The dipole moment of the electron carrier adrenodoxin is not critical for redox partner interaction and electron transfer

https://doi.org/10.1016/j.jinorgbio.2009.04.010Get rights and content

Abstract

Dipole moments of proteins arise from helical dipoles, hydrogen bond networks and charged groups at the protein surface. High protein dipole moments were suggested to contribute to the electrostatic steering between redox partners in electron transport chains of respiration, photosynthesis and steroid biosynthesis, although so far experimental evidence for this hypothesis was missing. In order to probe this assumption, we changed the dipole moment of the electron transfer protein adrenodoxin and investigated the influence of this on protein–protein interactions and electron transfer. In bovine adrenodoxin, the [2Fe–2S] ferredoxin of the adrenal glands, a dipole moment of 803 Debye was calculated for a full-length adrenodoxin model based on the Adx(4-108) and the wild type adrenodoxin crystal structures. Large distances and asymmetric distribution of the charged residues in the molecule mainly determine the observed high value. In order to analyse the influence of the resulting inhomogeneous electric field on the biological function of this electron carrier the molecular dipole moment was systematically changed. Five recombinant adrenodoxin mutants with successively reduced dipole moment (from 600 to 200 Debye) were analysed for their redox properties, their binding affinities to the redox partner proteins and for their function during electron transfer-dependent steroid hydroxylation. None of the mutants, not even the quadruple mutant K6E/K22Q/K24Q/K98E with a dipole moment reduced by about 70% showed significant changes in the protein function as compared with the unmodified adrenodoxin demonstrating that neither the formation of the transient complex nor the biological activity of the electron transfer chain of the endocrine glands was affected. This is the first experimental evidence that the high dipole moment observed in electron transfer proteins is not involved in electrostatic steering among the proteins in the redox chain.

Introduction

Protein–protein interactions are fundamental events for most cellular processes, and recently applied proteomics techniques have pointed out the importance of protein interaction networks as functional units or modules in biological cells [1], [2], [3]. A variety of these interactions support electron transfer reactions in central metabolic pathways, for example in photosynthesis, respiration, and enzymatic catalysis. Molecular defects in these protein interaction networks can lead to malfunction and severe diseases. Therefore, the challenge to predict and understand how proteins interact in the living cell is one of the major research subjects of the post-genomic era.

Protein–protein interactions during electron transfers produce transient complexes, which are characterised by a lifetime in the order of milliseconds and equilibrium dissociation constants in the millimolar to micromolar range [4]. The complexes exist just long enough to complete redox reactions between proteins and cycle between three elementary phases. In the first phase, an active donor–acceptor complex is formed, followed in the second phase by the electron tunnelling within the complex and the final phase in which the oxidised and reduced proteins dissociate [5]. The transient lifetime of this protein donor–acceptor complex makes it difficult to understand how electron transfer between the protein molecules occurs, because neither the docking geometries nor the conformations of these complexes can be predicted with high certainty. Nevertheless, recent approaches to determine transient electron-transfer protein–protein structures resulted in a better understanding of the mechanism of complex formation and electron transfer [6], [7], [8]. The proteins are believed to approach each other guided by electrostatic interactions, which involve the external electric field and the protein dipole moment. This electrostatic steering effect leads to the formation of a specific encounter complex [9], which evolves into the final electron transfer complex, ranging from well-defined geometry to highly dynamic arrangements of the donor and acceptor redox centers [10].

The redox center in bovine adrenodoxin (Adx) contains a [2Fe-2S] cluster and permits the protein to function as an electron carrier in the mitochondrial steroid hydroxylating systems of the adrenal gland. In order to perform this task, Adx forms complexes with the FAD containing adrenodoxin reductase (AdR) and the heme bearing cytochromes P450 CYP11A1, CYP11B1, and CYP11B2, respectively [11]. The three-dimensional structure of Adx displays a compact (α + β) fold typical for [2Fe-2S] ferredoxins [12], [13]. The polypeptide chain is structured as a large core domain, containing the iron–sulphur cluster, and a smaller 35 amino acid-containing domain (Fig. 1). This so called interaction domain includes the acidic region between residues 72 and 79 which was shown to be involved in the complex formation with the redox partners AdR and cytochrome CYP11A1 [14]. Additional interaction sites localised in the core domain have been identified in the crystal structure of a cross-linked 1:1 complex of Adx and AdR [15], [16] as well as by mutagenesis studies [17], [18].

The acidic amino acids of the interaction sites are part of a striking charge clustering on the Adx surface (Fig. 1) which renders one face of the molecule almost completely acidic and induces, together with the positively charged lysine residues on the opposite protein surface an asymmetric electric potential of the protein [19]. This charge distribution leads to the formation of a large molecular dipole moment and suggests that an electrostatic steering mechanism may accelerate the diffusion-limited interaction of Adx with its redox partners during complex formation. This hypothesis is supported by the presence of similar charge asymmetries on the surface of the redox partners AdR [20] and cytochromes P450 [21]. Adx is therefore an ideal candidate for mutagenesis, and the electron transfer chain AdR–Adx–CYP11A1 is an exemplary model system to probe the hypothesis of electrostatic steering induced by dipoles during protein interactions.

The influence of the dipole moment on this process is not yet understood, mainly due to the lack of experimental evidence. It has remained unclear to what extent dipole forces influence the long-range protein–protein encounter because of the counter ion shielding of involved charges. Once two protein molecules are in close proximity, however, the dipole moment could support the fine positioning of the molecules to form a productive complex.

In order to understand the particular role of the molecular dipole for the functional properties of the protein, a molecular model of Adx based on the crystal structure was constructed in the first step, and the dipole moments for different possible conformations of the flexible C-terminus were calculated. In the second step, a series of five mutants with changed magnitude of the dipole were generated by site-directed mutagenesis (Fig. 1). The recombinant Adx variants K6E, K98E, K6E/K98E, K6E/K22Q/K24Q, and K6E/K22Q/K24Q/K98E, were purified from Escherichia coli and characterised in detail in terms of complex formation and electron transfer.

Section snippets

Molecular modelling and calculation of the molecular dipole moment

The basic models are the crystal structures of the truncated Adx(4-108) (Protein Data Bank entry 1AYF) and wild type Adx in the complex with AdR (PDB entry 1E6E). In both structures Asp5 is well defined in the electron density but not Ser1–Ser3 and Glu4. Glu4 was modelled on the basis of revised low-electron-density maps generated from both datasets, assuming that modelled charge centers are more suitable for dipole calculations than missing ones. The rest of the N-terminus has been modelled

Calculation of the dipole moment

The dipole moment is a vector with a magnitude equal to the product of equal positive and negative charges with their distance. The approximate dipoles of full-length Adx and derived mutants were calculated from a model based on the Adx(4-108) and wild type Adx crystal structures. Several rotamers for the C-terminus beyond Pro108 were generated using the procedure described to cover potential conformations (Fig. 2). Full-length Adx has 35 charged residues, carrying 11 positive, and 24 negative

Discussion

In transient protein complexes involved in electron exchange it was hypothesised for several systems that electrostatic steering induced by dipoles plays a prominent role for the protein orientation and the resulting electron transfer [12], [21], [34]. So far, however, this suggestion to the best of our knowledge has not been proven experimentally. Such transient complexes of redox proteins show in general weak affinity and a short lifetime, but they still react with sufficient specificity to

Conclusions

Taken together, we present the first systematic experimental study on the influence of the dipole moment in electron transfer reactions. Therefore, single and multiple Adx mutants were carefully selected because mutations must neither influence the protein folding and the redox properties of the iron–sulphur cluster nor be positioned at the protein–protein interface. These mutants were investigated in a series of kinetic and thermodynamic measurements with the interesting result that none of

Abbreviations

Adxadrenodoxin
AdRadrenodoxin reductase
CYP11A1cytochrome P450 11A1
CDcircular dicroism
EDTAethylene diamine tetra-acetic acid
SDS–PAGEsodium dodecyl sulphate polyacrylamide gel electrophoresis
NADPHnicotinamide adenine dinucleotide phosphate
PDBprotein data bank
NMRnuclear magnetic resonance
HEPES4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid

Acknowledgments

The authors thank Wolfgang Reinle and Walter Klose for purifying proteins and Petra Weißgerber for PCR cloning of mutant K98E. The Deutsche Forschungsgemeinschaft (Grants Be 1343/12-3 and He 1318/19-3) and the Fonds der chemischen Industrie (Grants to R.B and U.H.) supported this work.

References (42)

  • P. Bork et al.

    Curr. Opin. Struct. Biol.

    (2004)
  • G. Drewes et al.

    Curr. Opin. Cell Biol.

    (2003)
  • K.P. Hofmann et al.

    Trends Biochem. Sci.

    (2006)
  • A. Müller et al.

    Structure

    (1998)
  • I.A. Pikuleva et al.

    Arch. Biochem. Biophys.

    (2000)
  • L.E. Vickery

    Steroids

    (1997)
  • J.J. Müller et al.

    J. Biol. Chem.

    (2001)
  • F. Hannemann et al.

    J. Biol. Chem.

    (2001)
  • A. Heinz et al.

    Biochem. Biophys. Res. Com.

    (2005)
  • G.A. Ziegler et al.

    J. Mol. Biol.

    (1999)
  • C.A. Hasemann et al.

    Structure

    (1995)
  • V.S. Lamzin et al.

    Method. Enzymol.

    (1997)
  • T. Kido et al.

    J. Biol. Chem.

    (1979)
  • S. Sugano et al.

    Anal. Biochem.

    (1989)
  • B. Schiffler et al.

    J. Biol. Chem.

    (2004)
  • J.D. Lambeth et al.

    J. Biol. Chem.

    (1983)
  • A. Zöllner et al.

    J. Inorg. Biochem.

    (2002)
  • C. Frazao et al.

    Structure

    (1995)
  • J.J. Müller et al.

    J. Mol. Biol.

    (1999)
  • W.H. Koppenol et al.

    J. Biol. Chem.

    (1982)
  • D.S. Bendall
  • Cited by (13)

    • Functional characterization of Fdx1: Evidence for an evolutionary relationship between P450-type and ISC-type ferredoxins

      2011, Journal of Molecular Biology
      Citation Excerpt :

      Another highly conserved region is the acidic interaction site between D72 and D79 of Adx, which is known to be involved in the interaction of this ferredoxin with its redox partners AdR and CYP11A1.34 Although it was shown recently that the dipole moment of Adx resulting from asymmetric charge distribution in the molecule is not critical for redox partner interaction and electron transfer,35 it is noteworthy that, in all ferredoxins, the D72–D79 region forms a cluster of negatively charged amino acids. Taking a closer look at the key residues D72, E73, D76 and D79 of Adx, a possible explanation becomes obvious why Fdx1 and Etp1fd can substitute for Adx in hydroxylase-dependent substrate conversions, whereas the human Adx homologue cannot: in both Fdx1 and Etp1fd, the four key residues are conserved, whereas in the case of human Fdx2 (as in all mammalian ISC-type ferredoxins considered here), a positively charged arginine residue interrupts the string of negatively charged residues (Fig. 5, Supplemental Fig. S1 and Supplemental Table S2).

    • Adrenodoxin: The archetype of vertebrate-type [2Fe-2S] cluster ferredoxins

      2011, Biochimica et Biophysica Acta - Proteins and Proteomics
    View all citing articles on Scopus
    View full text