Role of the surface charges D72 and K8 in the function and structural stability of the cytochrome c6 from Nostoc sp PCC 7119 ´ Christian Lange1, Irene Luque2, Manuel Hervas1, Javier Ruiz-Sanz2, Pedro L Mateo2 and Miguel A De la Rosa1 ´ ´ ´ Instituto de Bioquımica Vegetal y Fotosıntesis, Centro de Investigaciones Cientıficas Isla de la Cartuja, Seville, Spain ´ ´ ´ Departamento de Quımica Fısica e Instituto de Biotecnologıa, Universidad de Granada, Spain Keywords cytochrome c6; electron transfer; electrostatic interactions; protein folding; protein stability Correspondence Christian Lange, Martin-Luther-Universitat ă Halle Wittenberg, Institut fur ă Biotechnologie, Kurt-Mothes-Str 3, 06120 Halle (Saale), Germany Fax: +49 345 552 7013 Tel: +49 345 552 4948 E-mail: christian.lange@biochemtech uni-halle.de (Received 10 February 2005, revised 21 April 2005, accepted May 2005) doi:10.1111/j.1742-4658.2005.04747.x We investigated the role of electrostatic charges at positions D72 and K8 in the function and structural stability of cytochrome c6 from Nostoc sp PCC 7119 (cyt c6) A series of mutant forms was generated to span the possible combinations of charge neutralization (by mutation to alanine) and charge inversion (by mutation to lysine and aspartate, respectively) in these positions All forms of cyt c6 were functionally characterized by laser flash absorption spectroscopy, and their stability was probed by ureainduced folding equilibrium relaxation experiments and differential scanning calorimetry Neutralization or inversion of the positive charge at position K8 reduced the efficiency of electron transfer to photosystem I This effect could not be reversed by compensating for the change in global charge that had been introduced by the mutation, indicating a specific role for K8 in the formation of the electron transfer complex between cyt c6 and photosystem I Replacement of D72 by asparagine or lysine increased the efficiency of electron transfer to photosystem I, but destabilized the protein D72 apparently participates in electrostatic interactions that stabilize the structure of cyt c6 The destabilizing effect was reduced when aspartate was replaced by the small amino acid alanine Complementing the mutation D72A with a charge neutralization or inversion at position K8 led to mutant forms of cyt c6 that were more stable than the wild-type under all tested conditions In the oxygenic photosynthesis of cyanobacteria and many unicellular algae, cytochrome c6 acts as a soluble electron carrier between the cytochrome bf complex and photosystem I (PS I) [1,2] In recent years, considerable efforts have been undertaken to gain insight into the structure–function relationships of this small heme protein [3–7] The structures of the cytochromes c6 from various organisms have been determined [8–14] Study of the mutant series has contributed to our understanding of the functional roles of important individual amino acid residues within the protein [15–17] NMR studies with mutant bacterial c-type cytochromes [18– 20] have provided important insights into the role of heme–protein interactions for the structural stability of this class of proteins The mutant D72K of cytochrome c6 from Nostoc (formerly Anabaena) sp PCC 7119 (cyt c6) [21], was found to be a more reactive electron donor towards PS I than the wild-type [16] However, D72K was shown to have considerably reduced stability against Abbreviations cyt c6, cytochrome c6 from Nostoc sp PCC 7119; DSC, differential scanning microcalorimetry; E m pH , midpoint redox potentials at pH and at 25 °C; DGint, interaction Gibbs energies; kbim, bimolecular reaction rate constant; N, native folded state; PS I, photosystem I; U, ensemble of unfolded states; [urea]50, transition midpoint for urea-induced unfolding FEBS Journal 272 (2005) 3317–3327 ª 2005 FEBS 3317 C Lange et al Surface charge mutants of cytochrome c6 Results and Discussion Functional characterization of wild-type cyt c6 and its mutants Wild-type cyt c6 and its mutant forms were successfully expressed in Escherichia coli strain GM119 and purified to homogeneity The UV ⁄ Vis spectra of wildtype and mutants were virtually identical (not shown) Their midpoint redox potentials at pH (Em pH ) were very similar throughout (Table 1), as expected for the mutation of surface residues located far from the heme moiety Thus, the thermodynamic driving force for the reaction of cyt c6 with PS I was essentially unaltered by the introduction of the mutations Eventual differences in reactivity, therefore, have to be ascribed to structural factors that influence the formation and productivity of the encounter complex Indeed, such differences were observed All forms of cyt c6 were 3318 Table Redox potentials of wild-type cyt c6 and its mutant forms EmpH Wild-type D72A D72K D72N K8A K8D D72K ⁄ K8D D72K ⁄ K8A D72A ⁄ K8D D72A ⁄ K8A (mV) 335 336 339 340 341 336 338 338 331 342 highly active electron donors to PS I (Fig 1) The determined values for the bimolecular rate constant (kbim) for wild-type cyt c6 and the mutant D72K were in excellent agreement with the values that had been reported previously [16] Overall, a good correlation (R ¼ 0.83) between the proteins’ net charge and their reactivity towards PS I was observed (Fig 1) This confirms the importance of long-range charge interactions for the electron carrier function of cyt c6 However, closer inspection of the data revealed interesting deviations from the general trend A reversal of the charge effect of the mutation K8D in the mutant D72K ⁄ K8D did not lead to recovery of wild-type reactivity The same holds for the mutant pair K8A and D72A ⁄ K8A All three variants carrying the mutation K8D showed very similar kbim values, and the same was observed for all variants carrying 30 D72K 25 kbim (107 M-1s-1) unfolding by urea [22], and preliminary experiments indicated that its thermal stability was affected in a similar way Electrostatic interactions play a critical role in guiding and stabilizing functional protein–protein interactions [6], as well as in the structural stability of proteins [23–25] Inspection of a preliminary structural model of cyt c6, based on the known structure of cytochrome c6 from Synechococcus elongatus [11] and NMR data, revealed a spatial proximity between the side chain of D72, which is located in the C-terminal a helix, and a positively charged residue in the protein’s N-terminal helix, K8 In S elongatus cytochrome c6, the conserved residues K8 and D72 are located near the crossing region of the helices, and the Nx atom of K8 is found ˚ at a distance of 6.2 A from the nearest side-chain oxygen of D72 Taking the functionally and structurally interesting mutant D72K as a material starting point, we aimed to study the contribution of the electrostatic charges at positions D72 and K8 to the interaction of cyt c6 with PS I, as well as to its structural stability For this purpose, a series of mutant forms of cyt c6 was generated that spans the possible combinations of charge neutralization (by mutation to alanine) and charge inversion (by mutation to lysine and aspartate, respectively) at positions D72 and K8 After functional characterization by laser-flash absorption spectroscopy, ureainduced folding equilibrium relaxation experiments and differential scanning microcalorimetry (DSC) were used to assess the changes in the stability against denaturant-induced folding ⁄ unfolding as well as against thermal unfolding 20 D72N 15 WT 10 K8D K8A AD D72A AA KA KD -2 -1 relative nominal charge Fig Bimolecular rate constants, kbim, for the reduction of PS I The kbim values were determined as described in the text Data points are marked with the name of the corresponding mutants Error bars represent errors from the fit to determine kbim The solid line represents a linear fit to the data points, dashed lines mark the 95% confidence intervals Double mutants are abbreviated with the first letter representing the residue at position 72 and the second letter the one at position FEBS Journal 272 (2005) 3317–3327 ª 2005 FEBS C Lange et al )2.4 ± 0.2 m, DDGU0 ẳ )2.7 1.6 kJặmol)1) The mutant protein D72N was similarly affected (D[urea]50 ¼ )1.9 ± 0.1 m, DDGU0 ẳ )3.1 2.9 kJặmol)1) All other single mutations, including D72A, had a comparably minor effect Charge neutralization at position K8 has no significant effect on the structural A wild type [urea] final fluo signal (V) the mutation K8A This clearly indicates a specific functional significance of the charge in amino acid position for the formation of a productive encounter complex during the electron transfer from cyt c6 to PS I K8 is located at the edge of the contact surface of cyt c6 that directly interacts with PS I [26], and a positive charge in this position might be expected to play an important role by establishing specific electrostatic interactions during complex formation with PS I The variant carrying the single mutation D72N showed an increased kbim value compared with cyt c6 wild-type, whereas for the mutant D72A, which has the same net charge, no change in kbim was observed This result implies that changes in amino acid position 72 have additional specific effects on the interaction of cyt c6 with PS I, independent of the influence of global net charge Surface charge mutants of cytochrome c6 D72K [urea] final fluo signal (V) C AD [urea] final Preliminary experiments had shown that cyt c6 has maximum stability against unfolding by heat and denaturants around pH 5.5, in its physiological working range [27] Oxidized cyt c6 could not be fully unfolded by urea at pH and 30 °C Therefore, we chose to observe the urea-induced folding ⁄ unfolding of oxidized cyt c6 and its mutant forms at pH The study was carried out by performing folding equilibrium relaxation experiments In Fig 2, the results of typical experiments with wild-type cyt c6 and the mutant forms D72K and D72A ⁄ K8D are shown as examples, along with the derived equilibrium unfolding curves (Fig 2D) Analysis of all performed experiments yielded a set of parameters for comparison of the stabilities of wild-type and mutants (Table 2) As previously reported [22], the transition midpoint for urea-induced unfolding ([urea]50) of the mutant D72K was shifted to significantly lower urea concentrations ([urea]), and its Gibbs energy of unfolding at [urea] (DGU0) was significantly reduced (D[urea]50 ¼ B fluo signal (V) Urea-induced folding ⁄ unfolding of cyt c6: equilibrium stability 0.0 0.3 0.6 0.9 1.2 1.5 10 time (s) D Fig Folding equilibrium relaxation experiments The represented experiments were performed with wild-type cyt c6 (A), D72K (B) and D72A ⁄ K8D (AD) (C) Arrows on the right indicate increasing [urea]final (0–9.8 M) Percentages of unfolded protein at equilibrium as determined from the experiments shown in (A) to (C) for wildtype cyt c6 (squares), D72K (circles) and AD (triangles) are plotted as a function of [urea] in (D) The lines represent fits of two state transitions to the data for wild-type cyt c6 (solid), D72K (dashed) and AD (dotted) Double mutants are abbreviated with the first letter representing the residue at position 72 and the second letter the one at position FEBS Journal 272 (2005) 3317–3327 ª 2005 FEBS % unfolded 100 80 60 40 20 0 [urea] (M) 3319 C Lange et al Surface charge mutants of cytochrome c6 Table Evaluation of urea-induced folding equilibrium relaxation experiments Data represent means ± SD from 2–4 experiments for each protein [Urea]50 (M) Wild-type K8A K8D D72K D72N D72A D72K ⁄ K8D D72K ⁄ K8A D72A ⁄ K8D D72A ⁄ K8A DGU0 (kJỈmol)1) meq (kJỈmol)1ỈM)1) log (kU0Ỉs)1) 5.94 5.8 5.51 3.6 4.05 5.41 4.59 4.27 6.12 6.07 21.4 21.4 20.6 18.7 18.3 21.8 21.6 20.0 24.6 22.0 )3.6 )3.7 )3.74 )5.2 )4.6 )4.0 )4.7 )4.7 )4.0 )3.62 )1.5 )1.3 )1.26 )1.10 )1.13 )1.4 )1.3 )1.1 )1.6 )1.0 ± ± ± ± ± ± ± ± ± ± 0.07 0.2 0.02 0.1 0.06 0.05 0.02 0.03 0.02 0.07 ± ± ± ± ± ± ± ± ± ± 1.3 0.8 0.2 0.3 1.6 1.4 1.5 0.8 1.1 0.4 ± ± ± ± ± ± ± ± ± ± stability of the protein, whereas inversion of the charge at position 72 considerably reduces the stability against urea-induced unfolding, i.e electrostatic interactions of the negative charge at position D72 seem to play an important role for the stability of cyt c6 It is interesting to note that charge neutralization with a hydrogenbond-forming asparagine residue leads to destabilization of the protein, whereas neutralization with a small hydrophobic alanine residue does not Partial or total neutralization of the effect of the mutation D72K on the protein’s global charge was not sufficient to fully reverse the negative effect on the stability of cyt c6 The apparent stability of the mutants D72K ⁄ K8A (D[urea]50 ¼ )1.7 ± 0.1 m, DDGU0 ẳ )1.4 2.1 kJặmol)1) and D72K K8D (KD) (D[urea]50 ¼ )1.4 ± 0.1 m, DDGU0 ¼ +0.2 ± 2.8 kJỈmol)1) against urea-induced unfolding was still reduced when compared with wild-type cyt c6 This indicates that the electrostatic interactions of the side-chain of D72 that contribute to the stability of cyt c6 are at least partially local, and that the adverse effects of the mutations D72K and D72N cannot be entirely ascribed to global electrostatic repulsion within the positively charged protein However, although mutation D72A did not lead to a significant change in stability, the additional elimination of the positive charge K8 in the double mutants D72A ⁄ K8D and D72A ⁄ K8A resulted in an increased apparent stability against urea-induced unfolding, as well as an increased Gibbs energy of unfolding of these mutants (D[urea]50 ¼ +0.2 ± 0.1 m, DDGU0 ẳ +3.5 2.4 kJặmol)1 and D[urea]50 ¼ +0.2 ± 0.1 m, DDGU0 ¼ +0.5 ± 1.7 kJỈmol)1, respectively) The equilibrium urea interaction parameter (meq), which represents the steepness of the unfolding transition with respect to [urea] and may be interpreted as a measure of the change in solvent-accessible area 3320 0.2 0.2 0.02 0.2 0.5 0.2 0.3 0.2 0.2 0.05 ± ± ± ± ± ± ± ± ± ± mk1 (kJỈmol)1ỈM)1) 0.5 0.2 0.09 0.05 0.04 0.5 0.7 0.1 0.5 0.4 mk2 (kJỈmol)1ỈM)2) )1.7 )1.7 )1.6 )2.5 )2.0 )2.3 )2.8 )2.4 )2.4 )1.6 0.03 0.05 0.03 0.11 0.08 0.09 0.1 0.11 0.07 0.04 ± ± ± ± ± ± ± ± ± ± 0.9 0.3 0.2 0.2 0.1 0.7 1.3 0.4 0.9 0.6 ± ± ± ± ± ± ± ± ± ± 0.06 0.03 0.01 0.02 0.01 0.05 0.1 0.04 0.07 0.04 upon unfolding [28], was affected by changes at position D72, but not at position K8 It was significantly increased for all mutants in which position 72 had been modified, with the exception of D72A ⁄ K8A This finding might be partially explained by a prevalence of more extended conformations in the unfolded ensemble of states due to local electrostatic repulsion in the vicinity of the mutated position 72 Urea-induced folding ⁄ unfolding of cyt c6: kinetic parameters The performed folding relaxation experiments also allowed for the determination of kinetic parameters (Table 2) The most thermodynamically unstable mutants, D72N and D72K, were found to have the highest unfolding rate constant in absence of denaturant (kU0 ¼ 0.075 and 0.080 s)1, respectively), while the mutant with the highest DGU0 value, D72A ⁄ K8D, was found to show the lowest unfolding rate (kU0 ¼ 0.023 s)1) The value for wild-type cyt c6 lay between these extremes (kU0 ¼ 0.035 s)1) (Table 2) In general, a good correlation (R ¼ –0.73) was found between log kU0 and the Gibbs energy of unfolding at [urea] (DGU0) In the mutant series, the overall height of the energy barrier for the rate-determining step of the folding ⁄ unfolding transition is mainly determined by the overall difference in Gibbs energy between the native state (N) and the ensemble of unfolded states (U), in agreement with a single transition state with a disordered (unfolded-like) structure [28] This indicates that the folding ⁄ unfolding transition proceeds along a similar pathway for wild-type cyt c6 and for all mutants When the folding ⁄ unfolding traces were analysed individually, and the apparent rate constants (k) for the individual traces were plotted against [urea] in Chevron plots, pronounced deviations from linearity FEBS Journal 272 (2005) 3317–3327 ª 2005 FEBS C Lange et al Microcalorimetry studies In order to obtain the thermodynamic parameters of the thermal unfolding of cyt c6 and its mutants, DSC was performed At pH 7, oxidized cyt c6 was found to be redoxunstable in the absence of oxidizing agents, whose presence would affect the calorimetric trace of the DSC experiment It was also observed that the heating of reduced protein at pH gave rise to complex calorimetric traces, probably due to an oxidative process of the protein upon thermal unfolding It was found that at lower pH values ranging from 3.0 to 5.0, close to the physiological working range of cyt c6, the oxidized form was redox-stable in absence of oxidizing agents during DSC experimental time Consequently, DSC experiments were carried out at pH 3.0, 4.0 and 5.0 for all cyt c6 variants In all cases, heat-induced unfolding was highly reversible and independent of both the scanning rate and the protein concentration All calorimetric traces could be fitted very well to a twostate equilibrium model (NfiU) The experimental DSC curves, together with their best fits, for the thermal unfolding of wild-type cyt c6, and for the mutants D72K and D72A ⁄ K8D at pH 5.0 are shown in Fig 3A as typical examples The effect of pH on thermal unfolding of wild-type cyt c6 is FEBS Journal 272 (2005) 3317–3327 ª 2005 FEBS A Cp (kJ·K-1·mol-1) 60 50 40 30 20 10 30 50 70 90 110 90 110 temperature (ºC) B 60 Cp (kJ·K-1·mol-1) (‘roll over’ and ‘roll down’) were observed (not shown) These deviations were interpreted as indicative of a Hammond shift [29] in the structure of the transition state with increasing denaturant concentration For the global analysis, this was taken into account by introducing the parameter mk2 From the fit parameters, a values were calculated (as a function of urea) according to Eqn (3) (see Experimental procedures) These values may be interpreted as a measure of native-like structure in the transition state [22,28] Whereas a ¼ would indicate an all native-like transition state, a ¼ would indicate an all unfolded-like one Wild-type cyt c6 and all mutant forms showed a transition state shift to higher a-values with increasing [urea] (not shown) The single mutants D72K and D72N, as well as the double mutants D72K ⁄ K8D and D72K ⁄ K8A, showed the steepest increase in a with [urea] (i.e a higher mk2 value), and therefore significantly higher a-values at high denaturant concentrations than wild-type cyt c6 and the other mutant forms Their kinetic properties were clearly more susceptible to the effect of urea and the structure of their transition state shifted more strongly towards the native conformation under the influence of destabilization of the latter Surface charge mutants of cytochrome c6 50 40 30 20 10 30 50 70 temperature (ºC) Fig Temperature dependence of the partial molar heat capacity of cyt c6 (A) DSC traces at pH 5.0 for wild-type (h), D72K (s) and D72A ⁄ K8D (n) (B) DSC traces for wild-type at pH 5.0 (h), pH 4.0 (s) and pH 3.0 (n) Symbols correspond to experimental data, for clarity every third data point is plotted Solid lines correspond to best fits to the two-state equilibrium model shown in Fig 3B The Gibbs energy of unfolding as a function of temperature, DGU(T), could be obtained using the thermodynamic parameter values from the multiple fits of the DSC traces at the studied pH con30  ditions The values of DGU(T) at 30 °C (DGU C ) for )1 wild-type were 54.8 kJỈmol at pH 5.0, 49.3 kJỈmol)1 at pH 4.0 and 34.5 kJỈmol)1 at pH 3.0 It is clear from these results that there is a pronounced pH effect on the stability of cyt c6 Taking into account the value of 21.4 kJỈmol)1 obtained at pH 7.0 by the urea-induced folding ⁄ unfolding experiments (Table 2), it could be confirmed that the maximum stability of wild-type cyt c6 is around pH 5.0 As it is shown in Table 3, this behaviour was common to all studied mutants A similar dependence on the pH was observed for the Tm values obtained for all proteins (Table 3) By contrast, the effect of pH on the unfolding enthalpies 3321 C Lange et al Surface charge mutants of cytochrome c6 Table Thermal unfolding transition of wild-type cyt c6 and its mutant forms For each protein, thermodynamic parameters were determined by global fits of the two-state equilibrium model to the DSC traces recorded at all three pH conditions Tm values refer to the trans ition midpoint and DG U30 C values represent Gibbs energies of unfolding extrapolated to 30 °C The reported values are estimated to be  accurate within ±0.4 °C for Tm and ±10% for DG U30 C The DGU0 values for pH 7.0 from Table are included for comparison 21.4 21.4 20.6 18.7 18.3 21.8 21.6 20.0 24.6 22.0 82.1 80.5 77.6 75.1 75.9 81.8 79.1 77.9 83.6 83.7  C 54.8 53.2 48.7 46.0 50.2 51.5 51.3 50.6 54.7 53.9 seemed to be more complex (Fig 4) For many proteins the dependence of DHU,m on Tm is linear, indicating an entropic origin of the differences in stability [30–32] The slope of such a representation corresponds to DCp,U In our case, as shown in Fig 4, this dependence was not linear for many of the studied forms of cyt c6 Moreover, in the mutants for which a linear dependence was observed, e.g K8D, the obtained slope values were much greater than the DCp,U values calculated from the global analysis of the calorimetric traces As may be observed in Fig 4, the enthalpic variation with pH is most pronounced between pH and for the wild-type and most mutants, with the exception of D72A, for which the biggest pH effect was observed between pH and These results indicate that the stability differences found are not only of entropic nature but that there is a considerable enthalpic contribution Because the ionic strength was fixed in all experiments, the pH effects are exclusively due to changes in the electrostatic interactions The enthalpic contributions of protonation ⁄ deprotonation events are small compared with the changes in enthalpy observed in these mutants (up to 60 kJỈmol)1) So, it is likely that these ionization events are responsible for some kind of structural rearrangement that is associated with enthalpic changes The changes in the thermodynamic parameters of unfolding of each mutant with respect to the wild-type protein are summarized in Fig To avoid extrapolation over a large temperature range, the values of DDGU and DDHU are reported at 75 °C, which is at approximately the median unfolding temperature At all studied pH conditions, wild-type cyt c6 had a significantly higher Gibbs energy of unfolding than the mutants, with the exception of D72A at pH and 3, and of the double mutants D72A ⁄ K8D and 3322 Tm (°C) pH 4.0 DG U30 (kJỈmol)1) 77.3 76.2 74.0 71.9 72.0 78.1 74.5 74.3 79.3 80.1  49.3 48.0 41.7 42.6 42.5 39.0 44.7 46.7 49.8 50.8 C Tm (°C) pH 3.0 DG U30 (kJỈmol)1) 69.6 69.3 67.0 65.0 64.9 72.9 66.5 67.4 73.1 75.0  C 34.5 33.5 29.9 30.3 29.9 36.6 32.0 33.4 35.0 38.6 A 420 ∆HU,m (kJ mol-1) Wild-type K8A K8D D72K D72N D72A D72K ⁄ K8D D72K ⁄ K8A D72A ⁄ K8D D72A ⁄ K8A Tm (°C) pH 5.0 DG U30 (kJỈmol)1) 390 360 330 300 65 70 75 80 85 80 85 Tm (°C) B 420 ∆HU,m (kJ mol-1) pH 7.0 DGU0 (kJỈmol)1) 390 360 330 300 65 70 75 Tm (°C) Fig Temperature dependence of DHU,m Temperature dependence of the unfolding heat effect for wild-type cyt c6 and its mutant forms Symbols correspond to the DHU,m and Tm values obtained by multiple fits of the DSC traces at the three pH conditions for (A) wild-type ( ), K8A (s), D72A ⁄ K8D (n), D72K ⁄ K8D (,) and D72A ⁄ K8A (e), and for (B) K8D ( ), D72K (s), D72N (n), D72A (,) and D72K ⁄ K8A (e) The lines correspond to the DHU(T) functions obtained by multiple fits for wild-type (A) and for K8D (B) at pH 5.0 (solid line), pH 4.0 (dashed line) and pH 3.0 (dotted line) FEBS Journal 272 (2005) 3317–3327 ª 2005 FEBS C Lange et al Surface charge mutants of cytochrome c6 A B C K8D K8D K8D K8A K8A K8A D72K D72K D72K D72N D72N D72N D72A D72A D72A KD KD KD KA KA KA AD AD AD AA AA AA 10 ∆∆G U 75 ºC -5 (kJ mol-1) -4 100 50 75 ºC ∆∆H U ∆T (K) m -50 (kJ mol-1)  Fig Effect of mutations on the thermodynamic parameters of unfolding Difference of the thermodynamic parameters DDG U75 C (A), Tm  (B) and DDH U 75 C between wild-type and each mutant Black bars correspond to pH 5.0, grey bars to pH 4.0 and light grey bars to pH 3.0 Double mutants are abbreviated with the first letter representing the residue at position 72 and the second letter the one at position D72A ⁄ K8A at all three pH values (Fig 5A) This tendency was also observed when comparing the Tm values (Fig 5B) These results are in good agreement with the apparent order of stabilities obtained from the urea-induced unfolding experiments (Table 3) It is interesting to note that the stabilizing effect of substituting D72 to alanine (in single and double mutants) was neither observed when the position 72 was mutated to a charged lysine residue, nor when it was changed to the neutral and hydrogen bond forming asparagine With respect to DDH U 75  C values (Fig 5C) two groups of mutants could be distinguished The point mutations at position 8, for which no significant enthalpic difference was observed, and the rest (where position 72 is mutated), which presented lower enthalpies of unfolding than the wild-type protein Inspection of the values obtained for the Gibbs energies of unfolding of the mutant forms of cyt c6 revealed significant nonadditive effects in the stabilities of the double mutants when compared with the stabilities of the respective single mutants The interaction Gibbs energies (DGint) at 30 °C for the different double mutants were estimated by double-mutant cycle analysis [33,34] and are summarized in Table At pH 5.0, the DGint values obtained when K8 was substituted by aspartate were % kJỈmol)1 higher than when K8 was substituted by alanine, independent of the nature of the substitution at amino acid position 72 This additional interaction energy was reduced to kJỈmol)1 (for the double mutants of D72K ⁄ K8D and D72K ⁄ K8A) or even abolished (for the double FEBS Journal 272 (2005) 3317–3327 ª 2005 FEBS Table Double-mutant cycle analysis Interaction Gibbs energies  at 30 °C (DGint) were calculated from the DG U30 C -values reported  in Table according to DGint (D72X ⁄ K8Y) ¼ DG U30 C (wild30  C 30  C 30  C type) + DG U (D72X ⁄ K8Y) – DG U (D72X) – DG U (K8Y) DGint (kJỈmol)1) pH 5.0 D72K ⁄ K8D D72K ⁄ K8A D72A ⁄ K8D D72A ⁄ K8A pH 3.0 DGintpH5 – DGintpH3 (kJỈmol)1) 11.4 6.2 9.3 4.0 6.3 4.1 3.0 3.0 5.1 2.1 6.3 1.0 mutants D72A ⁄ K8D and D72A ⁄ K8A) at pH 3.0 Assuming a standard pKa value for the carboxyl group, the aspartate residue introduced at position would be almost completely neutralized at pH 3.0 Thus, replacement of K8 with a negatively charged residue resulted in stronger destabilizing interactions than a neutral residue, when the negative charge at position 72 was removed The influence of the nature of the introduced mutations and the complex pH dependence of the DGint values for the different mutant pairs strongly suggest that much more complex effects (multiple interactions and ⁄ or structural rearrangements) than a direct electrostatic interaction between the residues at amino acid positions and 72 play a role A refined model for the structure of cyt c6, incorporating NMR data, has become available during revision of this work [35] In this model (Fig 6), K8 is located relatively far from other charged residues, and its side chain is highly solvent exposed It forms backbone hydrogen bonds 3323 C Lange et al Surface charge mutants of cytochrome c6 Fig Structural model of cyt c6, showing secondary structure elements The side chains of amino acids K8 (blue), E68 (light red), Q69 (light green) and D72 (red), as well as the heme moiety (orange), are depicted as stick models The H-bond from the backbone amide group of G6 to the carboxyl oxygen of D72 is shown as dotted green line The image was generated with PYMOL 0.97 (DeLano Scientific, San Carlos, CA, USA) and is based on the structural model from [35] along the N-terminal a helix involving the amino acids V4 ⁄ N5, and S11 ⁄ A12 The side chain of D72 is part of an acidic patch, together with the residues D2 and E68 It forms backbone hydrogen bonds along the C-terminal a helix to E68 and Y76 Interestingly, one of the side chain oxygens of D72 appears to be hydrogen-bonded to the backbone N atom of G6, bridging the crossing between the N- and C-terminal a helices of cyt c6 (Fig 6) Any mutation leading to charge neutralization or inversion of a residue within this helix-crossing region may cause rearrangements and a reorganization of the local hydrogen-bond network It is, for example, conceivable that upon replacement of D72 by alanine one of the residues E68 or Q69 ‘takes over’ in forming a stabilizing hydrogen bond to an amino acid within a reoriented N-terminal a helix However, any such interpretation of the experimental data remains speculative until reliable structural information for cyt c6 and its mutant forms becomes available Conclusions The results of this study stress the importance of electrostatic interactions for the function, as well as for the stability, of cyt c6 Neutralization or inversion of the positive charge at position K8 significantly reduced 3324 the efficiency of electron transfer to PS I This effect was not reversed when the global charge of the protein was restored by additional mutations at position 72 This implies that residue K8 plays a role in the formation of the electron transfer complex between cyt c6 and PS I However, replacement of the negative charge at position D72 by asparagine or lysine increased the efficiency of electron transfer to PS I, at least partly, by favouring long-range electrostatic attraction between the reaction partners These mutations, however, were found to destabilize the protein significantly Again, this effect of the mutations could not be fully reversed by restoring the global charge balance The negative charge at position D72 apparently participates in electrostatic interactions that stabilize the structure of cyt c6, as indicated by the reduced enthalpy of unfolding of the corresponding mutants Interestingly, the negative effect of its deletion was reduced, due to entropic contributions, when aspartate was not replaced by the isosteric asparagine, or by lysine, but by the small amino acid alanine For the variants carrying the mutation D72A, the positive change in entropy upon unfolding was reduced with respect to other variants of cyt c6 This smaller unfolding entropy could be due to the effect of the mutation on the native state (higher flexibility and ⁄ or lower exposure of hydrophobic area to the solvent in the mutant D72A) and ⁄ or on the unfolded state (lower flexibility and ⁄ or higher solvation of hydrophobic surface area) Complementing the mutation D72A with a charge neutralization or inversion at position K8 led to mutant forms of cyt c6 that were more stable than the wildtype under all tested conditions Of all studied forms of cyt c6, AD was the thermodynamically most stable at pH 7, whereas AA was the most stable at pH and The price for this stabilization, however, was a reduction in catalytic efficiency by > 50% Experimental procedures Protein expression and purification Expression plasmids pEACwt and pEACD72K were kindly provided by F.P Molina-Heredia [16,36] Expression plasmids for the other mutants were generated by site-directed mutagenesis according to the QuikChange method (Stratagene, La Jolla, CA) using pEACwt as template Cyt c6 wild-type and its mutant forms were expressed under aerobic conditions in E coli GM119 cells that had been cotransfected with the plasmid pEC86 [37] (kindly provided by L Thony-Meyer, ETH Zurich) and were puried as desă ă cribed previously [16], with the exception of the mutants K8D, K8A and D72A ⁄ K8D (AD) These proteins were FEBS Journal 272 (2005) 3317–3327 ª 2005 FEBS C Lange et al purified by a combination of anion-exchange chromatography and gel filtration In brief, the periplasmic extracts containing the mutants K8D, K8A or AD were dialysed against 0.5 mm sodium phosphate buffer pH 7, oxidized by the addition of 50 lm potassium ferricyanide, and loaded onto a DEAE cellulose column (50 mL) that had been equilibrated with a mm sodium phosphate buffer, pH The mutant proteins were found in the flow-through After concentration under nitrogen pressure in an Amicon 8050 ultrafiltration device fitted with YM membranes (Millipore, Bedford, MA), the fractions containing the mutant forms of cyt c6 were applied to a Superdex 75 XK16 ⁄ 60 FPLC column The running buffer for the gel filtration was 10 mm Tricine ⁄ KOH pH 7.5, 100 mm NaCl After chromatography, the purified proteins were washed with 10 mm Tricine ⁄ KOH pH 7.5 in the ultrafiltration device, concentrated to 1–2 mm, and stored at )80 °C The purity of all preparations was > 95% as judged from the optical absorbance ratio at 554 and 280 nm, and from Coomassie Brilliant Blue-stained SDS ⁄ PAGE Protein concentrations were determined based on an absorptivity e553 of 26 200 m)1 cm)1 [36] for fully reduced cyt c6 Surface charge mutants of cytochrome c6 order to keep the protein in its oxidized state throughout the folding ⁄ unfolding experiment, were mixed in varying ratios (r) in a lSFM20 stopped-flow device (BioLogic SA, Grenoble, France) with a constant total flow rate of 5.6 mL s)1 in a constant total volume of 560 lL Protein folding ⁄ unfolding was monitored by tryptophan fluorescence Kinetic traces were recorded in duplicate or triplicate and averaged for analysis The data for all mutants were fitted globally to a two-state model assuming linear dependence of the Gibbs energy of unfolding (DGU) on the urea concentration ([urea]) according to DGU ẳ DGU0 ỵ meq ẵurea 1ị with DGU at [urea], DGU0, and the urea interaction parameter meq, and assuming a second-order dependence of the logarithm of the unfolding rate constant (kU) on [urea], according to ln kU ¼ ln kU0 mk1 ẵurea ỵ mk2 ẵurea2 ị; RT 2ị with kU at [urea], kU0, and the first- and second-order urea dependencies mk1 and mk2 To estimate the ureadependent shift in transition state structure, a values were calculated from the fit parameters according to Redox potentials Redox potentials were determined by potentiometric titration with sodium dithionite and potassium ferricyanide as described previously [38] The determinations were carried out in 50 mm sodium phosphate buffer pH and at 25 °C a¼ meq À ðmk1 ỵ mk2 ẵureaị : meq 3ị The global fitting of the data was carried out with the nonlinear analysis module of origin 5.0 (Microcal, Northampton, MA, USA) Laser-flash absorption spectroscopy Experiments were carried out as described previously [3] The reaction mixture was buffered with 20 mm Tricine ⁄ KOH at pH 7.5 and contained 10 mm MgCl2 The temperature was 25 °C For the determination of bimolecular rate constants (kbim), the kinetics of re-reduction of PS I from Nostoc sp PCC 7119 were measured at increasing concentrations of wild-type cyt c6 or of its mutant forms The kbim values were determined from the slope of a linear fit of the observed rate constants (kobs), plotted against donor concentration For most forms of cyt c6, fast kinetic phases were not observed, and it was possible to analyse all kinetic according to a simple bimolecular reaction mechanism without introducing significant deviations in kobs Stopped-flow experiments Folding equilibrium relaxation experiments were carried out as described previously [22] In brief, equal concentrations of cyt c6 (wild-type or mutant forms) in 20 mm sodium phosphate pH 7.0 (buffer) and in 9.8 m urea buffered with 20 mm sodium phosphate pH 7.0 (buffered urea solution), both containing 50 lm potassium ferricyanide in FEBS Journal 272 (2005) 3317–3327 ª 2005 FEBS Microcalorimetry The heat capacity of wild-type cyt c6 and its mutant forms was measured as a function of temperature with a VP-DSC differential scanning microcalorimeter (Microcal) Samples were dialysed overnight against the desired buffer and subsequently oxidized by addition of potassium ferricyanide to a final concentration of 2.5 mm Immediately before loading the samples into the calorimetric cell, ferricyanide was removed by gel filtration on a PD-10 column (Amersham Biosciences, Little Chalfont, UK) The buffers used in the experiments were 60 mm sodium acetate, pH 5.0, 250 mm sodium acetate, pH 4.0, or 45 mm sodium phosphate, pH 3.0, all giving rise to the same ionic strength (41 mm) as the 20 mm sodium phosphate buffer, pH 7, which was used in the urea-induced folding ⁄ unfolding experiments DSC experiments were performed at a heating rate of 1.5 KỈmin)1 between and 100 °C, using protein concentrations from 10 to 50 lm (0.1–0.5 mgỈmL)1) Baseline scans were performed with the corresponding dialysis buffer loaded in both calorimetric cells The partial molar heat capacity (Cp) was calculated assuming a value of 0.73 mLỈg)1 for the partial specific volume of the proteins 3325 Surface charge mutants of cytochrome c6 After transforming the DSC traces into partial molar heat capacity curves, they were subjected to individual and global curve fitting using the nonlinear analysis module of origin 5.0 with user-defined fit functions based on the equations corresponding to a two state model (NfiU) The temperature dependence of the heat capacity was described as a linear and quadratic function for the native and unfolded species, respectively [39] The corresponding parameters were left to float during the curve fitting, with the exception of the first- and second-order coefficients of the unfolded heat capacity, which were fixed and evaluated from the amino acid content as described previously [40,41] For each form of cyt c6, the values of the change in molar heat capacity upon unfolding (DCp,U) were determined by global fitting of the DSC traces recorded at all three studied pH values Chemicals C Lange et al Urea was SigmaUltra grade All other chemicals were at least analytical grade All solutions were prepared with MilliQ water Acknowledgements This work was supported by the Spanish Ministry of Science and Technology (MCYT grants BMC 2000444 and BIO2003-04274) and the Junta de Andalucı´ a (PAI, CVI-198) CL received a fellowship from the European Union’s Research Training Network program (HRPN-CT-1999-00095) IL was supported by a research contract from the University of Granada and ´ is the recipient of a Ramon y Cajal research contract from the Spanish Ministry of Science and Technology 10 11 12 References Wood PM (1978) Interchangeable cooper and iron proteins in algal photosynthesis: studies on plastocyanin and cytochrome c552 in Chlamydomonas Eur J Biochem 87, 9–19 Ho KK & Krogmann DW (1984) Electron donors to P700 in cyanobacteriae and algae: an instance of unusual genetic variability Biochim Biophys Acta 766, 310–316 ´ Hervas M, Navarro JA, Dı´ az-Quintana A, Bottin H & De la Rosa MA (1995) Laser-flash kinetic analysis of the fast electron transfer from plastocyanin and cytochrome c6 to photosystem I Experimental evidence on the evolution of the reaction mechanism Biochemistry 34, 11321–11326 ´ Hervas M, Navarro JA, Dı´ az-Quintana A & De la Rosa MA (1996) A comparative thermodynamic analysis by laser-flash absorption spectroscopy of photosystem I 3326 13 14 15 16 reduction by plastocyanin and cytochrome c6 in Anabaena PCC 7119, Synechocystis PCC 6803 and spinach Biochemistry 35, 2693–2698 Ullman GM, Hauswald M, Jensen A, Kostic NM & Knapp EW (1997) Comparison of the physiologically equivalent proteins cytochrome c6 and plastocyanin on the basis of their electrostatic potentials Tryptophan 63 in cytochrome c6 may be isofunctional with tyrosine 83 in plastocyanin Biochemistry 36, 16187–16196 Crowley PB, Diaz-Quintana A, Molina-Heredia FP, Nieto P, Sutter M, Haehnel W, De La Rosa MA & Ubbink M (2002) The interactions of cyanobacterial cytochrome c6 and cytochrome f, characterized by NMR J Biol Chem 277, 48685–48689 Dikiy A, Carpentier W, Vandenberghe I, Borsari M, Safarov N, Dikaya E, Van Beeumen J & Ciurli S (2002) Structural basis of the molecular properties of cytochrome c6 Biochemistry 41, 14689–14699 Frazao C, Soares CM, Carrondo MA, Pohl E, Dauter Z, ˜ ´ Wilson KS, Hervas M, Navarro JA, De la Rosa MA & Sheldrick GM (1995) Ab initio determination of the crystal structure of cytochrome c6 and comparison with plastocyanin Structure 3, 1159–1170 Kerfeld CA, Anwar HP, Interrante R, Merchant S & Yeates TO (1995) The structure of chloroplast cyto˚ chrome c6 at 1.9 A resolution Evidence for functional oligomerization J Mol Biol 250, 627–647 Banci L, Bertini I, De la Rosa MA, Koulougliotis D, Navarro JA & Walter O (1998) Solution structure of oxidized cytochrome c6 from the green alga Monoraphidium braunii Biochemistry 37, 4831–4843 Beißinger M, Sticht H, Sutter M, Ejchart A, Haehnel W & Rosch P (1998) Solution structure of cytochrome c6 from the thermophilic cyanobacterium Synechococcus elongatus EMBO J 17, 27–36 Schnackenberg J, Than ME, Mann K, Wiegand G, Huber R & Reuter W (1999) Amino acid sequence, crystallization and structure determination of reduced and oxidized cytochrome c6 from the green alga Scenedesmus obliquus J Mol Biol 290, 1019–1030 Yamada S, Park S-Y, Shimizu H, Koshizuka Y, Kadokura K, Satoh T, Suruga K, Ogawa M, Isogai Y, Nishio T et al (2000) Structure of cytochrome c6 from the red ˚ alga Porphyra yezoensis at 1.57 A resolution Acta Crystallogr D56, 1577–1582 Sawaya MR, Krogmann DW, Serag A, Ho KK, Yeates TO & Kerfeld CA (2001) Structures of cytochrome c549 and cytochrome c6 from the cyanobacterium Arthrospira maxima Biochemistry 40, 9215–9225 ´ De la Cerda B, Dı´ az-Quintana A, Navarro JA, Hervas M & De la Rosa MA (1999) Site-directed mutagenesis of cytochrome c6 from Synechocystis sp PCC 6803 J Biol Chem 274, 13292–13297 ´ Molina-Heredia FP, Dı´ az-Quintana A, Hervas M, Navarro JA & De la Rosa MA (1999) Site-directed FEBS Journal 272 (2005) 3317–3327 ª 2005 FEBS C Lange et al 17 18 19 20 21 22 23 24 25 26 27 28 29 mutagenesis of cytochrome c6 from Anabaena species PCC 7119 J Biol Chem 274, 33565–33570 ´ Molina-Heredia FP, Hervas M, Navarro JA & De la Rosa MA (2001) A single arginyl residue in plastocyanin and in cytochrome c6 from the cyanobacterium Anabaena sp PCC 7119 is required for efficient reduction of photosystem I J Biol Chem 276, 601–605 Banci L, Bertini I, Bren KL, Gray HB & Turano P (1995) pH-Dependent equilibria of yeast Met80Alaiso-1-cytochrome c probed by NMR spectroscopy: a comparison with the wild type protein Chem Biol 2, 377–383 Bartalesi I, Bertini I, Di Rocco G, Ranieri A, Rosato A, Vanarotti M, Vasos PR & Viezzoli MS (2004) Protein stability and mutations in the axial methionine loop of a minimal cytochrome c J Biol Inorg Chem 9, 600–608 Wen X & Bren KL (2005) Suppression of axial methionine fluxion in Hydrogenobacter thermophilus Gln64Asn cytochrome c552 Biochemistry 44, 5225–5233 Medina M, Louro RO, Gagnon J, Peleato ML, Mendes ´ J, Gomez-Moreno C, Xavier AV & Teixeira M (1997) Characterization of cytochrome c6 from the cyanobacterium Anabaena PCC 7119 J Biol Inorg Chem 2, 225– 234 ´ Lange C, Hervas M & De la Rosa MA (2003) Analysis of the stability of cytochrome c6 with an improved stopped flow protocol Biochem Biophys Res Commun 310, 217–223 Nakamura H (1996) Roles of electrostatic interactions in proteins Q Rev Biophys 29, 1–90 Mozo-Vilları´ as A, Cedano J & Querol E (2003) A simple electrostatic criterion for predicting the thermal stability of proteins Protein Eng 16, 279–286 Torrez M, Schultehenrich M & Livesay DR (2003) Conferring thermostability to mesophilic proteins through optimized electrostatic surfaces Biophys J 85, 2845– 2853 Dı´ az-Moreno I, Dı´ az-Quintana A, Molina-Heredia FP, Nieto PM, Hansson O, De la Rosa MA & Karlsson BG (2005) NMR analysis of the transient complex between membrane photosystem I and soluble cytochrome c6 J Biol Chem 280, 7925–7931 Kramer DM, Sacksteder CA & Cruz JA (1999) How acidic is the lumen? Photosynth Res 60, 151–163 Tanford C (1970) Protein denaturation C Theoretical models for the mechanism of denaturation Adv Protein Chem 24, 1–95 Hammond GS (1955) A correlation of reaction rates J Am Chem Soc 77, 334–338 FEBS Journal 272 (2005) 3317–3327 ª 2005 FEBS Surface charge mutants of cytochrome c6 30 Privalov PL (1979) Stability of proteins Small globular proteins Adv Protein Chem 33, 167–241 31 Ruiz-Sanz J, Simoncsits A, Toro I, Pongor S, Mateo ă ă PL & Filimonov VV (1999) A thermodynamic study of the 434-repressor N-terminal domain and of its covalently linked dimers Eur J Biochem 263, 246–253 32 Martı´ nez JC, Viguera AR, Bariso R, Willmanns M, Mateo PL, Filimonov VV & Serrano L (1999) Thermodynamic analysis of a-spectrin SH3 and two of its circular permutants with different loop lengths: discerning the reasons for rapid folding in proteins Biochemistry 38, 549–559 33 Serrano L, Bycroft M & Fersht AR (1991) Aromatic– aromatic interactions and protein stability: investigation by double-mutant cycles J Mol Biol 218, 465–475 34 Horovitz A & Fersht AR (1992) Co-operative interactions during protein folding J Mol Biol 224, 733–740 35 Dı´ az-Moreno I, Dı´ az-Quintana A, Ubbink M & De la Rosa MA (2005) An NMR-based docking model for the physiological transient complex between cytochrome f and cytochrome c6 FEBS Lett, doi:10.1016/j.febslet 2005.04.031 ´ 36 Molina-Heredia FP, Hervas M, Navarro JA & De la Rosa MA (1998) Cloning and correct expression in Escherichia coli of the petE and petJ genes respectively encoding plastocyanin and cytochrome c6 from the cyanobacterium Anabaena sp PCC 7119 Biochem Biophys Res Commun 243, 302–306 37 Arslan E, Schulz H, Zufferey R, Kunzler P & Thonyă ă Meyer L (1998) Overproduction of the Bradyrhizobium japonicum c-type cytochrome subunits of the cbb3 oxidase in Escherichia coli Biochem Biophys Res Commun 251, 744–747 ´ 38 Ortega JM, Hervas M & Losada M (1988) Redox and acid–base characterization of cytochrome b-559 in photosystem II particles Eur J Biochem 171, 449–455 39 Privalov PL, Tiktopulo EI, Venyaminov SYu, Griko YuV, Makhatadze GI & Kechinashvili NN (1989) Heat capacity and conformation of proteins in the denatured state J Mol Biol 205, 737–750 40 Privalov PL & Makhatadze GI (1990) Heat capacity of proteins II Partial molar heat capacity of the unfolded polypeptide chain of proteins: protein unfolding effects J Mol Biol 213, 385–391 ´ 41 Cobos ES, Filimonov VV, Galvez A, Maqueda M, Valdivia E, Martı´ nez JC & Mateo PL (2001) AS-48: a circular protein with an extremely stable globular structure FEBS Lett 505, 379–382 3327 ... protein–protein interactions [6], as well as in the structural stability of proteins [23–25] Inspection of a preliminary structural model of cyt c6, based on the known structure of cytochrome c6 from. .. oxygen of D72 Taking the functionally and structurally interesting mutant D72K as a material starting point, we aimed to study the contribution of the electrostatic charges at positions D72 and K8. .. traces at the three pH conditions for (A) wild-type ( ), K8A (s), D72A ⁄ K8D (n), D72K ⁄ K8D (,) and D72A ⁄ K8A (e), and for (B) K8D ( ), D72K (s), D72N (n), D72A (,) and D72K ⁄ K8A (e) The lines