Abstract
In plastids, conversion of light energy into ATP relies on cytochrome f, a key electron carrier with a heme covalently attached to a CXXCH motif. Covalent heme attachment requires reduction of the disulfide-bonded CXXCH by CCS5 and CCS4. CCS5 receives electrons from the oxidoreductase CCDA, while CCS4 is a protein of unknown function. In Chlamydomonas reinhardtii, loss of CCS4 or CCS5 yields a partial cytochrome f assembly defect. Here, we report that the ccs4ccs5 double mutant displays a synthetic photosynthetic defect characterized by a complete loss of holocytochrome f assembly. This defect is chemically corrected by reducing agents, confirming the placement of CCS4 and CCS5 in a reducing pathway. CCS4-like proteins occur in the green lineage, and we show that HCF153, a distant ortholog from Arabidopsis thaliana, can substitute for Chlamydomonas CCS4. Dominant suppressor mutations mapping to the CCS4 gene were identified in photosynthetic revertants of the ccs4ccs5 mutants. The suppressor mutations yield changes in the stroma-facing domain of CCS4 that restore holocytochrome f assembly above the residual levels detected in ccs5. Because the CCDA protein accumulation is decreased specifically in the ccs4 mutant, we hypothesize the suppressor mutations enhance the supply of reducing power through CCDA in the absence of CCS5. We discuss the operation of a CCS5-dependent and a CCS5-independent pathway controlling the redox status of the heme-binding cysteines of apocytochrome f.
Reduction of a disulfide formed between the heme-linking cysteines of apoforms of c-type cytochromes in the thylakoid lumen requires the provision of reducing power from stroma to apocytochrome c via thiol–disulfide exchange. Reducing power is supplied through two different pathways, pathways 1 and 2, with CCDA and CCS4 being common components to both pathways. In pathway 1, CCS5 directly reduces the disulfide-bonded heme-binding site. In pathway 2, one or several yet-to-be identified protein(s) (X) reduce(s) the disulfide. CCS4 functions in stabilizing CCDA. In the absence of CCS5, gain-of-function mutations altering the C terminus of CCS4 enhance the delivery of reducing power via CCDA to the unknown reductase(s) (X), underscoring the functional importance of the soluble domain in CCS4 (indicated by a yellow star). This domain might act by recruiting the stromal reductant to CCDA or control the reactivity of the redox-active cysteines in CCDA by providing an ideal chemical microenvironment for efficient thiol–disulfide exchange.
Introduction
Cytochromes c or c-type cytochromes are ubiquitous heme (ferroprotoporphyrin IX) containing proteins that function as one-electron carriers in energy-transducing membranes in bacteria, archaea, mitochondria, and plastids (Bertini et al. 2006; Ma and Ludwig 2019). Cytochromes c located on the positive side (or p-side) of energy-transducing membranes [i.e. the bacterial or archaeal periplasm, the thylakoid lumen in the plastid, and the mitochondrial intermembrane space (IMS)] carry a heme moiety covalently bound to two (or rarely one) cysteines in a typical CXXCH motif called the heme-binding site on the apoprotein (Alvarez-Paggi et al. 2017). The histidine residue acts as an axial ligand to the iron in the heme moiety (Dickerson et al. 1971). Conversion of apocytochrome c to its corresponding holoform requires the formation of the thioether linkage(s) between the vinyl carbons in heme and the cysteine thiol(s) in the heme-binding motif. The chemical requirements for thioether bond formation were formulated from the in vitro reconstitution of holocytochrome c formation alongside extensive genetic and biochemical dissection of cytochrome c maturation in bacterial and eukaryotic model systems (Cline et al. 2016; Das and Hamel 2021). In energy-transducing membranes, apo- to holocytochrome c conversion occurs on the p-side of the membrane and requires minimally: (1) the independent transport of apocytochrome c and heme substrates; (2) the chemical reduction of (a) ferriheme to ferroheme and (b) disulfide-bonded heme—linking cysteines in apocytochrome c to free thiols; and (3) the stereospecific ligation of heme to the free thiols of the CXXCH motif via formation of thioether bond linkages. In vivo, heme ligation to CXXCH cysteine thiols is a catalyzed process requiring at least 4 different maturation systems (systems I, II, III, and IV) which are defined by signature assembly factors (Mavridou et al. 2013; Travaglini-Allocatelli 2013; Babbitt et al. 2015; Gabilly and Hamel 2017; Belbelazi et al. 2021). The diversity in cytochrome c assembly routes challenges our understanding of the process considering we view the stereospecific attachment of cysteine thiols to the vinyl carbons of heme as a simple chemical reaction (Bowman and Bren 2008).
In plastids and several bacteria, cytochrome c maturation is assisted by the cytochrome c synthesis (CCS) factors, which were first identified through the isolation of the ccs mutants in the green alga Chlamydomonas reinhardtii (Gabilly and Hamel 2017; Das and Hamel 2021). The ccs mutants are specifically blocked in the attachment of heme to the apoforms of plastid cytochromes c, namely, membrane-bound cytochrome f and soluble cytochrome c6 (Karamoko et al. 2013; Gabilly and Hamel 2017; Das and Hamel 2021). Cytochrome f, one of the plastid c-type cytochromes, is a structural subunit of the cytochrome b6f complex and an essential electron carrier in photosynthesis (Kuras and Wollman 1994; Zhou et al. 1996). Consequently, the ccs mutants are deficient for cytochrome b6f assembly and photosynthesis (Cline et al. 2016; Gabilly and Hamel 2017).
In the thylakoid lumen, the redox state of the apocytochrome f CXXCH motif is under the control of (1) CCS5, a thylakoid-bound disulfide reductase catalyzing the reduction of the disulfide bond formed between the cysteines of the heme-binding site into thiols, and (2) CCDA, a thylakoid membrane protein delivering reducing power from stroma to lumen (Lennartz et al. 2001; Page et al. 2004; Motohashi and Hisabori 2006; Gabilly et al. 2010; Motohashi and Hisabori 2010). CCS5 is classified as a member of the thioredoxin-like family, a diverse group of proteins including molecules able to mediate thiol–disulfide exchange reactions via a redox-active WCXXC motif (Atkinson and Babbitt 2009). CCS5 displays high similarity to membrane-anchored lumen-facing HCF164, previously identified as being required for the assembly of cytochrome b6f in Arabidopsis (Lennartz et al. 2001; Gabilly et al. 2010). In vitro, a recombinant form of CCS5/HCF164 containing the WCXXC motif is redox-active and able to reduce a disulfide bond formed between the heme-linking cysteines of a soluble form of apocytochrome f (Lennartz et al. 2001; Gabilly et al. 2010). The cytochrome c assembly defect in the ccs5-null mutant can be chemically corrected by exogenously applied reducing agents, an observation demonstrating the physiological relevance of the disulfide reductase activity documented from the in vitro redox assays (Lennartz et al. 2001; Gabilly et al. 2010). Conversion of the disulfide-linked heme-binding cysteines to thiols is also dependent upon CCDA whose activity is required to maintain the catalytic WCXXC motif in CCS5 in the reduced form (Page et al. 2004; Bushweller 2020). The proposed model is that the thiol–disulfide oxidoreductase CCDA and the thioredoxin-like protein CCS5/HCF164 define a trans-thylakoid pathway that sequentially transfers reducing power via a cascade of thiol–disulfide exchanges from stroma to the oxidized CXXCH motif of apocytochrome f (Lennartz et al. 2001; Page et al. 2004; Motohashi and Hisabori 2006; Gabilly et al. 2010; Motohashi and Hisabori 2010; Gabilly et al. 2011). Maintenance of the CXXCH motif in the reduced form also relies on CCS4, a small thylakoid protein predicted to be membrane-bound with a soluble domain facing the stroma (Gabilly et al. 2011). However, CCS4 contains no motif or residue suggestive of a redox biochemical activity, and orthologs can only be detected in a subset of green algae (Gabilly et al. 2011). The placement of CCS4 in the disulfide reduction pathway for plastid cytochrome c assembly was inferred from the finding that provision of reductants or the expression of an additional ectopic copy of CCDA suppresses the photosynthetic deficiency of the ccs4-null mutant (Gabilly et al. 2011). In this manuscript, we provide genetic and biochemical evidence for the involvement of CCS4 in a CCS5-dependent and a CCS5-independent pathway for reducing the disulfide-bonded heme-binding cysteines of apocytochrome f and discuss the possible functions of this assembly factor.
Materials and methods
Strains, media, and growth conditions
All algal strains used in this study are listed in Supplementary Table 1. Strains CC-124 and CC-4533 were used as wild type (WT); strains CC-4129, CC-5922, and CC-5936 were used as ccs5 mutants. The CC-5936 strain is an insertional mutant (LMJRY0402.042908) from the CLiP library (Li et al. 2019) and carries the aphVIII cassette inserted at the end of exon 3 of the CCS5 gene (Cre17.g702150). The CC-4527 strain is the ccs5 mutant (CC-4129) complemented with the WT CCS5 gene. This strain is referred to as ccs5(CCS5) and was described in Gabilly et al. (2010). Strains CC-4519, ccs4(pCB412), and CC-4525 were used as ccs4 mutants. The ccs4Δccs5(9) and ccs4Δccs5(11) strains that displayed a tight photosynthetic deficiency had reverted at the end of the study. The suppressed strains were renamed SU9 and SU11 and deposited in the Chlamydomonas collection center as CC-5928 and CC-5929, respectively.
WT and mutant C. reinhardtii strains were maintained on Tris-acetate-phosphate, with Hutner's trace elements, 20 mM Tris-base and 17 mM acetic acid, or Tris-acetate-phosphate supplemented with arginine (1.9 mM) (TARG) (Harris 1989). For some strains, TAP/TARG supplemented with 25-µg/ml hygromycin B (TAP/TARG + Hyg) or 25-µg/ml paromomycin (TAP/TARG + Pm) was used. All algal strains were grown on a solid medium at 25°C in continuous light at 0.3 µmol/m2/s (Harris 1989). The solid medium contains 1.5% (w/v) agar (Select Agar, Invitrogen, 30391049).
Growth was tested on solid TAP/TARG or minimal medium, under 3 different light conditions. For protein immunodetection, cells were grown in liquid TAP/TARG in an environmental incubator with continuous light at 0.3 µmol/m2/s and shaking at 180–190 rpm at 25°C.
Chemo-competent Escherichia coli DH5α strains were used as hosts for molecular cloning. The bacterial strains were grown at 37°C in Luria–Bertani (LB) broth and LB medium solidified with bacteriological agar (Silhavy et al. 1984). When antibiotic selection was required, 100-µg/ml ampicillin or 50-µg/ml kanamycin was added.
Genetic crosses
The experimental details for sexual crosses are described in the former work (Subrahmanian et al. 2020a, 2020b). For molecular verification of the ccs5::ARG7Φ allele in the haploid progeny, the presence of the molecular tag (Phi “ϕ” flag) inserted in the ARG7 intron used for insertional mutagenesis was detected via diagnostic PCR using the Phi-1 (5′-GTCAGATATGGACCTTGCT-3′) and Phi-2 (5′-CTTCTGCGTCATGGAAGCG-3′) primers (Gabilly et al. 2010). The presence of the WT CCS4 allele, the ccs4-F2D8 mutation, or suppressor mutations, CCS4-2 and CCS4-3, in SU9 and SU11, respectively, was tested molecularly by diagnostic digestion of a PCR product. The alleles can be distinguished by virtue of a nucleotide polymorphism that abolishes the BanII restriction site in the nonsense (ccs4-1), missense (CCS4-2), and deletion mutations (CCS4-3) in the suppressor strains. The cell wall deficient strains CC-5925, CC-5922, and CC-5927 were generated by crossing the CC-4525 and CC-4517 strains and examining the resulting progeny for the cell wall deficient trait, by resuspending cells in 0.1% Triton X-100. Untreated cells served as a control. Cells without a cell wall lyse in the presence of Triton X-100 (Hoffmann and Beck 2005). The ccs4Δccs5 mutants were generated by crossing the ccs5 mutant (CC-4129) with the ccs4 mutant (CC-4520) and selecting the progeny based on arginine prototrophy and paromomycin resistance. For generating the vegetative diploids, the ccs5::aphVIII mutant from the CLiP collection was used. The CC-5923 and CC-5924 are haploid ccs5::aphVIII arg7 progeny originating from the cross of the CC-5936 strain by a WT arg7 strain (WT-PH2). WT-PH2 was obtained after four back-crosses of the CC-5611 (Subrahmanian et al. 2020a) by CC-4533 or CC-5155. The Δccs5/Δccs5 diploids were constructed by crossing the CC-4129 strain by the CC-5924 strain and selection on TAP + Pm. The Δccs5/Δccs5 CCS4-2 diploids were generated by crossing the CC-5931 strain by CC-4129 and selection on TAP + Pm. The CC-5931 strain is a haploid ccs5::aphVIII CCS4-2 arg7 progeny issued from the cross of CC-5928 by CC-5923. The (Δccs5/Δccs5; CCS4-3) diploids were generated by crossing CC-5930 by CC-5924 and selection on TAP + Pm. The CC-5930 strain is a haploid progeny (ccs5::ARG7Φ; CCS4-3) from the cross of CC-5929 by CC-5590. The WT diploids (+/+) were constructed by crossing CC-425 by CC-5590 and selection of the mitotic zygotes on TAP and already used in a former study (Barbieri et al. 2011).
Growth assay and thiol-dependent rescue
Cells were grown on a solid acetate-containing medium (TAP or TARG) for 5–7 days at 25°C under 0.3-µmol/m2/s illumination. Two loops of cells were resuspended in 500-µl water, and optical density was measured spectrophotometrically at 750 nm and normalized to OD750 = 2. This normalized suspension was used as the starting material (1) to make 5 serial 10-fold dilutions (10−1, 10−2, 103, 10−4, and 10−5). A volume of 5 µl from each dilution was plated on solid minimal medium (CO2) for photoautotrophic growth or TAP/TARG (acetate + CO2) for mixotrophic growth and incubated for 7–14 days under 30- to 50-µmol/m2/s illumination for growth assay. For thiol-dependent rescue, similar 10-fold serial dilutions were performed, and equal volume from each dilution was plated on solid minimal medium (CO2) containing (1) either no MESNA (2-MercaptoEthane Sulfonate Na, Sigma, M1511-25G) or (2) 0.75 or 1.5 mM MESNA. Ten-fold serial dilutions were also plated on TAP/TARG (acetate + CO2) and incubated for 14–21 days under 30–50 µmol/m2/s for thiol-dependent rescue assay and under 0.3 µmol/m2/s with no exogenously added thiols as control.
Fluorescence rise-and-decay kinetics
Fluorescence transients (also known as the Kautsky effect) were measured using Handy FluorCam from Photon Systems Instruments that provides actinic illumination and saturating flash of light. Cells were grown mixotrophically on solid TAP/TARG under 0.3-µmol/m2/s illumination for 5–7 days at 25°C. To measure the fluorescence, cells are dark-adapted for 15 min followed by a saturating flash of light. The fluorescence is recorded in arbitrary units (AUs) over an illumination period of 5 s.
Protein sample preparation
For heme staining and immunoblotting, two loops of cells grown on solid acetate-containing medium (TAP or TARG) for 5–7 days at 25°C under 0.3-µmol/m2/s illumination were collected and washed with 1–2 ml of 10 mM sodium phosphate buffer (NaH2PO4, pH 7.0), centrifuged at top speed and resuspended in 300 µl of the same buffer. Chlorophyll was extracted by mixing 10 µl of resuspended cells with 1 ml of acetone: methanol in a ratio of 80:20. The mixture was vortexed followed by centrifugation at 16,873 × g for 5 min. Chlorophyll concentrations were then determined by measuring the optical density of the suspension at 652 nm. Concentrations for different strains were normalized to 1 mg/ml of chlorophyll, followed by cell lysis by two freeze–thaw cycles (samples were frozen −80°C for 1 h and thawed on ice for ∼1 h). Cells were then separated into membrane and soluble fractions by centrifugation for 5 min at 16,873 × g (Howe and Merchant 1992). Membrane fractions corresponding to 10 µg of chlorophyll were separated electrophoretically by SDS–PAGE (12.5% acrylamide gel) for immunoblotting. For enhanced chemiluminescence (ECL)-based heme staining, an amount of 10–20 µg of chlorophyll for each sample was used, whereas for in-gel heme staining by 3,3′,5,5′-tetramethylbenzidine (TMBZ, Sigma, 860336-1G), a quantity corresponding to 30–40 µg of chlorophyll per sample was loaded.
Heme staining and immunoblotting
The presence of covalently bound heme in c-type cytochromes was detected by two methods, (1) by ECL (Thermo Scientific, 34094) and (2) in-gel TMBZ staining (TMBZ, Sigma, 860336-1G), both relying on the pseudo-peroxidase activity of heme (Wilks 2002). Detection of covalently bound heme in cytochrome c by ECL was done by separating proteins corresponding to 10–20 µg of chlorophyll on 12.5% acrylamide gel, followed by transfer to a PVDF membrane (Immobilon-P, IPVH00010) at 100 V (at 4°C) for 90 min and treatment of the membrane with ECL reagent. In-gel heme staining was performed by soaking the gel in a solution of 6.3 mM TMBZ dissolved in methanol in the dark, mixed with 1 M sodium acetate for 1–1.5 h in the dark with periodic shaking every 15–20 min. Bands revealing the heme peroxidase activity appeared immediately as a blue precipitate after addition of H2O2 (50% w/v) to a final concentration of 30 mM. For immunoblotting, membrane fractions corresponding to 10 µg of chlorophyll were separated by SDS–PAGE and transferred as described above. Immunodetection was done by incubating the membrane with Chlamydomonas α-CF1 (1:10,000), α-PsbO (1:5,000, Agrisera), α-cytochrome f (1:10,000), α-CCDA (1:1,000), and α-CCS5 (1:3,000). The CCDA and CCS5 antibodies were described in Motohashi and Hisabori (2010) and Gabilly et al. (2010), respectively. The anticytochrome f antibody is produced against a recombinant form of an apocytochrome f-GST fusion. The α-rabbit (1:10,000) was used as secondary antibody. Immunoblots were quantified by ImageJ by converting scanned images to gray scale and 8 bits (Stael et al. 2022). Each band was individually selected with the rectangular ROI selection, and area percentage was calculated using “Analyze > Gels > Label peaks.” Relative and adjusted density of each band was calculated for anti-CCDA immunodetection, and the quantification of the mean adjusted density of 3 independent blots is displayed as histograms. Statistical analysis was performed using the 2 sample t-test, and statistical significance was defined for P < 0.05.
Construction of the suppressor strains
The ccs4 strain expressing an additional copy of the CCDA gene was generated by introducing the construct pSL18-CCDA (Gabilly et al. 2011) in the CC-4525 strain via glass bead method (Kindle 1990) and selecting the transformants on TARG + Pm plates. The suppressor strains were reconstructed by introducing constructs containing the CCS4 gene from SU9 and SU11 into a ccs4Δccs5 recipient strain. For comparison, a ccs5 strain was also reconstructed by introducing the WT CCS4 gene in a ccs4Δccs5 mutant. The CCS4 gene (from ATG to stop including 165 bp of 3′ UTR) was cloned in front of the promoter of the beta-tubulin encoding gene (TUB2) in the pHyg3 plasmid (Berthold et al. 2002). The CCS4 sequence from WT (CC-124), SU9 (CC-5928), and SU11 (CC-5929) was amplified with pHyg3-CCS4-F1 (5′-GTCACAACCCGCAAACGGGCCCATGTCGACCGGCATTGAG-3′) and pHyg3-CCS4-R1 (3′-ATTGTACTGAGAGTGCACCATATGCCAGCTACCTACAGTC-5′), which are ApaI- and NdeI-engineered primers using purified genomic DNA as a template. The digested PCR products were cloned at ApaI and NdeI sites in pHyg3 to yield the constructs pHyg3-CCS4(WT), pHyg3-CCS4(SU9), and pHyg3-CCS4(SU11), which were introduced into the CC-4518 strain by the glass bead transformation method (Kindle 1990). Transformants were selected on TAP + Hyg based on the Hygromycin B resistance conferred by the iHyg3 cassette (containing the aphVII gene).
Construction of the HCF153-complemented strains
A chimeric CCS4-HCF153 gene in which the sequence corresponding to the first 36 residues of the CCS4 protein (including the predicted transmembrane domain followed by 6 residues) is fused to a sequence encoding the soluble domain of the Arabidopsis thaliana HCF153 protein (54 amino acids) was constructed (Supplementary Fig. 4). A 381-bp fragment containing part of intron 1 (42 bp), exon 2 (67 bp), intron 2 (103 bp), exon 3 (150 bp), and a short sequence following the stop codon with an XbaI site (18 bp) was synthesized (Genewiz). Part of the coding sequence in exon 2 and all codons in exon 3, which correspond to 54 residues in the Arabidopsis protein, were optimized to reflect the codon usage of a Chlamydomonas nuclear gene. The sequence was provided as a cloned fragment in pUC57-Kan, and the resulting plasmid was named pCCS4-HCF153. The 381-bp fragment was excised from pCCS4-HCF153 using the restriction enzymes NruI and XbaI and cloned at the corresponding sites in pSL18-ORF2L (Gabilly et al. 2011). In the resulting plasmid, named pSL18-HCF153, the chimeric CCS4-HCF153 gene is expressed under the PSAD promotor. The CC-4519 strain was transformed with pSL18-HCF153 using the glass bead transformation method (Kindle 1990), and transformants were selected on TARG + Pm under 0.3-µmol/m2/s. Transformants carrying the CCS4-HCF153 chimera were identified via diagnostic PCR using primers pSAD_Pr2 (5′-TTGGAGGTACGACCGAGATG-3′) and pSADT_R1 (5′-AGCTCTTCTCCATGGTACAG-3′). The amplicon was sequenced to verify the presence of the chimeric gene.
Glass bead transformation
Recipient strains were cultured in liquid TAP or TARG and grown until exponential phase. At a cell density of 2–5 × 106 cells/ml (measured by hemocytometer), cells were collected by centrifugation at 1,500 × g for 5 min at 25°C, followed by resuspension and incubation in autolysin to a density of ∼1 × 108 cells/ml for 45–90 min (depending on the strain). The efficiency of enzymatic digestion of the cell wall was tested with 0.1% Triton X-100 as described above in the section detailing genetic crosses. Autolysin was removed by centrifugation of the treated cells at 1,500 × g for 5 min at 25°C, and cells were resuspended in medium (TARG + 40 mM sucrose) to a final concentration of 1–2 × 108 cells/ml. After autolysin treatment, 300 µl of cells, 100 µl of 20% PEG 8,000, 300 mg of glass beads, and 1–2 µg of transforming DNA (linear or uncut) were mixed in a microfuge tube and vortexed for 30–45 s at maximum speed. After the glass beads settle, 300–350 µl of the mixture was transferred to 30 ml of TARG + 40 mM sucrose and incubated for 16–24 h at 25°C under 0.3 µmol/m2/s. After overnight recovery, cells are collected by centrifugation at 1,500 × g for 5 min at 25°C, resuspended in 0.5–1 ml of TAP, and plated on a selective medium. Successful transformation yields colonies in 5–7 days.
Bioinformatics
BLASTp, RPS-BLAST, and PSI-BLAST searches (Altschul et al. 1997) against the nonredundant database at NCBI and BLASTp search against 1KP database (Matasci et al. 2014) were carried out with default parameters (2020 February 1). Protein domains were identified using Pfam (El-Gebali et al. 2019) and conserved domain (Marchler-Bauer et al. 2015) databases. Profile–profile searches for domain identification were carried out using CDvist (Adebali et al. 2015) and HHpred (Soding et al. 2005) servers. Multiple sequence alignments were constructed using L-INS-I method in MAFFT program (Katoh et al. 2019). Sequence alignments were edited in Jalview (Waterhouse et al. 2009). Transmembrane regions and signal peptides prediction from amino acid sequences were carried out using Phobius (Käll et al. 2004). Sequence logos were generated using Skylign (Wheeler et al. 2014).
Results
CCS4 and CCS5 operate in functionally redundant pathways
Previously, we showed that CCS4 and CCS5 are specifically required to maintain the heme-binding cysteines in the reduced form, a critical step for plastid cytochrome c assembly (Xie et al. 1998; Gabilly et al. 2010, 2011). To define the contribution of each component in the assembly pathway of c-type cytochromes, the photosynthetic growth requiring the correct assembly of the membrane-bound plastid cytochrome f (Kuras et al. 1995; Zhou et al. 1996) was assessed in the ccs4 (ccs4), ccs5 (Δccs5), and ccs4ccs5 double mutants (ccs4Δccs5). Both ccs4 and ccs5 mutations correspond to complete loss of function. The ccs5 mutation is a 3-kb deletion encompassing the entire CCS5 coding sequence (Gabilly et al. 2010). The ccs4 mutation creates a stop codon at residue 50 of the CCS4 protein and results in a gene product which lacks the entire C-terminal soluble domain predicted to face the stroma. Moreover, the ccs4 mutation yields a drastic decrease in the CCS4 transcript accumulation, and we assume that there is probably no functional protein produced in the ccs4 mutant (Gabilly et al. 2011). To assess phototrophic growth, cells were grown under a light intensity of 30–50 µmol/m2/s, where they utilize atmospheric CO2 to perform photosynthesis. Compared to the ccs4 and Δccs5 single mutants that show residual photosynthetic growth, the ccs4Δccs5 double mutants are completely deficient for growth under phototrophic conditions (Fig. 1a). Under low light (0.3 µmol/m2/s), all mutant strains can grow in the presence of acetate and atmospheric CO2 due to the dominant contribution of respiration vs photosynthesis. Next, we performed chlorophyll fluorescence rise-and-decay kinetics, a noninvasive measurement of photosystem II (PSII) activity (Murchie and Lawson 2013). In WT, the initial rise indicates electron movement through PSII, which is followed by a decay that indicates electron transfer through the cytochrome b6f (Fig. 1b). When the energy absorbed by the chlorophyll in PSII cannot be utilized in the photosynthetic electron transport chain because of a complete block at cytochrome b6f, a saturating rise in fluorescence is observed with no decay phase thereafter. In ccs4Δccs5, the rise and plateau are a signature of a block at the level of cytochrome b6f complex due to the absence of holocytochrome f (Fig. 1b). In ccs4 and Δccs5, the decay phase is less pronounced than in WT, an indication that the strains retained cytochrome b6f functionality because some level of holocytochrome f is still assembled (Fig. 1b), which agrees with the residual photosynthetic growth (Fig. 1a). The level of holocytochrome f assembly in the different mutant strains was assessed biochemically via heme staining (Fig. 1c). Holocytochrome f assembly is completely blocked in ccs4Δccs5, whereas the single null mutants accumulate partial levels of holocytochrome f, ∼10% in Δccs5 and ∼>2% in ccs4, respectively. These results show that the synthetic photosynthetic growth defect of ccs4Δcss5 is indeed due to a complete block in holocytochrome f assembly. The photosynthetic growth of the different mutant strains (Fig. 1a) correlates with the level of assembled holocytochrome f (Fig. 1c). Loss of CCS4 function yields a more pronounced cytochrome f deficiency than loss of CCS5, whereas loss of both results in no cytochrome f assembly. Therefore, we conclude that CCS4 and CCS5 are partially redundant in the context of holocytochrome f assembly.
![The ccs4Δccs5 mutant exhibits a photosynthetic growth defect due to a complete loss of cytochrome c assembly. a). Ten-fold dilution series of WT (CC-124), Δccs5 (CC-4129), ccs4 (ccs4(pCB412)), and ccs4Δccs5 [CC-4517, CC-4518, and ccs4 ccs5 (9) and (11)] strains were plated on acetate-containing (right panel) and minimal (left panel) medium. Cells grown phototrophically (CO2) were incubated at 25°C for 7 days with 30–50 µmol/m2/s of light. Cells grown mixotrophically (acetate + CO2) were incubated at 25°C for 7 days with 0.3 µmol/m2/s of light. b) The fluorescence induction and decay kinetics of ccs4Δccs5 is shown in comparison to that of Δccs5, ccs4, and WT. c) Heme staining and α-PsbO immunodetection were performed on total protein extracts. Strains are as in 1a except for the ccs4 (CC-4519) and ccs4Δccs5 (CC-4518). Cells were grown mixotrophically on acetate with 0.3 µmol/m2/s of light at 25°C for 5–7 days. Samples correspond to 30 µg of chlorophyll. Immunodetection of PsbO was used as a loading control. The vertical line indicates cropping from the same gel and same exposure.](https://oup.silverchair-cdn.com/oup/backfile/Content_public/Journal/genetics/225/2/10.1093_genetics_iyad155/3/m_iyad155f1.jpeg?Expires=1749632497&Signature=y-7HQk75Opd9jmP~jMftiV8bN6ggbMq-J~YfGg6647TqyRNdem4cH0wO10VyyX2hDNHm7kvfZJhMrBEQH8af68WYSkDWfdFTsL653xmdFZncqcEzohQ0M9EUe3F4LHdvdJwERLckuPZR8fRCPS0zGL9vIWPcrL0IpokEleOOhdUBdcsZL1JgjxN--lVlqTYO~hJQOuXNJia-6MYBjw7sGKH~O7IUqXmARYyUcayEB0FxtmVDtiNVz9UIzeUIcn1rpCa~EEJXvVsgRUMlm5nQZYdA5D~xwktbtg2DboAjriITcNO-iwkleWlnADhjHcPIEQnYNpgHAmDwwmC6w3AW-w__&Key-Pair-Id=APKAIE5G5CRDK6RD3PGA)
The ccs4Δccs5 mutant exhibits a photosynthetic growth defect due to a complete loss of cytochrome c assembly. a). Ten-fold dilution series of WT (CC-124), Δccs5 (CC-4129), ccs4 (ccs4(pCB412)), and ccs4Δccs5 [CC-4517, CC-4518, and ccs4 ccs5 (9) and (11)] strains were plated on acetate-containing (right panel) and minimal (left panel) medium. Cells grown phototrophically (CO2) were incubated at 25°C for 7 days with 30–50 µmol/m2/s of light. Cells grown mixotrophically (acetate + CO2) were incubated at 25°C for 7 days with 0.3 µmol/m2/s of light. b) The fluorescence induction and decay kinetics of ccs4Δccs5 is shown in comparison to that of Δccs5, ccs4, and WT. c) Heme staining and α-PsbO immunodetection were performed on total protein extracts. Strains are as in 1a except for the ccs4 (CC-4519) and ccs4Δccs5 (CC-4518). Cells were grown mixotrophically on acetate with 0.3 µmol/m2/s of light at 25°C for 5–7 days. Samples correspond to 30 µg of chlorophyll. Immunodetection of PsbO was used as a loading control. The vertical line indicates cropping from the same gel and same exposure.
Loss of CCS4 and CCS5 can be chemically corrected by exogenous thiols
We previously showed that the partial deficiency in photosynthetic growth and holocytochrome f assembly in ccs4 and Δccs5 mutants could be rescued by the application of exogenous thiols (Gabilly et al. 2010, 2011). Application of reducing agents such as MESNA (2-mercaptoethane sulfonate sodium) or dithiothreitol (DTT, not shown) can also partially rescue the ccs4Δccs5 double mutant for phototrophic growth (Fig. 2a). Using fluorescence rise-and-decay kinetics, we documented that the chemical correction of the growth defect is due to a recovery of cytochrome b6f function, which indicates restoration of holocytochrome f assembly (Fig. 2b). Note that a higher concentration of MESNA (4 mM vs 1.5 mM) was necessary to evoke a rescue of the ccs4Δccs5 double mutant in the fluorescence transient experiments. We presumed that this difference is because cells are grown phototrophically under 30–50 µmol/m2/s for the growth assessment (Fig. 2a) vs mixotrophically under 0.3 µmol/m2/s for the measurement of fluorescence transients (Fig. 2b). The ΔpetA mutant, a null mutant in the plastid gene encoding apocytochrome f (Zhou et al. 1996), is completely blocked for photosynthetic growth and does not show any rescue with exogenous thiols, as expected.
The photosynthetic defect in ccs4Δccs5 is partially rescued by exogenous thiols. a) Ten-fold dilution series of WT (CC-4533), Δccs5 (CC-5922), ccs4 (CC-5925), ccs4Δccs5 (CC-5927), and ΔpetA (CC-5935) were plated on acetate and minimal medium with or without MESNA (2-mercaptoethane sulfonate sodium). Cells grown phototrophically (CO2) were incubated at 25°C for 20 days with 30–50 µmol/m2/s of light. Cells grown mixotrophically (acetate + CO2) were incubated at 25°C for 14 days with 0.3 µmol/m2/s of light. The horizontal lines indicate cropping from the same plate of serial dilution. b) Fluorescence transients were measured on cells grown on solid acetate-containing media for 5 days (with 0.3 µmol/m2/s of light) with or without MESNA. Strains are as in (a) except that the ccs4Δccs5 strain is CC-4518.
The abundance of CCDA is diminished in the ccs4 but not in the ccs5 mutant
Expression of an additional ectopic copy of CCDA encoding the trans-thylakoid thiol–disulfide oxidoreductase was shown to suppress the ccs4 mutant (Gabilly et al. 2011). This result led to the hypothesis that CCS4 might be involved in the stabilization of CCDA. To test this hypothesis, we monitored the abundance of CCDA by immunoblotting using an antibody against Arabidopsis CCDA but cross-reacting with the algal ortholog (Motohashi and Hisabori 2010). Because the original CCDA-suppressed strain (Gabilly et al. 2011) lost the suppression over time, the strain was reconstructed. The suppression is weak but is best seen when the strain with an extra ectopic copy of CCDA is grown in mixotrophic conditions under 30- to 50-mmol/m2/s illumination (Fig. 3a). The CCDA protein abundance decreased to ∼50% of the WT level in ccs4 but was restored to normal levels in the CCS4 complemented strain and enhanced in the strain expressing an ectopic copy of the CCDA gene (Fig. 3b and Supplementary Fig. 1a). Conceivably, the decrease in CCDA abundance is not specific to the loss of CCS4 but is a more general result of perturbing the function of the cytochrome b6f node. However, we showed that the loss of CCS5 has no effect on CCDA abundance (Fig. 3c and Supplementary Fig. 1b), which suggests that the decrease in CCDA steady-state level is specific to the loss of CCS4. From this result, we conclude that CCDA abundance is diminished when CCS4 function is compromised.
![The abundance of CCDA is diminished in the ccs4 but not in the ccs5 mutant. a) Suppression of the photosynthetic growth defect in the ccs4 mutant by ectopic expression of CCDA was assessed by 10-fold dilution series. The WT (CC-124), ccs4 (CC-4520), CCS4-expressing [ccs4 (CCS4), CC-4522], and CCDA-expressing [ccs4 (CCDA), CC-5926] ccs4 strains were plated on acetate medium and grown mixotrophically (with 30–50 µmol/m2/s or 0.3 µmol/m2/s of light) at 25°C for 14 days. The horizontal lines indicate cropping from the same plate. b) and c) CCDA was immunodetected on total protein extracts of strains described in a) and the Δccs5 (CC-4129) and corresponding complemented [Δccs5(CCS5), CC-4527] strains. Cells were grown mixotrophically (on acetate with 0.3 µmol/m2/s of light) at 25°C for 5–7 days. The vertical line indicates cropping from the same blot and same exposure. Samples (100%) correspond to 10 µg of chlorophyll. Immunodetection with antisera raised against Chlamydomonas CF1 (coupling factor 1 of chloroplast ATPase) and CCS5 was used as control (Gabilly et al. 2010).](https://oup.silverchair-cdn.com/oup/backfile/Content_public/Journal/genetics/225/2/10.1093_genetics_iyad155/3/m_iyad155f3.jpeg?Expires=1749632497&Signature=avPCHhvDAp3oXHsDAwe9cQPJP0OPVV~Q~RzeeHmcFdWn3LcI4yT-vmqutZMWx-hY5WiscW2Dxm4TqtpDhQtx6qesMEwmZnZSZGQ6N4PNIhNNcFB7pYaTCqj9QW7-d47RtOXAegoL1mgpNURRcIwawkrKUXXUEnYM262C~IlrgUsqtQjFSzZIF~ih-RMazfYk7Mb~duY13JoTvdrXfFBHO0jQZOo3cORccF-7eMviSt3BzPJwqk95-E5NgscLsFLuYVjAlJNGOuJZeuPJIldBL~umh-mEFBf6B5~FUTwBH6On~RElvSg6JDOYKywAj7XyWn9eTJetD9BldyqQMsHNGA__&Key-Pair-Id=APKAIE5G5CRDK6RD3PGA)
The abundance of CCDA is diminished in the ccs4 but not in the ccs5 mutant. a) Suppression of the photosynthetic growth defect in the ccs4 mutant by ectopic expression of CCDA was assessed by 10-fold dilution series. The WT (CC-124), ccs4 (CC-4520), CCS4-expressing [ccs4 (CCS4), CC-4522], and CCDA-expressing [ccs4 (CCDA), CC-5926] ccs4 strains were plated on acetate medium and grown mixotrophically (with 30–50 µmol/m2/s or 0.3 µmol/m2/s of light) at 25°C for 14 days. The horizontal lines indicate cropping from the same plate. b) and c) CCDA was immunodetected on total protein extracts of strains described in a) and the Δccs5 (CC-4129) and corresponding complemented [Δccs5(CCS5), CC-4527] strains. Cells were grown mixotrophically (on acetate with 0.3 µmol/m2/s of light) at 25°C for 5–7 days. The vertical line indicates cropping from the same blot and same exposure. Samples (100%) correspond to 10 µg of chlorophyll. Immunodetection with antisera raised against Chlamydomonas CF1 (coupling factor 1 of chloroplast ATPase) and CCS5 was used as control (Gabilly et al. 2010).
Photosynthetic revertants of ccs4Δccs5 double mutant are restored for cytochrome f assembly
Over the course of this study, we noticed that two slow-growing ccs4Δccs5 strains (#9 and #11, Fig. 1a) had evolved a faster growth phenotype. We suspected reversion to photosynthetic proficiency and named the strains SU9 and SU11 for suppressors of ccs4Δccs5#9 and #11, respectively. To test for the restoration of photosynthesis, we assessed the growth on acetate under high illumination. Under such conditions, cytochrome b6f-deficient strains display significant photosensitivity (Malnoë et al. 2014). A ccs4Δccs5 strain that had retained a tight photosynthetic deficiency due to loss of cytochrome f assembly displayed a photosensitive phenotype (Fig. 4a). Interestingly, both SU9 and SU11 were resistant to high light treatment, suggesting the original ccs4Δccs5#9 and ccs4Δccs5#11 strains have recovered the b6f function (Fig. 4a). Fluorescence transients confirmed restoration of the b6f function in SU9 and SU11 (Fig. 4b). We showed that photosynthetic proficiency in SU9 and SU11 is due to recovery of holocytochrome f assembly, as shown by heme staining and immunodetection (Fig. 4c). In SU9, cytochrome f assembly restoration is partial (∼25%), but in SU11, restoration appears to be WT-like (Fig. 4c). The level of holocytochrome f assembly in SU9 and SU11 (Fig. 4c) correlates with the recovery of photosynthetic proficiency assessed by growth in mixotrophic conditions and measurement of fluorescence transients (Fig. 4a and b).
Suppressors of ccs4Δccs5 are restored for phototrophic growth and cytochrome f assembly. a) Ten-fold dilution series of WT (CC-124), Δccs5 (CC-4129), ccs4 (CC-4519) and ccs4Δccs5 (CC-4518), SU9 (CC-5928), and SU11 (CC-5929) were plated on acetate medium and grown mixotrophically (with 250 or 0.3 µmol/m2/s of light) at 25°C for 14–18 days. The horizontal line indicates cropping from the same plate of serial dilutions. b) The fluorescence induction and decay kinetics observed in a dark-to-light transition of the SU9 and SU11 suppressors are shown in comparison to ccs4Δccs5, Δccs5, and WT. Strains are the same as in a). c) Heme staining and immunoblotting (α-cyt f and α-PsbO) were performed on total protein extracts prepared from cells (same strains as in (a)) grown mixotrophically (on acetate with 0.3 µmol/m2/s of light) at 25°C for 5–7 days. Samples (100%) correspond to 10 µg of chlorophyll. Immunodetection with antisera against PsbO was used as loading control.
Reversion of the ccs4Δccs5 is due to spontaneous dominant gain-of-function mutations in CCS4
Next, we sought to determine the molecular basis of the reversion in SU9 and SU11. We ruled out the possibility that a suppressor mutation maps to the CCS5 gene, considering a 3-kb region containing the entire coding sequence of CCS5 is missing in Δccs5 (Gabilly et al. 2010). Instead, we suspected that the point mutation (C to T) at a codon (CAG) encoding glutamine (Q50) in the CCS4 gene might have reverted in SU9 and SU11 (Gabilly et al. 2011). Sequencing of the CCS4 gene in SU9 revealed the nonsense mutation in the original ccs4 mutant (TAG) was changed to a missense mutation (TGG) resulting in a tryptophan (W) residue (Fig. 5a). Surprisingly, in SU11, the CCS4 gene carried an in-frame deletion of 3 codons removing the stop codon (GCTTAGATG), resulting in a CCS4 protein shorter by 3 residues (Fig. 5a). The intragenic mutations in the CCS4 gene alter residues in the soluble domain of CCS4, which is predicted to face the stromal side of the thylakoid membrane (Gabilly et al. 2011). We named the suppressor mutations in SU9 and SU11, CCS4-2, and CCS4-3, respectively (Supplementary Table 1).
Mutations in CCS4 suppress the cytochrome f assembly defect in the ccs5 mutant. a) Amino acid sequence of CCS4 deduced from the CCS4 gene in the original ccs4-F2D8 mutant; SU9 and SU11 are shown in black, green, and blue, respectively. The putative transmembrane domain, predicted using the TMHMM-2.0 prediction tool (Krogh et al. 2001), is underlined. Sequence analysis of the CCS4 gene shows that the original nonsense mutation in the ccs4 strain is changed to a missense mutation in SU9, resulting in the stop codon (at residue 50 of the protein) becoming a codon for tryptophan (WT codon is glutamine). In SU11, an in-frame deletion mutation encompassing the missense mutation resulted in a deletion of 3 residues A49Q50M51 in the CCS4 gene product. b) WT (CC-124), ccs5 (Δccs5, CC-4129), ccs4Δccs5 (CC-4518), and the reconstructed ccs5 (CCS4-WT, CC-5932), SU9 (CCS4-SU9, CC-5933), and SU11 (CCS4-SU11, CC-5934) strains were used for 10-fold dilution series. The Δccs5, SU9, and SU11 were reconstructed by introducing the constructs pHyg3-CCS4(WT), pHyg3-CCS4(SU9) and pHyg3-CCS4(SU11) into a ccs4Δccs5 recipient strain. Two independent transformants for each construct are shown. Cells were plated on minimal medium and grown phototrophically (with 30–50 µmol/m2/s of light) or mixotrophically (with 0.3 µmol/m2/s of light) at 25°C for 14–18 days. The horizontal lines indicate cropping from the same plate of serial dilutions. c) In-gel heme staining was performed on total protein extracts corresponding to 30 µg of chlorophyll. Samples from the Δccs5 and original SU9 (CC-5933) and SU11 (CC-5934) were loaded on the gel for comparison to the reconstructed strains. The strains are as in b) and ΔpetA is CC-5935. The level of LHC complex is used as a loading control. One-microgram equine heart cytochrome c (∼10 kDa; Sigma) is used as a control for the heme-dependent peroxidase activity.
To ascertain that the molecular changes identified in CCS4 were responsible for the suppression, we performed genetic analysis by crossing SU9 and SU11 to the original ccs4Δccs5, or ccs4 or Δccs5 mutants. However, despite several attempts, no complete tetrads and very few viable haploid progeny were recovered from our genetic crosses. We opted to reconstruct the suppressed strains in the background of the ccs4Δccs5 strain. The CCS4 gene from WT, SU9, and SU11 strains was cloned in a plasmid, and the resulting constructs were introduced into a ccs4Δccs5 mutant followed by analysis of photosynthetic growth and holocytochrome f assembly from representative transformants. In comparison to the recipient ccs4Δccs5 (that is completely blocked for photosynthetic growth), transformants expressing the WT CCS4 gene showed the phenotype of the Δccs5 strain, whereas transformants expressing CCS4-2 (from SU9) or CCS4-3 (from SU11) exhibit enhanced photosynthetic growth (Fig. 5b). The phenotypes of the reconstructed suppressed strains mirrored that of the original SU9 and SU11, with the reconstructed SU11 displaying a greater suppression of the photosynthetic defect than reconstructed SU9. We further demonstrated that the restoration of photosynthetic growth in the reconstructed suppressed strains is due to restoration of cytochrome f assembly, as shown by heme staining (Fig. 5c). We also noted the restoration of holocytochrome b6 accumulation, which is dependent upon the level of holocytochrome f, indicating that cytochrome b6f complex assembly is also restored (Fig. 5c). In both the original SU9 and SU11 and the reconstructed suppressed strains, the suppressor mutations restore the level of holocytochrome f assembly above that of the ccs5-null mutant. This result suggests the suppressor mutations CCS4-2 and CCS4-3 confer a gain of function. The photosynthetic growth of Δccs5 × SU9 and Δccs5 × SU11 diploids is enhanced compared to that of Δccs5 × Δccs5 diploids (Fig. 6) (Supplementary Table 1). This demonstrates that the CCS4-2 and CCS4-3 mutations in SU9 and SU11 are dominant, as expected for gain-of-function mutations. We conclude that CCS4-2 and CCS4-3 are dominant suppressor mutations and constitute the molecular basis for the restoration of cytochrome f assembly in the SU9 and SU11 strains.
Mutations in CCS4 in SU9 and SU11 are dominant. a) Ten-fold dilution series of diploid (+/+, Δccs5/Δccs5, Δccs5/Δccs5 CCS4-SU9, and Δccs5/Δccs5 CCS4-SU11) strains were plated on minimal medium and grown phototrophically (with 30–50 µmol/m2/s of light) at 25°C for 14–18 days (left panel) and mixotrophically in 0.3 µmol/m2/s for 20–21 days (right panel). Two representatives of each diploid constructed with a Δccs5 strain are shown. b) The fluorescence induction and decay kinetics of one representative diploid strain. Strains are the same as in a).
The abundance of CCDA is unchanged in the suppressor strains
Because loss of CCS4 function yields a decrease in the abundance of CCDA (Fig. 3), we wondered if the steady-state level of CCDA in the suppressor strains was restored with the regain of CCS4 activity. As shown in Supplementary Fig. 2, the CCDA level in SU9 and SU11 is restored compared to a ccs4 mutant. Immunodetection of CCDA indicated that the abundance of the protein in the suppressor strains is comparable to that of the WT or Δccs5. We conclude that alterations of the residues caused by either CCS4-2 or CCS4-3 mutations in the suppressors do not modify the accumulation of the CCDA protein in comparison to the WT strain or ccs5 mutant.
CCS4-like proteins occur in the green lineage
Except for orthologs detected in close relatives of Chlamydomonas, we previously reported that CCS4 does not appear to be evolutionarily conserved (Gabilly et al. 2011). CCS4-like proteins in algal species related to Chlamydomonas are small proteins (<100 residues) with a putative transmembrane domain suggesting anchoring to the membrane (Supplementary Fig. 3). Remarkably, the C-terminal domain of algal CCS4 orthologs is predicted to face the stroma and is characterized by an abundance of charged residues (Supplementary Fig. 3). To identify CCS4-like proteins that might have considerably diverged, we performed an exhaustive search using PSI-BLAST (Supplementary Data 1). CCS4-like proteins were identified in several species of green algae and land plants, which constitute the Viridiplantae or the entire green lineage (Supplementary Data 1, Supplementary Dataset 1A, and Fig. 7).
CCS4-like proteins are present throughout the green lineage. Sequences of Arabidopsis thaliana (NP_194884.1), Picea sitchensis (Sitka spruce, ABK22505.1), C. reinhardtii (ADL27744.1), Chara braunii (Braun's stonewort, GBG73810.1), and Selaginella moellendorffii (spikemoss, XP_024543966.1) were aligned using Clustal Omega (Sievers and Higgins 2014) and manually edited in Word. The transmembrane domains are highlighted in gray; residues with positive and negative charges are highlighted in blue and red, respectively. The percentage of similarity followed by the percentage of charged residues is indicated.
HCF153, a CCS4-like protein functions in plastid cytochrome c assembly
To test if the function of the CCS4-like proteins retrieved from our analysis (Supplementary Data 1, Supplementary Dataset 1A, and Fig. 7) is conserved in the green lineage, we chose to perform a heterologous functional complementation experiment. For our complementation experiment, we chose HCF153, the Arabidopsis CCS4-like protein. Because the CCS4 and HCF153 proteins are most divergent in their N-termini (Lennartz et al. 2006; Gabilly et al. 2011), we designed a CCS4-HCF153 construct encoding a chimeric protein in which the N-terminal domain including the putative transmembrane of CCS4 is fused to the soluble domain of HCF153 (Supplementary Fig. 4). As shown in Fig. 8, expression of the CCS4-HCF153 construct complements the defect due to loss of CCS4. The transformants complemented by the CCS4-HCF153 gene could not be distinguished from those expressing the CCS4 gene based on the assessment of photosynthetic growth (Fig. 8a), restoration of cytochrome f assembly, and cytochrome b6f function (Fig. 8b and c). We conclude that HCF153 functions in plastid cytochrome c assembly, a finding that supports the conservation of CCS4 function from green algae to land plants.
Arabidopsis HCF153 complements the ccs4 mutant. a) Ten-fold dilution series of ccs4 (ccs4_empty, CC-4520), ccs4 transformed by the CCS4 gene (CC-4522), and two independent ccs4 transformants carrying the CCS4-HCF153 gene (CC-5592 and CC-5593) were plated on minimal medium and grown phototrophically (with 30–50 µmol/m2/s of light) at 25°C for 14 days. b) The fluorescence induction and decay kinetics of the one representative CCS4-HCF153 transformant (CC-5592) is shown in comparison to that of a WT (CC-124) and ccs4 strains (CC-4520). c) In-gel heme staining was performed on total protein extracts corresponding to 30 µg of chlorophyll. The strains are as in a) and b), and the ccs4 strain is CC-4519.
CCS4—like proteins are related to mitochondrial Cox16
Interestingly, our analysis also revealed that CCS4 orthologs are related to COX16, a protein involved in the assembly of cytochrome c oxidase, a respiratory enzyme required for energy-transduction in mitochondria (Carlson et al. 2003) (Supplementary Data 1, Supplementary Dataset 1B, and Supplementary Fig. 5). First identified in yeast, COX16 is a mitochondrial inner membrane-bound protein with a C-terminal domain facing the IMS (Carlson et al. 2003). The biochemical activity of COX16 still remains enigmatic, but its placement in a disulfide-reducing pathway required for Cu incorporation into the Cox2 subunit of cytochrome c oxidase suggests COX16 and CCS4 might share some commonalities in their biochemical activity (Su and Tzagoloff 2017; Aich et al. 2018; Cerqua et al. 2018).
Discussion
CCS4 and CCS5 are functionally redundant in the delivery of reductants: In all bacteria and plastids, cytochrome c maturation relies on the transfer of reducing power from the n-side to the p-side via sequential thiol–disulfide exchanges involving a membrane thiol–disulfide oxidoreductase of the DsbD family and a thioredoxin-like protein (Gabilly and Hamel 2017; Bushweller 2020; Das and Hamel 2021). Operation of this transmembrane pathway is critical to counter the oxidation of the heme-linking cysteines in the CXXCH motif of apocytochrome c into a disulfide (Gabilly and Hamel 2017; Das and Hamel 2021). The disulfide-bonded heme-binding site is directly reduced by a membrane-anchored p-side facing thioredoxin-like named CcmG/ResA/CcsX in bacteria (Beckman and Kranz 1993; Beckett et al. 2000; Erlendsson et al. 2003) and CCS5/HCF164 in plastids (Lennartz et al. 2001; Gabilly et al. 2010). The Chlamydomonas ccs5-null and the Arabidopsis hcf164-null mutants are partially deficient in plastid holocytochrome f (Gabilly et al. 2010, 2011). In bacteria, loss of the thioredoxin-like CcmG/CcsX/ResA protein results in a complete block in cytochrome c maturation, suggesting that redundancy in the provision of reductants is specific to the assembly of c-type cytochromes in the plastid (Beckman and Kranz 1993; Schiött et al. 1997; Fabianek et al. 1998; Beckett et al. 2000; Deshmukh et al. 2000). In Chlamydomonas, the ccs4-null is still able to grow photosynthetically because it is partially deficient in plastid holocytochrome f (Gabilly et al. 2011). However, the ccs4-null ccs5-null double mutant displays a synthetic photosynthesis-minus phenotype characterized by a complete loss of holocytochrome f accumulation (Fig. 1). Evidence that ccs4Δccs5 can be rescued by application of exogenous thiols similarly to each single ccs4 and Δccs5 mutants (Gabilly et al. 2010, 2011 and Fig. 2) solidifies the view that the complete block in plastid cytochrome c maturation is attributed to a defect in the provision of reductants. Altogether, our results indicate that CCS4 and CCS5 operate in functionally redundant pathways for the supply of reducing power in the context of cytochrome c assembly.
CCS4 and CCDA function in a complex? The finding that ectopic expression of CCDA partially suppresses the photosynthetic deficiency of the ccs4 mutant led to the proposal that CCS4 might provide stabilization for the CCDA protein (Gabilly et al. 2011). CCDA levels are significantly decreased due to loss of CCS4 function and restored upon complementation with the CCS4 gene or suppression by CCDA (Fig. 3). Because the CCDA gene is not downregulated in the ccs4 mutant (Gabilly et al. 2011), a plausible scenario is that CCDA is still synthesized in the cytosol, imported into the plastid but presumably partially turned over in the absence of CCS4 (Fig. 3). One attractive model is that CCS4 and CCDA are in a complex and physically interact. In Arabidopsis, HCF153, the functional homolog of CCS4 is detected in a larger protein complex at the thylakoid membrane, but the presence of CCDA in this complex awaits experimental testing (Lennartz et al. 2006).
The soluble domain of CCS4 controls the delivery of reducing power by CCDA? The functional importance of the CCS4 C-terminal domain was evidenced with the isolation of dominant CCS4-2 and CCS4-3 mutations in photosynthetic revertants of the ccs4Δccs5 mutants (Figs. 4 and 5). In the CCS4-2 mutation, the original stop codon was changed to a codon encoding a tryptophan (instead of glutamine in the WT CCS4), while in the CCS4-3 mutation, an in-frame deletion produces a CCS4 protein which is shorter by 3 amino acids (Fig. 5a). Suppressor mutations in the CCS4 gene restore cytochrome f accumulation above the residual levels in the Δccs5 mutant, suggesting that these are gain-of-function mutations (Figs. 4 and 5). The mechanism of suppression is unclear, but one possible interpretation is that alteration of the stroma-facing domain of CCS4, where the changes occur, enhances the transduction of reducing power across the thylakoid membrane when CCS5 no longer operates. Indeed, the CCS5 gene is deleted in SU9 and SU11, and hence, the CCS5-dependent reduction of the disulfide-bonded CXXCH motif no longer takes place. In instance where CCS5 function is lost, the provision of reducing equivalents to the lumen is probably still dependent upon CCDA. The abundance of CCDA in the suppressor strains is like that of a Δccs5 mutant (Supplementary Fig. 2), and we established previously that CCDA overexpression does not compensate for the loss of CCS5 (Gabilly et al. 2011). It is conceivable that alteration of the residues in the C-terminal domain of CCS4, predicted to face the stroma, modifies the redox activity of CCDA such that delivery of reducing power to the lumen is increased in the suppressor strains.
Conservation in the green lineage and evolutionary relationship to COX16: In this study, we also provide evidence for the occurrence of highly divergent CCS4-like proteins in Viridiplantae. Interestingly, HCF153, the CCS4-like protein in Arabidopsis (Fig. 7), localizes to the plastid and is tightly bound to the thylakoid membrane (Lennartz et al. 2006). A possible function in holocytochrome f maturation is supported by the fact that loss of HCF153 elicits a cytochrome b6f assembly defect (Lennartz et al. 2006). That HCF153 is the functional equivalent to CCS4 is now established from our heterologous complementation experiments (Fig. 8). Our result suggests that despite little sequence conservation, the function of CCS4 in plastid cytochrome c maturation is maintained in the green lineage. CCS4-like proteins in the plastid are related to COX16-like proteins in mitochondria, which exhibit a transmembrane domain and a charged C-terminal moiety facing the IMS (Supplementary Fig. 5). Studies in yeast and humans have placed Cox16 in a disulfide-reducing pathway for insertion of Cu in cytochrome c oxidase, but its biochemical activity remains undeciphered (Su and Tzagoloff 2017; Aich et al. 2018; Cerqua et al. 2018). Considering there are no residues or motifs suggesting a reducing activity, CCS4 and COX16 might function in a disulfide-reducing pathway to recruit the source of reducing power or to create a chemical microenvironment for efficient use of the reductant by dedicated disulfide reductases. In plastids, stromal thioredoxin-m is the proposed source of electrons for CCDA based on in organello experiments (Motohashi and Hisabori 2010). In the mitochondrial IMS, the source of electrons in the disulfide-reducing pathway for Cox16-dependent metalation of Cox2 is currently unknown (Swaminathan and Gohil 2022).
Two routes for distribution of reducing power in cytochrome c assembly: We propose that maintenance of the heme-binding cysteines in the reduced state in the lumen depends on the delivery of reducing power via two routes (1 and 2) with CCDA and CCS4/HCF153 as shared components to both routes (see Graphical Abstract). In a simple model, CCDA transduces the source of reducing power from stroma to lumen across the thylakoid membrane via thiol–disulfide exchanges and physically interacts with CCS4/HCF153. CCS4/HCF153 is probably directly required for the CCDA-dependent transfer of reducing power, possibly by recruiting a stromal reductant to CCDA or controlling the thiol–disulfide chemistry involving the redox-active cysteines in CCDA. The CCS5-dependent pathway (route 1) is conserved in bacteria and relies on CCS5/HCF164 for the provision of reductants in the lumen. The other pathway is CCS5-independent (route 2) and must rely on one or several lumen resident enzyme(s) (X) whose identity remains unknown. If the operation of the trans-thylakoid disulfide reduction pathways is critical to counter the oxidation of the heme-linking cysteines in the CXXCH motif into a disulfide, the nature of the oxidant is unknown (Gabilly and Hamel 2017; Das and Hamel 2021). In bacteria, there is genetic evidence that the heme-linking cysteines are first targets of the disulfide bond-forming machinery and subsequently reduced via the transmembrane disulfide-reducing pathway prior to the heme ligation reaction (Erlendsson and Hederstedt 2002; Erlendsson et al. 2003; Turkarslan et al. 2008; Small et al. 2013). We envision that a similar mechanism exists in the thylakoid lumen. In Arabidopsis, a disulfide bond-forming enzyme named Lumen Thiol Oxidoreductase1 (LTO1) has been described (Feng et al. 2011; Karamoko et al. 2011). That the CXXCH of apocytochrome f is a relevant target of action of LTO1 still awaits experimental testing.
Data availability
Strains and plasmids used in this study can be requested at the Chlamydomonas collection center. The authors affirm that all data necessary for confirming the conclusions of the article are present within the article, figures, tables, and Supplementary material.
Supplemental material available at GENETICS online.
Funding
This work was supported by a U.S. Department of Energy (DOE), Office of Science, Basic Energy Sciences (BES) grant (DE-SC0014562) to P.H. and grant R35GM131760 from National Institutes of Health (to I.B.Z.).
Literature cited
Author notes
Conflicts of interest The authors declare no conflict of interest.
