Recent insights into the diversity, frequency and ecological roles of phenazines in fluorescent Pseudomonas spp.

Summary

Phenazine compounds represent a large class of bacterial metabolites that are produced by some fluorescent Pseudomonas spp. and a few other bacterial genera. Phenazines were first noted in the scientific literature over 100 years ago, but for a long time were considered to be pigments of uncertain function. Following evidence that phenazines act as virulence factors in the opportunistic human and animal pathogen Pseudomonas aeruginosa and are actively involved in the suppression of plant pathogens, interest in these compounds has broadened to include investigations of their genetics, biosynthesis, activity as electron shuttles, and contribution to the ecology and physiology of the cells that produce them. This minireview highlights some recent and exciting insights into the diversity, frequency and ecological roles of phenazines produced by fluorescent Pseudomonas spp.

Introduction

Phenazines represent a large class of natural antibiotics that are produced by diverse bacteria and exhibit unique redox properties and broad-spectrum antibiotic activity (Mavrodi et al., 2006). The blue-coloured pyocyanin (5-N-methyl-1-hydroxyphenazinium betaine) (PYO) is the oldest known natural phenazine compound and was first reported in 1859, when Fordos used chloroform to extract this microbial metabolite from purulent wound dressings (Fordos, 1859). A few years later, in 1882, Gessard realized that PYO was produced by an aerobic motile bacterium (Gessard, 1882), which he subsequently named ‘Bacillus pyocyaneus’. Gessard was, however, not the first to discover this organism: production of a blue-green pigment by bacteria, albeit without identification of the responsible compound, had already been described by Schroeter in 1872 (Schroeter, 1872). He called the responsible organism ‘Bacteridium aeruginosum’ after ‘aerugo’, the Latin word for verdigris, the blue-green coating that develops on the surface of copper exposed to air (Schroeter, 1872). In 1900, Migula (1900) finally replaced these long forgotten names of the pyocyanin-producing species by ‘Pseudomonas aeruginosa’, which is the name still in use today. Since then more than 100 different phenazine compounds of microbial origin have been reported in the literature (reviewed in Laursen and Nielsen, 2004; Mavrodi et al., 2006; Pierson and Pierson, 2010).

Phenazines were long considered to be pigments of uncertain importance – a stark contrast to current recognition that they are versatile metabolites involved in numerous aspects of bacterial physiology. Early studies revealed the broad cross-phylum inhibitory properties of these compounds, leading many phenazines to be classified as antibiotics. Antibiotic-producing species are common in microbial communities throughout nature, and natural antibiotics are traditionally associated with roles in microbial defence, fitness, interference competition, and the protection of plants and insects against pathogens (Weller et al., 2002; Currie et al., 2006; Mavrodi et al.,2006; Fajardo and Martinez, 2008; Little et al., 2008; Scott et al., 2008; Kroiss et al., 2010). More recently, research on natural antibiotics has focused on their role in microbial physiology, communication and gene regulation (Davies, 2006; Fajardo and Martinez, 2008; Martinez, 2008). However, few of these functions have been studied in the field and little is known about the frequency, amounts and physiology of antibiotic production in nature. In the past decade, phenazine-producing Pseudomonas spp. have emerged as a model system for addressing these fundamental questions.

This article is not intended to cover all aspects of phenazine biology, as thorough and comprehensive reviews are available elsewhere (Mavrodi et al., 2006; 2008; Pierson and Pierson, 2010). Rather, the aim of this mini-review is to briefly highlight some recent and exciting insights into the diversity, frequency and ecological roles of phenazines produced by fluorescent Pseudomonas spp.

Phz⁺Pseudomonas spp. represent a ubiquitous group of antibiotic-producing bacteria in the environment

Although phenazine-producing (Phz⁺) bacteria have long been isolated from diverse terrestrial, freshwater and marine environments, there are only a few systematic studies of their large-scale ecology. In contrast, the population structure of two distinct phylogenetic lineages of phenazine-producing fluorescent pseudomonads, P. aeruginosa and P. fluorescens, is well documented. Pseudomonas aeruginosa is an opportunistic pathogen associated with cystic fibrosis and common nosocomial infections. Pseudomonas aeruginosa is frequently recovered from hospital settings, and several epidemiological studies have reported a non-clonal population structure for this pathogen in the clinical habitat (Pirnay et al., 2002; 2009; Curran et al., 2004). Pseudomonas aeruginosa also is ubiquitous in the environment, and results of a recent study by Selezska and colleagues (2012) suggest that water represents the primary habitat of this important opportunistic pathogen. These authors conducted a 2-year survey of two river systems in northern Germany over a length of 150–200 km and recovered P. aeruginosa from most sampling stations. Genetic analysis of over 500 environmental isolates revealed the presence of several extended clonal complexes of P. aeruginosa (Selezska et al., 2012). Pseudomonas aeruginosa produces several phenazine compounds including PYO, which is used for phenotypic identification of this organism and represents an important virulence factor (Smirnov and Kiprianova, 1990; Pirnay et al., 2009; Selezska et al., 2012). The majority of environmental and clinical isolates of P. aeruginosa carry functional phenazine biosynthesis (phz) genes, and cystic fibrosis isolates often overproduce PYO (Finnan et al., 2004; Fothergill et al., 2007; Selezska et al., 2012).

The diversity and population biology of saprophytic P. fluorescens-like Phz⁺ bacteria have recently been studied by Mavrodi and colleagues (2010; 2012a), who carried out a survey of over 80 dryland (rain fed) wheat fields in the inland Pacific Northwest, USA, and detected an abundance of Phz⁺ pseudomonads in the rhizosphere of commercially grown cereals. Phz⁺Pseudomonas spp. were identified in all sampled fields and their presence was positively correlated with transient accumulation of high levels of phenazine-1-carboxylic acid (PCA) in the plant rhizosphere (Fig. 1). In a follow-up study, Parejko and colleagues (2012) isolated and characterized over 400 strains of Phz⁺Pseudomonas spp. from the roots of dryland cereals by BOX-PCR genomic fingerprinting, 16S rDNA- and recA-based sequence analyses, and Biolog substrate utilization profiling. Results of that study revealed that inland Pacific Northwest isolates form a unique Phz⁺ community dominated by species of the P. fluorescens phylogenetic lineage that are closely related to P. gessardii, P. orientalis, P. libanensis and P. synxantha.

Figure 1. Top panel: Distribution of Phz⁺ pseudomonads in wheat fields from eastern Washington, north-eastern Oregon and western Idaho. Map of the surveyed area is overlaid with mean annual precipitation values and the location of sampling sites is indicated by circles whose size is proportional to the frequency of rhizospheres colonized by Phz⁺Pseudomonas spp. (see inset for scale). Red circles indicate sites from which samples were extracted for PCA. Bottom panel: Relationship between the accumulation of PCA in field-grown cereals and populations of indigenous Phz⁺ rhizobacteria. Amounts of PCA were determined by extracting the antibiotic from 15 g samples of wheat roots followed by quantification of PCA by HPLC-Q-Tof-/MS/MS. In each location from which samples were extracted for PCA, the populations of Phz⁺ rhizobacteria were determined by analysing 8–16 individual wheat rhizospheres. Reprinted from Mavrodi and colleagues (2012a) Appl Environ Microbiol78: 804–812, doi:10.1128/AEM.06784-11, with permission of the publisher (ASM Press).

Among other surprising results obtained by Mavrodi and colleagues (2012a) are the arid conditions under which Phz⁺ populations flourish and produce PCA. In fact, the largest populations of Phz⁺ rhizobacteria were observed in the rhizosphere of wheat grown in the most arid parts of the surveyed area that receive less than 165 mm of annual precipitation (Fig. 1) (Mavrodi et al., 2012a). In a companion study, the authors reported the population levels and plant colonization frequencies of indigenous Phz⁺ rhizobacteria in adjacent dryland and irrigated wheat fields (Mavrodi et al., 2012b). Results of that study confirmed the affinity of Phz⁺ rhizobacteria for dryland field conditions and revealed that populations of these bacteria were negatively affected by irrigation. Collectively, these findings prompted the hypothesis that Phz⁺ rhizobacteria are uniquely adapted to the plant rhizosphere under conditions of water stress due to their ability to resist desiccation via formation of robust biofilms (see more about the link between phenazines and biofilms below).

In the wheat-producing region covered in the study by Mavrodi and colleagues (2012a) (an area of about 22 000 km²), most sampled fields supported Phz⁺ rhizobacteria with mean population sizes ranging from log 3.2 to log 7.1 cfu g⁻¹ fresh weight of root. Indigenous Phz⁺ rhizobacteria also were detected in a recent screen of soil samples collected from Australian wheat fields (O.V. Mavrodi, T.C. Paulitz and D.M. Weller, unpubl. data). Finally, Mazurier and colleagues (2009) demonstrated the abundance of indigenous Phz⁺ pseudomonads related to P. fluorescens and another phenazine-producing species, P. chlororaphis, in Fusarium wilt-suppressive soils of the Châteaurenard region of France. Taken collectively, these findings suggest that Phz⁺Pseudomonas spp. comprise a ubiquitously distributed group of antibiotic-producing bacteria in both natural and managed ecosystems.

Genetics and enzymology of phenazine biosynthesis

The structures of many naturally occurring phenazines were established in the 1960s and 1970s but the biochemistry of phenazine production remained elusive and was established only partially by label feeding experiments (reviewed in Turner and Messenger, 1986). The breakthrough in understanding the enzymology of synthesis of microbial phenazines occurred in the 1990s, when biosynthesis genes from two species of pseudomonads were cloned and analysed in detail (Pierson et al., 1995; Mavrodi et al., 1998). It is now clear that phenazine-producing bacteria utilize five enzymes to convert chorismic acid to PCA and/or phenazine-1,6-dicarboxylic acid (PDC) (Mentel et al., 2009). These enzymes are commonly encoded within a single or duplicated ‘core’ operon, which is accompanied by genes involved in regulation, transport, resistance and transformation of PCA/PDC to strain-specific phenazine derivatives (Mavrodi et al., 2010). Two recent phylogenetic studies concluded that horizontal gene transfer (HGT) has played an important role in the evolution of phenazine pathways in several distantly related groups of bacteria (Fitzpatrick, 2009; Mavrodi et al., 2010). The evidence of HGT is particularly evident in Phz⁺ enteric bacteria and Burkholderia spp., where phenazine genes often reside on plasmids or transposons (Mavrodi et al., 2010). In contrast, the core phenazine biosynthesis genes of fluorescent pseudomonads are highly conserved, and phylogenies inferred from housekeeping genes and the core phenazine biosynthesis gene phzF of Pseudomonas spp. exhibit a nearly perfect congruence (Mavrodi et al., 2010). The ratio of non-synonymous to synonymous substitutions (dN/dS) in phzF is slightly higher than that in housekeeping genes yet still significantly lower than 1.0, which suggests the presence of purifying selection (Mavrodi et al., 2010; Parejko et al., 2012). These findings reveal the presence of strong evolutionary pressure for the stable maintenance of phz gene function in Phz⁺Pseudomonas spp. and suggest that loss of the capacity to produce phenazines may significantly reduce the ability of these organisms to thrive in the environment.

Initial insight into the function of genes of the phz operon was provided by Floss and co-workers (McDonald et al., 2001), and their tentative pathway has been developed into a more complete view in recent years owing to detailed structural and biochemical studies (Fig. 2). Phenazine biosynthesis begins with the conversion of chorismic acid to 2-amino-2-desoxyisochorismate (ADIC) by ADIC synthase, PhzE. A recent study by Li and colleagues (2011) revealed that PhzE is bifunctional and hydrolyses glutamine in a type 1 glutaminase domain (GATase 1) to generate ammonia, which is transferred to a menaquinone, siderophore, tryptophan (MST) domain where ADIC is generated. In PhzE, both domains are fused into a single polypeptide and the enzyme forms an intertwined dimer in which the GATase 1 of one chain provides NH₃ to the MST domain of the other. Ligand binding induces structural transitions that establish an intramolecular, possibly gated, channel for the transport of ammonia between the two active sites. It is still unclear why PhzE releases ADIC without further elimination of pyruvate, as is observed in the closely related anthranilate synthases. ADIC is next hydrolysed to trans-2,3-dihydro-3-hydroxyanthranilic acid (DHHA) and pyruvate by PhzD, which employs an aspartic acid residue to catalyse the reaction (Parsons et al., 2003). DHHA is the last stable intermediate on the path to PCA/PDC. The enzyme PhzF subsequently isomerizes it to a reactive aminoketone (Blankenfeldt et al., 2004; Parsons et al., 2004b), which was identified as 6-amino-5-oxocyclohex-2-ene-1-carboxylic acid (intermediate 1, Fig. 2) by trapping experiments. Turnover in D₂O further indicated that PhzF initiates a suprafacial shift of the proton at C3 to C1 of DHHA (Blankenfeldt et al., 2004). The reaction may therefore have pericyclic character.

Figure 2. Enzymology of assembly of the phenazine core in bacteria. Abbreviations: ADIC, 2-amino-2-desoxyisochorismate; DHHA, trans-2,3-dihydro-3-hydroxyanthranilic acid; PDCH₂, 5,10-dihydrophenazine-1,6-dicarboxylic acid; PCAH₂, 5,10-dihydrophenazine-1-carboxylic acid; PDC, phenazine-1,6-dicarboxylic acid; PCA, phenazine-1-carboxylic acid.

Two molecules of intermediate 1 can spontaneously condense to the tricyclic product 2 (Fig. 2) that already resembles the phenazine moiety. This reaction is bimolecular and therefore exponentially dependent on substrate concentration, requiring relatively high levels of 1 to proceed uncatalysed. However, because 1 is expected to possess high non-specific reactivity towards, e.g. amines, sufficient accumulation of 1 to allow spontaneous tricycle formation is not possible in the cell and the condensation reaction is instead catalysed by PhzB, a small homodimeric enzyme that belongs to the nuclear transport factor 2/ketosteroid isomerase family. Interestingly, all phenazine-producing pseudomonads carry an additional copy of the gene immediately upstream in the operon. This gene, called phzA, encodes for a protein with approximately 80% sequence identity to PhzB, but due to two mutations in the active centre, PhzA does not display condensation activity. The function of PhzA in phenazine biosynthesis is therefore unclear at present (Ahuja et al., 2008).

Intermediate 2 gains stability through conjugation of its double bonds, leading to the asymmetric structure 2a (Fig. 2). In the presence of oxygen, this intermediate undergoes a rapid uncatalysed oxidative decarboxylation to intermediate 3, which may explain why in the majority of cases PCA and not PDC is found as a precursor for strain-specific phenazine derivatives (Ahuja et al., 2008). Two or even three two-electron oxidations are required to convert 3 into PCA and 2a into PDC respectively. However, the phenazine pathway contains only one oxidase, PhzG, which is related to the FMN-dependent pyridoxine 5′-phosphate oxidase PdxH, and it has been speculated that PhzG is involved in one of the steps following tricycle formation (Parsons et al., 2004a). It can be expected that PhzG oxidizes intermediate 3 to 5,10-dihydrophenazine-1-carboxylic acid (PCAH₂), the reduced form of PCA also involved in phenazine-mediated electron shuttling. PCAH₂, and not the fully aromatic PCA, could therefore be the true end-product of phenazine biosynthesis, an idea supported by the fact that PCAH₂, but not PCA, is readily converted by the phenazine-modifying enzymes of Streptomyces annulatus (Saleh et al., 2009). In the case of PDC biosynthesis, the spontaneous oxidative decarboxylation of intermediate 2a must be avoided, and it is conceivable that this is achieved by enzymatic catalysis with two consecutive oxidations by PhzG or by employing a second enzyme encoded outside of the phz operon before PhzG generates PDCH₂.

The accumulation, activity and turnover of phenazines in the environment

Phenazines are among the most copious secondary metabolites produced by fluorescent pseudomonads, due in part to the presence in the phenazine operon of phzC, a gene encoding a type II 3-deoxy-d-arabinoheptulosonate-7-phosphate (DAHP) synthase enzyme that, once the threshold population size for quorum sensing has been reached, is capable of diverting carbon metabolites into the shikimate pathway for phenazine synthesis (Pierson and Pierson, 2010). The documented extracellular concentrations of phenazines range from 275 µM PYO in batch cultures of P. aeruginosa (Price-Whelan et al., 2007) to yields as high as 18 mM PCA in optimized cultures of a gacA qscR regulatory mutant of Pseudomonas sp. M18 (Su et al., 2010). In the hundred micromolar range, most phenazines act as broad range antibiotics that are toxic to both pro- and eukaryotic organisms (Smirnov and Kiprianova, 1990). The effect of antibiosis is attributed to the redox properties of phenazines and their capacity to promote formation of toxic reactive oxygen species (ROS).

The molecular mechanism of phenazine toxicity is best understood in the case of PYO, which is toxic to bacteria, fungi, plants, nematodes and insects, and which acts as an important pathogenicity factor in both acute and chronic models of lung infection in mice (reviewed in Lau et al., 2004a; Liu and Nizet, 2009). PYO disrupts cellular respiration and inhibits vacuolar ATPase, which in turn leads to alterations in calcium homeostasis, vesicle trafficking and protein targeting, and increased inflammation at the tissue level. In addition, PYO can accept electrons from biological reducing agents and pass them on to oxygen, thus generating ROS and leading to elevated levels of oxidative stress in target organisms (Liu and Nizet, 2009). A recent study by Gibson and colleagues (2009) provided interesting additional details for the ROS-mediated mechanism of phenazine toxicity. The authors studied the capacity of 5-methyl phenazinium-1-carboxylate (5MPCA), a PYO intermediate from P. aeruginosa, to kill the yeast Candida albicans, which can coexist with P. aeruginosa in opportunistic infections of animals. Co-culturing C. albicans with P. aeruginosa resulted in accumulation in the yeast of a red-coloured phenazine derivative that could be reversibly oxidized and reduced, and its presence was correlated with loss of fungal viability (Gibson et al., 2009). Further analyses, facilitated by the use of the 5MPCA analogue phenazine methosulfate (PMS) (Morales et al., 2010), indicated that methylphenazinium compounds, either produced in the reduced form by P. aeruginosa or formed extracellularly, are taken up by the yeast, where they react with cellular amines such as proteins, leading to intracellular accumulation of the red methylphenazinium derivatives. These compounds retain redox activity and generate ROS, which, along with covalent modification of intracellular proteins by 5MPCA or PMS, are thought to be responsible for fungal cell death.

The production of natural antibiotics in situ is poorly understood due to the fact that recovery and detection of microbial metabolites in soil, water, or plant or animal tissues is an arduous task. However, it has been known for some time that phenazines can be produced under environmental conditions in biologically relevant amounts. Thus, Wilson and colleagues (1988) identified PYO and 1-hydroxyphenazine produced by P. aeruginosa in the sputum of cystic fibrosis patients at concentrations of up to 100 µM. At such concentrations pyocyanin becomes toxic to eukaryotic cells and may contribute to the inflammation and pathological changes observed in the lung tissues of CF patients colonized by P. aeruginosa (Lau et al., 2004a,b; Caldwell et al., 2009). Thomashow and colleagues (1990) recovered 0.2 nanomoles of PCA g⁻¹ of root from wheat grown in the greenhouse from seed that had been inoculated with the phenazine-producing strain P. fluorescens 2–79. More recently, Mavrodi and colleagues (2012a) reported recoveries of up to 2.5 nanomoles of PCA g⁻¹ of root produced by indigenous populations of fluorescent Pseudomonas spp. on wheat grown in the field under arid conditions in the Pacific Northwest, USA. The authors estimated that, based on the population size of the indigenous strains and their distribution in the rhizosphere, localized concentrations of PCA could easily reach 100 µM, which is sufficient for inhibition of other microorganisms in the rhizosphere environment. Indeed, isolates of these indigenous PCA producers were capable of suppressing root rot caused by Rhizoctonia solani AG-8, an important pathogen of wheat (Mavrodi et al., 2012a). Several other recent studies also have demonstrated the involvement of phenazines in the control of soil-borne fungal pathogens. These include the suppression by P. aeruginosa PNA1 of damping off and root rot, respectively, caused by Pythium spp. on bean and cocoyam (Perneel et al., 2008), the suppression by Pseudomonas sp. CMR12A of root rot caused by R. solani on bean (D'aes et al., 2011), and the reduction by P. chlororaphis Phz24 of hyphal growth and stem rot disease of groundnut caused by Sclerotium rolfsii (Le et al., 2012). It is worth noting that strains PNA1 and CMR12a also produce rhamnolipid and cyclic lipopeptide surfactants that functioned together with the phenazines in biological control, leading Perneel and colleagues (2008) to speculate that biosurfactants may synergistically enhance the activity of phenazines in the environment by increasing their solubility or facilitating their transport and entry into fungal cells.

While it is clear that phenazines can be produced in the environment, almost nothing is known about the dynamics of synthesis or decay of these metabolites. Y.-S. Kwak, R.F. Bonsall, D.M. Weller and L.S. Thomashow (unpubl. data) introduced 10 µg of PCA onto 10 g of roots that had been grown in a sandy loam soil and monitored recovery over 4 weeks by HPLC-coupled mass spectrometry. The half-life of PCA under these conditions was 3.4 days (Fig. 3), suggesting that phenazines are not long-lived under environmental conditions and that sustained synthesis may be required to maintain concentrations inhibitory to fungal root pathogens. It remains to be determined whether the decay values reflect irreversible binding to soil organic matter and mineral surfaces or microbial degradation, but the fact that phenazine-degrading bacteria can readily be isolated from soil (Chen et al., 2008; D.V. Mavrodi and O.V. Mavrodi, unpubl. data) favours the latter possibility.

Figure 3. Half-life of PCA in the rhizosphere. Ten micrograms of PCA in acetonitrile was spiked onto 10 g of wheat roots (cv. Penawawa) with rhizosphere soil in flasks. The roots were collected from seedlings that had been grown for 3 weeks in a Shano sandy loam collected in Quincy, Washington State, USA. The flasks were incubated at room temperature in the dark and sampled in triplicate (each flask served as a biological replicate) at 0 and 12 h, 1, 2 and 4 days, and 1, 2, 3 and 4 weeks. PCA was extracted from wheat roots essentially as described by Mavrodi and colleagues (2012a), and amounts recovered from the roots were quantified using a Q-Tof 2 mass spectrometer outfitted with an IonSabre atmospheric-pressure chemical ionization (APCI) probe, and coupled to a Waters 2695 liquid chromatograph with a 996 photodiode array detector (all from Waters). The separations were carried on a C₁₈ reverse-phase Symmetry column (5 µm, 150 mm × 3.5 mm, Waters) in a linear gradient of acetonitrile (1–10% v/v in 2% acetic acid). The gradient elution profiles were monitored for PCA at 248 nm, whereas Q-Tof two profiles were monitored at m/z 225.0664, the exact mass for phenazine-1-carboxylic acid under the APCI-positive ion mode (Mavrodi et al., 2012a). The half-life of the antibiotic, 3.4 days, was calculated by the experimental decay model y = ae^−bx. Error bars correspond to standard deviations of the mean.

Phenazines as molecular signals

Pseudomonas spp. often exist in nature as members of complex surface-attached communities enmeshed in exopolymers described as biofilms. Biofilms are thought to facilitate the development of a microhabitat that protects bacteria against abiotic and biotic stresses (Lopez et al., 2010), and are formed as a result of the production of various exopolysaccharides, lipopolysaccharides and proteinaceous attachment fibres by bacteria (Flemming and Wingender, 2010). Among the most interesting recent findings is the growing realization that phenazine production in Pseudomonas represents an adaptation to the biofilm lifestyle. Phenazines act both as molecular signals regulating the switch from the planktonic to the attached form of growth and as metabolites of crucial importance for the development of a mature biofilm.

Expression of the seven-gene phenazine biosynthesis (phz) operon is controlled in pseudomonads by homoserine lactone (HSL)-mediated quorum sensing (QS) (Brint and Ohman, 1995; Latifi et al., 1995; Wood and Pierson, 1996; Wood et al., 1997; Chancey et al., 1999; Khan et al., 2005; 2007) and is modulated by numerous transcriptional and translational regulators and small ncRNAs (Haas and Keel, 2003; Haas and Defago, 2005). QS regulation is the most complex in P. aeruginosa, where it consists of two hierarchical HSL-based circuits, Las and Rhl, whose activity is modulated by a third QS system based on 2-heptyl-3-hydroxy-4-quinolone, also known as the Pseudomonas quinolone signal (PQS) (Dubern and Diggle, 2008). In P. fluorescens and P. chlororaphis, QS regulatory genes are situated immediately upstream of the phenazine biosynthesis genes, as opposed to in P. aeruginosa, where las and rhl genes are not directly linked to the two copies of the phz gene cluster. The first evidence that phenazines themselves act as signalling compounds was provided by Dietrich and colleagues (2006) who demonstrated that PYO affects the expression of a small subset of genes in P. aeruginosa. That study also revealed that the HSL-, quinolone- and phenazine-based systems form a temporal regulatory cascade and operate in different growth phases, and that PYO acts downstream of the PQS system. However, Dietrich and colleagues (2006) carried out their experiments with planktonic cultures, and the connection between phenazine-mediated signalling and biofilm formation was demonstrated more recently by Ha and colleagues (2011). In P. aeruginosa, biofilm formation is inversely correlated with swarming motility on semisolid surfaces (Kuchma et al., 2007; Merritt et al., 2007), and Ha and colleagues (2011) demonstrated that this phenomenon involves 2-heptyl-4-quinolone (HHQ), an immediate precursor of PQS. They further revealed that HHQ induces the production of phenazines in P. aeruginosa, and that the phenazine PCA represses swarming and stimulates biofilm formation by a yet unknown mechanism. Currently it is unclear if similar regulatory mechanisms operate in other Pseudomonas spp., and preliminary studies indicate that the part of the transcriptome of P. chlororaphis that is induced by phenazines seems to be different from that of P. aeruginosa (Pierson and Pierson, 2010). However, at least some of the genes induced by phenazines in P. chlororaphis include those involved in surface-attached growth and biofilm development.

Phenazines and the physiology of biofilm growth

The importance of phenazines in the development of mature surface biofilms was first established in P. chlororaphis, in which no biofilms were formed in the absence of phenazines and biofilm architecture and bacterial dispersal rates were dependent on the identity and ratios of the phenazines produced (Maddula et al., 2006; 2008). Later experiments with P. aeruginosa revealed that although its Phz^- mutants formed biofilms, biofilm architecture was strongly influenced by the identity and amounts of amended phenazines (Ramos et al., 2010). The molecular mechanism explaining this phenomenon has emerged in a series of recent studies focused on the redox-cycling properties of phenazines in the context of the physiology of biofilm growth.

Pseudomonads are strict aerobes, and Price-Whelan and colleagues (2006) suggested that because the diffusion rate of oxygen through biofilms is thought to be slow, phenazines could help maintain the redox homeostasis of cells embedded in the film by acting as electron acceptors for the reoxidation of accumulating NADH. Indeed, in oxygen-limited stationary-phase cultures of P. aeruginosa, a decrease in intracellular NADH/NAD⁺ was correlated with the presence of pyocyanin in the culture (Price-Whelan et al., 2007), and phenazine-facilitated electron transfer promoted anaerobic survival but not growth under conditions of oxidant limitation (Wang et al., 2010). Interestingly, other electron shuttles that were reduced but not synthesized by P. aeruginosa did not facilitate survival, suggesting that sophisticated systems are needed to control the reactivity of these molecules within the cell and that mechanisms have evolved in pseudomonads to be specific for the phenazines they produce (Wang et al., 2010).

The changes in gene expression in response to PYO are mediated in P. aeruginosa by the transcription factor SoxR, which also has been implicated in the cellular responses of Pseudomonas spp. to other redox-cycling agents such as paraquat (Dietrich et al., 2006; Park et al., 2006). In Escherichia coli, SoxR, together with another transcription factor, SoxS, controls the expression of genes whose primary function is to protect the cell against redox-cycling agents and superoxide (Gu and Imlay, 2011). In contrast, the SoxR regulon of pseudomonads and other bacterial species producing redox-active metabolites is made up of genes encoding transporters, mono- and dioxygenases, and oxidoreductases. It has been suggested that the primary function of the SoxR response in Pseudomonas spp. is not defence against phenazine-induced superoxide stress, but rather, the proper shuttling of phenazines and other redox-cycling metabolites as electron acceptors that help to balance the intracellular redox state (Dietrich and Kiley, 2011).

A recent study by Koley and colleagues (2011) investigated the role of phenazines in the context of the spatial heterogeneity observed in bacterial biofilms. A better understanding of the chemical gradients within biofilms is critical for revealing how spatially isolated bacterial communities interact with the environment and neighbouring communities. Koley and colleagues (2011) carried out real-time spatial profiling of the pyocyanin gradient and its redox state in P. aeruginosa using a technique called scanning electrochemical microscopy, which allows a non-invasive determination of gradients of concentration and redox status of metabolites in bacterial biofilms. Results of the study revealed that P. aeruginosa biofilms actively produce PYO and maintain it in the reduced state. The layer of reduced PYO, which the authors called the ‘PYO electrocline’, has a thickness of over 400 microns and is formed under the electron acceptor-limiting conditions typical of mature biofilms. Results of that study also confirmed the hypothesis that PYO serves as an electron acceptor for P. aeruginosa (Price-Whelan et al., 2007). When the authors flooded the biofilm with an alternative electron acceptor, nitrate, they observed the rapid collapse of the PYO electrocline, and nitrate had no impact on the PYO electrocline in a nitrate reductase mutant that cannot reduce NO₃^- under aerobic conditions (Koley et al., 2011). Also, no electrocline was observed in a cytochrome bc1 mutant lacking the ability to reduce PYO (Price-Whelan, 2009). Finally, the formation of an active PYO electrocline was positively correlated with the accumulation of soluble iron in the medium surrounding the biofilm. These results further support earlier observations that, under aerobic conditions, PYO and other phenazines can reduce Fe³⁺ to Fe²⁺ (Wang and Newman, 2008; Wang et al., 2011).

Interestingly, elevated levels of iron trigger a transition from planktonic growth to biofilm formation in P. aeruginosa (Banin et al., 2005; Berlutti et al., 2005; Patriquin et al., 2008). A recent study by Wang and colleagues (2011) investigated this link between phenazine production, biofilm formation, and iron acquisition in greater detail. The authors demonstrated that siderophore mutants of P. aeruginosa, which are deficient in biofilm formation, could form normal biofilms in the presence of ferric iron and PCA, and that the effect was mediated by FeoB, an Fe²⁺ uptake protein. These results suggest that the PCA mediates reduction of Fe³⁺ to Fe²⁺ and that subsequent sequestration of the latter represents an important mechanism of iron acquisition for P. aeruginosa. The authors further speculated that Fe²⁺ and haem may be particularly important sources of iron for P. aeruginosa during the advanced phase of colonization of the CF lung.

Concluding remarks

Recent years have seen great progress in our understanding of the diverse biological functions of naturally occurring phenazines. In the past decade, phenazines produced by fluorescent Pseudomonas spp. have been implicated in electron shuttling, which contributes to iron acquisition, the modulation of intracellular redox homeostasis, regulation of colony morphology and gene expression, and the transition to biofilm growth and/or survival under conditions of low oxygen tension. New insights also have been gained into genetics and enzymology of assembly of the phenazine tricycle. Finally, ecological and phylogenetic studies have demonstrated the ubiquitous distribution of Phz⁺Pseudomonas spp. and the accumulation of high amounts of phenazines in terrestrial ecosystems, which is consistent with the proposed functions of these metabolites in natural settings. Taken collectively, these results demonstrate that bacterial phenazines are bioactive metabolites that play important roles in the physiology, ecological fitness, competitiveness, and the pathogenic and biocontrol activity of the producing strains.

In spite of these impressive advances, there remain important gaps in our understanding of the biology of phenazines that warrant further investigation. The first of these has to do with the fact that most studies focused on the ecological roles of phenazines continue to be conducted in vitro. It therefore seems fitting to validate the proposed diverse functions of bacterial phenazines (i.e. molecular signalling, promotion of biofilm formation and mineral reduction) in more natural settings.

Another gap in our understanding of phenazine biology concerns the fate of these bioactive metabolites in the environment. Pharmaceutical antibiotics, which are used to control or prevent infectious diseases in humans and animals, are ultimately excreted and contaminate surface, ground and drinking water, aquatic sediments and soil (Thiele-Bruhn, 2003). There is now extensive research focused on the amounts, fate, and non-target and long-term effects of pharmaceuticals in the environment (Cabello, 2006; Sarmah et al., 2006; Kim and Aga, 2007; Dantas et al., 2008). Phenazines possess broad-spectrum antibiotic activity and there is mounting evidence that many species in microbial communities have the capacity to produce these metabolites in biologically relevant amounts. However, in contrast to pharmaceutical antibiotics contaminating the environment, very little is known about the accumulation and turnover of phenazines produced in situ by indigenous microorganisms.

In recent years, the biosynthesis of the phenazine tricycle and functions of individual core enzymes have been elucidated in great detail but major gaps exist in the understanding of phenazine-modifying enzymes. Only a handful of these enzymes have been characterized experimentally (Bringmann et al., 2007; Parsons et al., 2007; Greenhagen et al., 2008; Saleh et al., 2009; 2012; Bera et al., 2010), and the biochemistry behind modification of PCA and the generation of species-specific phenazine compounds is poorly understood. Future research should focus on the mechanisms of selectivity of phenazine-modifying enzymes and activation of their substrates for modification, particularly in Actinobacteria that produce heavily substituted phenazines with complex chemical structures.

Finally, most of our recent knowledge about the diverse biological functions of phenazines and the effects they exert on eukaryotic hosts is based on a single model organism – the pyocyanin-producing opportunistic pathogen P. aeruginosa. Similar knowledge for other types of phenazines or other pathogenic and saprophytic species is severely lagging. Based on studies of the evolutionary history of phenazine clusters, we recently hypothesized that phenazines may play very different roles in different groups of Phz⁺ bacteria (Mavrodi et al., 2010). We proposed that the high degree of conservation of phz genes in fluorescent Pseudomonas spp. reflects the fact that in these bacteria, phenazines evolved as metabolites of key importance for the biofilm lifestyle. On the other hand, taxa that harbour horizontally transferred phz gene clusters may primarily employ phenazines as competition factors that help the producers to displace closely related Phz^- strains/species. Further studies are needed to better understand the role of these compounds in diverse groups of Phz⁺ bacteria and to prove or disprove this hypothesis.

Acknowledgements

The research was supported in part by Grant 2011-67019-30212 from the USDA-NIFA Soil Processes Program.