to nitrogen-fixing conditions after growth on NH₄Cl (Klugkist et al. 1986a). This flavodoxin is a so-called "long-chain" flavodoxin and has one of the lowest E₁ midpoint potentials found within the flavodoxin family (Steensma et al. 1996; see below).

Two subgroups have been identified within the flavodoxin family, the short-chain and the long-chain flavodoxins. The long-chain flavodoxins are up to 38 residues longer than the short-chain flavodoxins, mainly due to a long inserted loop of 22 residues in the final strand of the central -sheet, toward the end of the protein. Flavodoxin II from A. vinelandii contains this insertion and is therefore considered to be a member of this subgroup. The distinction between long- and short-chain flavodoxins appears to parallel a difference in redox potentials for these proteins. FMN, both when free in solution and bound to apoflavodoxin, can exist in three redox states, namely oxidized (ox), one-electron reduced semiquinone (sq), and two-electron reduced hydroquinone (hq). For FMN free in solution, the midpoint potential for the ox/sq redox couple, E₂, is lower than that of the sq/hq potential, E₁. However, complex formation with apoflavodoxin alters these potentials dramatically, inducing an inversion as well as a substantial separation between them. As a result, flavodoxin can transfer electrons in two discrete one-electron steps (Lostao et al. 1997).

Three-dimensional structures are known of oxidized flavodoxins from both subgroups, namely, from Clostridium beijerinckii (Burnett et al. 1974) and Desulfovibrio vulgaris (Watt et al. 1991) for the short-chain and from Anabaena variabilis (Rao et al. 1992), Chondrus crispus (Fukuyama et al. 1992), Escherichia coli (Hoover and Ludwig 1997), Helicobacter pylori (Freigang et al. 2002), Megasphaera elsdenii (van Mierlo et al. 1990), and Anacystis nidulans (Smith et al. 1983) for the long-chain flavodoxins. Additionally, structures of the sq and/or hq forms of the flavodoxins from A. nidulans, C. beijerinckii, and D. vulgaris are also available (Smith et al. 1977; Watt et al. 1991; Ludwig and Luschinsky 1992; Romero et al. 1996; Hoover et al. 1999). The combination of structural data with mutagenesis of identified key residues has resulted in a general picture where it appears that the separation of the midpoint potentials in flavodoxin is effected by the differential stabilization by the apoflavodoxin of the three oxidation states of FMN. In all three structures of flavodoxin semiquinones listed above, a striking difference in the flip of the 58–59 backbone peptide (A. vinelandii numbering) upon reduction to the sq form was observed (Watt et al. 1991; Ludwig and Luschinsky 1992). The hydrogen bond then formed between the carbonyl of residue 58 and N5 of the FMN stabilizes the neutral semiquinone in the protein and is purported to be the basis of the positive shift of the ox/sq potential compared with FMN free in solution (Smith et al. 1977; Ludwig et al. 1997; Hoover et al. 1999). The identity of one of the residues within this 58–59 peptide is also known to be a crucial determinant of E₂ (Ludwig et al. 1997; Chang and Swenson 1999). The exceptionally low sq/hq E₁ potentials in flavodoxins, on the other hand, are thought to be brought about by specific destabilizing electrostatic interactions between the anionic hq and the FMN binding site, which contains patches of uncompensated negatively charged residues (Zhou and Swenson 1996a,b). In addition, sandwiching of the FMN between two aromatic residues not only is unfavorable for the charged hq but also actively excludes solvent from the FMN binding pocket, thus maintaining an apolar environment (Swenson and Krey 1994; Lostao et al. 1997; McCarthy et al. 2002).

E₂ in particular appears to be higher in general for the short-chain compared with the long-chain flavodoxins (Fig. 1). The 22-residue insertion that differentiates the long- and short-chain subgroups does not, however, contact the FMN cofactor and, as such, is unlikely to be the source of the differences in potentials, as has previously been noted (Steensma et al. 1998). In fact, a recent study of shortened variants of the long-chain flavodoxin from Anabaena indicated that the loop was only indirectly responsible for stabilizing the protein complex with FMN and was more likely to be involved in the recognition of partner proteins (Lopez-Llano et al. 2004a, b).

View larger version (14K): [in this window] [in a new window]   Figure 1. The E1, sq/hq, and E2, q/sq, midpoint potentials for long-and short-chain flavodoxins from different sources at pH 7.0, in mV. From left to right: E. coli (Hoover and Ludwig 1997), A. nidulans (Hoover et al. 1999), A. chroococcum (Deistung and Thorneley 1986), A. vinelandii (Steensma et al. 1996), Anabaena 7120 (Paulsen et al. 1990), C. crispus (Sykes and Rogers 1984), D. vulgaris (Curley et al. 1991), C. beijerinckii (Mayhew 1971), M. elsdenii (Mayhew et al. 1969), and finally, FMN free in solution (Anderson 1983), with E1 and E2 inverted with respect to the flavodoxins. For the cases where the potentials were not measured at pH 7.0, a change of –59 mV/pH unit was applied to estimate E2 (E1 was assumed to be pH independent above pH 7.0).

Here we report the structure of the oxidized form of the Cys69Ala (C69A) mutant of flavodoxin II from A. vinelandii as solved by X-ray diffraction to a resolution of 2.25 Å using molecular replacement. While the basic fold of the flavodoxin resembles that of other known flavodoxins, several unique features stand out. An eight-residue loop purported by analogy to be involved in partner protein complex formation is identified at the surface of the protein. Focusing on the local structure around the FMN binding pocket, a striking difference was noted in the absence of a hydrogen bond to the N5 of FMN, contrary to all other structures of oxidized long-chain flavodoxins known. The presence of a leucine instead of a conserved tryptophan on the re face of the FMN results in a more polar FMN binding site, while a significantly larger cluster of negatively charged residues is found around the N1 of the FMN than in other flavodoxins. Both these factors are known determinants of E₁, the sq/hq potential. Despite the fact that only the E₁ redox couple is biologically relevant, close study of the structure of flavodoxins in the various redox states has been instrumental in furthering understanding of how variations in composition and structure modulate both E₁ and E₂ redox potentials of this protein family in particular, with implications for redox proteins in general. In this A. vinelandii flavodoxin, the structure suggests destabilization of the ox and hq forms of the protein, while stabilization of the sq form moves the ox/sq E₂ redox potential to a relatively high value. The discussion affords new insights into how variations in protein composition and structure can modulate their redox potentials.

	Results and Discussion

TOP Abstract Introduction Results and Discussion Materials and methods References

 
The crystal structure determination of A. vinelandii C69A flavodoxin II
The crystal structure of A. vinelandii C69A flavodoxin II was solved at 2.25 Å resolution using molecular replacement with a homology model based on the Anabaena flavodoxin structure (Rao et al. 1992; Burkhart 1995). The protein crystallized with two monomeric molecules in the asymmetric unit, called A and B, which are related by a 133° rotation. The refinement statistics and contents of the final model are given in Table 1. Except for some solvent-exposed side chains, the molecules are clearly defined in density. Superposition of the C atoms of the two molecules gives a root mean square (RMS) difference of 0.7 Å. Excluding residues that differ by more than 3 (residues 1, 59, 178, and 179), the RMS difference decreases to 0.38 Å. Taking into account only the nonglycine and nonproline residues and discounting the first and last residues of each molecule, 268 residues (88.2%) are in the most favored region of the Ramachandran plot, 36 residues (11.8%) are in the additional allowed regions, and none of the residues are in the generously allowed regions or in the disallowed regions.

View larger version (66K): [in this window] [in a new window]   Figure 2. The three-dimensional fold of A. vinelandii flavodoxin II, with -helices shown in red; -sheets, in yellow; and random coil, in green. The inserted loop of residues 61–71 is colored cyan, and the mutation Cys69Ala is shown in red sticks in the middle of that loop. The FMN cofactor is also represented in stick form at the top of the protein.

View this table: [in this window] [in a new window]   Table 2. RMS differences in angstrom between superpositions of the -carbon atoms of all long-chain flavodoxin structures

View larger version (27K): [in this window] [in a new window]   Figure 3. Structure-based sequence alignment using SEQUOIA v0.9.7 (Bruns et al. 1999) of the oxidized long-chain flavodoxins from A. vinelandii, Anabaena 7120 (PDB entry 1RCF [PDB] ; Burkhart 1995), H. pylori (PDB entry 1FUE [PDB] ; Freigang et al. 2002), and A. nidulans (PDB entry 1CZU [PDB] ; Drennan et al. 1999) and the oxidized short-chain flavodoxins from D. vulgaris (PDB entry 1J8Q [PDB] ; Artali et al. 2002) and C. beijerinckii (PDB entry 5NLL [PDB] ; Ludwig et al. 1997). The 22-residue loop that differentiates short- and long-chain flavodoxins is clearly visible inserted between strands 5a and 5b. The primary sequence of A. chroococcum flavodoxin was aligned with that of A. vinelandii using CLUSTALW version 1.81 (Thompson et al. 1994). The main secondary structure elements of the A. vinelandii protein were defined from the PDB coordinates using the DSSP program (Kabsch and Sander 1983) and are indicated above the amino acid sequence alignment.

A single phosphate group is present in the structure as part of the FMN cofactor. The gene for flavodoxin II used for this study is identical to that cloned from A. vinelandii strain OP Berkeley (Bennett et al. 1988), that has also been expressed heterologously in E. coli (Taylor et al. 1990). This crystal structure confirms an earlier notion that this flavodoxin form does not contain a covalently bound phosphate (Klugkist et al. 1986b), as was also found for the recombinant flavodoxin from A. vinelandii strain OP Berkeley (Taylor et al. 1990).

A total of seven solvent molecules were found buried in the interior of the protein, occupying comparable positions in both molecules in the asymmetric unit. The positions of these water molecules with respect to the three-dimensional fold of the flavodoxin molecule are illustrated in Figure 4A. Two of them, colored green, were found to interact with residues at the top ends of strands ₃ and ₄, and ₄ and ₅. Two others, colored light blue, bridge the loop connecting the end of ₂ to the start of ₃, and to the ¹⁵²D-L-D-N¹⁵⁵ loop between the end of ₅ and the start of ₅, respectively. A structurally conserved solvent molecule was also found between Leu93 on ₄ and the end of _5a, just at the start of the inserted loop common to long-chain flavodoxins; it is visible as the dark sphere toward the back of the molecule in Figure 4A. Finally, two more water molecules, colored red, were identified in the cavity adjacent to the more buried re face of the FMN. These latter two will be discussed in more detail later in the article. However, all mentioned solvent molecules appear to play a structural role in the packing of the protein. This is confirmed by the fact that they mostly occupy positions similar to those of water molecules found in other long-chain flavodoxins (Burkhart 1995; Hoover and Ludwig 1997; Drennan et al. 1999).

View larger version (27K): [in this window] [in a new window]   Figure 4. (A) Stereo views showing the positions of the structured solvent molecules, represented as spheres, with respect to the whole flavodoxin molecule, where the two light blue waters bridge -helices and -strands while the two green waters bridge the ends of two -strands. A single black water molecule is found toward the back of the flavodoxin, at the start of the 22-residue insertion in 5 strand of the sheet. (B) Zoom in on the FMN binding site flanked by Tyr102 and Leu57 at the top of the molecule in A where two red waters can be seen to lead from the cavity on the re face of the FMN created by the smaller leucine 57 side chain to the surface of the protein.

The different parts of the FMN cofactor associate with the flavodoxin polypeptide chain through a variety of interactions, including hydrogen bonds and hydrophobic stacking. Three main regions of apoflavodoxin implicated in the noncovalent binding of FMN in other flavodoxins are also closely associated with the flavin in A. vinelandii flavodoxin, as detailed below.

First, the extended phosphoribityl side chain acts to anchor the cofactor to the protein via an extensive network of hydrogen bonds, as summarized in Table 3 and shown in Figure 5A. The hydroxyl groups of the ribityl chain as well as the phosphate oxygen atoms form hydrogen bonds mainly to backbone amide and carbonyl groups of the protein and to two solvent molecules. In particular, numerous interactions are present between the phosphate oxygen atoms and backbone amides as well as side-chain hydroxyl groups of the ⁹S-N-T-G-K-T¹⁴ loop, as deduced from the H-bonding distances between the atoms. This loop, crucially requiring Gly at position 12, has been identified as a consensus sequence for phosphate binding in flavodoxins (Ludwig and Luschinsky 1992). It has a conformation very similar to that observed in other flavodoxins.

View larger version (29K): [in this window] [in a new window]   Figure 5. The FMN binding site in oxidized A. vinelandii flavodoxin II showing the hydrogen bonding partners from both main-chain and side-chain atoms (denoted by atom name, followed by the residue number) of the flavodoxin as well as waters (denoted W) to the phosphoribityl side chain of the FMN (including the phosphate-binding loop consisting of residues 9–14 including the crucial Gly12) (A) and the isoalloxazine part of the FMN (B).

View larger version (27K): [in this window] [in a new window]   Figure 6. Conformation of the four-residue turn in the 50s loop that contacts the FMN, in the same orientation from Anabaena (A), C. beijerinckii (B), and A. vinelandii (C) flavodoxin. The residue numbers are shown next to the backbone amides of the corresponding amino acids, where the backbone amide of residue 59 in A can clearly be seen to point toward the N5 of FMN, while in B and C the corresponding amide groups of residues 58 and 59, respectively, have geometries that are unsuitable for hydrogen bonding to the N5.

Distinctive features of FMN binding in A. vinelandii flavodoxin and consequences for its midpoint potentials
H-bonding to N5 of FMN destabilizes the ox form
The hydrogen bonding state of the N5 of FMN in the oxidized flavodoxin is also known to affect the E₂ redox potential by stabilizing the oxidized form of the protein (Rao et al. 1992; Hoover et al. 1999). In all structures of oxidized long-chain flavodoxins known to date, a hydrogen bond is found from a residue within the 50s loop to the unprotonated FMN N5. In Anabaena (Fig. 6A), A. nidulans, and E. coli, the donor is the backbone amide of residue 59, while in C. crispus it is the side-chain hydroxyl group of the adjacent Thr58. Although these hydrogen- bond donors are at between 3.1 Å and 3.6 Å from N5 in the different flavodoxins, their geometry is favorable, pointing directly toward the N5 (Burkhart 1995; Hoover and Ludwig 1997; Drennan et al. 1999). Additionally, the apolar environment of the FMN binding site may increase the strength of the interaction (Rao et al. 1992), and more compellingly, direct NMR evidence for hydrogen bonding to N5 exists (Stockman et al. 1988; Chang and Swenson 1999). Although the hydrogen bond is broken upon reduction of the flavodoxin to the semiquinone, N5 subsequently makes a hydrogen bond with the carbonyl oxygen atom as described above, which requires the backbone peptide plane to flip (see above). However, in the oxidized A. vinelandii flavodoxin, no candidate for a hydrogen-bonding interaction with N5 could be found (Fig. 6C). The closest backbone amide group is that of Glu59, at a distance of 4.0 Å from N5, but it is in an unfavorable position for hydrogen bonding. In this sense, the oxidized A. vinelandii protein resembles the short-chain flavodoxins from D. vulgaris (Romero et al. 1996) and C. beijerinckii (Fig. 6B) (Chang and Swenson 1999) more closely, where no potential hydrogen bond donor is near the N5 atom either. It is likely that the absence of a hydrogen bond to N5 in the oxidized form of the protein destabilizes the ox form of the FMN, making reduction comparatively easier and yielding a higher E₂ value.

Gly58 stabilizes the sq form
From sequence alignments, the residue at position 58 is most commonly a glycine in short-chain and an asparagine in long-chain flavodoxins. Uncommonly for a long-chain flavodoxin, the one from A. vinelandii has a glycine at this position. While the glycine can optimally accommodate the O-up conformation at that position of the type II' turn found in flavodoxin semiquinones (Hutchinson and Thornton 1994), the C of the asparagine found in the long-chain A. nidulans flavodoxin makes very close contact with the following amide group, raising the energy of the O-up conformation (Hoover et al. 1999). That this residue affects the stabilization of the semiquinone bound to flavodoxin and thus the redox potentials was confirmed by the Asn58Gly mutant of the long-chain A. nidulans flavodoxin, which had an ox/sq potential 46 mV more positive than that of the wild type, while its sq/hq potential was lowered by 26 mV (Hoover et al. 1999). On the other hand, mutants of the short-chain C. beijerinckii flavodoxin where the homologous Gly57 residue was replaced by alanine, asparagine, aspartate, and proline all had ox/sq potentials that were reduced by ~60 mV, while the Gly57Thr mutant with its bulkier -branched side chain had an E₂ that was even lower, by 180 mV (Ludwig et al. 1997). Direct measurement of the strength of the N5H···O57 hydrogen bond in this series of mutants by NMR definitively proved this effect, with a strong correlation between the temperature coefficients of the N5H and both E₂ and binding energy of the FMN semiquinone to the flavodoxin (Chang and Swenson 1999). Thus it is conceivable that the atypical glycine at this position in the flavodoxin from A. vinelandii may be a factor in shifting the ox/sq potential to a more positive value. The analogous residue in the H. pylori flavodoxin is also a glycine; however, its redox potentials are not known. The long-chain flavodoxin from E. coli may provide further support for the proposed correlation between E₂ and the nature of the residue at position 58: The ox/sq potential for this flavodoxin is among the lowest of all those known, where the protein has a bulky tyrosine residue here (Hoover and Ludwig 1997).

Additional charges close to the N1 of FMN further destabilize the hq form
Up to 11 acidic amino acids were found within 15 Å of N1 of FMN in the A. vinelandii flavodoxin, namely, the aspartates 68, 98, 108, 152, and 154 and the glutamates 59, 61, 104, 134, 136, and 139. Several of these residues are situated in the 100s loop, including the closest charged group at a distance of 3.0 Å, the carboxylic moiety of Asp98. Conversely, only three basic residues are present, and then only at the very periphery of the 15 Å radius from the FMN N1, with the positively charged side chains solvent-exposed at the surface of the protein instead of pointing toward the FMN binding pocket. This acidic character of the binding pocket around the N1 is known to be a major determinant of the sq/hq potential E₁. The FMN hydroquinone in flavodoxins from both M. elsdenii (Franken et al. 1984) and D. vulgaris (Vervoort et al. 1986) have been found to exist as anions, while evidence for sterical hindrance preventing protonation at N1 was found for the C. beijerinckii flavodoxin (Ludwig et al. 1990). The negative charges of the side chains destabilize the anionic hydroquinone, creating an unfavorable electrostatic environment that makes the reduction of the semiquinone more difficult. A patch of uncompensated negatively charged residues is commonly found grouped around the N1 of FMN in flavodoxins, and totals six to seven residues in both the D. vulgaris (Zhou and Swenson 1996b) and A. nidulans (Hoover et al. 1999) flavodoxins. Zhou and Swenson (1995) systematically mutated all six acidic residues in the D. vulgaris flavodoxin, with the effect that E₁ was raised by ~15 mV on average per neutralized amino acid. If by analogy the hq of the A. vinelandii flavodoxin is also negatively charged (Stockman et al. 1988), by this gauge, the five additional negative charges observed in this protein could potentially result in a negative shift in E₁ in the order of 75 mV.

Leu at the re face of FMN results in a more polar FMN binding site and stabilizes the hq form
While in most flavodoxins the FMN is stacked against a tyrosine on its si, or outer, face, more variability has been found in the amino acid present on its re, or inner, face. This is most commonly a tryptophan; however, nonaromatic residues have also been found. In A. vinelandii flavodoxin, a leucine residue is present. As a result, an open cavity is created at the re face of the cofactor, making it much more accessible to solvent. In fact, in both molecules of the asymmetric unit, two water molecules with low B-factors occupy this cavity in identical positions. They connect the innermost ring of the FMN to the surface of the flavodoxin (Fig. 4B). The FMN in the A. vinelandii flavodoxin is the most solvent-exposed one in the long-chain flavodoxins of known structure, where values as calculated by NACCESS (S.J. Hubbard and J.M. Thornton, University College London) for the surface accessibility for FMN in the five long-chain flavodoxins of known structure and potentials are as follows: A. vinelandii (averaged for molecule A and B), 114.8 Å; Anabaena, 93.9 Å; A. nidulans, 100.1 Å; C. crispus, 98.7 Å; and E. coli, 99.7 Å. The presence of smaller and/or polar residues stabilizes the anionic FMN hq in the protein, thus shifting the sq/hq potential to more positive levels. This was corroborated by the mutation of the analogous Trp60 in the D. vulgaris flavodoxin to alanine, which raised E₁ by 83 mV (Mayhew et al. 1996), while the Trp57Leu Anabaena flavodoxin mutant had an E₁ that was higher by 30 mV (Lostao et al. 1997). A methionine flanks the re face of the C. beijerinckii flavodoxin FMN, and a similar FMN-to- surface water channel is present. This flavodoxin has one of the highest E₁ potentials known (Ludwig et al. 1990). Intriguingly, however, despite the small leucine residue and the presence of two water molecules within close proximity of the FMN, at –458 mV (Steensma et al. 1996) the A. vinelandii flavodoxin appears to have one of the lowest E₁ values of all flavodoxins (Fig. 1). It is possible that the additional negative charges clustered around the N1 compensate for the increased polarity of the FMN binding site in this protein.

Comparisons between the two Azotobacter flavodoxins
The A. chroococcum flavodoxin has a sq/hq potential, which, at –522 mV, is even more negative than that of the A. vinelandii flavodoxin. Aligning the two protein sequences shows that the A. chroococcum flavodoxin also contains the additional negatively charged residues, as well as a phenylalanine instead of a tyrosine at position 106 (Fig. 3). In the A. vinelandii structure, this residue was found within 5 Å of the most buried ring of the FMN, in an almost coplanar conformation. No other differences could be mapped to locations close to the FMN that could account for the extremely low E₁ of the A. chroococcum flavodoxin, which reinforces the overriding influence of the acidic nature of the FMN binding site.

We report here the first crystal structure of flavodoxin II from A. vinelandii, of the monomeric C69A mutant. In general, the structure closely resembles that of other long-chain flavodoxins, with the exception of an eight-residue insertion between ₃ and ₃, which is found as a short 3₁₀ helix followed by a turn at the surface of the protein. The C69A mutation is solvent-exposed, as was previously proposed due to the propensity of the wild-type protein to form disulphide-bridged dimers.

On the whole, the interactions between the FMN cofactor and the apoflavodoxin are similar to those previously identified in other long- and short-chain flavodoxins, although variations in the amino acid sequence have resulted in several novel hydrogen bonds. More noteworthy, however, is the absence of a hydrogen bond to the N5 atom of the isoalloxazine ring, observed in all structures of oxidized long-chain flavodoxins determined to date. This is expected to destabilize the oxidized FMN and may in part explain the relatively high ox/sq E₂ potential of this flavodoxin. In addition, a glycine is present in the two-residue peptide that flips upon reduction of the protein, instead of the asparagine typical of long-chain flavodoxins. In both these characteristics, the A. vinelandii flavodoxin resembles the short-chain flavodoxins from C. beijerinckii and D. vulgaris more closely, where this is mirrored in their comparable E₂ values.

A cavity is found on the re face of the FMN, due to a leucine residue flanking the cofactor instead of the more common tryptophan. A short channel consisting of two water molecules leads from this cavity to the surface of the protein. As a result, the FMN in this flavodoxin is much more solvent-exposed, an effect which is known to stabilize the anionic hq form and raise the sq/hq E₁ potential. On the other hand, destabilization of the hq form is promoted by an exceptionally large cluster of 11 uncompensated negatively charged residues around the anionic FMN N1 atom, almost twice the number found in both short- and long-chain flavodoxins. The more polar FMN binding site counteracts the destabilization of the hq, and the net effect leads to the E₁ value of –458mV observed. However, differential stabilization of the sq form by the formation of a hydrogen bond to the N5 effectively raises E₂ to a value closer to that of short-chain flavodoxins.

In summary, this structure of the oxidized flavodoxin II of A. vinelandii is an important starting point for a deeper understanding of the factors that modulate the redox potential of flavodoxins in general, and this protein in particular. Comparing this structure with other long- and short-chain flavodoxins has allowed us to identify particular features (a more polar FMN binding site due to the Leu–Trp replacement, a large cluster of acidic residues close to the N1) and residues (Gly58 that does not hydrogen bond to N5 in the oxidized state) that may be key to fine-tuning the E₁ and E₂ potentials. Further mutational, biochemical, and structural studies involving these residues will be required to correlate the observed structural and functional differences. Modulations by the interactions of these residues with the FMN result in what may more usefully be considered as a continuous range of E₁ and E₂ values for the different flavodoxins, as opposed to stricter definitions of class and redox potential based on chain length. Finally, NMR techniques can identify residues important for complex formation between flavodoxin and its redox partners. Mapping these onto the three-dimensional structure of the protein may give new insights not only into the nature of the complex but also into the particular pathways that may be involved in the electron transfer reaction.

	Materials and methods

TOP Abstract Introduction Results and Discussion Materials and methods References

 
Expression and purification of recombinant A. vinelandii flavodoxin II
C69A flavodoxin II from A. vinelandii ATCC 478 was expressed heterologously in E. coli strain TG2 from a pUC19 plasmid containing the mutated Cys69Ala flavodoxin gene (van Mierlo et al. 1998). The large-scale culture was induced with 1 mM IPTG upon inoculation and grown for 24 h before harvesting. The protein was purified according to described protocols and was considered to be pure when the absorption peaks showed a ratio A₂₇₄/A₄₅₂ of 4.7 (Tollin and Edmondson 1980). The molar absorption coefficient at 452 nm of 11.3 mM^–1 cm^–1 (Klugkist et al. 1986b) was used to determine the concentration of flavodoxin.

Crystallization and X-ray data collection
Crystals were grown at room temperature using the sitting drop vapor diffusion method. The reservoir contained 2.8 M ammonium sulphate and 0.1 M Tris buffer (pH 7.0). The drops were made by mixing 2 µL of protein solution (15 mg/mL in 0.1 M Tris buffer [pH 8.0]) and 2 µL of the reservoir solution. Plates and clusters of plate-like crystals grew over a period of between 1 and 2 wk. The crystal used for data collection was transferred to a 4:1 solution of reservoir solution and water. The crystal was subsequently dipped in Al’s oil (1:1 paraffin oil and silicon oil) and flash frozen in a stream of evaporating nitrogen gas of 100 K. The crystal belongs to space group P2₁2₁2₁ with cell dimensions a=39.01 Å, b=70.54 Å, and c=132.64 Å. It diffracted to 2.25 Å resolution. Data were collected in house with a MAR CCD detector with CuK X-rays from a NONIUS FR591 rotating anode generator (Table 1). Data were processed with DENZO and SCALEPACK (Otwinowski and Minor 1997). Reduction to structure factor amplitudes was performed with TRUNCATE (French 1978).

Model building and crystallographic refinement
The structure of flavodoxin II C69A was solved by molecular replacement. The search model was based on the structure of the 47% sequence identical Anabaena flavodoxin (PDB code 1RCF [PDB] ; Burkhart 1995) using the 3D-PSSM server (Kelley et al. 2000). This resulted in a model in which the side chains had been modeled by placement with a rotamer library (Bower et al. 1997). Two molecules were located in the asymmetric unit using the molecular replacement program EPMR (Kissinger et al. 1999), yielding a solvent content of 50%. Using data between 10 Å and 4 Å resolution, the correlation coefficient and R-factor of EPMR became 0.227 and 57.0% for the first molecule only and improved to 0.387 and 50.4% when both molecules in the asymmetric unit were included. The two molecules were rigid body refined with CNS (Brunger et al. 1998), resulting in an R-factor of 48% (10–3 Å). After a first round of simulated annealing (30–2.25 Å, R-factor 37.0%, R-free 42.7%), the electron density map was clear and showed the FMN group that had been omitted from the molecular replacement model. Side chains pointing in the wrong direction and main-chain at insertion points were manually rebuilt using the programs O (Jones et al. 1991) and QUANTA (Accelrys). The improved model was further refined with REFMAC5 (Murshudov 1997). No NCS constraints were used. Water molecules were added, and electron density for six sulphate ions was also present. The final R-factor is 22.3%; the free R-factor, 27.8%. The refinement statistics and information about the final model are listed in Table 1. The geometry of the final model was analyzed using PROCHECK (Laskowski et al. 1993). Figures were prepared using PyMOL (http://pymol.sourceforge.net). The coordinates have been deposited in the PDB (Berman et al. 2000) under accession number 1YOB.

Calculation of surface accessibility of FMN in flavodoxins
The program NACCESS (S.J. Hubbard and J.M. Thornton, University College London) was used to calculate the surface accessibility of the FMN cofactor in Ų for a number of both long- and short-chain flavodoxins, using the PDB files of each of the proteins as input and the default probe size of 1.4 Å.

	References

TOP Abstract Introduction Results and Discussion Materials and methods References

 
Anderson, R.F. 1983. Energetics of the one-electron reduction steps of riboflavin, FMN and FAD to their fully reduced forms. Biochim. Biophys. Acta 722: 158–162.[Medline]

Artali, R., Bombieri, G., Meneghetti, F., Gilardi, G., Sadeghi, S.J., Cavazzini, D., and Rossi, G.L. 2002. Comparison of the refined crystal structures of wild-type (1.34 Å) flavodoxin from Desulfovibrio vulgaris and the S35C mutant (1.44 Å) at 100 K. Acta Crystallogr. D. Biol. Crystallogr. 58: 1787–1792.[CrossRef][Medline]

Barman, B.G. and Tollin, G. 1972. Flavine-protein interactions in flavoenzymes: Thermodynamics and kinetics of reduction of Azotobacter flavodoxin. Biochemistry 11: 4755–4759.[CrossRef][Medline]

Bennett, L.T., Jacobson, M.R., and Dean, D.R. 1988. Isolation, sequencing, and mutagenesis of the nifF gene encoding flavodoxin from Azotobacter vinelandii. J. Biol. Chem. 263: 1364–1369.[Abstract/Free Full Text]

Berman, H.M., Westbrook, J., Feng, Z., Gilliland, G., Bhat, T.N., Weissig, H., Shindyalov, I.N., and Bourne, P.E. 2000. The Protein Data Bank. Nucleic Acids Res. 28: 235–242.[Abstract/Free Full Text]

Bollen, Y.J., Sanchez, I.E., and van Mierlo, C.P. 2004. Formation of on-and off-pathway intermediates in the folding kinetics of Azotobacter vinelandii apoflavodoxin. Biochemistry 43: 10475–10489.[CrossRef][Medline]

Bower, M.J., Cohen, F.E., and Dunbrack Jr., R.L. 1997. Prediction of protein side-chain rotamers from a backbone-dependent rotamer library: A new homology modeling tool. J. Mol. Biol. 267: 1268–1282.[CrossRef][Medline]

Brunger, A.T., Adams, P.D., Clore, G.M., DeLano, W.L., Gros, P., Grosse-Kunstleve, R.W., Jiang, J.S., Kuszewski, J., Nilges, M., Pannu, N.S., et al. 1998. Crystallography & NMR system: A new software suite for macromolecular structure determination. Acta Crystallogr. D. Biol. Crystallogr. 54: 905–921.[CrossRef][Medline]

Bruns, C.M., Hubatsch, I., Ridderstrom, M., Mannervik, B., and Tainer, J.A. 1999. Human glutathione transferase A4–4 crystal structures and mutagenesis reveal the basis of high catalytic efficiency with toxic lipid peroxidation products. J. Mol. Biol. 288: 427–439.[CrossRef][Medline]

Burkhart, B.M. 1995. Structure of the trigonal form of recombinant oxidized flavodoxin from Anabaena 7120 at 1.40 Å resolution. Acta Crystallogr. D. Biol. Crystallogr. 51: 318–330.[CrossRef][Medline]

Burnett, R.M., Darling, G.D., Kendall, D.S., LeQuesne, M.E., Mayhew, S.G., Smith, W.W., and Ludwig, M.L. 1974. The structure of the oxidized form of clostridial flavodoxin at 1.9- Å resolution. J. Biol. Chem 249: 4383–4392.[Abstract/Free Full Text]

Chang, F.C. and Swenson, R.P. 1999. The midpoint potentials for the oxidized-semiquinone couple for Gly57 mutants of the Clostridium beijerinckii flavodoxin correlate with changes in the hydrogen-bonding interaction with the proton on N(5) of the reduced flavin mononucleotide cofactor as measured by NMR chemical shift temperature dependencies. Biochemistry 38: 7168–7176.[CrossRef][Medline]

Cheddar, G., Meyer, T.E., Cusanovich, M.A., Stout, C.D., and Tollin, G. 1986. Electron-transfer reactions between flavodoxin semiquinone and c-type cytochromes: Comparisons between various flavodoxins. Biochemistry 25: 6502–6507.[CrossRef][Medline]

Curley, G.P., Carr, M.C., Mayhew, S.G., and Voordouw, G. 1991. Redox and flavin-binding properties of recombinant flavodoxin from Desulfovibrio vulgaris (Hildenborough). Eur. J. Biochem. 202: 1091–1100.[Medline]

De Francesco, R., Tollin, G., and Edmondson, D.E. 1987. Influence of 8 -imidazole substitution of the FMN cofactor on the rate of electron transfer from the neutral semiquinones of two flavodoxins to cytochrome c. Biochemistry 26: 5036–5042.[CrossRef][Medline]

Deistung, J. and Thorneley, R.N.F. 1986. Electron transfer to nitrogenase: Characterization of flavodoxin from Azotobacter chroococcum and comparison of its redox potentials with those of flavodoxins from Azotobacter vinelandii and Klebsiella pneumoniae (nifF gene product). Biochem. J. 239: 69–75.[Medline]

Drennan, C.L., Pattridge, K.A., Weber, C.H., Metzger, A.L., Hoover, D.M., and Ludwig, M.L. 1999. Refined structures of oxidized flavodoxin from Anacystis nidulans. J. Mol. Biol. 294: 711–724.[CrossRef][Medline]

Drummond, M.H. 1986. Structure predictions and surface charge of nitrogenase flavodoxins from Klebsiella pneumoniae and Azotobacter vinelandii. Eur. J. Biochem. 159: 549–553.[Medline]

Duyvis, M.G., Wassink, H., and Haaker, H. 1998. Nitrogenase of Azotobacter vinelandii: Kinetic analysis of the Fe protein redox cycle. Biochemistry 37: 17345–17354.[CrossRef][Medline]

Franken, H.D., Ruterjans, H., and Müller, F. 1984. Nuclear-magnetic-resonance investigation of ¹⁵N-labeled flavins, free and bound to Megasphaera elsdenii apoflavodoxin. Eur. J. Biochem. 138: 481–489.[Medline]

Freigang, J., Diederichs, K., Schafer, K.P., Welte, W., and Paul, R. 2002. Crystal structure of oxidized flavodoxin, an essential protein in Helicobacter pylori. Protein Sci. 11: 253–261.[Abstract/Free Full Text]

French, S.W.K. 1978. On the treatment of negative intensity observations. Acta Crystallogr. A 34: 517–524.[CrossRef]

Fukuyama, K., Matsubara, H., and Rogers, L.J. 1992. Crystal structure of oxidized flavodoxin from a red alga Chondrus crispus refined at 1.8 Å resolution: Description of the flavin mononucleotide binding site. J. Mol. Biol. 225: 775–789.[CrossRef][Medline]

Hoover, D.M. and Ludwig, M.L. 1997. A flavodoxin that is required for enzyme activation: The structure of oxidized flavodoxin from Escherichia coli at 1.8 Å resolution. Protein Sci. 6: 2525–2537.[Abstract]

Hoover, D.M., Drennan, C.L., Metzger, A.L., Osborne, C., Weber, C.H., Pattridge, K.A., and Ludwig, M.L. 1999. Comparisons of wild-type and mutant flavodoxins from Anacystis nidulans. Structural determinants of the redox potentials. J. Mol. Biol. 294: 725–743.[CrossRef][Medline]

Hutchinson, E.G. and Thornton, J.M. 1994. A revised set of potentials for -turn formation in proteins. Protein Sci. 3: 2207–2216.[Abstract]

Jones, T.A., Zou, J.Y., Cowan, S.W., and Kjeldgaard, M. 1991. Improved methods for building protein models in electron density maps and the location of errors in these models. Acta Crystallogr. A 47: 110–119.

Kabsch, W. and Sander, C. 1983. Dictionary of Protein Secondary Structure: Pattern-recognition of hydrogen-bonded and geometrical features. Biopolymers 22: 2577–2637.[CrossRef][Medline]

Kelley, L.A., MacCallum, R.M., and Sternberg, M.J. 2000. Enhanced genome annotation using structural profiles in the program 3D-PSSM. J. Mol. Biol. 299: 499–520.[Medline]

Kissinger, C.R., Gehlhaar, D.K., and Fogel, D.B. 1999. Rapid automated molecular replacement by evolutionary search. Acta Crystallogr. D. Biol. Crystallogr. 55: 484–491.[CrossRef][Medline]

Klugkist, J., Haaker, H., and Veeger, C. 1986a. Studies on the mechanism of electron transport to nitrogenase in Azotobacter vinelandii. Eur. J. Biochem. 155: 41–46.[Medline]

Klugkist, J., Voorberg, J., Haaker, H., and Veeger, C. 1986b. Characterization of three different flavodoxins from Azotobacter vinelandii. Eur. J. Biochem. 155: 33–40.[Medline]

Laskowski, R.A., Moss, D.S., and Thornton, J.M. 1993. Main-chain bond lengths and bond angles in protein structures. J.Mol. Biol. 231: 1049–1067.[CrossRef][Medline]

Lopez-Llano, J., Maldonado, S., Bueno, M., Lostao, A., Angeles-Jimenez, M., Lillo, M.P., and Sancho, J. 2004a. The long and short flavodoxin: I. The role of the differentiating loop in apoflavodoxin structure and FMN binding. J. Biol. Chem. 279: 47177–47183.[Abstract/Free Full Text]

Lopez-Llano, J., Maldonado, S., Jain, S., Lostao, A., Godoy-Ruiz, R., Sanchez-Ruiz, J.M., Cortijo, M., Fernandez-Recio, J., and Sancho, J. 2004b. The long and short flavodoxins: II. The role of the differentiating loop in apoflavodoxin stability and folding mechanism. J. Biol. Chem. 279: 47184–47191.[Abstract/Free Full Text]

Lostao, A., Gómez-Moreno, C., Mayhew, S.G., and Sancho, J. 1997. Differential stabilization of the three FMN redox forms by tyrosine 94 and tryptophan 57 in flavodoxin from Anabaena and its influence on the redox potentials. Biochemistry 36: 14334–14344.[CrossRef][Medline]

Lostao, A., El Harrous, M., Daoudi, F., Romero, A., Parody-Morreale, A., and Sancho, J. 2000. Dissecting the energetics of the apoflavodoxin-FMN complex. J. Biol. Chem. 275: 9518–9526.[Abstract/Free Full Text]

Ludwig, M.L. and Luschinsky, C.L. 1992. Structure and redox properties of clostridial flavodoxin. In Chemistry and biochemistry of flavoenzymes (ed. F. Müller), Vol. 3, pp. 427–466. CRC Press, Boca Raton, FL.

Ludwig, M.L., Schopfer, L.M., Metzger, A.L., Pattridge, K.A., and Massey, V. 1990. Structure and oxidation-reduction behavior of 1-deaza-FMN flavodoxins: Modulation of redox potentials in flavodoxins. Biochemistry 29: 10364–10375.[CrossRef][Medline]

Ludwig, M.L., Pattridge, K.A., Metzger, A.L., Dixon, M.M., Eren, M., Feng, Y., and Swenson, R.P. 1997. Control of oxidation-reduction potentials in flavodoxin from Clostridium beijerinckii: The role of conformation changes. Biochemistry 36: 1259–1280.[CrossRef][Medline]

Mayhew, S.G. 1971. Studies on flavin binding in flavodoxins. Biochim. Biophys. Acta 235: 289–302.[Medline]

Mayhew, S.G. and Ludwig, M.L. 1975. Flavodoxins and electron-transferring flavoproteins. In The enzymes (ed. P.D. Boyer), Vol. 12, pp. 57–118. Academic Press, New York.

Mayhew, S.G. and Tollin, G. 1992. General properties of flavodoxins. In Chemistry and biochemistry of flavoenzymes, Vol. 3 (ed. F. Müller), pp. 389–426. CRC Press, Boca Raton, FL.

Mayhew, S.G., Foust, G.P., and Massey, V. 1969. Oxidation-reduction properties of flavodoxin from Peptostreptococcus elsdenii. J. Biol. Chem. 244: 803–810.

Mayhew, S.G., O’Connell, D.P., O’Farrell, P.A., Yalloway, G.N., and Geoghegan, S.M. 1996. Regulation of the redox potentials of flavodoxins: Modification of the flavin binding site. Biochem. Soc. Trans. 24: 122–127.[Medline]

McCarthy, A.A., Walsh, M.A., Verma, C.S., O’Connell, D.P., Reinhold, M., Yalloway, G.N., D’Arcy, D., Higgins, T.M., Voordouw, G., and Mayhew, S.G. 2002. Crystallographic investigation of the role of aspartate 95 in the modulation of the redox potentials of Desulfovibrio vulgaris flavodoxin. Biochemistry 41: 10950–10962.[CrossRef][Medline]

Murshudov, G.N. 1997. Refinement of macromolecular structures by the maximum-likelihood method. Acta Crystallogr. D. Biol. Crystallogr. 53: 240–255.[CrossRef][Medline]

Otwinowski, Z. and Minor, W. 1997. Processing of X-ray diffraction data collected in oscillation mode. Methods Enzymol. 276: 307–326.

Paulsen, K.E., Stankovich, M.T., Stockman, B.J., and Markley, J.L. 1990. Redox and spectral properties of flavodoxin from Anabaena 7120. Arch. Biochem. Biophy. 280: 68–73.[CrossRef][Medline]

Peelen, S., Wijmenga, S., Erbel, P.J., Robson, R.L., Eady, R.R., and Vervoort, J. 1996. Possible role of a short extra loop of the long-chain flavodoxin from Azotobacter chroococcum in electron transfer to nitrogenase: Complete ¹H, ¹⁵N and ¹³C backbone assignments and secondary solution structure of the flavodoxin. J. Biomol. NMR 7: 315–330.[Medline]

Pueyo, J.J., Curley, G.P., and Mayhew, S.G. 1996. Kinetics and thermodynamics of the binding of riboflavin, riboflavin 5'-phosphate and riboflavin 3',5'-bisphosphate by apoflavodoxins. Biochem J. 313 (Pt. 3): 855–861.

Rao, S.T., Shaffie, F., Yu, C., Satyshur, K.A., Stockman, B.J., Markley, J.L., and Sundarlingam, M. 1992. Structure of the oxidized long-chain flavodoxin from Anabaena 7120 at 2 Å resolution. Protein Sci. 1: 1413–1427.[Abstract]

Romero, A., Caldeira, J., Legall, J., Moura, I., Moura, J.J., and Romao, M.J. 1996. Crystal structure of flavodoxin from Desulfovibrio desulfuricans ATCC 27774 in two oxidation states. Eur. J. Biochem. 239: 190–196.[Medline]

Smith, W.W., Burnett, R.M., Darling, G.D., and Ludwig, M.L. 1977. Structure of the semiquinone form of flavodoxin from Clostridum MP: Extension of 1.8 Å resolution and some comparisons with the oxidized state. J. Mol. Biol. 117: 195–225.[CrossRef][Medline]

Smith, W.W., Pattridge, K.A., Ludwig, M.L., Petsko, G.A., Tsernoglou, D., Tanaka, M., and Yasunobu, K.T. 1983. Structure of oxidized flavodoxin from Anacystis nidulans. J. Mol. Biol. 165: 737–753.[CrossRe