Keywords

1 Overview and Relevance of Pseudomonas aeruginosa

Pseudomonas aeruginosa is a Gram-negative opportunistic pathogen that causes devastating diseases in immunocompromised individuals. P. aeruginosa is currently one of the most common causes of nosocomial infection, and is frequently isolated from chronic pulmonary, wound, and medical device-associated infections [1]. The incidence of multidrug resistance in P. aeruginosa has increased drastically in recent years, and P. aeruginosa is now ranked by the World Health Organization (WHO) as one of the most serious threats to human health [2].

Aside from its ability to establish infections, P. aeruginosa also thrives in marine and soil environments. Given the remarkable range of environments P. aeruginosa inhabits, it is no surprise that this bacterium has one of the largest bacterial genomes (up to 7.1 Mbp). Approximately 9% of its genes are predicted to function as transcriptional regulators or as part of two-component systems [3, 4]. These features emphasize the adaptability of P. aeruginosa, to respond to different environmental conditions. Perhaps the most prominent of these survival mechanisms is the formation of protective biofilm communities, which requires a complex coordination of regulatory mechanisms to transition from a planktonic to biofilm lifestyle. During biofilm formation, P. aeruginosa encases itself within a matrix composed of exopolysaccharides, extracellular DNA, and proteins. The biofilm enhances bacterial persistence, and P. aeruginosa becomes more tolerant toward environmental stress, antimicrobial treatment, and the host immune system [5, 6]. The exopolysaccharides alginate, Psl, and Pel are major structural components of the biofilm matrix and contribute to antimicrobial tolerance and defense against the host immune system [7,8,9]. Thus, biofilms have important clinical implications for chronic infection and antimicrobial tolerance, and it is apparent that understanding this bacterial process is essential for controlling infection.

It is now well-established that the transition from a motile to sessile biofilm lifestyle is primarily driven by modulating levels of the secondary messenger molecule cyclic dimeric (3′–5′) GMP (cyclic di-GMP) [10, 11]. High intracellular cyclic di-GMP concentrations generally activate production of various adhesins and biofilm matrix components while simultaneously downregulating motility [11]. Conversely, low levels of cyclic di-GMP promote biofilm dispersal and motility function [11]. The levels of cyclic di-GMP in the cell are controlled by the competing activity of two classes of enzymes known as diguanylate cyclases (DGCs) and phosphodiesterases (PDEs). DGCs are identified by a conserved GGDEF domain and synthesize cyclic di-GMP from two molecules of GTP. PDEs can be identified by an EAL or HD-GYP domain and degrade cyclic di-GMP into pGpG or two molecules of GMP, respectively [10]. Furthermore, many of these enzymes contain both DGC and PDE associated domains; however, typically only one of these domains exhibits catalytic activity. P. aeruginosa is predicted to produce up to 38 proteins (17 GGDEF, 5 EAL, and 16 GGDEF/EAL) involved in the metabolism of cyclic di-GMP depending on the strain [12]. Thus, a complex regulatory network must be in place to control cyclic di-GMP production in response to environmental signals. This chapter will discuss the regulation and function of the currently identified P. aeruginosa DGCs and PDEs, which are summarized in Table 28.1.

Table 28.1 Identified P. aeruginosa DGCs and PDEs
Full size table

2 Principals of Cyclic di-GMP Regulation and Signaling Inputs

The transition between a motile and biofilm lifestyle requires the concerted regulation of many systems, and cyclic di-GMP serves as a master regulator of this process (Fig. 28.1). To exert its function, cyclic di-GMP binds to a protein or RNA, and the resulting conformational change alters activity. Depending on the target, cyclic di-GMP is capable of regulating at both a transcriptional or posttranslational level. Thus, cyclic di-GMP regulation is predominately controlled by modifying its concentration within the cell. One way this occurs is by activation or deactivation of DGCs and PDEs by a regulatory sensor, and in many P. aeruginosa cyclic di-GMP systems such sensors harbor an N-terminal domain that perceives signals. In these systems, signal transduction often occurs through a phosphotransfer event between the sensor histidine kinase and the response regulator, leading to its activation. Through this mechanism, it is possible for P. aeruginosa to modify DGC and PDE activity in response to environmental signals recognized by the sensor kinase. In Gram-negative bacteria, conserved signaling domains respond to a variety of factors including oxygen, phosphorylation, nutrients, antibiotics, protein/cyclic nucleotide binding, and light [43, 44]. In the context of cyclic di-GMP, multiple regulatory components with Per-Arnt-Sim (PAS) and receiver (REC) sites have been identified which respond to oxygen, surface attachment, and nutrient conditions (Table 28.1) [26, 28, 44]. This suggests environmental signals are crucial for controlling DGC and PDE activity depending on whether conditions are preferable for biofilm or planktonic growth.

Fig. 28.1
figure 1

Overview of cyclic di-GMP activity in P. aeruginosa. DGCs synthesize cyclic di-GMP from two molecules of GTP, while PDEs degrade cyclic di-GMP into pGpG or GMP. DGCs can be self-regulated at an allosteric I-site that binds to cyclic di-GMP to inhibit function. Once produced, cyclic di-GMP can regulate many bacterial processes by directly binding to RNA or proteins. Conformational changes in response to cyclic di-GMP binding regulate processes at the transcriptional, translational, and posttranslational level

Full size image

Additionally, DGCs and PDEs frequently contain either a degenerate or nonfunctional GGDEF or EAL site, or an allosteric RxxD inhibitory site (I-site) [44]. These degenerate domains are enzymatically inactive but retain binding affinity. These serve as allosteric inhibitor or activator sites, where cyclic di-GMP or GTP can bind, leading to a conformational change that alters enzyme activity. P. aeruginosa encodes 16 proteins with predicted dual DGC and PDE domains, and often one of these sites is degenerate [12, 44]. These I-sites serve as an efficient self-regulatory mechanism, allowing these enzymes to control their own activity based on local concentrations of cyclic di-GMP and GTP.

3 P. aeruginosa Cyclic di-GMP Synthesis, Degradation, and Regulation

Currently, the function and regulation of only a few of the predicted P. aeruginosa DGC and PDE enzymes have been determined (Table 28.1). These proteins have a number of critical regulatory functions, and their identification has vastly improved our understanding of motility, biofilm formation, and dispersion. The following sections will summarize the function and regulation of the currently identified P. aeruginosa DGCs and PDEs.

3.1 Cell Motility

The regulation of motility is a crucial step in controlling the transition between a motile and biofilm lifestyle, and cyclic di-GMP has an important role in regulating flagella-mediated swarming motility and type IV pili-mediated twitching motility. One system involved in controlling swarming motility is the Gac/Rsm network. This system is a complex regulatory system involving multiple sensor kinases and more than 500 downstream targets [45, 46] (Fig. 28.2a). The activity of this system is controlled by two sensor kinase proteins, LadS and RetS [47, 48]. RetS recognition of an unknown carbohydrate signal represses the sensor kinase GacS [49]. Conversely, LadS calcium signaling activates GacS [50]. Once activated, GacS-mediated phosphorylation of GacA results in the production of two noncoding small RNAs, rsmY and rsmZ. These factors bind and sequester the translational repressor RsmA [51]. Inhibition of RsmA ultimately results in reduced motility and virulence factor production, but increased production of biofilm factors, including Psl and Pel [51, 52]. GacS-mediated inhibition of RsmA promotes translation of three DGCs, SiaD, GcbA (AdcA), and SadC [23, 53, 54]. SiaD activity promotes biofilm matrix production [18], while GcbA and SadC activity promotes attachment and downregulates swimming and swarming motility, respectively [19, 23]. Thus, these DGCs contribute toward biofilm formation.

Fig. 28.2
figure 2

Pseudomonas aeruginosa regulation of cyclic di-GMP. Schematic of three of the main cyclic di-GMP regulating pathways in P. aeruginosa. Green arrows indicate activation, while red bars indicate inhibition. (a) GacS activity is promoted by LadS and inhibited by RetS interaction. GacS activates GacA, which leads to the production of the sRNAs rsmY and rsmZ. These sRNAs bind and inhibit RsmA translational regulation, allowing for unimpeded translation of the DGCs, GcbA, SiaD, and SadC. DGC activity results in increased exopolysaccharide production and attachment, as indicated. AmrZ inhibits GcbA transcription, and PilY1 promotes GcbA activity. (b) WspC and WspF regulate WspA activity through transfer of a methyl group. In response methylation state and surface signals, WspA modulates WspE autophosphorylation activity, which in turn phosphorylates and activates the DGC, WspR. Activation of WspR leads to cyclic di-GMP production and increased production of the exopolysaccharides Pel and Psl [13,14,15, 17]. (c) RocS1 and RocS2 phosphorylate multiple response regulators in response to unknown environmental signals. Phosphorylation and activation of the PDE RocR results in reduced levels of cyclic di-GMP and CupC fimbriae production leading to reduced biofilm formation

Full size image

Furthermore, GcbA and SadC are regulated by additional factors independent of the Gac/Rsm network. The transcriptional factor AmrZ regulates a number of virulence factors, and it directly inhibits transcription of GcbA, resulting in reduced cyclic di-GMP levels and increased motility [55]. SadC activity is further promoted by PilY1, which is part of the type IV pili (T4P) machinery [56]. T4P is necessary for twitching motility and biofilm maturation in P. aeruginosa [40, 57]. PilY1 is an important adhesion protein, containing a conserved von Willebrand A mechanosensory domain, which indicates it may have a signaling function in response to surface sensing [20]. Since PilY1 production is regulated by the FimS-AlgR two-component system and Pil-Chp chemotaxis complex, an intricate regulatory network involving interactions between all these systems is likely involved in modulating SadC activity [20]. While the details of this regulatory network remain to be elucidated, it is clear that SadC and PilY1 interaction enhances surface attachment, while inhibiting swarming motility during biofilm formation [20, 56].

Another role for cyclic di-GMP in T4P regulation has been reported. The PDE FimX promotes localization of T4P components to the cell pole, which is a critical step in pili formation. FimX contains both a GGDEF and EAL domain, but it appears to exhibit only PDE activity [40,41,42]. While PDE activity was required for T4P function, mutation of the nonfunctional GGDEF domain also inhibited T4P localization [58]. This could indicate the GGDEF site functions as an allosteric activation site that promotes PDE activity. Interestingly, the necessity of FimX for T4P production is limited to low cyclic di-GMP conditions, and fimX deletion mutants still produce T4P in high cyclic di-GMP conditions [40].

3.2 Biofilm Matrix Production

The Wsp (wrinkly spreader) chemosensory system is one of the most well-defined cyclic di-GMP regulatory systems in P. aeruginosa. It has an active role in regulating swimming motility and the production of biofilm matrix components [13, 59,60,61]. This system is encoded by the wspABCDEFR (PA3708-PA3702) operon, and its components closely resemble the enteric chemotaxis signal transduction system (Che) (Fig. 28.2b). WspA functions as a membrane-bound methyl-accepting chemotaxis protein (MCP), which responds to surface contact and environmental stresses that alter membrane fatty acid composition, such as high NaCl concentrations [14,15,16]. WspA is activated by the transfer of a methyl group via the methyltransferase WspC. Conversely, WspA is deactivated by the methylesterase WspF, which removes this methyl group [13]. WspB and WspD act as linkers to connect the sensor histidine kinase, WspE, to WspA. Activation of WspA results in a conformational change promoting WspE autophosphorylation. WspE is then able to phosphorylate and activate two response regulators, WspF and WspR [62]. Phosphorylation of WspF promotes methylesterase activity leading to WspA deactivation, while phosphorylation of the DGC WspR stimulates cyclic di-GMP production. WspR activity is further regulated through its oligomerization. At high concentrations of cyclic di-GMP, the molecule binds to the WspR RxxD allosteric I-site, which results in a conformational change that promotes the assembly of enzymatically inactive WspR tetramers, thus limiting its DGC activity [17].

Loss-of-function mutations inhibiting the regulatory activity of WspF result in constitutive WspR activity, providing insight into the downstream targets of WspR activity [13, 15, 60]. These studies suggest WspR activity promotes biofilm production and persistence by increasing Pel, Psl, and CdrA production [61, 63, 64]. The transcriptional regulator, FleQ, directly binds to both the psl and pel promoters [65, 66]. FleQ promoter binding represses pel transcription. In the presence of cyclic di-GMP, the molecule binds FleQ, leading to a conformational change, which elevates pel transcription due to derepression [65, 66]. While FleQ binding to the psl promoter was reported, a regulatory role for FleQ has not yet been directly demonstrated [65]. However, considering cyclic di-GMP is required to activate Psl production, it is likely FleQ regulates Psl similarly. Additionally, cyclic di-GMP promotes Pel production posttranslationally by directly binding to PelD, a component of the synthesis complex [67]. While it is unknown if WspR DGC activity is involved in alginate production, cyclic di-GMP similarly promotes alginate production by binding and activating the synthesis protein Alg44 [68].

Another DGC that promotes exopolysaccharide production is the DGC YfiN, which induces Pel and Psl production [69]. YfiN is embedded within the inner membrane, and is part of the YfiBNR three-component system. YfiN activity is regulated by interaction with YfiR, which binds to the PAS periplasmic domain of YfiN. Interaction with YfiR causes a conformational change of YfiN that disrupts its DGC activity. YfiB is located within the outer membrane and binds and sequesters YfiR, allowing for YfiN activity and cyclic di-GMP production [25]. This system is activated by the detergent SDS or high concentrations of Na-succinate, indicating that membrane stress promotes YfiN activity [25]. Combined with the Wsp system, cyclic di-GMP may be produced by multiple DGCs in response to membrane stress. However, whether YfiN activity is directly integrated with the Wsp system is unclear.

Lastly, two other DGCs are also involved in biofilm matrix production. RoeA activity promotes Pel production, likely through the interaction of cyclic di-GMP with FleQ and PelD [21]. MorA contains both active DGC and PDE domains and is capable of synthesizing and degrading cyclic di-GMP [29]. Upregulation of MorA was observed in a clinical P. aeruginosa isolate from a chronic lung infection, which indicates it may have an important role in bacterial persistence [30]. Deletion of morA in Pseudomonas putida enhanced motility and reduced biofilm formation. However, morA deletion in P. aeruginosa only resulted in decreased biofilm formation, with no effect on motility [70]. While this indicates MorA DGC activity may be important for biofilm formation in P. aeruginosa, the function of its PDE activity remains unclear.

3.3 Biofilm Dispersal

Biofilm dispersion is often an active bacterial process, triggered in mature biofilms by poor nutrient and anoxic conditions within the biofilm [71, 72]. Therefore, this process encourages the bacteria to seek nutrient-rich environments better suited for survival. The initiation of biofilm dispersal has been linked to low levels of cyclic di-GMP and the activity of a number of PDEs, including MucR, NbdA, RbdA, and DipA [31, 33, 38]. MucR and NbdA are anchored to the inner membrane and each contains a GGDEF and EAL domain. MucR exhibits both PDE and DGC activity, while NbdA only possesses PDE activity. However, deletion of either gene inhibits biofilm dispersion [31]. Nitric oxide (NO) accumulates in the biofilm during anaerobic growth, and it is a major stimulator of P. aeruginosa biofilm dispersal [72]. NbdA and MucR contain an MHYT domain, which is predicted to recognize diatomic gases including NO. NbdA-mediated biofilm dispersion was found to be specifically tied to NO signaling, and increased transcription of NbdA was observed following NO exposure [31]. Similarly, the PDE RbdA also promotes biofilm dispersal [73]. This protein contains both GGDEF and EAL domains; however, only exhibits PDE activity [73]. RbdA activity is regulated by the GGDEF domain, which acts as an allosteric activation site when bound to GTP [38, 73]. RbdA also contains a PAS domain and disruption of this domain inhibits biofilm dispersal, indicating that RbdA may be important for sensing anoxic conditions and promoting dispersal via PDE activity [73].

DipA (Pch) is another PDE that promotes biofilm dispersion, flagella-mediated motility, and chemotaxis in response to nutrient and environmental conditions, including glutamate, NO, ammonium chloride, and mercury chloride [33, 74]. DipA localizes to the flagellated cell pole leading to asymmetrical PDE distribution following cell division. This results in heterogeneous subpopulations that exhibit a range of motility phenotypes, which is predicted to promote bacterial survival in dynamic environments [41]. As an additional level of control, DipA activity is regulated by the inner membrane-bound DGC, NicD, and the chemotaxis protein, BdlA. In the presence of extracellular glutamate, NicD is dephosphorylated leading to increased DGC activity [28]. BdlA is subsequently phosphorylated and activated by NicD-produced cyclic di-GMP and nonprocessive proteolysis via ClpP and GcbA [34, 39, 75]. Once activated, BdlA promotes DipA PDE activity resulting in biofilm dispersion. Thus, NicD and GcbA appears to temporarily increase cyclic di-GMP in response to extracellular nutrients, which leads to DipA PDE activation and biofilm dispersal through reducing cyclic di-GMP levels. As mentioned previously, GcbA DGC activity also promotes surface attachment by inhibiting flagella-mediated motility. These seemingly opposing dual functions for GcbA further highlight the intricate regulatory system required to transition between sessile and motile lifestyles.

The RocSAR system in PAO1 and SadARS in PA14 are homologous three-component systems with multiple response regulators (Fig. 28.2c). This system regulates the production of cup fimbriae adhesins and the Type III secretion system (T3SS) [35, 36]. One of the response regulators in each of these systems (RocR/SadR) contains an active EAL domain [76]. RocS1 functions as the sensor kinase, which has a domain associated with solute-binding proteins (Sbp3) and a PAS domain in the periplasm [35]. The presence of a PAS domain suggests oxygen or redox signaling activates the phosphorelay in this system [37]. A paralog of RocS1, named RocS2, has also been identified and functions similarly to RocS1 [77]. RocS1 and RocS2 activate RocR, RocA1, and RocA2 via phosphorylation at the histidine-containing phosphotransfer (HPt) domain [35, 76]. Activation of RocA1 results in increased Cup fimbriae formation and inhibited T3SS gene expression, while RocA2 phosphorylation leads to decreased mexAB-oprM-dependent efflux activity [36, 77]. RocR contains an active PDE domain and functions antagonistically on the Roc system [76, 77]. RocR activation by RocS1 and RocS2 inhibits cupC fimbriae expression and attachment, likely due to reduced cyclic di-GMP levels.

4 Conclusions and Moving Forward

Our understanding of the regulation and function of cyclic di-GMP in P. aeruginosa has improved considerably since its original discovery in 1987 [78]. As discussed here, a significant amount of work has gone into identifying cyclic di-GMP regulators, and the effect these systems have on the cell. However, considering that the P. aeruginosa genome encodes 38 proteins with conserved GGDEF and EAL domains, many proteins involved in the synthesis and degradation of cyclic di-GMP remain to be discovered. Moreover, there remains a large disparity in the number of DGC/PDEs compared to cyclic di-GMP effectors. To date, only four cyclic di-GMP effectors that bind cyclic di-GMP have been identified, Alg44, PelD, FimX, and FleQ [42, 65, 67, 68]. The role of these effectors is related to the transition between motile and sessile lifestyles. However, a recent study determined that the DGC PA3177 promotes antibiotic tolerance independent of biofilm formation [79]. These findings suggest that a broader role for cyclic di-GMP effectors may exist for P. aeruginosa, and given the quantity of DGCs and PDEs, it is likely that many cyclic di-GMPs effectors are still unidentified. Two studies have attempted to address this, and using cyclic di-GMP binding affinity assays have identified a number of predicted possible effectors [22, 80]. Future studies that verify and identify the function of cyclic di-GMP effectors will be necessary to fully understand the extent of cyclic di-GMP regulation in P. aeruginosa.

Another major hurdle in understanding cyclic di-GMP regulation is related to signal specificity. Cyclic di-GMP controls a range of processes which effectively share the same signaling molecule, and it is unclear how a specific process is targeted by cyclic di-GMP produced by any given DGC. In the simplest model, cyclic di-GMP could be thought of as a global pool within the cell. Here, all active DGCs and PDEs would contribute to raising or lowering the total pool of cyclic di-GMP, which in turn alters cyclic di-GMP effector activity. The relative activity of a given cyclic di-GMP target could then be further controlled by differences in binding affinity among individual effectors. In this system, reaching various cyclic di-GMP concentrations would result in either motility or biofilm phenotypes. A global pool model could explain the observed redundancy among DGC effects, such as WspR and YfiN, which both upregulate Pel and Psl production [25, 64]. Further evidence supporting a global pool model is observed in regards to T4P formation. As described previously, T4P assembly at low cyclic di-GMP concentrations requires functional FimX PDE activity, but high levels of cyclic di-GMP bypass the requirement of FimX [40]. Jain and colleagues (2012) also report that FimX could be bypassed in P. aeruginosa by overexpression of an unrelated DGC isolated from Caulobacter crescentus. This suggests the source of cyclic di-GMP may be irrelevant in this system and high global levels of cyclic di-GMP are sufficient to stimulate T4P production. Contrary to this model, there is also evidence suggesting cyclic di-GMP signaling is highly specific to a particular target. In a local pool model, DGCs and PDEs are trafficked to the vicinity of their target effector. This would allow for the creation of local cyclic di-GMP pools that act only on their nearby effectors. Merritt and colleagues (2010) report that the activity of two DGCs, RoeA and SadC, have different and distinct effects on biofilm formation, despite producing similar total levels of cyclic di-GMP [21]. Furthermore, SadC and RoeA appear to be differentially distributed throughout the cell. While both proteins are localized to the inner membrane, SadC forms distinct foci around the cell periphery, and RoeA forms a diffuse patchy background. Differing localization patterns could explain why SadC controls swarming motility, and RoeA controls Pel production, despite producing similar levels of cyclic di-GMP [20]. It is possible that a combination of these models function in P. aeruginosa, depending on the effector, but future studies will be necessary to fully understand cyclic di-GMP signaling.

In conclusion, we have made great strides in understanding cyclic di-GMP regulation, but many of the details regarding its function and regulation remain elusive. Cyclic di-GMP controls a number of virulence factors associated with chronic infection, and the continued study of cyclic di-GMP in P. aeruginosa will provide important insights into biofilm formation, development, and dispersal.