Cytochrome c_M Is Probably a Membrane Protein Similar to the C Subunit of the Bacterial Nitric Oxide Reductase

Abstract

Cytochrome c_M was first described in 1994 and its sequence has been found in the genome of manifold cyanobacterial species ever since. Numerous studies have been carried out with the purpose of determining its function, but none of them has given place to conclusive results so far. Many of these studies are based on the assumption that cytochrome c_M is a soluble protein located in the thylakoid lumen of cyanobacteria. In this work, we have reevaluated the sequence of cytochrome c_M, with our results showing that its most probable 3D structure is strongly similar to that of the C subunit of the bacterial nitric oxide reductase. The potential presence of an α-helix tail, which could locate this protein in the thylakoid membrane, further supports this hypothesis, thus providing a new, unexpected role for this redox protein.

1. Introduction

Electron transport chains play a fundamental role in the metabolism of all living beings. They consist of a series of redox reactions, spatially separated, in which electrons are transferred from a donor molecule to an acceptor one. In phototrophic eukaryotes, the respiratory and photosynthetic electron transport chains are spatially separated into specialized organelles (mitochondria and chloroplasts). However, in cyanobacteria, both electron transport chains are located in the thylakoid membranes and share some components, such as plastoquinone (PQ), plastocyanin (Pc) or cytochrome (Cyt) c₆ and the Cyt b₆-f complex.

In cyanobacteria, two c-type cytochromes involved in the electron transport chains have been described: Cyt c₆ and Cyt c₅₅₀. While Cyt c₆ is a soluble electron carrier involved both in photosynthesis and in respiration, Cyt c₅₅₀ is structurally bound to photosystem (PS) II and is not directly involved in electron transport [1]. One further c-type cytochrome has been described in cyanobacteria: the so-called Cyt c_M [2]. The gene encoding for Cyt c_M is conserved in nearly every sequenced cyanobacterium [3,4], and it has been consistently studied in the unicellular organism Synechocystis sp. PCC 6803 [5,6]. In Synechocystis, the annotated gene contains two methionine codons (Figure 1), of which the distal one was proposed to be the actual start codon as it was preceded by a GA-rich motif, i.e., a probable ribosome binding site [2,6]. The sequence further comprised a hydrophobic amino-terminal region that was assumed to constitute either a membrane anchor or a signal peptide. However, due to its close similarity to the signal peptide of Cyt c₆ from Anabaena sp. PCC 7119, the latter was regarded as the most probable option [6].

Cho et al. [5] were the first to successfully detect the protein in cyanobacteria through Western blot analysis. Nevertheless, their results were not clear enough to conclude whether the protein featured the N-terminal hydrophobic domain in vivo or not, since the molecular weight of the protein detected by Western blotting did not match either of the two possible options [6]. As for the protein location, even though Bernroitner et al. [3] claimed to have determined that the thylakoid and plasma membranes are the exact places where Cyt c_M is located in the cell, these results have later been questioned [7] given the high likelihood that cross-contamination between membranes occurred. Ever since, no other author has declared to have successfully established the cellular region where the protein is placed, probably due to its extremely scarce expression levels under normal conditions [8].

Based on the hypothesis that the actual Cyt c_M protein sequence starts with the second methionine and that it contains a signal peptide to the thylakoid lumen—which would be subsequently released thereby generating a soluble globular protein—several authors have tried to determine its physiological function. As an alleged electron carrier similar to Cyt c₆, speculations about the biological role of Cyt c_M have been mainly linked it to a hypothetical role in respiratory and photosynthetic electron transport chains. The fact that deletion mutants were able to carry out both processes without any significant phenotype [2] dismissed the possibility of Cyt c_M being a main component of any of these systems. However, there was a chance that Cyt c_M participated in respiration or photosynthesis as a secondary element, i.e., as a backup piece that is only expressed under certain conditions to substitute or complement a major constituent. Interestingly, the expression levels of Cyt c_M sharply increase in stress conditions such as low temperature and high-intensity light whereas those of the two main electron carriers, plastocyanin (Pc) and Cyt c₆, decrease [9]. Despite some further evidence indicating a “spare component” role for Cyt c_M in photosynthesis [10,11], we ruled this out by showing that the surface electrostatic potential of Cyt c_M is too different from those of Pc and Cyt c₆ and by proving that the kinetic interaction constant between Cyt c_M and PS I, to which Pc and Cyt c₆ donate electrons, is negligible [6]. Moreover, the low redox potential of Cyt c_M (+150 mV) makes it thermodynamically unlikely that it can be reduced by Cyt f (with a redox potential of +320 mV), as is the case with Pc and Cyt c₆ [6,12]. Consequently, only respiration remained as a real possibility. Initially, it was suggested that Cyt c_M could reduce respiratory terminal oxidases (COX) since it was kinetically plausible [3] and it was impossible to create a ΔcytM/PSI double mutant [10]. However, the hypothesis of Cyt c_M working as an electron donor in respiration was dismissed once it was observed that ΔcytM mutants have in fact a higher respiration rate under dark heterotrophic conditions [13], with the lack of COX not affecting this whatsoever [8].

In conclusion, none of the hypotheses formulated to this day regarding the physiological role or the cellular location of Cyt c_M has proved solid, and this matter remains altogether an open debate. Since most attempts to determine any of these have been based on the assumption that it is a small soluble protein located in the lumen of the thylakoid, we have reevaluated the sequence and structure of Cyt c_M in the present work. Our results suggest that, contrarily to what was previously thought, Cyt c_M could be a membrane protein with a structure similar to that of the C subunit of bacterial nitric oxide reductase.

2. Materials and Methods

For the sequence collection, we searched for protein sequences corresponding to the cytM gene appearing in either of the two main repositories of annotated cyanobacterial genes: CYORF [14] (http://cyano.genome.jp/, accessed on 11 May 2021) and CyanoBase [15] (http://genome.microbedb.jp/cyanobase/, accessed on 11 May 2021).

Secondary and 3D protein structure prediction was made with the tools Phyre² [16] (http://www.sbg.bio.ic.ac.uk/phyre2/html/page.cgi?id=index, accessed on 11 May 2021), I-TASSER [17] and AlphaFold [18].

The free software UCSF Chimera [19] (https://www.cgl.ucsf.edu/chimera/, version 1.15, accessed on 11 May 2021) was used to visualize protein 3D structures and to determine hydrophobicity and electrostatic potential, as well as for distance measurements and structure matching. For the electrostatic potential, the “Coulombic Surface Coloring” option was chosen. For comparison with NorC, we used the structure of nitric oxide reductase of Pseudomonas aeruginosa [20], which we obtained from the Protein Data Bank (https://www.rcsb.org/; PDB entry: 3O0R, accessed on 14 May 2021).

3. Results and Discussion

With the purpose of determining whether the N terminal part of the protein encodes for a signal peptide or a transmembrane helix, we retrieved all the protein sequences corresponding to the cytM gene that appear in either of the two main repositories of annotated cyanobacterial genes: CYORF and CyanoBase. A total of 18 sequences was obtained (

Table 1. Cyt c_M protein sequences retrieved from CyanoBase and/or CYORF.

Organism	Strain	Phyre2 Structure Matching	DB of Retrieval
Gloeobacter violaceus	PCC7421	Yes	CyanoBase
Prochlorococcus marinus	MIT9211	Yes	CyanoBase
Synechococcus elongatus	PCC6301	No *	CyanoBase & CYORF
Synechococcus sp.	WH7803	No *	CyanoBase & CYORF
Synechocystis sp.	PCC6803	Yes	CyanoBase & CYORF
Nostoc sp.	PCC7120	Yes	CyanoBase & CYORF
Prochlorococcus marinus	NATL1A	Yes	CyanoBase & CYORF
Prochlorococcus marinus subsp. pastoris	CCMP1986	Yes	CyanoBase & CYORF
Synechococcus elongatus	PCC7942	No **	CyanoBase
Prochlorococcus marinus	MIT9303	Yes	CyanoBase & CYORF
Synechococcus sp.	RCC307	Yes	CyanoBase & CYORF
Synechococcus sp.	WH8102	Yes	CyanoBase & CYORF
Acaryochloris marina	MBIC11017	Yes	CyanoBase & CYORF
Cyanothece sp.	ATCC51142	Yes	CyanoBase & CYORF
Synechococcus sp.	PCC7002	Yes	CyanoBase & CYORF
Thermosynechococcus elongatus	BP-1	Yes	CyanoBase & CYORF
Microcystis aeruginosa	NIES-843	Yes	CYORF
Prochlorococcus marinus	MIT9313	Yes	CYORF

The “Phyre²” column states whether the structure predicted for that specific sequence by the tool Phyre² matched most of the structures predicted for the other sequences or was different from the rest. * First model proposed by Phyre² did not match the rest, but the second one (the next most probable according to the software’s algorithm) did. ** None of the models proposed for this sequence by Phyre² matched the rest.

), and they were then subjected to a structural prediction through Phyre², which uses the homology recognition (threading) approach.

The first striking result was that 15 out of 18 sequences produced almost identical structures with a long, protruding α-helix (or “tail”) corresponding to the initial section of the protein (Figure 2A). This was the precise same region whose nature—signal peptide or transmembrane domain—has not been clearly established to this day. In fact, the VLA motif, repeatedly alleged as the cleavage site of the signal peptide, is located exactly at the junction of the tail with the “body” of the protein in the newly proposed structure (Figure 2B). The length of this helix motif of 5.1 nm (the average membrane being between 5 and 10 nm wide; Figure 2C) and the remarkably hydrophobic nature of this section (Figure 2D) agree with a membrane anchoring function.

At this point, it is important to bring up the fact that the signal peptides of Cyt c₆ and Cyt c₆-like proteins from the cyanobacterium Anabaena sp. PCC 7119, having a processing site (AXA) downstream from the N-terminal end, are recognized by E. coli which results in transport of the proteins to periplasmic space, a subcellular compartment equivalent to thylakoid lumen, where the signal peptide is correctly processed [21,22,23]. However, the hydrophobic tail of the Synechocystis Cyt c_M is not recognized by E. coli as a signal peptide. For this reason, it has only been possible to express the soluble region of Synechocystis Cyt c_M in E. coli by substituting its hydrophobic tail for the signal peptide of Anabaena Cyt c₆ [6].

The template protein used by Phyre² for calculating those 15 structures happened to be the c subunit of nitric oxide reductase (NorC) from Pseudomonas aeruginosa. Thus, the similarity between the returned models and the original template were compared. Surprisingly, the overlap between both proteins was clear and almost total (Figure 3A). To confirm this, a second, independent prediction for the same 15 sequences was performed using I-TASSER, a different structure prediction software that incorporates the ab initio approach in its algorithm. In this new test, the structures obtained were, once again, almost identical to that of NorC (Figure 3B). We obtained similar results once more using AlphaFold, a novel deep learning algorithm that incorporates physical and biological knowledge about protein structure and that leverages multi-sequence alignments [18].

Crucially, NorC is a subunit of nitric oxide reductase that has never been found in cyanobacteria, while the core subunit NorB is present in several cyanobacterial species [24] and its expression decreases in ΔcytM mutants [7]. Furthermore, Cyt c_M sharply increases its expression in oxidative stress conditions [9], a condition in which nitric oxide reductase is highly active [25].

In order to determine similarities in charges between Cyt c_M and NorC, we subsequently evaluated their surface electrostatic potential (Figure 4).

Also in this aspect, we found similarities between Cyt c_M and NorC that might be indicative of a similar function, with Cyt c_M acting as a functional part of a nitric oxide reductase and therefore being involved in the oxidative stress response. These included a positively charged tip at the end of the tail, a negatively charged region at the body closest to the tail, another positively charged region on the opposite side of the body and yet one more negatively charged pocket on the bottom. On the other hand, there are some clear differences too, such as a more extensive positive zone in the case of NorC, but the resemblances between both structures here are again undeniable. Moreover, the midpoint redox potential of Cyt c_M and NorC, a fundamental parameter to establish the function of redox proteins, are relatively close with +150 mV and +183 mV, respectively [6,26]. This is interestingly opposed to the major divergences found previously between the surface electrostatic potential and midpoint redox potential of Cyt c_M and the ones of Cyt c₆ and Pc [6].

4. Conclusions

Here we show simple yet solid evidence that Cyt c_M most probably has a transmembrane domain, and that it might well be a protein very similar to—or simply the equivalent of—the C subunit of nitric oxide reductase in cyanobacteria. If this last was the case, it would explain why it has been so difficult to assign any clear function to the protein so far, since all previous studies have focused on the soluble part of it when in fact it could hardly carry out any task without the rest of the protein it belongs to.