Enzyme repurposing of a hydrolase as an emergent peroxidase upon metal binding† †Electronic supplementary information (ESI) available: Experimental details, additional characterization and catalytic data. See DOI: 10.1039/c5sc01065a

Adding a metal cofactor to a protein bearing a latent metal binding site endows the macromolecule with nascent catalytic activity.

Crystallization of 6-PGLac (Apo-form without CuSO 4 ) and Cu·6-PGLac (6-PGLac with CuSO 4 ). Crystals of both forms were obtained by the sitting drop vapor diffusion method at 20 °C. Crystallization droplets were prepared by mixing the precipitant solution (0.2 µl) and protein solution (0.2 µl). 10 and 25 mg/ml solution were used for without and with CuSO 4 , respectively. Precipitant contains Mix A (20% polyethylene glycol 1000, 0.2 M Li 2 SO 4 , and 0.1 M phosphate/citrate (pH 4.2)) and Mix B (22.5% polyethylene glycol 4000, 15% glycerol, 150 mM (NH 4 ) 2 SO 4 , and 3 mM CuSO 4 ) for without and with CuSO 4 , respectively. Crystals of 6-PGLac were grown in 3 days and were cryoprotected in a reservoir solution supplemented with ethylene glycol in a stepwise fashion (ethylene glycol final concentration (v/v) of 10%, 15%, 20% and 25%, 5 min each) and directly transferred into liquid nitrogen. Crystals of Cu·6-PGLac were obtained in 4 days and were cryoprotected in a reservoir solution supplemented with 3 mM CuSO 4 and 20 mM Tris-HCl (pH 8.0) and 0.15 M NaCl, and ethylene glycol in a stepwise fashion (ethylene glycol final concentration (v/v) of 10%, 15%, 20% and 25%, 5 min each). The crystals were subsequently soaked into the same solutions with 12 mM CuSO 4 for 2h and directly transferred into liquid nitrogen to freeze them.
Data Collection of X-Ray Diffraction and Structure Determination. All diffraction data were collected at the PXI-Xo6SA and the PXIII-Xo6DA beamlines at the Swiss light source (SLS) synchrotron facility at the Paul Scherrer Institute (Villigen, Switzerland). X-ray diffraction images were collected on pilatus 2M and 6M detectors equipped with a cryo-system at 100 K using a wavelength of 1.0 Å. The data were processed and scaled with the XDS program package. The data collection and refinement statistics are summarized in Table S4. The two crystal structures of 6-PGLac were solved by molecular replacement with Phaser using the model of the reported coordinate (PDB code: 3Q6C) as search model. The resulting model covered 243 residues of the one subunit in an asymmetric unit with 243 side chains assigned. The structure model was manually rebuilt with the program COOT 3 followed by refinement calculations with the program Phenix.refine. 4 This procedure was iterated until the model did not further improve. In the structures, Ramachandran analysis with the Molprobity program showed no outliers. 5 The final structure of 6-PGLac and Cu·6-PGLac contains residues Ala0-Leu244 and Ser2-Leu244 respectively. The electron density at the N-terminal Strep-tag II regions was too weak/absent to allow reliable building of residues Trp(-10)-Lys(-3). The missing residues and atoms are summarized in Table S5. The anomalous difference Fourier maps were calculated from the data set of Cu·6-PGLac using the Phenix suite. All figures of protein structures were prepared using PyMOL (Version 1.3 Schrödinger, LLC).
Substrate Docking Simulation. The crystal structure coordinations of Cu·6-PGLac (PDB code: 4TM7) and o-dianisidine (PDB ligand entry: DDJ) were used in the substrate-protein docking studies. Hydrogens were added to the Cu·6-PGLac structures. The o-dianisidine was docked into cavity around Cu1 with GOLD Suite. All of the side chains were not allowed to rotate during docking and twenty dockings were visually inspected.
Michaelis-Menten Kinetics. The steady-state kinetic experiments were carried out in 2 mL disposable cuvettes and the absorbance was measured with a Varian Cary 50 Bio UV/Vis Spectrophotometer every 4 sec at room temperature (~ 25 °C). All experiments were performed in triplicates. The reaction mixture was prepared by adding CuSO 4 (210 µL, 250 µM) and t-butyl hydroperoxide (t-BuOOH 800 µL, 100 mM) to the protein solution (12 mL, 1.25 µM) in MESbuffer (pH 6.5, 50 mM, 150 mM NaCl). To trigger the reactions, various amounts of o-dianisidine solutions were added (3.3, 6.6, 10, 13, 17, 20, 26 µL of 2.5 mM and 10, 13, 17, 25 µL of 10 mM). The amount of product was determined using the molar extinction coefficient at 460 nm ε 460 nm = 11.3 cm -1 ·mM -1 . 6, 7 The reaction velocity was calculated from the initial slope in triplicate. Steadystate kinetic parameters (catalytic turnover number (k cat ) and Michaelis constant (K M ) were determined by nonlinear fitting of Michaelis-Menten plots. The background reaction caused by free copper was subtracted from the raw data for fitting purposes (Fig. S12B).
Detection of Native Lactonase Activity. The NMR experiments were recorded with a 600 MHz Bruker spectrometer. The detection of the lactonase activity was performed as described previously. 10 500 µL of a solution containing NADP + (3 mM), Glucose-6-phosphate (20 mM), MgCl 2 (5 mM) and CuSO 4 (12 µM) in TEA-buffer (pH 6.7, 1 M) was prepared for the first NMR experiment (t = 0 min). After 10 min a second NMR experiment was done (t = 10 min). Then Glucose-6-phosphate dehydrogenase (final concentration, 2 µM) was added and the NMR experiments were repeated (up to 20 min). Finally, 6-PGLac (final concentration, 3.5 µM) was added and another NMR experiments were measured (up to 45 min).
Electron Paramagnetic Resonance Spectroscopy. The EPR measurements at 9 GHz were performed with a JEOL JES-FA100 by using highly purified quartz tubes (I.D. 4.0 mm, Radical Research Inc.) with a rubber septum under an N 2 saturated conditions at 77 K. The microwave power and the modulation amplitude were 1 mW and 0.3 mT, respectively. 500 µM CuSO 4 was added into 1 mM (final) protein solution. The spectral simulation was carried out using a simulation software, JEOL AniSimu/FA ver 2.0.0. The simulation parameters are given in Fig.  S14C.
To investigate the effect of Cu 2+ ion its native activity, the experiment was repeated in the presence of 12 µM CuSO 4 in an NMR-tube. Under these conditions, the conversion from the 6phosphogluconolactone to 6-phosphogluconate upon addition of 6-PGLac was still apparent.  Close-up views of two fully occupied copper binding sites (green spheres). (C) Other unspecific bound copper binding sites (orange spheres). The crystal structure of 6-PGLac was determined by molecular replacement using the structure of the reported 6-phosphogluconolactonase (PDB code: 3OC6) as a search model. 6-PGLac (apo form) was crystallized with one monomer in the crystallographic asymmetric unit (space group I4 1 22) and the final structure was refined to a resolution of 1.81 Å with excellent statistics (Table S4). The protomer consists of 256 amino acid residues deduced from DNA sequence and MS analysis (Fig. S4F) and most of the residues were well defined in the final electron density, except for some disordered loops in the N-terminal flexible regions (mainly Strep-tag II, Table S5). The overall structure is nearly identical to the reported structure (PDB code: 3OC6, RMSD = 0.57 (243 Cα atoms)).
In the crystal co-crystallized with 3 mM CuSO 4 and then soaked with 12 mM CuSO 4 (Cu·6-PGLac), there is one monomer in the crystallographic asymmetric unit (space group P3 2 12). The final structure was refined to a resolution of 1.39 Å with excellent statistics. In this crystal structure, most of the residues were also well defined in the final electron density, except for some disordered loops in the N-terminal flexible regions (mainly Strep-tag II, Table S5) as it is the case with 6-PGLac. Three copper ions were identified by strong peaks in anomalous difference Fourier map contoured at 15 σ as threshold value including in the putative metal binding site. The alternative positions for copper binding (Cu2 and Cu3, 22 Å and 32 Å apart from Cu1) were observed in the crystal lattice contacts (Fig. S9B). On the edge of the pseudo βsheet structure, Cu2 was supported by carboxyl His9 and Asp3' (prime refers to residues from the adjacent protomer). Cu3 was also located on the interface between protomers in the different crystal lattice and supported by H95, Asp158' and Asp220'. Around Cu2, a tight inter-protomer interaction was observed. These two protomers are likely to be held together through 4-hydrogen bonds between Asp3-Arg8 forming an anti-parallel pseudo-β sheet structure on each other. Under physiological conditions, 6-PGLac was found to exist mainly as a monomer.
Detailed scrutiny of the anomalous difference Fourier map contoured at 3 σ as threshold value, 13 other copper binding sites were found on the protein surface including the lactonase active site (Cu7 , Table S6 and Fig S9C). All of these copper ions were > 12 Å far from Cu1 and have much lower occupancies and higher isotropic B-factors (0.299 and 42.7 (average), respectively) as compared to those of the three copper binding sites described above (0.970 and 24.5 (average), respectively). The detailed B-factors of each copper are summarized in Table S6.  The determined masses are almost identical to the calculated masses derived from the corresponding amino acid sequence without the initial Met (Fig. S2). Analysis based on a two metal binding scheme provides a good fit with the following dissociation constants: K d1 = 0.83 ± 0.11 µM, K d2 = 130 ± 3.3 µM. The corresponding fluorescence quenching values (ΔF 1 and ΔF 2 ) are 22 ± 1.0 % and 57 ± 4.1 %, respectively.