How the Local Geometry of the Cu-Binding Site Determines the Thermal Stability of Blue Copper Proteins

1 How the local geometry of the Cu binding site determines the thermal stability of blue copper proteins. Jesús Chaboy1,4,*, Sofía Díaz-Moreno2, I. Díaz-Moreno3, M. A. De la Rosa3, A. Díaz- Quintana3,* 1Instituto de Ciencia de Materiales de Aragón, Consejo Superior de Investigaciones Científicas – Universidad de Zaragoza, 50009 Zaragoza, Spain. 2Diamond Light Source Ltd, Harwell Science and Innovation Campus, Didcot, Oxfordshire, OX11 0DE, U.K. 3Instituto de Bioquímica Vegetal y Fotosíntesis, Consejo Superior de Investigaciones Científicas – Universidad de Sevilla, 41092 Sevilla, Spain. 4Dpto. Física de la Materia Condensada, Universidad de Zaragoza, 50009 Zaragoza, Spain. Contact: Antonio Díaz-Quintana Phone: +34 954 489 507 Fax: +34 954 460 065 e-mail: [email protected] Jesús Chaboy Phone: +34 976 761 222 Fax: +34 976 761 229 e-mail: [email protected] *Manuscript Click here to view linked References 2 Running title Copper site geometry and BCP thermal stability. Summary Identifying the factors that govern the thermal resistance of cupredoxins is essential for understanding their folding and stability, and for improving our ability to design highly stable enzymes with potential biotechnological applications. Here we show that the thermal unfolding of plastocyanins from two cyanobacteria ─ the mesophilic Synechocystis and the thermophilic Phormidium ─ is closely related to the short-range structure around the copper centre. Cu K-edge X-ray absorption spectroscopy shows that the bond length between Cu and the S atom from the cysteine ligand is a key structural factor that correlates with the thermal stability of the cupredoxins in both oxidised and reduced states. These findings were confirmed by an additional study of a site-directed mutant of Phormidium plastocyanin showing a reverse effect of the redox state on the thermal stability of the protein. Highlights  The local geometry of Cu determines the thermal stability of blue copper proteins.  The melting point of homologous BCPs correlates with the Cu site geometry.  Remote mutations affect the local structure of Cu in BCPs. 3 Introduction In recent years, our knowledge of the thermal stability of proteins has improved considerably. However, rational design of thermo stable proteins remains elusive. Thermophilic organisms thrive at temperatures at which proteins from mesophilic organisms are often completely unfolded and nonfunctional. Understanding the mechanisms by which proteins develop thermal stability will help to optimise and design thermostable functional proteins for a variety of biotechnological applications (Adams and Kelly, 1995; Persidis, 1998; Szilagyi and Zavodszky, 2000; Lehmann et al., 2000; Tsujimura et al., 2008; Miura et at, 2009). For this purpose, we need to clarify what factors make proteins from thermophilic organisms (thermophilic proteins) different from their mesophilic homologues. Previous work has shown that the overall structure of the native conformation is not a dominant factor, since homologous proteins invariably adopt the same fold (Szilagyi and Zavodszky, 2000). Blue Copper Proteins (BCPs) are of outstanding industrial interest, as putative components of biological fuel cells (Tsujimura et al., 2008; Miura et al., 2009). In fact, the multicopper oxidases used for this purpose comprise functional BCP domains (Nakamura and Go, 2005; Kosman, 2010). Despite its role is barely understood, the copper site plays an essential part in the thermal stability of BCPs and in the kinetics of their unfolding process (Leckner et al., 1997; Pozdnyakova and Wittung-Stafshede, 2001; Pozdnyakova et al., 2001; Alcaraz and Donaire, 2004; Alcaraz and Donaire, 2005). Oxidised species of plastocyanin from thermophilic cyanobacteria (Feio et al., 2004) are more thermally stable than their reduced forms, whereas the contrary occurs in plastocyanin from plants (Sandberg et al., 2003) or mesophilic cyanobacteria (Feio et al., 2006). These different 4 behaviours are unexpected, since aminoacids at the surroundings of the metal centre are highly conserved in all these proteins (Fig. 1A,1B and 1C). Plastocyanins contain a mononuclear copper site (Type I) (Gough and Chotia, 2004) wherein the first coordination sphere of the metal ion is formed by two nitrogen atoms and two sulphur atoms in a distorted trigonal pyramid (Solomon et al., 2004; Solomon, 2006). In the reduced species, ligands to metal (dative) interactions involve 2p and 3p atomic orbitals (AO) from ligands and the 4p AO from copper. Minor, if any, back-bonding from metal has been estimated (Guckert et al., 1995). In oxidised BCPs, the anisotropy of the highest occupied molecular orbital (HOMO) of Cu2+ determines the physical and chemical properties of the metal centre (Randall et al, 2000; Solomon, 2004; Solomon, 2006). This orbital lies over the equatorial plane containing the two nitrogen atoms and S-Cys. Among the metal-ligand bonds, that involving S-Cys shows the largest degree of covalence, resulting from the combination of the copper dx2─y2 AO with the 3p AO of S-Cys (Solomon et al. 2007; Pavelka and Burda, 2008). Opposite, the long axial Cu–S-Met bond shows little covalent character. Even though the Cu-S-Cys bond provides a high electronic coupling for electron transfer, the physiological electron exchange port locates in the histidine imidazol ligand that is partially exposed to solvent (Canters et al., 2000). Theory also shows that the copper site geometry is somehow modulated by the overall protein scaffold (Comba, 2000). Studies on model systems show that there a reciprocal relationship between Cu–S-Cys and Cu–S-Met bond strengths. Distortion of the geometry alters the overlap between copper and ligand orbitals: the Cu dx2─y2 AO tilts, and the Cu–S-Cys bond weakens, whereas that between Cu and S-Met strengthens or vice 5 versa (LaCroix et al., 1996). Moreover, mutations of charged residues at the protein surface affect the Cu–S-Met bond length while modulating the stability of BCPs (Sato and Denisson, 2002). Namely, a conserved lysine residue is essential for the correct folding of plant phytocyanins, which display glutamine instead of methionine as the axial ligand (Harrison et al., 2005). Of outstanding biotechnological interest is that the copper site environment can be modified to tune its acid-base properties and redox potential (Remenyi et al., 2001; Narshell et al., 2009), to lodge a different kind of copper centre (Jonas et al., 2003), or to improve the electron transfer rate of these proteins (Lancaster et al., 2009). We have investigated by X-ray Absorption Spectroscopy (XAS) the relationship between the local geometry of the Cu site and the thermal unfolding behaviour of the wellknown copper-containing protein plastocyanin (Pc) from thermophilic (Phormidium laminosum, Pho-WT) and mesophilic (Synechocystis sp. PCC 6803, Syn-WT) cyanobacteria. We have previously shown that the higher thermal stability of Pho-WT is not related to typical structural adaptation patterns (Muñoz-López et al., 2010a). This study has also been extended to the double mutant P49G/G50P of Phormidium Pc (Pho-Mut) (Muñoz-López et al., 2010b). This mutation reverses the relationship between the oxidation state of copper and the thermal stability of the protein (Fig. 1D). As it lies far from the metal site (Fig. 1A), it illustrates how the protein matrix affects the metal site. Our results show that the thermal stability of plastocyanin is closely related to the bond length between Cu and the S atom from the cysteine acting as a ligand. Notably, the replacement of this residue is essential to enhance the capability of the protein for direct electron transfer to electrodes (Lancaster et al., 2009). 6 Results and discussion Pho-WT shows a larger thermal stability than Syn-WT (Feio et al., 2004; Feio et al., 2006). The midpoint unfolding temperatures (Tm) represented in Fig. 1D (see Table 1 for numerical values) show that reduced Syn-WT is more stable than its oxidised species. Opposite, in Pho-WT the oxidised form is the most stable one. Notably, Tm values of the Pho-Mut are, in both reduced and oxidised forms, closer to the Tm values of Syn-WT than to those of Pho-WT (Muñoz-López et al., 2010b). Interestingly, the largest Tm differences are found among the oxidised forms in all cases. In fact, reduced plastocyanins exhibit similar Tm values lying in between the two extreme corresponding to the oxidised forms . Overall, Cu K-edge X-ray Absorption Near Edge Spectroscopy (XANES) spectra of the oxidised forms of Pho-WT, Syn-WT and Pho-Mut plastocyanins (Muñoz-López et al., 2010b) are quite similar but showing well defined differences (Fig. 2, upper panel). Indeed, the XANES spectrum of Pho-Mut perfectly overlaps with that of Synechocystis through all the spectral range. However, the spectrum of Pho-WT differs in the shape, intensity and energy position of the shoulder-like feature (A) at the absorption threshold, the peak B, at ~9.5 eV above the edge, and of the main absorption line C ( ~18.5 eV). Actually, the B peak is clearly observed in both Syn-WT and Pho-Mut but it is missing in the case of Pho- WT. The spectral shape and the intensities of the main absorption features in the nearedge region of the XAS spectra are extremely sensitive to changes of the local structure (Chaboy, 2009). Therefore, these results indicate that the local structural environment of Cu is similar in both Syn-WT and Pho-Mut, as the thermal unfolding behaviour is too, being unlike the one of Pho-WT. Moreover, despite the mutation of Phormidium taking place at distances greater than 20 Å away from the copper site, it induces a structural 7 modification that is sensed by the metal ion. In this way, the local structure around Cu in Pho-Mut is similar to that of Syn-WT, and is different from that found in the unmodified Pho-WT’s one. It should be noted that the mutation locates at a loop (L5) settled at the end of the -strand that links it to a His residue acting as a copper ligand. However, the mutation does not affect the secondary structure substantially (Muñoz-López, 2010b). Noteworthy, Molecular Dynamics reports show concerted motions between loop L5 and loops surrounding the copper site (Arcangeli et al., 2001), which are in agreement with the observed modification of the Cu environment obtained by XAS. Bearing in mind the similarity of the Tm values of the oxidised forms of both Syn-WT and Pho-Mut, all these results suggest that the local environment details of Cu are linked to the thermal unfolding properties. Hence, we have performed ab initio computations of the Cu K-edge XANES spectra of these oxidised species, to determine the origin of the differences observed in the XANES spectra, and to identify which individual structural component (histidines, cysteine, methionine) of the Cu environment is affected by the mutation of Phormidium that brings its behaviour into a comparative agreement with Synechocystis. As shown in Fig. 2 and in the Supplemental Information, we have obtained an accurate reproduction of the experimental spectra of both Syn-WT and Pho-WT. Hence, our next step was to determine how the Cu environment found in Pho-WT has to be modified in order to reproduce the experimental XANES spectra of its mutant form and of Syn-WT. The Cu binding site in BCPs features a distorted trigonal pyramid. Two nitrogen atoms (N1,HisN and N2,HisC) from separate His residues and a sulphur atom (S,Cys) from a Cys ligand represent the trigonal plane of the pyramidal base. A second sulphur (SMet) 8 atom from an axial Met residue forms the apex. Distortion occurs in the contact lengths between the copper and sulphur ligands being the Cu-SCys bond shorter than the Cu- S,Met one. According to X-ray diffraction studies (Romero et al., 1998; Bond et al., 1999), the Cu-SCys inter-atomic distance is 2.10 Å in oxidised Pho-WT (Bond et al., 1999; PDB code 1baw), shorter than that in oxidised Syn-WT (Romero et al., 1998; PDB code 1pcs), which is 2.25 Å long. Conversely, the Cu-SMet bond length in Pho-WT, 2.73 Å, is larger than in Syn-WT, 2.65 Å. Then, we have computed the Cu K-edge XANES of Pho-WT but replacing different coordination shells by those found in Syn-WT. In this way, we have constructed three modified Pho-WT clusters containing 151 atoms in which the first 5, 14 and 112 next neighbours of Cu have been substituted by those found in Synechocystis. The results of these calculations are reported in Fig. 3. Replacing the first 14 neighbours of Cu, (atoms within 4 Å from the metal) allows us to reproduce the experimental spectra of Syn-WT. Exchanging further Synechocystis-like shells up to 8 Å has no significant effect on the calculated spectra. Next, we have calculated the Cu K-edge of a Pho-WT cluster in which we have specifically substituted the neighbours within the first 4 Å around the Cu ion coming from the His, Cys and Met ligands. As shown in Fig. 3 (middle panel) the substitution of Met coordinates has no effect on the spectrum. However, the modification of His or Cys neighbours changes the original Pho-WT spectrum to the one of Syn-WT. Specifically, after modifying the Cys ligand the spectrum perfectly fits to the Syn-WT’s one at the high-energy region. Finally, we have performed four similar computations in which only one atom is changed at each time: the N1 atoms associated to the two His and the S and S of Cys and Met, respectively. The results of these computations show that the modification of the Cu-SCys bond accounts for the differences found between Syn-WT and Pho-WT. 9 The picture emerging from the XANES results for the oxidised forms remarks that the length of the Cu-S,Cys bond is the essential parameter in determining the thermal stability of these BCPs. If this is the case, the Cu-S,Cys should also play a fundamental role into determining the thermal properties of the reduced forms. As shown in Fig. 1c, the Tm values of the three reduced species are similar and their values are between those of the oxidised forms of Pho-WT and Syn-WT or Pho-Mut. The above hypothesis entails that the lower Tm, the longer Cu-S,Cys bond length. In fact, oxidised Syn-WT shows the lowest Tm and the longest Cu-S,Cys (2.25 Å) bond length, whereas the highest Tm value and the shortest Cu-S,Cys bond (2.10 Å) corresponds to oxidised Pho-WT. Then, it would be expected that the Cu-S,Cys bond length to be similar in the three reduced forms and, in addition, intermediate between the two extremes of the oxidised species. The similarity of the Cu K-edge XANES of the reduced forms of Pho-WT, Syn-WT and Pho-Mut, reported in Fig. 4, supports this hypothesis as differences found in the near-edge region are less marked than for the oxidised forms. This is further supported by computations assuming identical Cu-S,Cys bond lengths in the three proteins. Starting from the Cu environment of oxidised Pho-WT and Syn-WT, changes of the Cu-S,Met bond length do not affect the spectral profile. However, when the Cu-SCys length is fixed to an intermediate value (2.17Å) the spectra of both Pho-WT and Syn-WT perfectly match at the threshold region in which the major differences in the Cu K-edge absorption of the oxidised form were observed. In summary, our results provide evidence of the relationship between the thermal unfolding behaviour and the different local geometry of the Cu binding site in mesophilic and thermophilic proteins. This finding shows how the interplay between the metal centre 16 Pavelka, M., and Burda, J.V. (2008) Computational study of redox active centres of blue copper proteins: a computational DFT study. Mol. Phys. 106, 2733─2748. Persidis, A. (1998) Extremophiles: unusual and useful molecules are found in life on the edge of environmental tolerance. Nat. Biotechnol. 16, 593–594. Pettersen, E.F., Goddard, T.D., Huang, C.C., Couch, G.S., Greenblatt, D.M., Meng E.C., and Ferrin, T.E. (2004) UCSF Chimera - A Visualization System for Exploratory Research and Analysis, J. Comput. Chem. 25, 1605-1612. Pozdnyakova, I., and Wittung-Stafshede, P. (2001) Biological relevance of metal binding before protein folding. J. Am. Chem. Soc. 123, 10135–10136. Pozdnyakova, I., Guidry, J., and Wittung-Stafshede, P. (2001) Probing copper ligands in denatured Pseudomonas aeruginosa azurin: Unfolding His117Gly and His46Gly mutants. J. Biol. Inorg. Chem. 6, 182–188. Randall, D.W., Gamelin, D.R., LaCroix, L.B. and Solomon, E.I. (2000) Electronic structure contributions to electron transfer in blue Cu and Cu(A). J. Biol. Inorg. Chem. 5, 16–19. Remenyi, R., Jeuken, L.J.C., Comba, P. and Canters, G.W. (2001). An amicyanin C- terminal loop mutant where the active-site histidine donor cannot be protonated. J. Biol. Inorg. Chem. 6, 23-26. Romero, A., De la Cerda, B., Varela, P.F., Navarro, J.A., Hervás, M., and De la Rosa, M.A. (1998) The 2.15 A crystal structure of a triple mutant plastocyanin from the cyanobacterium Synechocystis sp. PCC 6803. J Mol Biol. 275, 327-36. Sandberg, A., Harrison, D.J., and Karlsson, B.G. (2003) Thermal denaturation of spinach plastocyanin: Effect of copper site oxidation state and molecular oxygen. Biochemistry 42, 10301-10310. 17 Sato, K., and Dennison, C. (2002) Effect of histidine 6 protonation on the active site structure and electron-transfer capabilities of pseudoazurin from Achromobacter cycloclastes. Biochemistry 41, 120-130. Solomon, E.I., Szilagyi, R.K., George, S.D., and Basumallick, L. (2004) Electronic Structures of Metal Sites in Proteins and Models: Contributions to Function in Blue Copper Proteins. Chem. Rev. 104, 419-458. Szilagyi, A., and Zavodszky, P. (2000) Structural differences between mesophilic, moderately thermophilic and extremely thermophilic protein subunits: results of a comprehensive survey. Struct. Fold. Des. 8, 493–504. Tsujimura, S., Miura, Y., and Kano, K. (2008) CueO-immobilized porous carbon electrode exhibiting improved performance of electrochemical reduction of dioxygen to water. Electrochim. Acta 53, 5716-5720. 18 Supplementary Information accompanies the paper in Chemistry and Biology. Acknowledgements. This work was partially supported by CICYT-MAT2008-06542-C04 and BFU2009-07190 grants of the Spanish Ministry of Science and Innovation, and by the Andalusian Government (BIO-198 and P06-CVI-01713). The authors declare that they have no competing financial interests. 19 Figure 1. Comparison between plastocyanins from Phormidium and from Synechocystis. Robertson diagrams of A) Pho-WT (pdb code: 1baw) and B) Syn-WT (1pcs) plastocyanins (21, 22). Residues are coloured according to conservation patterns from ConSurf (31) phylogenetic computations, which include all the BCP family. Average residues are in white. Colour scale corresponds to residues with increasing variability as compared to average. Red corresponds to highly conserved residues and blue to variable aminoacids. ConSurf (Landau et al., 2005) calculations are unreliable for residues coloured in yellow. Graphics are performed with UCSF Chimera (Petersen et al., 2004). The arrow in A points to the location of the site-directed mutation. C) Detailed view of the copper site. Residues within 5Å from the metal ion are represented in sticks. Those containing first-sphere atoms are labelled. Backbone ribbon is coloured as in A and B. D) Comparison of Tm values among plastocyanin species (Feio et al., 2004; Feio et al., 2006; Muñoz-López et al., 2010b). Figure 2. XANES experiments. Top: Detailed comparison of the experimental XANES spectrum at the Cu K-edge of plastocyanin in Syn-WT (●) and Pho-WT (○). The dotted blue line corresponds to the spectrum recorded in the case of Pho-Mut. In all the cases the zero energy has been taken as E0 = 8981eV. Bottom: Comparison of the experimental XANES spectrum at the Cu K-edge in Syn-WT plastocyanin (●) and the theoretical spectra (green, solid line) calculated by using a complex Dirac-Hara ECP potential within the two channel approach (Chaboy et al., 2005; Chaboy et al., 2006a; Chaboy et al., 2006b). Figure 3. XANES of the oxidised forms. Comparison of the theoretical spectra calculated for Cu in Syn-WT (●) and Pho-WT (○) and those of: a 151 atoms Pho-WT cluster in which the first 15 (blue, dots) and 113 (purple, dash) atoms have been 20 substituted by the structural arrangement of Syn-WT (Top panel); a 151 atoms Pho-WT cluster in which the next-neighbours associated to the histidines (blue, dots), cysteine (red, dash) and methionine (green, dot-dash) within the first 4 Å around the absorbing Cu have been substituted for those found in Syn-WT (Middle panel); and of a 151 atoms Pho-WT cluster in which the N atoms associated to the histidines (black, solid line and red, dots), and the sulphur ones associated to cysteine (green, dash) and methionine (blue, dot-dash) have been substituted for those found in Syn-WT (Bottom panel). Figure 4. XANES of the reduced forms. Top: Detailed comparison of the experimental XANES spectrum at the Cu K-edge of the reduced Pc in Synechocystis (●) and Phormidium (○). The dotted blue line corresponds to the spectrum recorded in the case of Pho-Mut. Bottom: Comparison of the theoretical spectra calculated for Syn-WT (○) and Pho-WT (○) by adopting the structural Cu environment of the oxidised forms and those in which the Cu-S,Cys and Cu-S,Met bond lengths are fixed to intermediate values between those found in the oxidised forms. Figure 1 Click here to download high resolution image Figure 2 Click here to download high resolution image Figure 3 Click here to download high resolution image Figure 4 Click here to download high resolution image Table 1. Melting pointsa (ºC) of BCPs in this work. Protein Oxidised (CuII) species Reduced (CuI) species Pho-WTb 81.8  0.4 75.7  1.0 Pho-Mutb 69.1  1.0 72.3  1.1 Syn-Pcc 65.0  1.0 71.3  1.0 a values obtained by UV fluorimetry under the same buffer conditions used in the X-ray absorption experiments. b Muñoz-López et al. (2010); Feio et al. (2004). c Feio et al. (2006). See Figure 1D for a graphic representation. Table 1 6 68% (3d10L) and 32% (3d9), as previously found3-5 in the case of Cu(II) complexes. This procedure has been applied to all the computations performed for the complete set of different ECP potentials. The results of these double-channel computations are shown in Fig. 2 of the main paper and in Supplementary Fig. S2. They clearly show an improvement in the agreement with the experimental data as compared with the single-channel computations. By contrast only one channel is needed in the case of the reduced Cu(I) species. 7 SUPPLEMENTAL REFERENCES Chaboy, J., and Quartieri, S. (1995) X-ray absorption at the Ca K-edge in natural-garnet solid solutions: A full-multiple-scattering investigation. Phys. Rev. B 52, 6349-6357. Chaboy, J. (2009) Relationship between the structural distortion and the Mn electronic state in La1-xCaxMnO3: a Mn K -edge XANES study. J. Synchrotron Rad. 16, 533-544. Chaboy, J., Muñoz-Páez, A., Carrera, F., Merkling. P., and Sánchez-Marcos, E. (2005) Ab initio x-ray absorption study of copper K-edge XANES spectra in Cu(II) compounds. Phys. Rev. B 71, 134208. Chaboy, J., Muñoz-Páez, A., and Sánchez-Marcos, E. (2006) The interplay of the 3d9 and 3d10L electronic configurations in the copper K-edge XANES spectra of Cu(II) compounds. J. Synchrotron Rad. 13, 471–476. Chaboy, J., Muñoz-Páez, A., Merkling, P., and Sánchez-Marcos, E. (2006) The hydration of Cu2+: Can the Jahn-Teller effect be detected in liquid solution? J. Chem. Phys. 124, 064509. Wu, Z., Benfatto, M., and Natoli, C.R. (1996) Electronic structure and final-state effects in Nd2CuO4, La2CuO4, and Ca2CuO3 compounds by multichannel multiple-scattering theory at the copper K edge. Phys. Rev. B 54, 13409-13412. Natoli, C.R., Benfatto, M., Brouder, C., Ruiz-López, M.F., and Foulis. D.L. (1990) Multichannel multiple-scattering theory with general potentials. Phys. Rev. B 42, 1944- 1968. 8