Active-site plasticity revealed in the asymmetric dimer of AnPrx6 the 1-Cys peroxiredoxin and molecular chaperone from Anabaena sp. PCC 7120

Peroxiredoxins (Prxs) are vital regulators of intracellular reactive oxygen species levels in all living organisms. Their activity depends on one or two catalytically active cysteine residues, the peroxidatic Cys (CP) and, if present, the resolving Cys (CR). A detailed catalytic cycle has been derived for typical 2-Cys Prxs, however, little is known about the catalytic cycle of 1-Cys Prxs. We have characterized Prx6 from the cyanobacterium Anabaena sp. strain PCC7120 (AnPrx6) and found that in addition to the expected peroxidase activity, AnPrx6 can act as a molecular chaperone in its dimeric state, contrary to other Prxs. The AnPrx6 crystal structure at 2.3 Å resolution reveals different active site conformations in each monomer of the asymmetric obligate homo-dimer. Molecular dynamic simulations support the observed structural plasticity. A FSH motif, conserved in 1-Cys Prxs, precedes the active site PxxxTxxCp signature and might contribute to the 1-Cys Prx reaction cycle.

grown by the vapour diffusion hanging drop method at 291 K 2 . The crystallization drops were made by mixing 2 l AnPrx6 protein at 13 mg/ml with 2 l crystallization buffer, (0.2 M sodium acetate trihydrate, 0.1 M sodium cacodylate trihydrate pH 6.5 and 20% (w/v) polyethylene glycol 8000). The drops were equilibrated against 500 l of the same buffer. Crystals grew in about one month to a size of approximately 0.6 • 0.1 • 0.03 mm 3 .
Size-exclusion chromatography. The oligomeric state of AnPrx6 at room temperature was determined by analytical size exclusion chromatography (SEC) using a high resolution Superdex 200 10/300 column (GE Healthcare, Sweden). Prior to each run, the column was equilibrated with the buffer used to store the AnPrx6 protein, containing 50 mM Tris-HCl pH 7.5, 150 mM NaCl, and 1 mM CaCl2. For each run, a sample loop with a capacity of 100 µl was filled with AnPrx6 protein at a concentration of either 0.1 mg ml -1 , 1.0 mg ml -1 or 10.0 mg ml -1 , and loaded onto the column, resulting in 0.01 mg, 0.1 mg or 1.0 mg of protein passing through the column. Retention curves at a wavelength of 280 nm were recorded in the absence or presence of 20 mM H2O2 in the buffer. A standard curve for the Superdex 200 10/300 column using the above buffer condition was generated from the elution profile of a standard protein mix (BioRad Gel Filtration Standard, catalog # 151-1901). The standard curve was used to estimate the molecular weight of the eluted AnPrx6 protein.
Small Angle X-ray Scattering (SAXS) data collection: SAXS measurements were carried out in a buffer containing 50 mM TRIS-HCl pH 8.0, 150 mM NaCl, 1 mM CaCl2, without DTT, at two AnPrx6 concentrations (1.6 mg/ml and 4.8 mg/ml). All samples were centrifuged prior to measurements. The SAXS data sets of AnPrx6 were collected at the MAX IV Laboratory, MAX II beamline I911-SAXS 20 . The intensity data was recorded on a hybrid pixel Pilatus 1M detector as a function of the scattering vector q (q = (4 sin)/). Data reduction, normalization to the intensity of the transmitted beam, and buffer subtraction were performed using beamline tools and scripts. The Primus program in the ATSAS software package 21,22 was used to estimate forward scattering I(0), radius of gyration, pair-distance distribution function, and the excluded volume of the hydrated particle (Porod volume) 21,23 . Molecular weight estimations were made using Lysozyme as a molecular weight standard, as well as the Porod volume.
Peroxidase activity assay. Peroxidase activity of AnPrx6 was determined using the ammonium ferrous sulphate and xylenol orange (FOX) assay 24 . The assay was performed in triplicates at 23°C in 100 μl reaction buffer containing 20 mM NaPi pH 7.0, 150 mM NaCl, and 1 mM DTT and a protein concentration of 4 M. The reaction was initiated by adding hydrogen peroxide (H2O2) to a final concentration of 600 μM. After exactly one minute, 5 μl of the reaction mixture was removed and added to 500 μl of the FOX reagent solution, consisting of 250 μM ammonium ferrous sulphate, 100 μM xylenol orange, 25 mM H2SO4 and 100 mM sorbitol, which terminated the reaction. After incubation for one hour in the dark, 500 μl of deionized water was added to the FOX assay mixture and the absorbance of the final solution was measured at a wavelength of 560 nm. The enzymatic consumption of hydrogen peroxide due to the AnPrx6 protein in the sample was determined with the help of a standard curve that shows the absorption at 560 nm plotted as a function of increasing H2O2 concentrations. Control experiments were performed using samples containing no AnPrx6 in order to account for the potential background signal originating from the constituents of the reaction buffer ( Supplementary Fig. S8).
To determine the rate of H2O2 consumption with respect to the enzyme concentration, FOX assays in the same buffer as before were performed at AnPrx6 concentrations increasing from 0 M to 60 M and a fixed, saturating H2O2 concentration of 600 μM ( Supplementary Fig.  S9). Tables   Supplementary Table S1. Data Figure S1. Structure anchored multiple sequence alignment (saMSA). AnPrx6 and thirteen Prx proteins with known crystal structures and at least 30% sequence identity, compared to AnPrx6, were simultaneously superimposed. The respective PDB codes precede each sequence, except for AnPrx6. Identical amino acids are indicated by their single letter code above the alignment. Similar amino acids are marked with a hash (#). Highlighted in yellow are the conserved residues of the catalytic PxxxTxxC motif 25,26 , and the fully conserved active site arginine. In AnPrx6 these are residues Pro38, Thr42, Cys45 and Arg122. The peroxidatic cysteine (CP) is shown in red and the blue frame marks the residues of the active site CP-loop. CP was mutated to Ala and Ser in 1QMV and 2V2G, respectively and was oxidized to Cysteine sulfonic acid (X) in AnPrx6 and 2CV4. If present, the distal resolving Cys (CR) is marked red. Additional amino acids that contribute to the active site environment are Phe35, His37, Pro43, Val44, Glu48, Leu49, Trp80, Pro141 and Arg145. Of these, Glu48, Trp80, and Arg145 are fully conserved and highlighted in green. Highly conserved among the Prx6 proteins is the FSH motif, highlighted in pink, just before PxxxTxxC motif. The four regions (I-IV) marked with red above the alignment indicate those residues of 2-Cys Prxs that are involved in decamer formation 27 . Highlighted in light blue are seven residues of Prx6 proteins that extend helix 3 by almost two turns compared to other Prxs, which causes the loop between 3 and 5 to adopt a conformation that interferes with decameric ring formation. The succession of secondary structure elements for each protein is indicated below the saMSA: aquamarine arrows depict -strands, red tubes represent -helices and magenta tubes 310 helical turns. The loop connecting 4 with 6. The loop spanning residues Asp109 to Arg122 of AnPrx6 contains an -helical turn in monomers A and C (purple in Fig.  1a and 1b), while the same loop in monomer B and D (orange in Fig. 1b) adopts an extended, irregular shape which resembles the corresponding loop of Aeropyrum pernix Prx6 (2CV4). Other Prx proteins form a short two stranded -sheet as can be seen in PDB entries 2V2G and 4LLR.     The SAXS intensity data for AnPrx6 at 1.6 mg/ml (black circles) and 4.8 mg/ml (magenta circles), were scaled to visualize the lack of concentration dependent protein-protein interactions. Due to the low noise level, the actual experimental data up to 3 nm -1 displayed instead of the fitted curves. The inset shows the pair-distance distribution function calculated using Primus. Figure S8. The AnPrx6 peroxidase activity was measured at increasing concentrations of H2O2 in the range from 0 M to 600 M.All measurements were carried out in triplicates. The average value is marked by a dot and the error bars mark the uncertainty of the measurements. Figure S9. H2O2 Consumption rate vs. AnPrx6 concentration. The graph shows a linear dependency of the H2O2 consumption rate with respect to the enzyme concentration, which suggests that no structural changes, such as oligomerization, occur as the protein concentration increases from 0 M to 60 M. The error for the first three points is very small and which results in error bars that are not clearly visible.