Anne Volbeda

Crystal structure of hexameric haemocyanin from Panulirus interruptus refined at 3.2 A resolution.

The use of non-crystallographic symmetry restraints in the refinement of the haemocyanin hexamer from Panulirus interruptus at 3.2 A resolution has resulted in a final model with a very reasonable geometry and a crystallographic R-factor of 20.1%, using 59,193 observed structure factor amplitudes between 8.0 and 3.2 A. The mean co-ordinate error is approximately 0.35 A. The six subunits appear to be related by symmetry operations that differ slightly from 32 point group symmetry. The six subunits have essentially maintained the same structure. The hexamer, with point group 32, is best described as a trimer of tight dimers. The contacts between the subunits in such a dimer are more numerous, and better conserved during evolution than contacts in a trimer. The interface of a tight dimer is separated by an internal cavity into two contact areas. The contact area nearest to the centre of the hexamer is most extensive and consists mainly of residues that are quite conserved among arthropodan haemocyanins. All these residues are provided by the second domain of each subunit. Hence, this second domain may play a crucial role in the allosteric functioning of this oxygen transport protein. The dinuclear copper oxygen-binding site resides in the centre of domain 2. This oxygen-binding centre is not fully accessible from the solvent. Three large cavities occur, however, within each subunit at the interfaces of the three domains. All three cavities contain ordered water molecules, and two of them are accessible from the surrounding solvent. These cavities may play a role in facilitating fast movement of dioxygen towards the binding site, which is situated in a highly conserved, rather hydrophobic core. A detailed definition of the geometry of the copper site is, of course, not possible at the limited resolution of 3.2 A. Nevertheless, it is possible to conclude that each copper is co-ordinated by two, more or less tightly bound, histidine ligands and one more distant histidine residue. The six histidine residues utilize their N epsilon atoms for copper co-ordination, while their N delta atoms are engaged in hydrogen bonds with conserved residues or water molecules. The two distant histidine ligands are located in apical positions and are on opposite sides with respect to the plane approximately defined by the four more tightly bound histidine ligands and the two copper ions. The copper-to-copper distance is 3.5 to 3.6 A in four of the subunits, but this distance deviates considerably in two others.(ABSTRACT TRUNCATED AT 400 WORDS)

Structure determination of Panulirus interruptus haemocyanin at 3.2 Å resolution: Successful phase extension by sixfold density averaging

The three-dimensional structure of Panulirus interruptus haemocyanin has been determined at 3.2 A resolution by X-ray diffraction techniques. Starting from a double isomorphous replacement map at 4 A resolution, the phases of the 32,709 reflections were first improved by six cycles of sixfold molecular averaging and solvent flattening. This also generated phase for 3078 reflections with no isomorphous replacement data. Next, phases for the reflections between 4.0 A and 3.2 A were obtained by a stepwise expansion procedure. In each expansion step 2000 to 3000 new reflections were added to the set of already phased reflections, followed by a few cycles of density averaging and solvent flattening at constant resolution. The eventual map at 3.2 A was calculated with 63,843 reflections with an average Sim weight of 0.75 and an overall R-factor of 23.5%. The polypeptide chain could be traced without any problems, while the dinuclear copper site, disulphide bridges and the first three moieties of the carbohydrate chain were clearly visible. Each subunit consists of three distinct domains. The first domain is mainly helical, containing one disulphide bridge and the carbohydrate chain. The second domain is also predominantly helical and contains the dinuclear copper site at its centre. The core of the third domain is an anti-parallel beta-barrel with the same topology as in the immunoglobulins and Cu,Zn-superoxide dismutase. This domain contains two disulphide bridges. Two long loops extend from the beta-barrel and have numerous interactions with the other two domains. The two copper ions are at approximately 3.7 +/- 0.25 A from each other and co-ordinated by six histidines, which are strictly conserved in all seven arthropodan haemocyanins with known amino acid sequences. No bridging protein ligand is present in the electron density distribution. Hydroxyl groups from tyrosines, which are invariant among arthropodan haemocyanins, are 17.4 A or more removed from the centre of the two copper ions. The closest Panulirus tyrosine hydroxyl is 10.6 A from the copper ions. So, it appears unlikely that tyrosine is involved in the co-ordination of the coppers either in the deoxy or in the oxy form of arthropodan haemocyanins.

Pseudo 2-fold symmetry in the copper-binding domain of arthropodan haemocyanins: possible implications for the evolution of oxygen transport proteins

Investigation of the copper-binding centre of Panulirus interruptus haemocyanin led to the discovery of a pseudo 2-fold axis relating two helical pairs surrounding and co-ordinating the two copper ions. The pseudo 2-fold symmetry relating one helical pair, co-ordinating Cu-A, to the second helical pair co-ordinating Cu-B is quite precise with 31 equivalent C alpha atoms having a root-mean-square deviation of only 1.47 A. The 2-fold consists of a rotation of 174.6 degrees and a translation parallel to the rotation axis of 0.7 A. After superposition of the helical pairs, the two copper ions are within 1.1 A and the three C alpha atoms of the histidine ligands of Cu-A are within a root-mean-square deviation of 1.0 A from the C alpha atoms of the histidine residues co-ordinating Cu-B. Of the superimposed residues, 26% are identical in sequence. These data suggest that the current oxygen-binding centre of arthropodan haemocyanins is the result of dimerization, gene duplication and gene fusion of an ancestral mono-copper-binding helical pair. This suggestion is supported by the recent discovery that in the sequence of functional domains of molluscan haemocyanins only amino acid sequence homology with the arthropodan Cu-B helical pair has been found and no evidence for similarity with a Cu-A binding helical pair was observed. This provides strong evidence that a mono-copper-binding helical pair has been the ancestor of both the arthropodan and molluscan haemocyanins. Turning to the Fe-binding helical pairs in haemerythrins, it appears that they are less similar to each other than the two Cu-binding helical pairs in arthropodan haemocyanins. Nevertheless, the Fe-B haemerythrin helical pair superimposes well onto the Cu-A helical pair of Panulirus haemocyanin. A root-mean-square deviation of 1.9 A for 24 equivalent C alpha carbon atoms is obtained, while Fe-B deviates 1.4 A from Cu-A after superposition of the helices. Moreover, the three histidine ligands of the Cu-A helical pair are equivalent with three histidine ligands of the Fe-B pair. The structural similarity and correspondence in metal-binding ligands suggests that both haemocyanins and haemerythrins have originated from an ancestral mono-metal-binding helical pair having two ligands provided by the first helix and one ligand by the second helix.(ABSTRACT TRUNCATED AT 400 WORDS)

Crystallisation and preliminary crystallographic analysis of P1 nuclease from Penicillium citrinum

P1 nuclease, a zinc-dependent single-strand specific endonuclease from Penicillium citrinum, has been crystallized in three different space groups using either ammonium sulphate or polyethylene glycol 4000 as the precipitating agent. The crystals diffract to between 3 A and 2.2 A. A 4.5 A electron density map has been calculated for a tetragonal crystal form, based on a platinum derivative, and was improved by solvent flattening. The boundaries of the two molecules in the asymmetric unit are clearly visible in most regions and the presence of rod-like density features are indicative of a rather high alpha-helix content. The highest density peaks in the map were identified as a trinuclear zinc cluster present in each monomer by a difference Fourier of an EDTA-soaked crystal.

Carbon Monoxide Dehydrogenase Reaction Mechanism: A Likely Case of Abnormal CO2 Insertion to a Ni−H− Bond

Ni-containing carbon monoxide dehydrogenases (CODH), present in many anaerobic microorganisms, catalyze the reversible oxidation of CO to CO(2) at the so-called C-cluster. This atypical active site is composed of a [NiFe(3)S(4)] cluster and a single unusual iron ion called ferrous component II or Fe(u) that is bridged to the cluster via one sulfide ion. After additional refinement of recently published high-resolution structures of COOH(x)-, OH(x)-, and CN-bound CODH from Carboxydothermus hydrogenoformans (Jeoung and Dobbek Science 2007, 318, 1461-1464; J. Am. Chem. Soc. 2009, 131, 9922-9923), we have used computational methods on the predominant resulting structures to investigate the spectroscopically well-characterized catalytic intermediates, C(red1) and the two-electron more-reduced C(red2). Several models were geometry-optimized for both states using hybrid quantum mechanical/molecular mechanical potentials. The comparison of calculated Mössbauer parameters of these active site models with experimental data allows us to propose that the C(red1) state has a Fe(u)-Ni(2+) bridging hydroxide ligand and the C(red2) state has a hydride terminally bound to Ni(2+). Using our combined structural and theoretical data, we put forward a revised version of an earlier proposal for the catalytic cycle of Ni-containing CODH (Volbeda and Fontecilla-Camps Dalton Trans. 2005, 21, 3443-3450) that agrees with available spectroscopic and structural data. This mechanism involves an abnormal CO(2) insertion into the Ni(2+)-H(-) bond.

The active site and catalytic mechanism of NiFe hydrogenasesBased on the presentation given at Dalton Discussion No. 6, 9?11th September 2003, University of York, UK.

This Perspective describes, from our own personal experiences, how the architecture of the NiFe hydrogenase active site has been elucidated by a combination of protein crystallography, Electron paramagnetic resonance and Fourier transform infrared spectroscopic studies. Thus within a period of eight years our perception of the active center has changed from a mononuclear Ni center with S and N/O coordination to a binuclear NiFe unit with thiolate (to Ni and Fe) and CO and CN− n(to Fe) ligands. This biologically unusual organometallic cluster poses a real challenge in terms of understanding the role of its different components. Current ideas concerning the NiFe hydrogenase catalytic mechanism are discussed in this context.

Three-Dimensional Structure of Haemocyanin from the Spiny Lobster, Panulirus Interruptus, at 3.2 Å Resolution

Haemocyanins are the non-haem, copper-containing oxygen transporting molecules occurring freely dissolved in the haemolymph of a large number of invertebrate species. The molecular architectures of the two known classes of haemocyanins are entirely different. Molluscan haemocyanins have the form of cylinders, with 10–20 subunits forming the complete molecules with molecular weights up to about ten million daltons. The subunits are made up by maximally 8 “repeated” domains, each of which has one dinuclear copper site. Arthropodan haemocyanins are composed of hexamers, or multi-hexamers, with individual subunits having molecular weights in the order of 75.000 daltons, each subunit containing one pair of copper ions. Complete molecules range from single hexamers of ~ 460.000 daltons to octa-hexamers with molecular weights of something like 3.7 million. All haemocyanins are thus large molecules, but some are larger than others (1–3).

Crystal Structures of Quinolinate Synthase in Complex with a Substrate Analogue, the Condensation Intermediate, and Substrate-Derived Product.

The enzyme NadA catalyzes the synthesis of quinolinic acid (QA), the precursor of the universal nicotinamide adenine dinucleotide (NAD) cofactor. Here, we report the crystal structures of complexes between the Thermotoga maritima (Tm) NadA K219R/Y107F variant and (i) the first intermediate (W) resulting from the condensation of dihydroxyacetone phosphate (DHAP) with iminoaspartate and (ii) the DHAP analogue and triose-phosphate isomerase inhibitor phosphoglycolohydroxamate (PGH). In addition, using the TmNadA K219R/Y21F variant, we have reacted substrates and obtained a crystalline complex between this protein and the QA product. We also show that citrate can bind to both TmNadA K219R and its Y21F variant. The W structure indicates that condensation causes dephosphorylation. We propose that catalysis by the K219R/Y107F variant is arrested at the W intermediate because the mutated protein is unable to catalyze its aldo-keto isomerization and/or cyclization that ultimately lead to QA formation. Intriguingly, PGH binds to NadA with its phosphate group at the site where the carboxylate groups of W also bind. Our results shed significant light on the mechanism of the reaction catalyzed by NadA.