Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

2.1 Introduction

The dramatic coloration of phenazine-producing bacterial cultures has attracted researchers in many disciplines for over a century (Fordos 1859; Gessard 1894). It may have contributed to the early recognition and classification of Pseudomonas chlororaphis (Guignard and Sauvageau 1894), P. aeruginosa, and other bacteria whose species epithets derive from their pigmentation (Schroeter 1872). Furthermore, it has inspired researchers to ask many different types of questions about the biological relevance of phenazines, and a variety of physiological effects have been demonstrated for these compounds in both the organisms that produce them and the organisms exposed to them (see Chap. 3).

For microbiologists cultivating phenazine-producers, it is apparent that phenazine biosynthesis can vary unpredictably among cultures, suggesting that it is sensitive to subtle environmental variations (Fig. 2.1). Under many conditions, the precise molecular cues that interact with regulatory proteins to control phenazine biosynthetic gene expression are not known. However, downstream mechanisms controlling their expression have been identified in several species, and themes of phenazine regulation have emerged, including control by two component systems, quorum sensing (QS), and small noncoding RNAs (sRNAs) (Fig. 2.2). In this chapter, we will summarize the phylogenetic distribution of phenazine biosynthetic clusters and cite examples from phenazine-producing pseudomonads that illustrate specific regulatory mechanisms. In addition, we will discuss some of the environmental signals that control phenazine production in various isolates. We will focus on the regulation of phenazine biosynthesis in members of the genus Pseudomonas, where the bulk of studies on this topic have been conducted.

Fig. 2.1
figure 1

P. aeruginosa PA14 grown on an agar plate containing a gradient of tryptone

Full size image
Fig. 2.2
figure 2

Conceptual hierarchy of the phenazine regulation network

Full size image

2.2 Phylogenetic Distribution and Mechanisms of Phenazine Biosynthesis

Phenazine-producing organisms have been identified that belong to the bacterial phyla Actinobacteria and Proteobacteria and the archaeal phylum Euryarcheota (Mavrodi et al. 2010). The gene cluster encoding biosynthetic enzymes for the archaeal electron carrier methanophenazine, produced by Methanosarcina mazei, and the conditions controlling its biosynthesis are unknown. In bacterial species for which phenazine production has been observed, variation exists at the level of the isolate, such that some strains within a species produce phenazines while others lack the biosynthetic genes. Bacterial phenazine biosynthetic pathways identified to date proceed via chorismate to the formation of the core phenazine structure and ultimately, to production of the common phenazine precursor phenazine-1-carboxylic acid (PCA) (Mentel et al. 2009; Seeger et al. 2011; see also Chap. 1 of this volume). Whether and how PCA is modified to produce other phenazines varies among organisms and depends to some extent on environmental conditions. A variety of functional groups can be added to the core structure to produce phenazines in a range of colors with diverse chemical properties (Turner and Messenger 1986; Laursen and Nielsen 2004; Mavrodi et al. 2006; Pierson and Pierson 2010) (Table 2.1). Many of the decorating enzymes responsible for PCA transformation have been identified and characterized, and regulation of their activities determines the complement of phenazines produced by a given strain under particular conditions.

Table 2.1 Phenazine derivatives produced by Pseudomonas spp. and other bacteria
Full size table

The archetypical core phenazine operon is found in pseudomonads such as P. chlororaphis, P. aeruginosa, Pseudomonas sp. CMR12a, and P. fluorescens 2-79 and contains seven genes; the operon structure is more variable in other Proteobacteria and in Actinobacteria (Mavrodi et al. 2010). P. aeruginosa strains appear to be unique among these organisms in that their genomes contain two phenazine biosynthetic operons, which we will refer to as phz1 and phz2 and which are nearly identical at the DNA level. Although the contributions of the core phz operon products to phenazine biosynthesis are generally known, this is an area of active research (see Chap. 1 of this volume). Genes for decorating enzymes can lie adjacent to the core operon or elsewhere in the genome. In some cases, the products of adjacent genes play roles in regulation of the core operon or phenazine transport.

2.3 Mechanisms and Conditions Controlling Phenazine Biosynthesis in Pseudomonads

Researchers in disparate subdisciplines of microbiology have long been interested in elucidating the mechanisms that control phenazine biosynthesis. Phenazine production is critical for the biocontrol properties of certain agriculturally important root-colonizing pseudomonads that protect food crops from attack by pathogenic fungi (Chin-A-Woeng et al. 2003; Haas and Défago 2005; Mavrodi et al. 2006; Mavrodi et al. 2012). In the clinical setting, phenazine production contributes to virulence during acute and chronic P. aeruginosa infections (Lau et al. 2004; Caldwell et al. 2009; Hunter et al. 2012). Regulation of phenazine biosynthesis has therefore been studied in diverse Pseudomonas isolates. Although the precise linkages between regulatory mechanisms may vary among species and even strains, general mechanisms and their overall hierarchy are often shared. We will highlight these commonalities and focus on specific systems that exemplify general principles. Figure 2.3 summarizes the main regulatory systems and molecules that ultimately control phenazine production: two component systems, QS, sRNAs and environmental cues. These mechanisms and cues can act indirectly by influencing activities far upstream of phenazine biosynthetic gene expression or RNA translation, or they can directly control these processes. Additional details regarding the complex relationships between and within these regulatory systems can be found in recent reviews that summarize the literature (Mavrodi et al. 2006; Williams and Camara 2009; Pierson and Pierson 2010; Sonnleitner and Haas 2011; Balasubramanian et al. 2013).

Fig. 2.3
figure 3

Proteins and sRNAs that control phenazine production, with targets indicated. X indicates which pseudomonad contains a particular element or in which an environmental effect on phenazine production has been observed. Shading in the table indicates the location of each type of element in regulatory cascades; a sample cascade is shown. *Only found in P. aeruginosa

Full size image

2.3.1 Two Component Systems

In both biocontrol and pathogenic pseudomonads, two component systems were among the first regulatory mechanisms identified that play critical roles in phenazine regulation (Reimmann et al. 1997; Chancey et al. 1999; van den Broek et al. 2003; De Maeyer et al. 2011). They lie conceptually at the top of signaling hierarchies because they have the potential to directly sense environmental cues and then modulate the activities of downstream regulatory mechanisms or directly control gene expression (Fig. 2.4). Such systems typically consist of a membrane-bound sensor protein and a cytoplasmic regulatory protein. The phosphorylation status of the sensor protein is altered through binding of a small molecule or other environmental cue that triggers conformational changes and affects activity. The phosphate group is then transferred to and activates the response regulator protein (Bourret and Silversmith 2010).

Fig. 2.4
figure 4

Summary of two component, quorum sensing, and sRNA systems that control phenazine production in P. aeruginosa and P. chlororaphis. The P. fluorescens regulatory system is likely similar to that of P. chlororaphis

Full size image

GacS/GacA, which is required for wild-type phenazine production in many isolates, is the best-characterized two component system controlling this process (Heeb and Haas 2001; Haas and Defago 2005). In phenazine-producing species, it occupies a position between environmental cues and downstream, intracellular regulatory mechanisms such as those dependent on sRNAs. In describing this system and others below, we will include references to the regulatory cascade in P. fluorescens strains that do not produce phenazines in cases where it is possible that the same cascade operates and affects phz gene expression in P. fluorescens 2-79. In addition, we note here that we use ORF numbers from P. aeruginosa PAO1 (starting with "PA") for some of the proteins described below.

The cue that activates GacS has not been identified, but additional membrane proteins that control GacS activity in some isolates are known. These include RetS and LadS, hybrid sensor kinases that contain the unusual arrangement of periplasmic sensor domains linked to cytoplasmic histidine kinase and response regulator receiver domains (Goodman et al. 2004; Ventre et al. 2006). In strains of P. fluorescens and P. aeruginosa, RetS interacts with GacS and inhibits the phosphorelay (Goodman et al. 2009; Workentine et al. 2009; Vincent et al. 2010). A physical interaction between LadS and GacS is also predicted, however, this hybrid sensor potentiates phosphotransfer from GacS to GacA (Workentine et al. 2009). Although the GacS/GacA system does not directly control expression of phz genes, it does modulate the activities of sRNA- and QS-dependent regulatory mechanisms, which can then directly interact with phz gene promoters or transcripts (Fig. 2.4). These systems are discussed in further detail below.

The two component system CzcS/CzcR was recently implicated in regulation of phz gene expression in P. aeruginosa PAO1 (Dieppois et al. 2012). CzcS/CzcR is activated by zinc, cadmium, and cobalt and induces expression of an efflux pump that confers resistance to these metals. Dieppois et al. (2012) observed that mutants lacking functional CzcS/CzcR overproduce the phenazine pyocyanin (5-N-methyl-1-hydroxyphenazine, PYO), despite the fact that this two component system positively regulates QS. Chromatin immunoprecipitation assays suggest that CzcR binds to the phz1 promoter when the system is activated by zinc. In this way, CzcR could inhibit phz1 expression directly, negating the positive control of this operon by QS. This mechanism would constitute an unusual example of a direct link between a two component system and transcriptional control of phz genes.

Several other two component systems have been shown to affect phenazine production in various isolates, and evidence reported thus far indicates that this regulation is mediated via additional proteins and mechanisms. The RpeB/RpeA system positively regulates phenazine production in P. chlororaphis 30-84, and homologues to this system are present in other biocontrol strains but not in P. aeruginosa (Wang et al. 2012a). Proteins identified that positively regulate, or would be expected to regulate, phenazine production or phz genes in P. aeruginosa include the sensor/regulator pair BfiS/BfiR (acting via an indirect effect on levels of the sRNA RsmZ) (Petrova and Sauer 2010); the sensor/regulator pair CbrA/CbrB, which induces the expression of sRNA CrcZ in response to nonpreferred carbon sources (Sonnleitner et al. 2009); the sensor PA2573, which affects PYO production through an unknown regulator (McLaughlin et al. 2012); and the individual sensors PA1611, PA1976, and PA2824, which can all control the activation state of the response regulator HptB (Hsu et al. 2008). HptB would be expected to affect phenazine production indirectly through a complex regulatory cascade that ultimately controls expression of the sRNA RsmY. The positions of RsmZ and RsmY in the regulatory network controlling phenazine production are discussed further below.

2.3.2 Quorum Sensing

Phenazine production in liquid batch cultures typically occurs after the period of most rapid growth, and phenazines accumulate in the culture in stationary phase. This is in part due to the regulation of phz gene expression by QS. In QS, bacteria excrete small molecule or peptide signals into the environment which can then affect gene expression in the producer. Their regulatory effects become apparent after they have reached a minimum concentration, often after a decrease in culture growth rate. Molecules with diverse structures have been implicated in this behavior, but acyl homoserine lactones (AHLs) and quinolone derivatives in particular are most relevant for phz gene expression.

Systems that support AHL-dependent QS control of phz gene expression contain homologues of LuxR, a DNA binding sensory protein whose activity is controlled by AHL, and, typically, homologues of LuxI, an AHL synthase. These proteins were first identified in the recently reclassified luminescent bacterium Aliivibrio fischeri (Meighen 1991) and their homologues have since been characterized in a broad diversity of species. LuxR homologues vary in their specificity for AHL derivatives, with some proteins requiring a single signal for activation and others responding to several variations on a core structure. In P. chlororaphis strains 30-84 and PCL1391, P. fluorescens 2-79, and Pseudomonas sp. CMR12a, the LuxR/I homologues PhzR and PhzI are encoded by ORFs that lie adjacent to the phenazine biosynthetic genes but are each individually transcribed (Pierson et al. 1994; Wood and Pierson 1996; Chin-A-Woeng et al. 2001; Khan et al. 2005; De Maeyer et al. 2011). The PhzR/I systems in these strains produce and are controlled by N-(3-hydroxy-hexanoyl)-HSL (3–OH–C6–HSL). Although additional LuxR/I type systems in some of these strains produce and respond to other AHLs, 3–OH–C6–HSL is the main such signal relevant for phz gene regulation (Khan et al. 2007). The promoter regions of the phz operons in the root-colonizing strains that have phzR and phzI contain canonical binding sites for LuxR-type regulators; these are near-perfect inverted repeats that can be referred to as lux, las/rhl, or phz boxes (Egland and Greenberg 1999; Chin-A-Woeng et al. 2001; Khan et al. 2005).

The P. aeruginosa genome encodes at least three LuxR homologues called LasR, RhlR, and QscR. The cognate AHL synthases for LasR and RhlR produce N-(3-oxododecanoyl) homoserine lactone (3–O–C12–HSL) and N-butanoyl-L-homoserine lactone (C4–HSL), respectively (Pearson et al. 1994; Pearson et al. 1995). Interestingly, QscR has no obvious cognate AHL synthase, but it responds most effectively to 3–O–C12–HSL (Lee et al. 2006). Whether these LuxR homologues activate or repress gene expression depends on the location of the binding site relative to the transcription start site of the target gene. In contrast to the QS circuits in root-colonizing pseudomonads that control phenazine production, which are not known to regulate loci other than the phz operons, the AHL-controlled regulatory networks in P. aeruginosa affect expression of countless targets (Whiteley et al. 1999; Wagner et al. 2003).

Although their genomes share high sequence similarity, the P. aeruginosa strains PAO1, PA14, and M18 exhibit strain-dependent differences with respect to QS-dependent regulation of phz gene expression, and in many cases the mechanisms underlying these activities have not been thoroughly characterized. Often, the PCA derivative PYO is used as an indicator molecule in studies evaluating phenazine production because it is easier to detect than the other P. aeruginosa phenazines. In PAO1 and PA14, Las- and Rhl-defective mutant strains lose the ability to produce PYO, while in M18, the Las and Rhl systems are apparently negative regulators of phenazine production. Recently, Wurtzel et al. (2012) used gel mobility shift assays to confirm the presence of a LasR/RhlR binding site in the promoter region of phz1 in PA14. This binding site also influences expression of phzM, which encodes an enzyme that catalyzes the first step in the transformation of PCA to PYO, and is divergently transcribed from the phz1 operon. No las/rhl box has been identified in the promoter region of phz2, although interestingly, the gene encoding QscR lies adjacent to this operon. QscR is a negative regulator of phz1 and phz2 expression and appears to act through repression of lasI (Chugani et al. 2001; Ledgham et al. 2003).

Many additional regulators have been identified that affect QS, thereby altering phenazine production; in some cases they may affect phenazine production both indirectly through QS and through direct regulation of phz gene expression (Beatson et al. 2002; Juhas et al. 2004; Xu et al. 2005; Liang et al. 2009; Rampioni et al. 2009; Siehnel et al. 2010). An important class of regulators that can influence transcription are sigma factors, which associate with RNA polymerase and control preferences for specific promoters. The sigma factors σ S (RpoS) and σ 54 (RpoN) both affect QS-dependent regulation. In P. aeruginosa, σ S participates in mutual regulation with AHL-dependent QS, in which σ S stimulates a moderate induction of lasR and rhlR, and these QS systems subtly induce rpoS (Schuster et al. 2004). Despite this, P. aeruginosa PAO1 rpoS mutants overproduce PYO, suggesting that σ S also acts independently of AHL to modulate phenazine biosynthesis (Suh et al. 1999). In P. chlororaphis 30-84, σ S is positively regulated by the GacS/GacA two component system and activates phenazine inducing protein (Pip). Pip positively regulates the PhzR/I QS system, which ultimately upregulates the P. chlororaphis phenazine operon, making σ S a positive regulator of phenazine production in this strain (Girard et al. 2006a). Although the downstream effects of σ S on phenazine production differ in P. aeruginosa and P. chlororaphis, the regulator PsrA positively controls σ S activity in both strains (Kojic and Venturi 2001; Girard et al. 2006b). Conflicting results have been reported regarding the effects of σ 54 on the Rhl QS system in P. aeruginosa (Heurlier et al. 2003; Thompson et al. 2003). PA14 mutants lacking functional σ 54 produce less PYO (Hendrickson et al. 2001); this may be because the expression of the CrcZ sRNA (discussed below) is σ 54-dependent (Abdou et al. 2011).

One important target of the AHL-controlled regulatory network in P. aeruginosa is the operon pqsABCDE. This locus is required for the production of another class of chemical signaling molecules called quinolones, and together, the AHL and quinolone signaling pathways form the core of the P. aeruginosa QS signaling cascade (Pesci et al. 1999). pqsA-D encode biosynthetic enzymes that are required for production of 2-heptyl-4-quinolone (HHQ). PqsH, encoded elsewhere in the genome, is a monooxygenase that converts HHQ to Pseudomonas Quinolone Signal (PQS) (Deziel et al. 2004). Both HHQ and PQS activate the transcriptional regulator PqsR (also known as MvfR), but PQS does so with greater efficiency (Xiao et al. 2006; Diggle et al. 2007). PqsR itself activates expression of pqsABCDE; therefore, HHQ/PQS and PqsR participate in an autoregulatory positive feedback loop in which the quinolones potentiate their own production (Fig. 2.5).

Fig. 2.5
figure 5

Quinolone-mediated expression of the pqs operon and its influence on transcription of the two phz operons in the presence and absence of oxygen

Full size image

pqsE encodes a putative metallo-β-hydrolase of unknown function (Yu et al. 2009) that appears to be “caught” in the positive feedback loop controlling HHQ production: it is induced as a result of this mechanism but is not required for HHQ synthesis. Nevertheless, PqsE is required for phenazine production in P. aeruginosa PAO1 and PA14 (Gallagher et al. 2002; Recinos et al. 2012). Constitutive expression of PqsE in a pqsA mutant background is sufficient to promote phenazine production (Farrow et al. 2008), suggesting that, in the context of phz operon expression, the positive feedback loop that promotes HHQ production serves the sole purpose of tangentially upregulating pqsE (Williams and Camara 2009). The mechanism whereby PqsE promotes phz operon expression remains completely undefined, as PqsE itself does not contain a DNA binding domain. PqsE may be involved in the transformation of an unknown signal (Yu et al. 2009). We hypothesize that this signal controls the activity of an unidentified regulator of phz gene expression.

Until recently, studies examining the roles of quinolones in the regulation of phz gene expression focused on aerobically grown, well-mixed planktonic cultures. Under this condition, PQS is required for phz1 expression and phz1 is a major contributor to phenazine biosynthesis. Our group has evaluated the relative contributions of phz1 and phz2 to phenazine production in aerobic liquid cultures and biofilms in P. aeruginosa PA14 (Recinos et al. 2012). We have reported that, although phz1 is expressed at much higher levels than phz2 in P. aeruginosa PAO1 grown in aerobic liquid cultures (Whiteley et al. 1999; Chugani et al. 2001), in strain PA14 both operons make significant contributions to phenazine production when it is grown under the same conditions.

Interestingly, when P. aeruginosa PA14 is grown as a colony biofilm on agar plates, phz2 alone is sufficient for wild-type phenazine production. Furthermore, HHQ is sufficient to fully activate expression of phz2 in this context. The observation that HHQ rather than PQS is the major regulator of phz gene expression in biofilms is intriguing when considered in the context of oxygen availability. The conversion of HHQ to PQS is catalyzed by PqsH and requires molecular oxygen (Schertzer et al. 2010). In biofilms, which become anoxic at depth due to limited diffusion and oxygen consumption by cells closer to the surface (Dietrich et al. 2013), HHQ is likely produced in greater abundance than PQS. phz2 expression is mediated through PqsE and downstream regulators that are apparently specific for this operon. The mechanism whereby PqsE controls expression of phz1 and phz2, and the mechanisms that confer differential, condition-dependent expression of phz1 and phz2 are currently under investigation but likely include AHL-dependent regulation (Farrow et al. 2008).

2.3.3 Post-transcriptional Regulation

Several regulatory mechanisms have been identified that control, or would be expected to control, pseudomonad phenazine production post-transcriptionally (Fig. 2.6). These mechanisms are diverse and include mRNA binding by proteins and mRNA base pairing with 5′-leader RNA sequences, both of which can affect translation (Sonnleitner and Haas 2011). Additional sRNAs can further modulate the binding of such proteins and cis-acting regulatory RNAs to mRNA. Expression of these protein and RNA regulators is often controlled by two component systems or QS. They can indirectly control phz gene expression by modulating earlier steps in the regulatory cascade, or directly control translation of phz mRNAs. Extensive characterization of post-transcriptional regulators has been conducted in P. aeruginosa phenazine-producing strains, but also in P. fluorescens strains that do not contain phz operons. Our discussion includes references to these P. fluorescens strains as their post-transcriptional regulatory mechanisms may be relevant for the regulation of phenazine production in P. fluorescens 2-79.

Fig. 2.6
figure 6

sRNAs and proteins that regulate phenazine production post-transcriptionally in P. aeruginosa. Some sRNAs regulate target genes by base pairing with target mRNAs, while others sequester proteins that interact with mRNAs. Known environmental cues that regulate these elements are indicated

Full size image

The proteins RsmA and RsmE modulate secondary metabolism, QS-dependent activities and phenazine production in diverse pseudomonads. Both proteins are found in P. fluorescens and P. chlororaphis, while only RsmA is found in P. aeruginosa (Blumer et al. 1999; Reimmann et al. 2005). In P. fluorescens, RsmA and RsmE have been shown to interact with and inhibit the translation of target mRNAs that contain unpaired ANGGA motifs near the ribosomal binding site; by stabilizing a stem loop that contains the ANGGA motif, RsmA/E prevents the ribosome from binding (Lapouge et al. 2007). In P. aeruginosa strains, RsmA expression increases with cell density and regulates the LasR/I and RhlR/I QS circuits in a post-transcriptional manner (Pessi et al. 2001). It is thought that RsmA binds to lasR and rhlR mRNAs, inhibiting their translation. Though RsmA and RsmE have not been studied in the phenazine-producer P. fluorescens 2-79 (containing the LuxR/I homologues PhzR/I), they may have a similar effect on QS in this strain.

As RsmA is a negative regulator of lasR and rhlR mRNAs, one would predict that mutations in rsmA would lead to phenazine overproduction (Reimmann et al. 2005). Burrowes et al. (2006) found, however, that the phenotype of an rsmA mutant was condition-dependent: the mutant showed decreased PYO production in LB but increased PYO production in a defined medium containing glycerol and alanine (Burrowes et al. 2006). Furthermore, differing phenazine production phenotypes in rsmA mutants in strains PAO1 and M18 may indicate that temperature is an additional environmental parameter that affects this regulatory cascade. In M18, which is typically grown at 28 °C, RsmA is a positive regulator of phenazine production (Zhang et al. 2005). Interestingly, RsmA consensus sequences are present near the ribosomal binding sites of the phzA1 and phzA2 promoters, raising the possibility of direct control of the phenazine biosynthetic operons by RsmA. Preliminary evidence suggests that RsmA and RsmE also regulate expression of the phz operon in P. chlororaphis, although whether this occurs via direct interaction with phz operon mRNA, through regulation of the PhzR/I QS system, or both has not been reported (Wang et al. 2012a).

sRNAs containing the ANGGA motif can compete with target mRNAs for binding sites on RsmA and RsmE, thereby controlling the extent to which these proteins repress their targets. In P. aeruginosa, two of these sRNAs, called RsmY (sometimes referred to as RsmB) and RsmZ, have been identified (Heurlier et al. 2004; Burrowes et al. 2005). P. fluorescens strains produce homologues of these plus an additional sRNA called RsmX (Heeb et al. 2002; Valverde et al. 2003; Kay et al. 2005). Evidence suggests that the genes encoding these sRNAs are directly regulated by GacA. When phosphorylated GacA activates transcription of rsmX, rsmY, and rsmZ, the sRNA products bind to and sequester the translational repressors RsmA and RsmE, allowing expression of RsmA/E target genes (Heurlier et al. 2004; Kay et al. 2006). In P. aeruginosa, additional regulators have been reported to control expression of the RsmY and RsmZ sRNAs. HptB indirectly represses RsmY expression through a complex regulatory cascade (Hsu et al. 2008; Bordi et al. 2010). Furthermore, the BfiS/BfiR two component system induces expression of RNaseG, which specifically degrades RsmZ (Petrova and Sauer 2010). Levels of RsmA and RsmE are also regulated by complicated networks involving two component sensors, sRNAs, and QS systems. These mechanisms further contribute to the complexity of regulation of phz gene expression, but their ultimate effects on phenazine production have not been measured.

The two component system CbrA/CbrB controls expression of phzM via a mechanism analogous to the network linking GacS/GacA to phz operon expression. When CbrB is activated, it induces expression of CrcZ, an sRNA that binds to and sequesters the translational repressor Crc. The CrcZ and phzM transcripts both contain an A-rich motif that is recognized by Crc. CrcZ therefore limits the ability of Crc to inhibit phzM translation, and a crc mutant overproduces PYO due to increased PhzM levels (Huang et al. 2012).

In contrast to RsmX/Y/Z and CrcZ, which are controlled at the transcriptional level by two component systems, the sRNAs Lrs1 and Lrs2 are regulated by QS (Wurtzel et al. 2012). Using P. aeruginosa PA14 as a model strain, Wurtzel et al. identified las boxes in the promoter regions of the lrs1 and lrs2 genes and confirmed their regulation by LasR. In addition, they generated an lrs1 deletion mutant and found that it was defective in PYO production. RNA-seq analysis revealed two major differences in transcript levels between this mutant and the wild-type parent: increased abundance of transcript from the antABC operon, and increased abundance of the PrrF1 and PrrF2 sRNAs (discussed further below). The authors hypothesized that the PYO production defect in the lrs1 mutant arose from increased flux through an anthranilate-catechol conversation pathway (mediated by the products of the antABC operon). Anthranilate and phenazines are produced by pathways that branch from chorismate as a common precursor (Mentel et al. 2009). Notably, increased conversion of anthranilate to catechol also diverts it away from the quinolone biosynthetic pathway. Given that quinolones regulate phz operon expression, indirect Lrs1-dependent downregulation of anthranilate degradation may be important for wild-type levels of quinolone, and therefore phenazine production.

The sRNAs PrrF1 and PrrF2 have been characterized in further detail in P. aeruginosa PAO1. PrrF1 and PrrF2 are encoded by adjacent loci and repressed by the iron-dependent regulator Fur when iron is abundant (Wilderman et al. 2004; Oglesby et al. 2008). They are expressed during iron limitation and base-pair with target mRNAs, preventing their translation. One such target is the transcript of sodB, which encodes superoxide dismutase and is, for unknown reasons, required for PYO production in P. aeruginosa PAO1 (Hassett et al. 1995). Also in this strain, a prrF1/prrF2 mutant shows increased expression of the antABC operon, an effect that seems to contradict the simultaneous upregulation of antABC and prrF1/prrF2 transcripts in the lrs1 mutant of strain PA14 (Wurtzel et al. 2012). This may represent a strain-dependent difference in this branch of the P. aeruginosa sRNA-dependent regulatory network.

Expression of PhrS, an sRNA that positively controls translation of pqsR mRNA, is also controlled by a regulator that responds to an environmental cue: the oxygen-sensitive transcription factor ANR. In the absence of PhrS, pqsR mRNA adopts an intramolecular secondary structure in which an upstream open reading frame base-pairs with the pqsR transcript and inhibits translation. Under oxygen-limited conditions, ANR is activated and induces expression of PhrS. PhrS competes with the pqsR transcript for binding of its 5′ untranslated region, and via this anti-antisense mechanism, exposes the pqsR mRNA to allow for ribosome binding and translation. This regulatory cascade was elucidated in P. aeruginosa PAO1, where a PhrS-overexpressing strain shows increased PYO production due to elevated PqsR levels and quinolone production (Sonnleitner et al. 2011).

Hfq, an abundant mRNA-binding protein found in diverse bacteria, also affects phenazine production through post-transcriptional mechanisms. In P. aeruginosa M18, Hfq binds qscR and phzM mRNA transcripts via AU-rich sequences present in their 5′-leader sequences and inhibits their translation (Wang et al. 2012b). As qscR is a negative regulator of the phz operons, and phzM is required for the conversion of PCA to PYO, Hfq would be expected to enhance phenazine production overall but limit PYO production. In mutants lacking functional Hfq, Wang et al. observed increased PYO production and decreased PCA production, consistent with decreased production of the PhzA-G biosynthetic enzymes, but increased translation of phzM. Formation of the active, hexameric form of Hfq is promoted by the RelA enzyme, a critical regulator of the stringent response to amino acid starvation (Argaman et al. 2012). P. aeruginosa PAO1 relA mutants also overproduce PYO, suggesting that Hfq may regulate translation of the phzM transcript in this strain according to a mechanism similar to the one described for M18 (Erickson et al. 2004).

2.3.4 Environmental Signals and Conditions Affecting phz Gene Expression

Many studies characterizing the conditional dependence of phenazine production have revealed environmental cues that affect the regulation of this process and, in some cases, mechanisms linking the condition to the response. These studies have evaluated the effects of environmental parameters such as temperature, pH, salinity, and oxygen availability. They have also examined how phenazine production is influenced by the availability of carbon and nitrogen sources, phosphate, sulfate, iron, and magnesium. These environmental variables can affect phenazine production by indirectly or directly altering expression of Phz proteins (for example, through their effects on the production of signals upstream in the regulatory cascade (van Rij et al. 2004; Farrow and Pesci 2007)), or they can alter the availability of substrates and thus, flux through the relevant metabolic pathways that support phenazine biosynthesis.

The effect of temperature on phenazine production has been investigated in P. chlororaphis PCL1391, P. fluorescens 2-79, and multiple strains of P. aeruginosa. P. chlororaphis PCL1391 produces the PCA derivative phenazine-1-carboxamide (PCN) at comparable levels when grown at temperatures ranging from 21 to 31 °C, but production is almost undetectable when it is grown at 16 °C (van Rij et al. 2004). In P. fluorescens 2-79, PCA production was found to inversely correlate with temperature in a survey of temperatures ranging from 25 to 37 °C (Slininger and Shea-Wilbur 1995). In P. aeruginosa M18, transcription of phz1 and phz2 is elevated at 28 °C compared to 37 °C, and this correlates with a large increase in PCA production (Huang et al. 2009). In P. aeruginosa PA14, PYO production increases modestly when this strain is grown at 37 °C compared to 28 °C. Using RNA-seq, Wurtzel et al. (2012) found that the transcript abundances of both phz1 and phz2 are elevated at the higher temperature, with a larger effect on phz1 than phz2. These results also indicated the presence of a temperature-dependent transcriptional start site upstream of phzB1. The differential regulation of phzA1 and phzB1 is interesting because these two genes encode highly similar proteins that form heterodimers required for in vivo formation of the phenazine core. Temperature-dependent differences in expression may have consequences for PhzA/B dimerization (Ahuja et al. 2008).

Ambient oxygen levels also influence the production of different phenazine derivatives. In P. aeruginosa, PCA can be biosynthesized anaerobically (Dietrich et al. 2006; Mentel et al. 2009; Recinos et al. 2012). However, oxygen is required for the conversion of 5-methylphenazine-1-carboxylic acid (the product of PhzM, 5-MCA) to PYO by the PhzS monooxygenase. Therefore, PYO production is inhibited in the absence of oxygen. Interestingly, Holliman (1969) reported increased production of the red phenazines aeruginosin A and B in low oxygen conditions; inefficient conversion of 5-MCA to PYO may shunt the biosynthetic pathway toward the production of these alternative phenazines when oxygen is limited. An effect of oxygen limitation on phenazine biosynthesis has also been observed in P. chlororaphis PCL1391, where growth in low oxygen conditions leads to PCN overproduction (van Rij et al. 2004).

The effects of pH and salinity on phenazine production have been tested in biocontrol strains, where optimization of soil conditions could facilitate the application of such strains for crop growth promotion. P. chlororaphis PCL1391 produces PCN when grown at pH 7 or pH 8, but not at pH 6 (van Rij et al. 2004). For P. fluorescens 2-79, however, PCA production was maximized at pH 7, partially decreased but still substantial at pH 6, and abolished at pH 8 (Slininger and Shea-Wilbur 1995). Increasing concentrations of salts decreased PCN production in P. chlororaphis PCL1391, but this effect was specific to ionic solutes as xylose did not affect PCN production when introduced at isoosmotic levels, and osmoprotectants did not restore PCN production in a high-salt medium.

Variations in the availability of minerals and the compounds that provide the major elements for biomass can have dramatic effects on phenazine biosynthesis. In a survey of carbon sources for growth of P. chlororaphis PCL1391, van Rij et al. (2004) found that glucose, glycerol, and L-pyroglutamic acid gave rise to the highest levels of PCN production. The amount of PCN produced did not correlate with growth rate, and the most stimulatory carbon sources were not the most abundant organic compounds in the rhizosphere, where the organism is commonly found. Glucose and glycerol have also been found to stimulate PCA production in P. fluorescens 2-79. That glucose and glycerol promote the highest levels of phenazine production is surprising because they are not preferred carbon sources for pseudomonads; unlike E. coli, Pseudomonas species typically utilize organic acids such as succinate before utilizing sugars (Behrends et al. 2009; Rojo 2010; Valentini and Lapouge 2012).

Given that phenazine structures, and particularly that of PCN, contain multiple nitrogen atoms, one would predict that the type of nitrogen source provided would affect phenazine production. Generally, supplementation with amino acids stimulates phenazine production, but the effects of individual amino acids and inorganic nitrogen sources on phenazine production vary widely between species and conditions. Increasing levels of nitrogen provided as ammonium sulfate stimulated PCN production in P. chlororaphis PCL1391, but did not stimulate PCA production in P. fluorescens. Although glutamine is used to form the carboxamide functional group in PCN, the addition of this amino acid to the medium did not stimulate PCN production any more than other individual amino acids such as leucine. All aromatic amino acids stimulate PCN production in P. chlororaphis PCL1391, whereas the effects of phenylalanine, tryptophan, and tyrosine on PYO production in P. aeruginosa appear to be strain- and condition-dependent (Burton et al. 1947, Palmer et al. 2007). The effect of tryptophan in particular is at least partially related to its ability to serve as a precursor for quinolone biosynthesis (Farrow and Pesci 2007).

In both P. aeruginosa and P. chlororaphis PCL1391, PYO and PCN production, respectively, are maximized when the medium contains an intermediate level of phosphate; this is apparently not an artifact of effects on growth (Burton et al. 1947; van Rij et al. 2004). Iron and magnesium supplementation at micromolar levels is required and optimal for growth and phenazine production by P. aeruginosa and P. chlororaphis PCL1391. Because iron and magnesium are often provided as sulfate salts, it can be difficult to decouple their effects from that of varying the sulfur source. The importance of sulfate has been thoroughly evaluated in P. chlororaphis PCL1391, however, where millimolar concentrations are required for maximum production of PCN (van Rij et al. 2004).

2.4 Regulation of Phenazine Biosynthesis in Other Genera

In addition to the Pseudomonas species we have discussed, many diverse species belonging to other genera also produce phenazines with highly derivatized chemical structures (Table 2.1). Relatively little is known about the regulation of phenazine biosynthesis in these species, but recent studies have identified regulators that affect the process in strains of Burkholderia and Streptomyces (Ramos et al. 2010; Saleh et al. 2012). In Burkholderia cenocepacia K56-2, wild-type phenazine production requires a regulator called phenazine biosynthesis regulator (Pbr), which is encoded by a gene that lies near phzF and phzD homologues on the chromosome (Ramos et al. 2010). Pbr binds to the promoter region of the phzF-phzD operon and is required for wild-type expression. In Streptomyces anulatus 9663, regulators of phz gene expression have been identified through characterization of a large gene cluster that includes all of the genes required for PCA biosynthesis and genes required for transformation of PCA to the prenylated phenazines endophenazines A and E (Saleh et al. 2012). One of these regulators, encoded by the gene ppzV, is similar to a putative TetR-family regulator called EpzV found in S. cinnamonensis, another phenazine-producer. Inactivating the ppzV gene in a strain expressing the large phenazine biosynthetic cluster led to loss of the ability to produce prenylated phenazines but an increase in the amount of unprenylated phenazines, suggesting that the ppzV product regulates PCA derivatization. The second regulator, encoded by ppzY, is similar to transcriptional regulators of the LysR family. Inactivation of ppzY led to a nearly complete defect in all phenazine production, suggesting that the ppzY product is required for expression of PCA biosynthetic genes in S. anulatus 9663.

2.5 Conclusion

Characterization of the regulation of phenazine biosynthesis in diverse Pseudomonas isolates has revealed common mechanisms and hierarchies. As more of the mechanisms regulating phenazine biosynthesis in other genera are uncovered, it will be interesting to compare them to the Pseudomonas paradigm and evaluate their physiological relevance in these new contexts. The intricacy of the networks controlling phenazine production in Pseudomonas is becoming clear at a time when phenazines themselves are gaining recognition for their roles in bacterial physiology, which include intercellular signaling and redox balancing. The multilayered cascades that modulate phenazine biosynthesis are consistent with their new status as primary players in cellular metabolism and communication. Indications that not just the core genes for PCA synthesis, but also the genes for PCA modification, are regulated at multiple levels may suggest that different phenazines perform different physiological roles, consistent with their unique chemistries.

Although our understanding of the complicated networks controlling phenazine production is still developing, a hint at this complexity has long been evident in the variability of phenazine production that is apparent among species, isolates, and even repeat cultivations of the same strain. Differences in phenazine production among strains of the same species likely arise in part from subtle discrepancies in regulatory networks and sensing mechanisms. On the other hand, differences between repeat experiments imply that, although many of the conditions and regulators that affect phenazine production have been identified, unrecognized variables can still alter phenazine production in unpredictable ways. Elucidating the parameters and mechanisms that affect this process has the potential to facilitate the use of beneficial phenazine-producing pseudomonads in agriculture, support the development of therapeutics for patients suffering from P. aeruginosa infections, and allow us to learn new techniques for controlling antibiotic production in diverse species.