To evaluate the importance of the predicted atypical kinase/ATPase domain of ABC1K1 protein for its function, we transformed an abc1k1 T-DNA mutant (SALK_068628) with a construct expressing a C-terminally YFP-HA-tagged mutated version of ABC1K1, in which the highly conserved aspartic acid residue in the catalytic domain (Motif VIIb, position 400) had been replaced by an asparagine (abc1k1 ABC1K1D400N-YFP-HA (abbreviated K1 D400N)) ( Figures 1A, B ). As a control, we complemented abc1k1 with a C-terminally YFP-HA-tagged wildtype version of ABC1K1 (abc1k1 ABC1K1-HA-YFP (abbreviated K1 WT)). Transformed plants were selected by segregation and verified by PCR, genotyping, and sequencing ( Supplementary Figure S1 ).

The protein expression level was evaluated by SDS-PAGE followed by Western blotting ( Figure 2 ). By using an anti-HA and anti-ABC1K1 antibody, we observed that the two K1 WT lines (K1 WT 1 and 2) and the third line of K1 D400N (K1 D400N 3) gave strong signals for recombinant ABC1K1 protein under control white light whereas those for the other two lines of K1 D400N (K1 D400N 1 and 2) were around 5 to 10 times weaker ( Figure 2A , Supplementary Figure S2 ).

Interestingly, the amount of recombinant ABC1K1 protein varied depending on the light conditions. Under red light, the level of ABC1K1 in the two K1 WT lines (K1 WT1 and 2) lines decreased compared to control light while it increased in all K1 D400N lines (K1 D400N 1, 2 and 3) ( Figure 2A ).

To evaluate the levels of the endogenous ABC1K1 protein in Col-0, we used the anti-ABC1K1 antibody. In order to detect the weak signal of endogenous ABC1K1 protein in Col-0, we did western blotting experiments ( Figure 2B ), without the lines expressing the ABC1K1 protein the most (K1 WT1, K1 WT2 and K1D400N 3) that would obscure the signal of endogenous ABC1K1 in Col-0. We observed that under control light there was 1.3 times more ABC1K1D400N-YFP-HA in the K1 D400N1 line compared to endogenous ABC1K1 in Col-0 whereas K1 D400N2 expressed 1.4 times less ( Figure 2B ). Similar to what we previously observed for ABC1K1-YFP-HA in K1 WT 1 and 2 lines, endogenous ABC1K1 in Col-0 was decreased around 5 times under red light ( Figure 2B ) compared to control light conditions. The opposite, an increase in ABC1K1 levels under red light, was observed instead in K1 D400N1 and K1 D400N2 lines, having levels of around 1.5 and 6 times higher compared to control light ( Figure 2B ). Under high light, the level of ABC1K1 in Col-0 and ABC1K1D400N-YFP-HA in K1 D400N2 and K1 D400N1 remained constant or decreased slightly ( Figure 2B ).

To determine whether the pale green phenotype of abc1k1 under red and high light would be complemented by the wildtype or the mutated version of ABC1K1, plants were illuminated for 5 days with constant white light (80 µE), red light (680 nm, 60 µE), or high light (500 µE). In the two K1 WT1 and 2 lines the abc1k1 phenotype under red and high light fully reverted to wildtype whereas the three K1 D400N lines maintained the abc1k1 phenotype ( Figure 3A ). Chlorophyll quantifications of these lines confirmed full complementation in K1 WT 1 and 2 but not in K1 D400N 1, 2 and 3 ( Figure 3B ).

abc1k1 is characterized by a strong photosynthetic defect manifested in reduced PSII efficiency and non-photochemical quenching (NPQ) (Martinis et al., 2014; Pralon et al., 2019). To assess whether the predicted active site D400 of ABC1K1 is required for photosynthetic activity plants were grown under constant control white light (CL, 80 µE), red light (RL, 60 µE) or high light (HL, 500 µE). We determined the maximum quantum yield of PSII (φ _max (= F_V/F_M)), as well as NPQ using chlorophyll fluorescence analysis under increasing light intensity in all complemented lines. The K1 D400N1, 2 and 3 lines and the abc1k1 mutant showed a very similar decrease of PSII maximum quantum yield compared to Col-0 regardless of the growth light conditions while K1 WT lines (1 and 2) fully complemented the defect ( Figure 4A ). PSII efficiency in abc1k1 was already affected when grown under control light, whereas no visible phenotype was observed under this condition. Line K1 D400N 3, which expresses the most ABC1K1 protein, appeared to partially restore PSII efficiency, when grown under control light.

NPQ was strongly decreased in abc1k1 and K1 D400N 1, 2 and 3 particularly when grown under control and red light while the NPQ values in K1 WT 1 and 2 were higher than in Col-0. When grown under high light, NPQ in Col-0 and K1 WT 1 and 2 was strongly diminished compared to plants grown under control or red light. NPQ in K1 D400N 1, 2 and 3 as well as abc1k1 remained comparatively unchanged. The unexpected differences observed between red and high light may be due to differential developmental effects and increased zeaxanthin production under high light (Chauhan et al., 2023).

ABC1k1 and the K1 D400N lines showed pale green phenotypes when grown under red or high light. In a separate study, we observed incompletely processed forms of thylakoid lumen proteins that accumulated in abc1k1 under red light (including PsbP and PsbQ) while the mature forms of these protein as well as other photosynthesis-associated proteins were strongly down-regulated (Collombat et al., 2024). To determine whether this was also the case when the active site residue was mutated as in the K1 D400N lines, we analyzed the levels of PsbA, Lhcb1, PsbO1, PsbP and PsbQ in all complemented lines under control light, red light and high light by Western blot. Down regulation of photosynthesis-associated proteins in abc1k1 was observed primarily under red light conditions ( Figure 5 ), less so under high light and not at all control light ( Figure 5 ). In the three K1 D400N lines the levels of PsbA, PspQ and PsbP under red light were not restored while in K1 WT 1 and 2 they were fully recovered. The presence of one or several additional bands of higher molecular mass for PsbQ and PsbP in abc1k1 and in the three K1 D400N lines under red light was observed and may reflect effects on preprotein processing (Collombat et al., 2024).

Arabidopsis thaliana wild-type refers to var. Columbia-0 (Col-0). abc1k1 mutant is a T-DNA insertion line (SALK_068628) obtained from the Nottingham Arabidopsis Stock Centre (NASC, http://arabidopsis.info). The abc1k1 mutant was complemented with 35S:ABC1K1-YFP-HA or 35S:ABC1K1D400N-YFP-HA. The pEarlyGate101-ABC1K1 plasmid was created by introducing the ABC1K1 coding sequence between the 35S promoter and the YFP-HA tag of pEarleyGate101 plant expression vector (Earley et al., 2006) by GeneCust’s (Boynes, France). The resulting plasmid was used to transform Agrobacterium tumefaciens C58 strain by electroporation. The modified strain was used to transform abc1k1 mutant plants using the Floral Dip method (Clough and Bent, 1998). Seeds from transgenic plants were harvested and the mutation selected by resistance to 30 mg.l-1 glufosinate ammonium in ½ MS medium. Segregation analysis was performed to obtain homozygous 35S:ABC1K1-YFP-HA and 35S:ABC1K1D400N-YFP-HA lines with a single transgene insertion. Protein expression was confirmed by immunoblot in selected complemented plants. The next generation of seeds was sterilized and spread on 0.5x MS plates then placed in the dark for 24 hours at 4°C.Seeds were moved to 22-24°C and exposed for 1 hour to white light (80 µmol m−2 s−1), afterwards were kept for 5 days under continuous white light (80 µmol m−2 s−1) or moved to continuous red light (60 µmol m−2 s−1) or high light (500 µmol m−2 s−1). 5-day old seedlings were then collected under the light, immediately frozen in liquid nitrogen and stored at -20°C.

Total chlorophyll was extracted from 5-day old seedlings (minimum of 20 mg of fresh weight) by adding 10 µl per mg FW of DMF (Dimethylformamide). Samples were centrifuged 1 minute at 16,000 g and kept overnight at 4°C in the dark. Extracts were once again centrifuged for 3 minutes at 16,000 g before measuring the absorbance at 664 nm and 647 nm with a Nanodrop spectrophotometer (NanoDrop ND-1000 413 spectrophotometer, Witec AG). Total chlorophyll concentrations were calculated according to (Porra et al., 1989).

Maximum quantum yield of PSII (Φmax) and NPQ were measured using a Fluorcam MF800 (Photon System Instrument, Czech Republic, http://www.psi.cz). The actinic light was provided by blue LED (470 nm). The protocol starts by the measurement of the Φmax = (FM–FO)/FM where FM is the maximal fluorescence in dark acclimated plants, measured during a saturating light pulse, and FO the fluorescence in the dark. The actinic light intensity was increased by 1 minute steps. At the end of each light intensity step we determined the non-photochemical quenching NPQ = (FM – FM’)/FM’. FM’ is the maximal fluorescence at the end of each light intensity step. The experiment was performed on three independent biological replicates composed of 10 to 30, 5-day old seedlings per genotype.

Frozen 5-day old seedlings were ground in 400 µl of lysis buffer (100 mM Tris-HCl pH8.5, 2% SDS, 10 mM NaF, 0.05% of protease inhibitor cocktail for plants (Sigma)) with a micro-pestle in a 1.7 ml microtubes. Samples were heated at 37°C for 30 minutes and centrifuged for 15 minutes at 16,000 g at room temperature. Protein concentration in each sample was determined using the Pierce BCA protein assay kit (Thermo Scientific, cat. No. 23225) following manufacturer instructions. Proteins were precipitated by chloroform-methanol, then resuspended in sample buffer (50 mM Tris-HCl pH6.8, 100 mM Dithiothreitol, 2% SDS, 0.1% Bromophenol Blue, 10% Glycerol) at a final concentration of 1 µg.µl-1 and denatured for 10 minutes at 65°C. 5 µl was loaded on a 16% polyacrylamide SDS gel and proteins were separated by electrophoresis before transfer to a nitrocellulose membrane for immunoanalysis using the following antibodies: anti-HA (Agrisera AS15 2924), anti-PsbA (Agrisera AS05 084) anti-Lhcb1 (Agrisera, AS01 004), Anti-PsbO1 (Agrisera AS14 2824), Anti-PsbP (Agrisera AS06 142-23), anti-PsbQ (Agrisera AS06 142-16). Primary antibodies were decorated with horseradish peroxidase-conjugated anti-rabbit (Merck, AP132P) or anti-mouse (Sigma, A5278) antibodies. Chemiluminescent signals were detected using a home-made reaction mixture (luminol 1.25 mM, cumaric acid 0.20 mM, mixed with 0.01% H₂O₂ just before the reaction) using a CCD camera (Amersham Imager 600, AmershamBiosciences, Inc.). The quantification of protein signals was done with the ImageQuant TL software and the quantification of amido black coloration was performed using ImageJ software.