In current research we tried to clarify the predicted role of cytochrome c oxidase in phenazine antibiotics synthesis in strains-producers P. chlororaphis subsp. aurantiaca B-162. These chemical compounds are considered to be the promising sources of next-generation antimicrobial and anticancer medicines. The development of such medicines is due to the increasing antibiotic resistance of pathogenic microorganisms and the emergence of multidrug resistant tumors. The introduction of new compounds for medical uses requires an accurate preliminary analysis of mechanisms of their action on cells of pro- and eukaryotic organisms, as well as the prediction of possible modes of resistance development. The natural resistance to phenazines in strains-producers might be an additional excellent model for prognosis of key steps of resistance development. The mechanism of phenazines` resistance in phenazine-synthesizing bacteria is not characterized well enough, as well as the mechanism of high productivity formation. We strongly believe that these two mechanisms are directly related.
In our former researches chemical mutagenesis followed by selection was used to derive the mutant strain P. chlororaphis subsp. aurantiaca B-162/17, capable of producing phenazines on minimal media (Shapira et al. 2021). We have also shown that this mutant strain demonstrated lower level of productivity on rich PCA-medium than the parental strain B-162/255, but it was still much higher than the wild type strain B-162 (Shapira et al. 2021). To identify mutations which could lead to such phenotype we made whole-genome sequencing and comparison of genomes of strains B-162/255 and B-162/17. The comparison of genome sequences of B-162/255 and B-162/17 strains helped us to discover the potentially novel candidate genes, which products might be involved in phenazine biosynthesis (Liaudanskaya et al 2022). Of the most interesting mutations there were two deletions directly influenced genes sequences (Liaudanskaya et al 2022). These deletions interfere with the sequence of cytochrome c oxidase subunit I gene (coxA gene) and arginine N-succinyltransferase, α-subunit gene (astA gene). In current research we focused on the coxA gene and its influence on phenazines production in P. chlororaphis subsp. aurantiaca.
It is known that Pseudomonas species possess one A-type (caa3, cox-genes products) and multiple C-type (cbb3) cytochrome c oxidases as well as two quinol oxidases for aerobic respiration (Osamura et al. 2017). These enzymes are the components of energy-converting chain that catalyzes the reduction of oxygen to water, thus participating in the buildup of the proton gradient across the cell membrane required for ATP synthesis (Kadenbach,2018). Cytochrome c oxidase is a large transmembrane protein, composed of several peptide chains possessing different activities. This complex can be found in prokaryotes and inside the mitochondria of eukaryotes.
In Pseudomonas aeruginosa A-type cytochrome c oxidases (cox-genes products) are expressed only under specific growth conditions (for example, nutrient starvation, and combination of starvation and high oxygen conditions) (Kadenbach 2018). A-type proteins are the members of the heme-copper oxidase superfamily and are closely related to the mitochondrial terminal oxidases (Pereira et al. 2001). Functionally active A-type enzyme in Pseudomonas species consists of three subunits. Subunit I is the main catalytic subunit of the enzyme, which forms the binuclear heme-CuB center. This center is responsible for O2 binding and reduction. Subunit I (su I) also contains a second heme, which is necessary to facilitate the transfer of electrons to the binuclear center (Garcia-Horsman et al. 1994). The expression of cox-genes cluster is regulated by stationary phase sigma factor, RpoS (Kawakami et al. 2010). The same transcriptional activator is required for optimal phenazines` synthesis (Girard et al. 2006). It means that the highest concentration of Cox-proteins will correlate with the pick of phenazines` concentration.
3.1. Analysis of CoxA proteins in P. chlororaphis subsp. aurantiaca B-162 and B-162/17 strains.
According to the data of NGS-sequencing, the mutation in coxA of B-162/17 strain caused the deletion of 12 nucleotides without a frameshift, which led to the loss of 4 amino acids in the middle part of the protein (Fig. 2) (Liaudanskaya et al 2022).
The amino acids (228-231 positions) which were deleted in B-162/17 strain are marked in red
The deleted amino acids are the part of one of the transmembrane regions of cytochrome c oxidase su I which is made up of hydrophobic α-helices (Fig. 3).
Deleted amino acids (228-231 positions) are marked with the blue square. Flanking amino acids V and M are underlined with red. Small pink squares on 3-D models mark the V (position 227 in normal protein sequence) and M (position 232 in normal protein sequence) amino acids flanking the deletion site.
Correct α-helices formation is important for overall proteins structure, dynamics and regulation of their functions. Even subtle changes in the amino acid sequence of these structural elements can alter protein function and its interaction with the membrane and with the other parts of polypeptide cytochrome c oxidase complexes. In cytochrome c oxidase su I these elements of secondary structure are necessary for the formation of 12 transmembrane segments which form one of the key parts of the minimal functional unit (oxygen reducing part) (Branden et al. 2006). We clarified that according to the data of Protein Blast sequence analysis, deleted amino acid residues TM and L are highly conserved among the most well studied in this regard prokaryotes Paracoccus denitrificans and Rhodobacter sphaeroides. The same residues are highly conserved among yeast, which enzyme is similar to the human enzyme (Bratton et al. 2003), and Homo sapiens cytochrome c oxidase su I (48 % of identity). It means that these residues have an important biological function and might be engaged into assembly or functionality of the minimal functional unit. In R. sphaeroides these residues are located in helix V, which are supposed not to be involved into metal ligands binding and redox-active centers formation (Calhoun et al. 1994). But this subunit is still might be engaged into H-channel formation, which is the unique proton pump pathway in the mammalian mitochondrial oxidase and prokaryotic enzyme as well (Wikström et al. 2018). Plenty of mutations in cytochrome c oxidase su I reported in human patients suffering from a range of disorders had no direct effect on the catalytic activity of the enzyme but supposed to have influence on the whole enzyme complex assembly and interaction among subunits (Bratton et al. 2003). As it is obvious from 3-D models (Fig. 3), the distance between flanking amino acid residues V and M became shorter in mutant strain. We speculated that these changes might alter su I folding or its interaction with su II-III or membrane. Such changes in CoxA structure of B-162/17 strain potentially might be the reason for its lower phenazines yield on rich media comparing to parent strain B-162/255.
3.2. The role of cytochrome c oxidase subunit I in phenazines synthesis.
To further clarify the possible role of cytochrome c oxidase su I in phenazines synthesis in P. chlororaphis subsp. aurantiaca we knockouted the coxA gene in wild type strain B-162. The overlap PCR was used for this purpose. The schematic description of the experiment and the primers design are shown in Fig 4 (a–c).
Building the deletion of full-length coxA gene required two primer pairs. The first primer pair (F1 (Fragment 1.FOR) and R1 (Fragment 1.REV)) was target the genomic region upstream of the desired deletion point, and the second primer pair (F2 (Fragment 2.FOR) and R2 (Fragment 2. REV)) was target the genomic region downstream of the desired deletion point. The pair of primers F1 (Fragment 1.FOR) and R2 (Fragment 2. REV)) was used to create full-length construction for coxA knockout and for further identification of deletion mutants. The first primer pair was used to receive Fragment 1 (602 b.p.) and the second primer pair was used to receive Fragment 2 (654 b.p.).
Primers F1 (Fragment 1.FOR) and R2 (Fragment 2. REV) were used to receive Fragment 3. Fragment 1 and 2 were used as the templates in this PCR reaction. Fragment 3 included last 475 bp of 3`-end of cytochrome c oxidase su II gene, whole gene of cytochrome oxidase biogenesis protein Cox11-CtaG and several nucleotides of intergenic region between Cox11-CtaG gene and cytochrome c oxidase su III gene, which flanked coxA gene, but there was no coxA gene in its structure. Fragment 3 was cloned into pKNG101 plasmid into the BamHI sites, resulting pKNG101/coxA- recombinant plasmid. This recombinant plasmid was transferred into E. coli BW 19851 and clones were selected for ampicillin resistance.
The resulting plasmid pKNG101/coxA- was introduced into the P. chlororaphis subsp. aurantiaca B-162 by biparental mating. Then P. chlororaphis subsp. aurantiaca B-162 strains with the plasmid pKNG101/coxA- were selected on LB agar plates with streptomycin and ampicillin. The final step of knockout procedure was to select against merodiploids and to identify the desired mutants. For this purpose colonies from the previous stage were streaked on no-salt LB agar that contains 15% (wt/vol) sucrose. We managed to select 6 strains on no-salt LB agar with 15% sucrose (presumably B-162/coxA- strains). Then each single colony isolated on sucrose agar was screened for the coxA deletion. PCR analysis was carried out to identify the mutant allele. PCR products, which corresponded to the size of the mutant allele and that had been generated from sucrose resistant and antibiotic sensitive colonies, were subsequently sent for Sanger sequencing to confirm the mutation.
It was shown that all mutant strains have a culture density and CFU comparable to that of the wild type strain, so the viability of the mutant strain was comparable to that of wild type strain. We speculate that the normal viability of mutant strains could be explained by the existence and activity of the other types of cytochrome c oxidases in Pseudomonas genome, which are usually active under the normal growth conditions (Osamura et al. 2017). These enzymes might completely or partially replace the function of the knockouted gene in such conditions. We next analyzed the level of phenazines production in 6 resulting B-162/coxA--strains. Analysis revealed that practically all of the B-162/coxA--strains lost the ability to produce phenazines in rich PCA-medium. Only one of B-162/coxA—strains (strain № 6) generated up to 26.5 mg/L after 5 days of cultivation that is 38 times lower than the wild-type strain B-162 (Fig. 5). This mutation didn`t let to acquire the ability to produce phenazines in M9 minimal salts medium as well.
The loss of the ability to produce phenazines in rich PCA-medium by the B-162/coxA-–strains demonstrated the existence of the connection between phenazines biosynthesis and energy-converting chain. Previously it was shown that cbb3-type cytochrome oxidase subunit had supported Pseudomonas aeruginosa biofilm growth (Jo et al. 2017). Biofilm formation is controlled by the PhzR/I QS that is also critical for phz-operon activity (Pierson 3rd and Pierson 2010). So, the mature biofilm formation strongly correlates with enhanced phenazines production. Jo et al. (2017) demonstrated that cbb3-type cytochrome oxidases were required for phenazine reduction in hypoxic biofilm subzones, but till now there were no evidences that the disruption of cytochrome с oxidase genes could block phenazines production.
How exactly cytochrome c oxidase is involved in the production of phenazines is not yet fully understood. It is unlikely that this protein is directly involved in the reactions of shikimate pathway or the reactions of phenazine pathway. The involvement of this protein can occur at the level of antioxidant defense systems of the bacterial cell, which are also very sensitive to changes in the concentration of phenazines and reactive oxygen species formed in their presence. A possible violation of the redox balance upon deletion of coxA leads to the activation of direct or indirect mechanisms of blocking phenazines` synthesis. It should be mentioned that the direct relationship between phenazines` production and activity cytochrome c oxidase su I has not been shown till current research.
To identify the possible interconnection between cytochrome c oxidase su I and phenazines production a protein-protein network was built using a web resource STRING (version 11.5). The most probable functional partners of cytochrome c oxidase su I are shown at Fig. 6. The majority of functional partners are the components of cytochrome c oxidase complex or the other proteins of energy-converting chain. One of the most interesting issues is EY04_29170-protein which associates with cytochrome c oxidase and belongs to a large group of transporter proteins found in all organisms – major facilitator superfamily (MFS). It is known that the multidrug efflux pumps of this superfamily help to excrete a large variety of microbial metabolites from bacterial cell, including antibiotics and redox-active substances (hydroperoxide, potassium superoxide, many singlet oxygen-generating compounds) (Chen et al. 2017).
H. Huang et al. showed that the exposure of pathogenic fungus Phellinus noxius to low-level phenazine-1-carboxylic acid (PCA) cause up-regulated gene expression of seven tested genes, among which there were genes of MFS-transporters and cytochrome c oxidase, su I (cox1, homologue of coxA) (Huang et al. 2016). The authors suggested that those genes could be involved in PCA detoxification and improvement of the pathogen cell viability. The similar pattern of transporters gene expression was observed in Pseudomonas aeruginosa. The genes of efflux pomp MexGHI-OpmD, which was proved to be a main transporter of 5-methylphenazine-1-carboxylate, were induced in the presence of this phenazine species (Sakhtah et al. 2016). So, EY04_29170-protein might be involved into phenazines transport form P. chlororaphis subsp. aurantiaca cells, protecting them from toxic phenazines effects. Taking into account the fact that this protein is functionally connected with cytochrome c oxidase we speculate that the disruption of coxA gene might abolish this level of producer`s defense that leads to complete block of phenazines synthesis. If so, this level of defense could play a key role in providing the ability of bacterial cells to synthesize phenazines and, possibly, the other redox-active secondary metabolites.
In early research it was shown that some phenazines were tend to accumulate in mitochondria of rat liver cells (French et al. 1973). This accumulation depends on electron flow through an energy-converting chain and coupled with phenazines` reduction and consequent storage of reducing equivalents in the mitochondria (French et al. 1973). A-type proteins are closely related to the mitochondrial terminal oxidases (Pereira et al. 2001). It can be assumed that a similar process also takes place in bacterial cells, despite the absence of mitochondria. The key role at this process seems to be played by the electron transport chain and its components, primarily A-type cytochrome c oxidase. Phenazines are redox-active compounds and cytochrome c oxidases are involved in maintaining the redox balance in the cell. Like inside rat liver cells, A-type cytochrome c oxidase of bacterial cells can participate in phenazines reduction and detoxification. It might be the second point of cytochrome c oxidase engagement into producers` self-resistance development to phenazines. This level of protection might function in tight bound with MFS-transporters. Based on our data, we believe that the intracellular reduction of phenazines in bacterial cells can lead to their conversion into redox-inactive forms and thus ensures their detoxification and safety for principal cell molecules and metabolites. Then the inactivated phenazines might be excreted through different types of MFS-proteins. The proposed protection mechanism would be especially effective under the shortage of stronger oxidizing agents (for example, oxygen), which can be observed into the microbial biofilms. This mechanism is also might be engaged in eukaryotic cells defense and should be taken into account while developing phenazines-containing medicines. According to the results of our research, cytochrome c oxidase might play the leading role at this process. The absence of this protein reduces the effectiveness of given protective mechanism that leads to a complete blockage of phenazine production in bacteria, particularly in P. chlororaphis subsp. aurantiaca.