Calredoxin represents a novel type of calcium-dependent sensor-responder connected to redox regulation in the chloroplast

Abstract

Calcium (Ca(2+)) and redox signalling play important roles in acclimation processes from archaea to eukaryotic organisms. Herein we characterized a unique protein from Chlamydomonas reinhardtii that has the competence to integrate Ca(2+)- and redox-related signalling. This protein, designated as calredoxin (CRX), combines four Ca(2+)-binding EF-hands and a thioredoxin (TRX) domain. A crystal structure of CRX, at 1.6 Å resolution, revealed an unusual calmodulin-fold of the Ca(2+)-binding EF-hands, which is functionally linked via an inter-domain communication path with the enzymatically active TRX domain. CRX is chloroplast-localized and interacted with a chloroplast 2-Cys peroxiredoxin (PRX1). Ca(2+)-binding to CRX is critical for its TRX activity and for efficient binding and reduction of PRX1. Thereby, CRX represents a new class of Ca(2+)-dependent 'sensor-responder' proteins. Genetically engineered Chlamydomonas strains with strongly diminished amounts of CRX revealed altered photosynthetic electron transfer and were affected in oxidative stress response underpinning a function of CRX in stress acclimation.

Figure 1. In vitro characterization of calredoxin.

(a) K_d estimation for Ca²⁺-binding by MST measurements. Fluorescently labelled recombinant WT CRX (CRX) was incubated with defined concentrations of free Ca²⁺ (filled circles) or Mg²⁺ (open circles) and the ratio of detected fluorescence before and after the thermophoretic movement was plotted against the corresponding cation concentration. Data for CRX incubated with Ca²⁺ were fitted according to the law of mass action (black line) and gave a K_d of 88.2 nM±16.5 nM. Each data point represents the mean value of at least three experiments (±s.d.). (b) CRX shows Ca²⁺-dependent redox activity. 10 μM recombinant WT CRX (closed circles) was reduced by E. coli NTR and NADPH in defined Ca²⁺ concentrations for 10 min at RT. 200 μM DTNB were added as substrate for reduction by CRX and the increase in absorption at 412 nm was recorded to calculate the redox activity (slope 0–80 s after addition of DTNB). Data were normalized on the highest activity measured for each protein purification and fitted by Michaelis–Menten kinetics (K_d: 281.1±153.8 nM). Error bars represent s.d. of three independent measurements. Assay modified after ref. . (c) Oxidation–reduction titration of WT CRX. The disulfide/dithiol redox state at each E_h value was monitored using the monobromobimane fluorescence method. The line represents a fit of the data to a two-electron Nernst curve and yielded an E_h of −288.2±5.3 mV. Data were acquired in two independent experiments.

Figure 2. Chloroplast localization of calredoxin in C. reinhardtii.

Microscopy images of a transgenic strain expressing ble-2A-calredoxin-mVenus. (a,d) YFP fluorescence (detected with a 525–555 nm filter) and (b,e) chlorophyll fluorescence (AUF, detected with a 690–740 nm filter). Merged images (YFP/AUF) are shown in the bottom row (c,f). The second row (d–f) shows the fluorescence signals in a single cell at higher magnification. Scale bars, 20 μm.

Figure 3. Potential calredoxin interaction partners.

(a–d) Results from the CRX affinity chromatography. WT and mutated versions of recombinant CRX were immobilized on a CNBr-activated resin and a whole-cell lysate of heterotrophically grown C. reinhardtii was added to the column to supply potential CRX target proteins. The log2 protein intensities after label-free quantification (LFQ) of the C241S sample were plotted either against the intensities of the WT (a,c) or the C238S (b,d) sample. Two proteins were repeatedly (two experiments) significantly more abundant in the C241S sample when Ca²⁺ was present on the column (a,b): PRX1, Cre06.g257601.t1.2 (open square) and another 2-cys peroxiredoxin, Cre02.g114600.t1.2 (open circle). Elimination of Ca²⁺ (c,d) reduced the abundance of these proteins and led to identification of a third potential target protein: TRXL1 (Cre03.g157800.t1.1, open triangle). (e) Interaction of CRX and PRX1 in vitro. 5 μM recombinant CRX was reduced by E. coli TRXR and NADPH in the presence of 40 μM H₂O₂ at RT. The NADPH absorbance at 340 nm was monitored until a steady decrease was observed. After subsequent addition of 1 μM oxidized recombinant PRX1 (indicated by the arrow) NADPH oxidation was increased in the presence of Ca²⁺ (black filled circle, 44.3 μmol NADPH min⁻¹μmol PRX1⁻¹) in contrast to CRX without Ca²⁺ (open circle) or without reductase (grey filled circle). CRX C238S, C241S (black filled square) was not able to reduce PRX1. One exemplary measurement is shown. Assay modified after ref. . (f) Titration of electron transfer between PRX1 and CRX in dependence of Ca²⁺. Fitting the data to Michaelis–Menten kinetics revealed a half-maximal rate of NADPH oxidation at a concentration of 122.3±64.5 nM free Ca²⁺. Scale bars give s.d. of three measurements.

Figure 4. X-ray structure of calredoxin.

(a) Overall structure of CRX. Sequence diagram of CRX is shown at the top. The N- and C-subdomains of the CaM domain are displayed in orange and magenta, respectively. Bound Ca²⁺ ions are represented as green spheres. The TRX domain is highlighted in marine-blue with the disulfide bridge shown as yellow ball-and-stick model. (b) Open-book representation of interactions between the CaM and the TRX domains. Residues involved in direct inter-domain interactions except for water-mediated hydrogen bonds are shown. (c) Inter-domain networks between the Ca4 calcium ion in the CaM domain and the disulfide bridge in the TRX domain. Purple labels indicate the residues of structure based mutagenesis. (d) The stereo image of 2|F_o |-|F_c | electron density maps (2σ level) showing amino acid residues around Ca3 (upper) and the disulfide bridge between Cys238 and Cys241 (lower).

Figure 5. Calredoxin function in vivo.

(a,b) Immunoblot analysis of WT versus IMcrx (a) or empty vector (EV) versus amiRNA-crx-12/23 (KD#12/KD#23) (b) whole-cell extracts. ATPB, LHCSR3 and CRX protein expression were examined in cells grown under photoheterotrophic (TAP) LL (30 μE m⁻² s⁻¹), which were shifted to photoautotrophic (HSM) HL (180 μE m⁻² s⁻¹) growth conditions. Chlorophyll (1.5 μg) were loaded per lane and equate 100%. ATPB was used as loading control. (c,d) Comparison of linear photosynthetic electron flow and CEF in WT versus IMcrx (c) or EV versus KD#12 and KD#23 (d) cells grown under TAP/LL, which were shifted for 6 h either to HSM/LL or HSM/HL growth conditions. LL and HL data (±s.d.) refer to analyses of three biological replicates for WT and IMcrx, with 5 and 11 as well as 6 and 12 measurements, respectively; KD LL Data (±s.d.) refer to analyses of three biological replicates for EV, KD#12 and KD#23 with three measurements each; HL Data (±s.d.) refer to analyses of three (EV), three (KD#12) and seven (KD#23) biological replicates, with three, nine and 10 measurements, respectively. For statistical analysis of indicated data t-test with *P<0.05, **P<0.01 and ***P<0.001 was performed. (e,f) Measurement of lipid peroxidation in calredoxin-deficient strains under HL conditions. WT versus IMcrx (e) and of empty vector (EV) versus KD#23 (f) cells after transition from TAP/LL to HSM/HL growth conditions. Malondialdehyde equivalents were measured from whole cells (4 μg Chl ml⁻¹) before (0 h, grown in TAP at 30 μE m⁻² s⁻¹) and after HL treatment (shifted to HSM for 6, 24, 30 h at 180 μE m⁻² s⁻¹). WT versus IMcrx data refer to four biological replicates with eight measurements each, EV versus KD#23 data refer to analyses of three biological replicates with six measurements each. Error bars represent s.d.'s. Statistical comparison of indicated data was done using t-test with *P<0.05, **P<0.01 and ***P<0.001.

Figure 6. TRX f is diminished in HL on depletion of CRX.

MS-based ¹⁵N metabolic labelling-based quantitation from CRX (Cre03.g202950.t1.1) (a) and TRX f (Cre01.g066552.t1.1) (b). Each pie slice represents a quantified peptide. The area of a pie slice is proportional to the log2 ratio of the corresponding peptide between the indicated conditions. Colours represent pyQms quantification score; 0.7 (yellow, false discovery rate (FDR)≤1%) to 1 (blue, prefect match). (c) Selected knowledge based TRX community. The heat map represents the ratios of the proteins ranges from yellow to green indicating an upregulation and from yellow to red indicating a downregulation. The s.d. is visualized by the size of the box, smaller the box higher the s.d. of the protein ratio (see legend). For quantitative MS analysis, ¹⁴N-/¹⁵N-labelled WT and ¹⁵N-/¹⁴N-labelled IMcrx were mixed based on equal protein amount (total protein amount 100 μg). Eight different conditions were examined as followed: (1) ¹⁴N-labelled WT TAP LL versus ¹⁵N-labelled IM TAP LL, (2) ¹⁵N-labelled WT TAP LL versus ¹⁴N-labelled IM TAP LL, (3) ¹⁴N-labelled WT HSM HL versus ¹⁵N-labelled IM HSM HL, (4) ¹⁵N-labelled WT HSM HL versus ¹⁴N-labelled IM HSM HL, (5) ¹⁴N-labelled WT TAP LL versus ¹⁵N-labelled WT HSM HL, (6) ¹⁴N-labelled IM TAP LL versus ¹⁵N-labelled IM HSM HL, (7) ¹⁴N-labelled WT TAP LL versus ¹⁵N-labelled IM HSM HL and (8) ¹⁴N-labelled IM TAP LL versus ¹⁵N-labelled WT HSM HL.