An ancient glaucophyte c₆-like cytochrome related to higher plant cytochrome c_6A is imported into muroplasts

Cyanophora paradoxa is a model flagellate for the studies of the endosymbiotic cyanobacterial origin of the primary plastids bounded by two membranes. It belongs to the glaucophytes, which are the most basal branch of Archaeplastida, the eukaryotic supergroup also containing rhodophytes, chlorophytes and streptophytes (Adl et al., 2019; Rodríguez-Ezpeleta et al., 2005). Alternative phylogenies suggest rhodophytes being the most basal archaeaplastidial lineage instead of glaucophytes (Sánchez-Baracaldo et al., 2017; Sibbald and Archibald, 2020). It was proposed that the cyanobacterial ancestor of primary plastids was a relatively complex nitrogen-fixing (heterocyst-forming) cyanobacterium (Dagan et al., 2013; Deschamps et al., 2008; Deusch et al., 2008) or more likely a freshwater/terrestrial cyanobacterium related to Gloeomargarita lithophora (Moore et al., 2019; Ponce-Toledo et al., 2017).

The plastids of glaucophytes are termed muroplasts or cyanelles, because they possess primordial cyanobacterial features absent from green algae and land plants, such as phycobilisomes and the peptidoglycan (murein) wall between the inner and outer plastid membranes (Pfanzagl et al., 1996). The C. paradoxa plastid genome is gene rich (Löffelhardt and Bohnert, 1994) and its plastid protein import apparatus is the most reminiscent of the ancestor of Archaeplastida (Steiner et al., 2005b).

The stromal precursors of Archaeplastida use translocon of inner membrane (Tic) and translocon of outer membrane (Toc) protein complexes to traverse these membranes (Knopp et al., 2020; Soll and Schleiff, 2004). In glaucophytes, stromal proteins also have to pass through the plastid murein wall. In contrast to stroma-targeting peptides (STPs) in land plants, STPs in C. paradoxa require a phenylalanine in their N-terminal domain (Price et al., 2012; Steiner et al., 2005b; Wunder et al., 2007). Thylakoid lumen proteins use additional import pathways that have mainly been described in detail for green algae and land plants (Albiniak et al., 2012; Richter et al., 2010). The secretory (Sec) pathway is responsible for the thylakoid import of proteins in an unfolded state, whereas the so-called ‘twin-arginine translocation’ (Tat) pathway allows translocation of folded proteins. Thylakoid protein precursors therefore possess a specific thylakoid-targeting peptide (TTP) for import into the thylakoid lumen via either the Sec or Tat pathway. Tat passengers are generally characterized by the twin-arginine motif followed by a hydrophobic core in their TTP (Frain et al., 2019). In Cyanophora, genes encoding homologs of all key components of the plastid import machinery have been found (Price et al., 2012; Steiner et al., 2012). The muroplast and thylakoid protein import, including its dependency on energy requirements (such as ATP or light), was previously studied using in vitro import systems and various inhibitors of different import pathways followed by subsequent fractionation of chloroplasts into stroma and thylakoids (Steiner et al., 2005a).

The copper-binding plastocyanin and the heme-binding cytochrome c₆ are thylakoid lumen proteins crucial for the electron transfer from cytochrome b₆f to photosystem I. It has been suggested that the iron-dependent cytochrome c₆ is more ancient, descending from cyanobacterial lineages living in a poorly oxidized environment (De la Rosa et al., 2006). It has been shown that many cyanobacterial species and certain chlorophytes possess both of these electron carriers, and that they utilize plastocyanin in response to copper availability (Merchant and Bogorad, 1986; Sandmann and Böger, 1980). In land plants, plastocyanin is the main carrier transferring electrons from cytochrome b₆f to photosystem I. Cytochrome c₆ had been thought to be absent from land plants, until the cytochrome c₆-like proteins designated as c_6A were discovered in these organisms (Gupta et al., 2002; Wastl et al., 2002). The diversity of the cytochrome c₆ family has increased also by the discovery of two additional classes of cytochrome c₆-like proteins, c_6B and c_6C, in cyanobacteria (Bialek et al., 2008). The function of cytochromes c_6A, c_6B and c_6C remains unclear, but it is unlikely that they are directly involved in photosynthetic electron transport (Bialek et al., 2016; Howe et al., 2016). However, in the cyanobacterium Nostoc sp. PCC 7119, a cytochrome c_6C can indeed serve as a putative electron donor to photosystem I but cannot accept electrons from cytochrome f (Reyes-Sosa et al., 2011).

It has been speculated that c₆-like proteins might function in redox balance maintenance and protection against over-reduction of electron carriers and reactive oxygen species production. They might also serve as safety-valves under light stress conditions or as sensors of O₂, NO or CO (Díaz-Moreno et al., 2016). Such a regulatory role has been shown for cytochrome c_6c (Cyt c_6-3) in heterocyst differentiation (Torrado et al., 2016).

C. paradoxa utilizes the cytochrome c₆ (PetJ1) and plastocyanin is absent (Steiner et al., 2000). In this study, we describe a novel cytochrome c₆-like protein PetJ2 present in C. paradoxa, propose the evolutionary relationship between PetJ2 and other c₆-like proteins, and characterize its import into plastids and thylakoids.

The endosymbiotic cyanobacterial ancestor of primary plastids likely possessed both plastocyanin and cytochrome c₆ (Bialek et al., 2016; Howe et al., 2006). Differential loss of plastocyanin, c₆ and c₆-like cytochromes likely occurred in different lineages of Archaeplastida.

We have searched for cytochrome c₆-like proteins in the C. paradoxa genomic data (Price et al., 2012) and we have found the gene (Contig 12985; Cyanophora Genome Project, http://cyanophora.rutgers.edu/cyanophora_v2018/) encoding a cytochrome c₆-like protein, which is designated as PetJ2 in this study. Fig. 1A shows the alignment of PetJ1 and PetJ2. The glaucophyte PetJ1 and PetJ2 precursors are only 31.1% identical and show 46.3% similarity, while the mature proteins are 37.5% identical and 54.2% similar, respectively.

A BLASTP search using the PetJ2 mature sequence as a query revealed a sequence from the cyanobacterium Stanieria sp. NIES-3757 as the best hit with 51% identity and 67% similarity. Thus, the glaucophyte PetJ2 is more related to protein homologs from cyanobacteria than to glaucophyte PetJ1, strongly suggesting that PetJ2 from C. paradoxa is a novel cytochrome c₆-like protein. Both glaucophyte PetJ precursors possess the phenylalanine residue in their STP that is essential for import into muroplasts (Steiner and Löffelhardt, 2002; Steiner et al., 2005b).

Based on bioinformatic predictions and import experiments, PetJ2 is targeted into the thylakoid lumen and therefore comprises three isoforms – cytoplasmic precursor, stromal intermediate and thylakoid mature protein. The calculated molecular masses for the PetJ2 precursor, intermediate and the mature protein were 16.5 kDa, 13.2 kDa and 10 kDa, respectively. PetJ2 possesses all residues necessary for heme binding (red in Fig. 1A) which is a basic feature of cytochromes. The residues shown in Fig. 1A in yellow boxes are indicative of c₆ or c₆-like cytochrome groups. The presence of leucine (L) and tyrosine (Y) residues in these positions is characteristic for the cytochrome c_6BC group (Bialek et al., 2008). PetJ2 is thus more similar to the cytochrome c_6BC group than to cytochrome c₆ (PetJ1).

We also searched for rhodophyte PetJ2 homologs using the glaucophyte PetJ2 as a query. We have found its homologs in Pyropia haitanensis, Pyropia umbilicalis, Acrochaetium secundatum and Rhodochorton sp. 1SJuan. In contrast to the plastid-encoded cytochrome c₆ (PetJ1) in these rhodophytes, PetJ2 is nucleus-encoded as in C. paradoxa.Fig. 1B shows the alignment of cytochromes c_6BC from cyanobacteria, PetJ1 and PetJ2 proteins from Cyanophora and four rhodophytes, and cytochromes c_6A from green algae and land plants. PetJ2 proteins from all four rhodophytes contain two insertions of amino acid residues in the same position as C. paradoxa in comparison to cytochrome c₆ and cyanobacterial c_6B and c_6C. These insertions are shorter (2 and 1 amino acids) in these four rhodophytes than in the glaucophyte (4 and 3 amino acids; Fig. 1B). The second insertion (VEG in C. paradoxa and D/G in rhodophytes) is located in the same position as the 12-amino-acid insertion forming a loop (loop insertion peptide, LIP) in c_6A cytochromes.

Cytochrome c₆ contains a glutamine residue in the position 54 of the mature protein, which is critical for the midpoint redox potential of the heme group, while cytochromes c_6A, c_6B and c_6C contain isoleucine, leucine or valine in this position (Bialek et al., 2008; Worrall et al., 2008) (Fig. 1B). Cyanophora PetJ2 has a leucine residue in this position (Fig. 1A). Another characteristic of the cytochrome c_6BC subgroup is the presence of a conserved tyrosine residue in the position 65 of the mature protein, while c₆ and c_6A cytochromes usually possess phenylalanine or tryptophan in this position (Bialek et al., 2008). Tyrosine is present in this position in Cyanophora PetJ2, again indicating its relationship to the cytochrome c_6BC subgroup (Fig. 1B).

All cytochromes c₆ as well as cytochromes c_6A possess a crucial arginine residue in position 71 (Fig. 1B), which is essential for efficient reduction of photosystem I (Molina-Heredia et al., 2001). In PetJ2 from C. paradoxa, there is a lysine in this position, rendering an efficient interaction with photosystem I unlikely.

Fig. S1 shows the phylogenetic position of Cyanophora PetJ1 and PetJ2. The cytochromes not belonging to the c₆ and c₆-like groups cluster separately (c_M, c₅₅₀, c₅₅₅). The land plant c_6A cytochromes form a monophyletic group with c_6A from C. reinhardtii. The two Cyanophora cytochromes were placed in two separate clades. While PetJ1 clusters with cytochrome c₆ from green algae and Euglena gracilis, the c₆-like PetJ2 clusters with rhodophyte PetJ2 homologs. Glaucophyte and rhodophyte PetJ2 proteins are more closely related to the cytochromes c_6A, c_6B and c_6C than to c₆. Although the exact branching order of cytochromes c₆, c_6A, c_6B and c_6C cannot be inferred from the tree, due to low bootstrap support and hence low resolution in the basic nodes, the topology of the tree is similar to phylogenetic trees of c₆ and c₆-like cytochromes published previously (Bialek et al., 2016; Howe et al., 2016). Based on the presence of the second insertion in the same position in glaucophyte and rhodophyte PetJ2 as in green algal and land plant cytochrome c_6A, we propose that green algal and land plant cytochromes c_6A evolved from the glaucophyte/rhodophyte-like PetJ2.

The TTPs present in glaucophyte PetJ1 and PetJ2 precursors are different. PetJ1 possesses a TTP typical for Sec passengers and it uses the Sec pathway for the translocation into thylakoids in muroplasts (Steiner et al., 2005a), while PetJ2 possesses a twin arginine motif upstream of a hydrophobic domain, which is generally characteristic for protein precursors imported into thylakoids via Tat pathway (Fig. 1A, green box). However, a cytochrome of c₆-type would be rather an unusual Tat passenger, because it normally binds its heme group in the thylakoid lumen only after its translocation as an unfolded protein through the Sec translocon (Gabilly and Hamel, 2017).

We performed in vitro import experiments of radioactively labeled PetJ2 into isolated muroplasts. Fig. S2 shows the import of the PetJ2 precursor into isolated muroplasts with and without the addition of Sec- and Tat- pathway inhibitors (sodium azide and nigericin, respectively) after 20 min. The radiolabeled 16.5 kDa PetJ2 precursor was efficiently imported into the isolated muroplasts and processed to the 10 kDa mature protein. In contrast to the precursor protein, the mature protein is protected from the protease treatment after the import experiment, confirming its import into the muroplast. However, when muroplasts were treated with the Sec pathway inhibitor, sodium azide (Az), prior to the import experiment, the two–step processing of the PetJ2 precursor was negatively affected and led to the accumulation of the PetJ2 intermediate. The addition of nigericin (Nig), an inhibitor of the Tat pathway, did not lead to the accumulation of the PetJ2 intermediate, but slightly reduced the import of PetJ2 into muroplasts (Fig. S2).

PetJ2 is thus imported into thylakoids via Sec pathway. This suggests that not all proteins possessing a twin arginine motif in a signal peptide have to be Tat passengers. The observation that nigericin slightly inhibits the import of PetJ2 into muroplast (Fig. S2) is consistent with previous import experiments using the C. paradoxa OEC33 (PsbO) precursor and the Rieske protein precursor, in which the addition of nigericin also decreased import efficiency into the isolated muroplasts (Steiner et al., 2005a).

To analyze the thylakoid import requirements of PetJ2 in more detail, we interrupted the import assay at two different time points (Fig. 2). In the control experiment without the addition of azide, the radioactively labeled PetJ2 precursor was imported into the plastid stroma and processed to the intermediate and the mature form after 5 min as indicated by bands with corresponding molecular mass (Fig. 2A). After 11 min, the bands corresponding to the mature protein were slightly more intense, indicating that the import into thylakoids and processing of the PetJ2 intermediate into mature protein is the rate-determining step. The addition of the SecA inhibitor sodium azide led to the accumulation of the intermediate in the muroplast stroma and reduced the amount of the mature protein, confirming that PetJ2 is a Sec passenger. The inhibitory effect of sodium azide accompanied by the accumulation of mature protein was higher after 5 min than after 11 min, suggesting that sodium azide loses its inhibitory potency in time (Fig. 2A). This is likely a result of intraplastidial ATP generation during the import reaction, since the SecA inhibition by sodium azide can be reduced by the addition of ATP, which binds the same SecA sites as sodium azide (Knott and Robinson, 1994).

In parallel, a control experiment was performed with the PsbO precursor (OEC33 kDa protein) from C. paradoxa, which has been shown to use the Sec pathway (Steiner et al., 2005a). When the PsbO precursor was incubated with isolated muroplasts, its import was also time dependent. The inhibition of the Sec translocon with azide also resulted in the accumulation of the PsbO intermediate form, while only a minor fraction was processed to the mature protein (Fig. 2B). The thermolysin protease digested the residual amount of the precursor protein bound to the peptidoglycan-containing muroplast envelope (Fig. 2).

Muroplast fractionation after PetJ2 import to confirm its localization in the thylakoid lumen was inconclusive, since muroplast thylakoids do not readily form closed vesicles after muroplast isolation as the thylakoids of plant chloroplasts do (Steiner et al., 2000). A possible explanation for this is the special architecture of muroplast thylakoids, which are composed of cyanobacterial-like long layers instead of smaller, more compact stacks in plants (Giddings et al., 1983). Therefore, the localization of imported proteins into muroplast thylakoids can only be shown for larger proteins like PsbO, but not for smaller proteins like PetJ1 or PetJ2 (Steiner et al., 2005a).

Isolated plant thylakoids are more stable and more suited for thylakoid in vitro import experiments than those of Cyanophora (Steiner et al., 2005a). Pea thylakoids and radiolabeled glaucophyte PetJ2 were used for another import experiment to test whether the plant thylakoid import machinery can recognize and import PetJ2 into the pea thylakoid lumen, and to verify that the transport requires the Sec pathway (Fig. 3).

The radiolabeled PetJ2 precursor was incubated with pea thylakoids with the addition of HM buffer or pea stromal extract, the latter containing ATP and soluble components needed for the Sec pathway (Albiniak et al., 2012). Although the import, processing into mature protein and protease protection was observed even without the addition of pea stromal extract, the efficiency of these processes was highly increased after the addition of the pea stromal fraction, indicating that components present in the stroma facilitate PetJ2 import. Sodium azide completely inhibited processing to the mature protein under both conditions, strongly suggesting that the PetJ2 import into thylakoids is SecA dependent (Fig. 3). The experiments with pea thylakoids have thus confirmed that glaucophyte PetJ2 is imported into thylakoids via Sec pathway, that the pea heterologous thylakoid system is suitable for the import experiments with glaucophyte thylakoid proteins and that the thylakoid import mechanisms are highly conserved in Archaeplastida.

C. paradoxa strain CCMP329 (‘Pringsheim’) was grown in 50 ml Erlenmeyer flasks in mineral medium described previously (Jakowitsch et al., 1993) under constant illumination (30 μEm⁻² s⁻¹) at 25°C. To gain a high yield of biomass necessary for muroplast isolation, C. paradoxa was grown in the same medium under the same conditions in 1.5 l ‘Kniese’-type glass column with the perfusion of CO₂-enriched air (4%) to increase the growth rate. P. sativum (var. Feltham first) was grown in soil using a 16 h light–8 h dark regime as described previously (Brock et al., 1995).

Total RNA was isolated from C. paradoxa culture in exponential growth phase using RNeasy Plant Mini Kit (Qiagen). cDNA was synthesized using Superscript II Reverse Transcriptase (Thermo Fisher Scientific). PetJ2 was amplified by PCR using cDNA as a template and Pfu Turbo Polymerase (Agilent) with the following primers: PetJ2-For, 5′-AAGAATTCATGAACGCCGCAGCCTTCTC-3′ and PetJ2-rev, 5′-AAGGATCCGGCCACGTCGCTTTAATTAGA-3′. The product of correct size was cloned into pBAT-vector (Annweiler et al., 1991) via EcoRI and BamHI restriction sites and verified by sequencing.

Radioactively labeled PetJ2 and PsbO precursor proteins were synthesized using an in vitro transcription–translation system based on wheat germ extract in the presence of [³⁵S]methionine as described previously (Marques et al., 2003).

Muroplasts were isolated from C. paradoxa in the exponential growth phase as described previously (Steiner et al., 2000). Muroplast isolation and import experiments were done as described previously (Steiner et al., 2005a). Chloroplasts and thylakoids from 1-week-old Pisum sativum seedlings were isolated as described previously (Hou et al., 2006). Stroma acquired during the osmotic lysis of chloroplasts (thylakoid isolation step) was also kept for the experiments.

Import into muroplasts and time-dependent import experiments were performed according to the protocol described previously (Steiner et al., 2005a). To block the Sec and Tat pathways, muroplasts were treated prior to the import experiment with 10 mM sodium azide for 10 min at 4°C and 2.5 μM nigericin for 10 min at 4°C, respectively. Representative examples of three independent transport experiments are shown.

Thylakoid vesicles equivalent to an amount of 15 µg chlorophyll were incubated with the radioactively labeled PetJ2 precursor protein for 15 min at 25°C in light as described previously (Hauer et al., 2017). The import procedure was stopped by cooling the reagents to 4°C and the addition of 1 equal volume of HM-Buffer (10 mM HEPES-KOH pH 8.0 and 5 mM MgCl₂) followed by centrifugation (4 min, 20,000 g) and subsequent resuspension in HM buffer. This washing step was repeated, and resuspended thylakoids were divided into two reaction tubes (equivalent to an amount of 7.5 μg chlorophyll each). One of the aliquots was treated with thermolysin (P1512; Sigma-Aldrich) for 30 min on ice followed by the addition of 1 volume of HM buffer and 10 mM EDTA to stop the reaction. After centrifugation the thylakoids were analyzed using 10–17.5% SDS-polyacrylamide gradient gels under denaturing conditions as described previously (Laemmli, 1970). The gels were then dried and visualized using phosphorimaging with Fujifilm FLA-3000 (Fujifilm, Düsseldorf, Germany) utilizing BAS-Reader software (version 3.14) and AIDA (version 3.25; Raytest, Straubenhardt, Germany). To block the Sec pathway, thylakoids were treated prior to the import experiment with 10 mM sodium azide for 10 min at 4°C. Representative examples of three independent transport experiments are shown.

EMBOSS Needle (Rice et al., 2000) was used for the alignments of PetJ1 and PetJ2 proteins from C. paradoxa. BLASTP (Altschul et al., 1990) was used for searching for homologs of PetJ2 in other organisms (i.e. cyanobacteria and rhodophytes). The processing sites for the stroma processing peptidase (SPP) and the thylakoid processing peptidase (TPP) were calculated with ChloroP (Emanuelsson et al., 1999) and PRED-TAT (Bagos et al., 2010), respectively. The molecular mass of precursor, intermediate and mature PetJ2 were calculated using the Compute pI/Mw tool from the ExPASy Bioinformatics Resource Portal (expasy.org).

The same cytochrome c₆ and c₆-like sequences as analyzed previously (Bialek et al., 2008) (except for some redundant ones) were used to infer the phylogenetic tree with the additional inclusion of rhodophyte and glaucophyte cytochrome c₆ and c₆-like sequences (see legend to Fig. S1 for the list of accession numbers). Only mature proteins were used for analysis. After alignment in MAFFT (Katoh and Standley, 2013) phylogenetic analyses was conducted using MEGA7 program (Kumar et al., 2016). The evolutionary history was inferred by using the Maximum Likelihood method based on the JTT matrix-based model (Jones et al., 1992). The bootstrap consensus tree inferred from 1000 replicates was taken to represent the evolutionary history of the taxa analyzed (Felsenstein, 1985). Initial tree(s) for the heuristic search were obtained automatically by applying Neighbor-Join and BioNJ algorithms to a matrix of pairwise distances estimated using a JTT model, and then selecting the topology with superior log likelihood value. Arabidopsis mitochondrial cytochrome c was used as an outgroup. The resulting phylogenetic tree was visualized with iTOL v4 (Letunic and Bork, 2019).

The peer review history is available online at https://journals.biologists.com/jcs/article-lookup/DOI/10.1242/jcs.255901