X-ray structure of a blue-copper nitrite reductase in two crystal forms. The nature of the copper sites, mode of substrate binding and recognition by redox partner.

Denitrification is one of the main steps of the global nitrogen cycle that is sustained by prokaryotic organisms. Denitrifying bacteria use two entirely different enzymes in this process, one based on haem cd1 prosthetic groups and the other on type 1-type 2 Cu centres. Copper-containing nitrite reductases (NiRs) are sub-divided into blue and green NiRs, which are respectively thought to be redox partners of azurins and pseudo-azurins. Crystallographic structures of the blue nitrite reductase from Alcaligenes xylosoxidans (AxNiR) are presented in the oxidised hexagonal form and the substrate-bound orthorhombic form to 2.1 A and 2.8 A resolution, respectively. The complete amino acid sequence of AxNiR has been determined by conventional chemical analysis. A 3 A structure of AxNiR has been published where the modelling was based on the sequence of another blue NiR. The higher resolution of the hexagonal form together with the correct sequence allows a detailed comparison with the crystallographic structures of the green NiRs. There is a striking difference in the overall surface charge distribution between the two sub-groups, providing a neat structural explanation for their different reactivities to pseudoazurin or azurin and supporting the view that electron transfer proceeds via complex formation. A detailed examination of the type-1 Cu site, the site responsible for the colour, reveals several subtle differences, including a lateral displacement of 0.7 A for Smet. The structure of the type-2 Cu site, and changes that occur upon substrate binding are discussed in terms of the catalytic mechanism. The similarity of the type 2 Cu site to the catalytic Zn site in carbonic anhydrase and the catalytic Cu site of superoxide dismutase is re-examined in view of the high-resolution (2.1 A) structure.

Structural and kinetic evidence for an ordered mechanism of copper nitrite reductase.

The crystallographic structures of several copper-containing nitrite reductases are now available. Despite this wealth of structural data, no definitive information is available as to whether the reaction proceeds by an ordered mechanism where nitrite binds to the oxidised type 2 site, followed by an internal electron transfer from the type 1 Cu, or whether binding occurs to the reduced type 2 Cu centre, or a random mechanism operates. We present here the first structural information on both types of Cu centres for the reduced form of NiR from Alcaligenes xylosoxidans (AxNiR) using X-ray absorption spectroscopy. The reduced type 2 Cu site EXAFS shows striking similarity to the EXAFS data for reduced bovine superoxide dismutase (Cu2Zn2 SOD), providing strong evidence for the loss of the water molecule from the catalytic Cu site in NiR on reduction resulting in a tri-coordinate Cu site similar to that in Cu2Zn2 SOD. The reduced type 2 Cu site of AxNiR is shown to be unable to bind inhibitory ligands such as azide, and to react very sluggishly with nitrite leading to only a slow re-oxidation of the the type 1 centre. These observations provide strong evidence that turnover of AxNiR proceeds by an ordered mechanism in which nitrite binds to the oxidised type 2 Cu centres before electron transfer from the reduced type 1 centre occurs. We propose that the two links between the Cu sites of AxNiR, namely His129-Cys130 and His89-Asp92-His94 are utilised for electron transfer and for communicating the status of the type 2 Cu site, respectively. Nitrite binding at type 2 Cu is sensed by the proton abstracting group Asp92 and the type 2 Cu ligand His94, and relayed to the type 1 Cu site via His89 thus triggering an internal electron transfer. The similarity of the type 2 Cu NiR catalytic site to the reduced Cu site of SOD is examined in some detail together with the biochemical evidence for the SOD activity of AxNiR.

Direct-method structure determination of the native azurin II protein using one-wavelength anomalous scattering data.

The one-wavelength anomalous scattering (OAS) X-ray diffraction data of azurin II, a copper-containing protein from Alcaligenes xylosoxidans were collected at the Photon Factory, Japan at a 'routine' wavelength of 0.97 A. The structure had been originally solved by the molecular-replacement method [Dodd, Hasnain, Abraham, Eady & Smith (1995). Acta Cryst. D51, 1052-1064]. As a technique of ab initio structure determination, the direct method [Fan, Hao, Gu, Qian, Zheng & Ke (1990). Acta Cryst. A46, 935-939] was attempted to break the phase ambiguity intrinsic to OAS data. The phases were then improved using the solvent-flattening method. The final electron-density map clearly shows most Calpha positions and many side chains and it is traceable without prior knowledge of the structure. It is concluded that the direct method is capable of phasing anomalous scattering data collected at one wavelength from moderate-sized native proteins (M(w) approximately 20 kDa) which contain copper or atoms with a similar scattering power.

Structures of oxidized and reduced azurin II from Alcaligenes xylosoxidans at 1.75 A resolution.

Crystallographic structures of oxidized and reduced forms of azurin II are reported at 1.75 A resolution. Data were collected using one crystal in each case and by translating the crystal after each oscillation range to minimize photoreduction. Very small differences are observed at the Cu site upon reduction and these cannot be determined with confidence at current resolution. A comparison with the three-dimensional EXAFS reveals a good correspondence for all the ligand distances except for Cu-His46, where a larger deviation of approximately 0.12-0.18 A is observed, indicating that this ligand is more tightly restrained in the crystallographic refinement at the current resolution.

Ab initio phasing using molecular envelope from solution X-ray scattering.

Solving the phase problem is the crucial and quite often the most difficult and time-consuming step in crystallographic structure determination. The traditional methods of isomorphous replacement (MIR or SIR) and molecular replacement require the availability of an isomorphous heavy-atom derivative or the structure of a homologous protein, respectively. Here, a method is presented which utilizes the low-resolution molecular shape determined from solution X-ray scattering data for the molecular search. The molecular shape of a protein is an important structural property and can be determined directly by the small-angle scattering technique. The idea of locating this molecular shape in the crystallographic unit cell has been tested with experimental diffraction data from nitrite reductase (NiR). The conventional Patterson search proved to be unsuccessful, as the intra-envelope vectors are uniformly distributed and do not match those of intra-molecular (atom-to-atom) vectors. A direct real-space search for orientation and translation was then performed. A self-rotation function using 2.8 A crystallographic data yielded the polar angles of the non-crystallographic threefold axis. Knowledge of the orientation of this axis reduces the potential six-dimensional search to four (Eulerian angle gamma and three translational parameters). The direct four-dimensional search within the unit cell produced a clear solution. The electron-density map based on this solution agrees well with the known structure, and the phase error calculated from the map was 61 degrees within 20 A resolution. It is anticipated that the low-resolution envelope can be used as a starting model for phase extension by the maximum-entropy and density-modification method.

Structure of a new azurin from the denitrifying bacterium Alcaligenes xylosoxidans at high resolution.

It has been reported previously that Alcaligenes xylosoxidans (NC1MB 11015) grown under denitrifying conditions produces two azurins instead of the single previously identified azurin [Dodd, Hasnain, Hunter, Abraham, Debenham, Kanzler, Eldridge, Eady, Ambler & Smith (1995). Biochemistry. In the press]. The new azurin, called azurin II, has been crystallized as blue elongated rectangular prisms with the tetragonal space group P4(1)22 and unit-cell parameters a = b = 52.65, c = 100.63 A. X-ray crystallographic data extending to 1.9 A resolution were collected by the Weissenberg method using 200 x 400 mm image plates and synchrotron X-rays of wavelength 0.97 A. The three-dimensional structure of azurin II has been solved by the molecular-replacement method using the structure of azurin from Alcaligenes denitrificans NCTC 8582 with which this new azurin shows a close homology. The quality of the initial map was sufficient to predict a number of sequence differences. The model is currently refined to an R-factor of 18.8% with X-ray data between 8.5 and 1.9 A. The final model of 961 protein atoms, one Cu atom and 50 water molecules has r.m.s. deviations from ideality of 0.009 A for bond lengths and 1.7 degrees for bond angles. The overall structure is similar to that of the azurin from A. denitrificans NCTC 8582. It has a beta-barrel structure with the Cu atom located near the top end of the molecule. The Cu atom is coordinated to Ndelta of His46 and His117 at 2.02 A and to Sgamma of Cys112 at 2.12 A, while the carbonyl O atom of Gly45 and Sdelta atom of Met121 provide the additional interactions at 2.75 and 3.26 A, respectively.

Structure Solution of Azurin II from Alcaligenes xylosoxidans using the Laue Method: Possibility of Studying In Situ Redox Changes using X-rays.

We have recently demonstrated that X-rays can be used for changing the redox states of the metal centre in metalloproteins [Murphy et al. (1995). J. Synchrotron Rad. 2, 64-69]. The possibility of using the Laue method for studying the structural changes associated with such X-ray-induced reactions is explored by applying the method to the structure determination of a new azurin (hereafter referred to as azurin II) from the denitrifying bacterium Alcaligenes xylosoxidans. Laue X-ray diffraction data of azurin II were collected at station 9.7 of the SRS Daresbury. Three diffraction patterns were recorded on film packs at three different crystal orientations. The data were processed using the Daresbury Laue Software Suite to give 2224 independent single reflections (R(merge) = 0.136) in the wavelength range 0.36-1.40 A. The data completeness was 44% at 2.55 A resolution. Phase determination for the data was undertaken using the molecular-replacement method; the top peak was chosen in both the rotation function and the subsequent translation function. This solution agreed well with the molecular-replacement solution achieved independently using monochromatic data. The electron-density map showed reasonably good agreement with the model and the copper site was readily recognizable as it had the highest density. To see if the electron-density map could be improved, ;the doublets in the diffraction data were then deconvoluted. This added 26% data in the region infinity-2d(min) resulting in an improvement in the data completeness to 50% and thus in improved continuity of the electron-density map. The quality of these maps is discussed from the point of view of the suitability of this approach for studying redox-induced structural changes.

Structures of a blue-copper nitrite reductase and its substrate-bound complex.

Copper-containing nitrite reductases (NiR's) have been conveniently subdivided into blue and green NiR's which are thought to be redox partners of azurins and pseudo-azurins, respectively. Crystal structures of two green NiR's have recently been determined. Alcaligenes xylosoxidans has been shown to have a blue-copper nitrite reductase (AxNiR) and two azurins with 67% homology both of which donate electrons to it effectively. The first crystal structure of a blue NiR (AxNiR) in its oxidized and nitrite-bound forms, with particular emphasis to the Cu sites, is presented. The Cu-Smet distance is the same as those in the green NiR's. Thus, the length of this interaction is unlikely to be responsible for differences in colour. Crystallographic data presented here taken together with structural data of other single Cu type-1 proteins and their mutants suggest that the displacement of Cu from the strong ligand plane is perhaps the cause for the differences in colour observed for otherwise 'classical' blue Cu centre. Nitrite is observed binding to the catalytic Cu in a bidentate fashion displacing the water molecule, offering a neat rationalization for the XAFS observation that the type-2 Cu-ligand distances increase on nitrite binding as a result of increased coordination. These results are discussed in terms of enzyme mechanism.

The substrate-binding site in Cu nitrite reductase and its similarity to Zn carbonic anhydrase.

Here we investigate the structure of the two types of copper site in nitrite reductase from Alcaligenes xylosoxidans, the molecular organisation of the enzyme when the type-2 copper is absent, and its mode of substrate binding. X-ray absorption studies provide evidence for a fourth ligand at the type-2 Cu, that substrate binds to this site and indicates that this binding does not change the type-1 Cu centre. The substrate replaces a putative water ligand and is accommodated by a lengthening of the Cu-histidine bond by approximately 0.08 A. Modelling suggests a similarity between this unusual type-2 Cu site and the Zn site in carbonic anhydrase and that nitrite is anchored by hydrogen bonds to an unligated histidine present in the type-2 Cu cavity.

X-ray structure of a blue copper nitrite reductase at high pH and in copper-free form at 1.9 A resolution.

Copper-containing nitrite reductases possess a trimeric structure where the catalytic Cu site, located at the monomer-monomer interface, resembles the catalytic sites of a number of Zn enzymes. Nitrite reductase from Alcaligenes xylosoxidans has optimum activity at pH 5.2 which decreases to a negligible level at pH 8. The structure of this nitrite reductase has previously been determined at pH 4.6. It has now been crystallized under new conditions at pH 8.5. Its crystallographic structure provides a structural explanation for the greatly reduced activity of the enzyme at high pH. Characterization of overexpressed protein in solution by EXAFS suggested that the protein lacked Cu in the catalytic type 2 Cu site and that the site was most probably occupied by Zn. Using the anomalous signals from Cu and Zn, the crystal structure revealed that the expressed protein was devoid of Cu in the catalytic site and that only a trace amount (<10%) of Zn was present at this site in the crystal. Despite the close structural similarity of the catalytic site to a number of Zn enzymes, these data suggest that Zn, if it binds at the catalytic copper site, binds weakly in nitrite reductase.

Evidence for two distinct azurins in Alcaligenes xylosoxidans (NCIMB 11015): potential electron donors to nitrite reductase.

We have isolated two type 1 copper-containing proteins (M(r) approximately 13K) from Alcaligenes xylosoxidans (NCIMB 11015) grown under denitrifying conditions. Amino acid sequence analysis of these two proteins shows one to be the previously identified azurin (Ambler, 1971), which we shall call azurin I, and the other to be a related, but previously undescribed, blue copper protein which we show to also be an azurin and propose to call azurin II. Thus, NCIMB 11015 becomes the second system where two distinct azurins are found, the other being Methylomonas J (Ambler & Tobari, 1989). On isoelectric focusing, azurin I migrates very similarly to the previously identified azurin from this organism while azurin II migrates similarly to azurin purified from Alcaligenes denitrificans NCTC 8582. The sequence of azurin II is 33% different than the azurin I sequence but is only 11% different than the azurin from Alcaligenes denitrificans NCTC 8582. Optical spectra for the two proteins are very similar with epsilon mM values of 6.27 and 5.73 mM-1 cm-1 for azurin I and II, respectively, at lambda max approximately 620 nm. The 291 nm shoulder normally ascribed to the hydrophobic nature of tryptophan 48 is clearly observed in azurin I but is missing in azurin II. Amino acid analysis confirms that this tryptophan is missing in azurin II. Azurin I and azurin II show essentially the same redox potential of 305 +/- 10 mV at pH 7.5 and are equally effective electron donors to the purified dissimilatory nitrite reductase of Alc. xylosoxidans in vitro.(ABSTRACT TRUNCATED AT 250 WORDS)