For transcriptional regulation, the selectivity of RNA polymerase is controlled by interaction with two types of regulatory proteins: sigma factors and transcription factors (Ishihama, 2012). The utilization of sigma factors alternative to the housekeeping σ70 provides a mechanism for bacterial responses to many stresses, by redirecting transcription initiation to the simultaneous regulation of large numbers of genes. Among these alternative sigma factors, RpoS increases under conditions of nutrient stress, or during the stationary phase, leading to general stress resistance (Battesti et al., 2011). PHB accumulation and mobilization are related to the nutrient status of bacteria (López et al., 2015), so RpoS has a role in the regulation of PHAs metabolism in some bacteria (Brigham et al., 2012).

In R. eutropha, inactivation of rpoS induced no change during PHB production, however, it exhibited an increased PHB depolymerization when polymer utilization was induced, in comparison to the wild-type strain H16. Therefore, RpoS was proposed to have a role in PHB utilization (Brigham et al., 2012).

A similar role was found in P. putida (Raiger-Iustman and Ruiz, 2008). In this organism inactivation of rpoS had no effect in PHAs accumulation, however, when entering the stationary phase, the polymer content of the mutant diminished faster than in the wild type, resulting in a lower PHAs accumulation. The transcript containing phaC1 and phaZ (coding for one of the two PHA synthases present in this bacterium and the PHA depolymerase) was overexpressed in the stationary phase, probably causing a higher depolymerization activity. The authors concluded that RpoS participates in the negative regulation of the promoter of phaC1-phaZ, probably in an indirect manner and suggested that PhaD, a transcriptional regulator of the TetR family, could be an intermediary.

In Pseudomonas aeruginosa, Choi et al. (2011) reported that rpoS inactivation in PA14 and PAO1 strains showed a strong negative effect on PHAs accumulation when grown on fructose, however, on decanoic acid, only the PA14 mutant showed a similar negative effect. These results suggest RpoS sigma factor participates in the transcription of some PHAs biosynthetic genes with differences depending on the strain. It is interesting to note that, besides the observed effect on the amount of polymer accumulated, inactivation of rpoS shifted the monomeric composition of the PHAs to longer hydroxyalkanoates (Choi et al., 2011). PHAs depolymerization has also been linked to rpoS expression and to tolerance to oxidative and thermal stresses, through an unknown mechanism (Ruiz et al., 2004). This could be related to the reported need of PHB utilization for ATP and ppGpp synthesis (Ruiz et al., 2001), or as part of the role that PHAs metabolism has in stress resistance in several bacteria (López et al., 2015).

The role of sigma factor RpoS has also been elucidated in A. vinelandii UW136. In this bacterium, the synthesis of PHB occurs mainly during the stationary growth phase. Accordingly, RpoS was shown to participate in the transcription of the PHB biosynthetic operon phbBAC and of phbR, which codes for the transcriptional activator of the phb genes (Hernández-Eligio et al., 2011).

In Pseudomonas chlororaphis PA23, inactivation of rpoS negatively affected the amount of PHAs, due to down-regulation of phaC1, phaC2, and the two phasin genes phaF and phaI (Sharma et al., 2018).

Because the alternative sigma factor RpoN (sigma 54) is involved in the regulation of nitrogen metabolism and nitrogen limitation induces PHAs biosynthesis, the participation of this sigma factor on the accumulation of these polymers was evaluated in P. putida KT2440 and in P. aeruginosa PAK by Hoffmann and Rehm (2004, 2005). They studied PHAs accumulation in the wild type and rpoN-negative mutant strains of these bacteria under nitrogen excess or nitrogen limitation. In P. aeruginosa, inactivation of rpoN abrogated PHAs accumulation on gluconate and octanoate, suggesting that RpoN participates in the expression of a gene involved in PHAs metabolism. The expression of transacylase PhaG (now reclassified as a 3-hydroxyacyl-ACP tioesterase), needed for the biosynthesis of PHAs from carbohydrates, was found to be inducible under nitrogen limitation in both species and was dependent on RpoN in P. aeruginosa, but not in P. putida (Hoffmann and Rehm, 2004). It was also suggested that RpoN is a negative regulator of phaF, which was thought to be a negative regulator of phaC1 (Hoffmann and Rehm, 2005). In the case of P. putida KT2440 it is interesting to note that, although rpoN inactivation had no effect on PHAs accumulation on gluconate, it increased their accumulation 40% on octanoate, reaching the highest accumulation in these experiments (Hoffmann and Rehm, 2004). A similar result was obtained by Mozejko-Ciesielska et al. (2017) while analyzing the effect of rpoN inactivation in the same P. putida strain. They also found a significant increase in the accumulation of mcl-PHAs in the rpoN mutant when growing on octanoate under nitrogen limitation. Therefore, the inactivation of rpoN could be further studied as a candidate for strain improvement in P. putida, which is the model bacterium for mcl-PHA synthesis.

Another example illustrating the participation of an alternative sigma factor on the control of PHB metabolism is found in the non-nitrogen-fixing cyanobacterium Synechocystis sp. PCC 6803 (Osanai et al., 2013). In this microorganism, PHB accumulation increases under phosphorous or nitrogen starvation. Transcription levels of PHB biosynthetic genes increase after nitrogen depletion. Coincidentally, the expression of the sigma factor named SigE also increased under this condition. A strain overexpressing sigE showed higher mRNA levels of the phb biosynthetic genes, with respect to those of the wild type, during the nitrogen-replete stage. This overexpression of sigE also increased the levels of the PHB biosynthetic enzymes during nitrogen limitation. This strategy increased 2.4 times the accumulation of PHB (Osanai et al., 2013).

Also in Synechocystis sp. PCC 6803, the response regulator Rre37 of a two-component system, has been shown to participate in the control of PHB accumulation (Osanai et al., 2014). This protein plays a role in carbon metabolism because the overexpression of rre37 up-regulated genes related to sugar catabolism, under both nitrogen-replete and nitrogen-depleted conditions and decreased glycogen accumulation. The expression of the phb biosynthetic genes was also up-regulated by Rre37, mainly under nitrogen depletion, and almost doubled PHB accumulation. Interestingly, overexpression of both Rre37 and SigE further increased PHB accumulation to 2.9 times. The Mr and monomer composition of PHB were not altered by the double overexpression of SigE and Rre37. Thus, Rre37 is a regulatory protein that changes the metabolic flow from glycogen to polyhydroxybutyrate and is probably participating in the same regulatory circuit with SigE, through an unknown mechanism. Besides suggesting the participation of sigma factor SigE and Rre37 in the control of transcription of the phb biosynthetic genes, these are good examples of how changing the regulatory networks affecting PHB metabolism can be a powerful approach, because not only the genes participating in PHB accumulation altered their expression when overexpressing sigE or rre37, but a whole set of genes related with carbon metabolism (glycogen catabolism), among other changes, so PHB production improvement was due to a higher expression of the phb genes, but also to a change in carbon distribution from glycogen to PHB (Osanai et al., 2013, 2014). It is noteworthy that increasing the enzymatic activities involved in the synthesis of PHAs by introducing the biosynthetic genes of R. eutropha into Synechocystis 6803 did not really enhance PHB accumulation, suggesting this is not the only rate-limiting point for PHB synthesis (Sudesh et al., 2002).

PsrA, a positive regulator of rpoS expression that also represses the fatty acid β-oxidation operon fadBA5 in P. aeruginosa (Kojic and Venturi, 2001; Kang et al., 2008), has been implicated in the fatty acid - PHA metabolic network in P. putida KT2440. In this bacterium, inactivation of psrA reduced PHAs accumulation between 16 and 54%, depending on the carbon source. This effect was explained as a consequence of a more active β-oxidation pathway induced by the psrA inactivation, although direct regulation of pha genes by PsrA or changes in their expression caused by a negative effect on rpoS expression, were not ruled out. In fact, overexpression of phaC2 gene was observed in the psrA mutant. Interestingly, a change in the composition of the polymers was also observed in this mutant, with a higher content of short chain length monomers (Fonseca et al., 2014).

Because PHAs in many organisms are produced under conditions of nutritional limitation, and adaptation to nutrient deprivation is mediated in many bacteria by the so-called stringent response (Boutte and Crosson, 2013), the participation of this stress response regulatory mechanism in the control of PHAs metabolism has been analyzed in several bacteria (López et al., 2015). Stringent response mediates adaptation to changes in nutrient availability by reprograming transcription of a high percentage of genes. This response is mediated by the nucleotides guanosine tetraphosphate (ppGpp) and guanosine pentaphosphate (pppGpp), the so-called alarmone, whose levels are controlled by the enzymes RelA (ppGpp synthetase), and SpoT (ppGpp synthetase/hydrolase) (Boutte and Crosson, 2013).

The first report showing a possible connection between stringent response and PHAs metabolism was found in P. aeruginosa GPo1 through the analysis of a phbZ minus mutant (Ruiz et al., 2001). It was shown that utilization of the PHB reserve is needed for the proper synthesis of ppGpp, and this, in turn, has an effect on stress resistance to ethanol and high temperature, however, no regulatory relationships between stringent response and PHAs metabolism were demonstrated.

Later, the analysis of a Rhizobium etli mutant unable to accumulate (p)ppGpp, showed that it was affected in amino acids and nitrate utilization, nodulation, nitrogen fixation and also was unable to accumulate PHB (Calderón-Flores et al., 2005).

In R. eutropha, the role of stringent response was investigated by inactivating the (p)ppGpp synthetase/hydrolase and (p)ppGpp synthetase genes of this bacterium. Double mutants, unable to synthesize (p)ppGpp, accumulated a minor amount of PHB, whereas unusually high levels of this polymer were obtained in a strain with increased levels of (p)ppGpp, generated by overexpression of the (p)ppGpp synthase SpoT2 in the absence of (p)ppGpp hydrolase. The artificial induction of stringent response through the addition of the amino acid analogs norvaline or serine hydroxamate, increased the contents of PHB during the stationary phase. Therefore, a positive correlation between the amount of ppGpp and the accumulation of PHB was suggested. The composition of the polymer was not affected by the induction of the stringent response. The authors concluded that the concentrations of (p)ppGpp affect PHB accumulation by negatively regulating PHB depolymerization, and suggested this could be done through allosteric regulation and covalent modification of the enzymes involved. They identified the main PHB depolymerase PhaZa1 as the target for this control (Juengert et al., 2017).

In P. putida KT2440 accumulation of PHAs is also induced under nitrogen limiting conditions. Interestingly, disruption of strict response by the inactivation of relA and spoT genes eliminated the need for nitrogen limitation, so the polymers were accumulated to the same level under both conditions. The relA/spoT mutant was able to accumulate 2.8 times the PHAs produced by the wild-type strain under no nitrogen limitation (Mozejko-Ciesielska et al., 2017). However, lower accumulation of PHAs was observed under nitrogen limitation, in comparison to the wild-type strain. In this case, the expression of the PHA depolymerase gene (phaZ) was lower in the absence of (p)ppGpp and could explain in part the phenotype observed. This altered regulation could be further investigated in order to promote PHAs accumulation under culture conditions allowing growth to high cell densities, with no need for a change to nutritional limitation to start producing the polymers.

More recently, a report in P. chlororaphis PA23, a bacterium used for biocontrol, showed that mutations in the regulatory genes relA/spoT significantly decreased PHAs accumulation, and also resulted in altered PHAs gene expression. In this study, it was shown that the two PHA synthase genes phaC1, phaC2, and the two phasin genes phaF, and phaI were also significantly down-regulated in gacS and rpoS mutants, reducing the amount of PHAs (Sharma et al., 2018). Interestingly, this work illustrates that the modification of stringent response and of other regulatory circuits can also change the monomeric composition of PHAs. For the three regulatory mutants (relA/spoT, gacS, rpoS), the PHAs produced on glucose contained significantly higher amounts of 3-hydroxyhexanoate and/or 3-hydroxyoctanoate, and the content of 3-hydroxydecanoate was lower, as compared to the parental strain. For the PHAs synthesized on octanoate, only in the case of the gacS mutant the monomer composition changed (Sharma et al., 2018).

Bacterial quorum sensing (QS) is a communication process that involves self-produced extracellular signaling molecules, called autoinducers, which can accumulate into levels that are required to collectively alter global patterns of gene expression (Papenfort and Bassler, 2016). The bioluminescent marine bacterium Vibrio harveyi is a good example of a canonical QS circuit, with a LuxR-type receptor that detects autoinducers. This bacterium is capable to synthesize PHB only at a high cell density and this production is positively controlled by the LuxR-autoinducers circuit (Sun et al., 1994; Miyamoto et al., 1998). Although the exact molecular mechanism is unknown, it is proposed that V. harveyi autoinducer leads to the increased expression of luxR; the LuxR activator in turn stimulates at high cell density the expression of PHB synthesis and luminescence (Miyamoto et al., 1998). Similarly, the physiological implication of PHB synthesis in relation to the luminescence is not clear, but it is possible that the polymer in V. harveyi serves as an energy source for maintaining cell viability in the stationary growth phase.

Interestingly, it has been found that the PHB synthesis of Rhodobacter sphaeroides, a facultative photosynthetic bacterium, is controlled in a manner analogous to the induction of cell luminescence. However, unlike the positive regulation of the PHB synthesis of V. harveyi, the PHB synthesis of R. sphaeroides is under negative regulation. In this bacterium, the cerR/cerI QS circuit codes for LuxR-type regulator and the enzyme that synthesize the autoinducer molecule, respectively (Ye et al., 2013). Under aerobic growth conditions, mutations in the cerR/cerI genes resulted in a 2-fold increase in the cellular PHB content, with respect to the wild type, and this increase is coincidental with a proportional increase of the PHB synthase enzymatic activity. Moreover, the transcriptional levels of phaC gene were also 2-fold higher than the corresponding values of the wild type (Kho et al., 2003). The potential of mutations in cerR/cerI genes has not been fully explored.

Another example illustrating the participation of QS in the control of PHA gene expression is found in P. chlororaphis PA23. There are three distinct QS regulatory circuits in this bacterium, PhzRI, CsaRI, and AurRI, but only the Phz system has been characterized. An AHL-deficient strain (comparable to a triple phzI/csaI/aurI QS mutant) resembles stringent response-defective P. chlororaphis PA23: it shows a significantly decreased PHAs accumulation, a radical change in PHAs composition, and down-regulation of phaC1, phaZ, phaC2, phaD, phaF, and phaI. The search for regulatory sequences in these genes revealed the presence of a region with 15/20 conserved nucleotides upstream of phaZ and phzF, named “phz-box”; nevertheless, direct QS regulation of phaZ and phaF has not been demonstrated (Mohanan et al., 2019).

In addition to the previously described allosteric regulatory mechanism controlling PHB synthesis in response to oxygen limitation in R. eutropha (Oeding and Schlegel, 1973) and A. beijerinckii (Senior and Dawes, 1973), other mechanisms for the control by oxygen have been described, but operate at the level of transcription.

In Pseudomonas extremaustralis PHB production is induced under micro aerobic and anaerobic cultures. The redox transcriptional regulator Anr (Fnr-like), which regulates several genes during the transition from aerobic to anaerobic growth, is involved in this control. The Anr protein contains a [4Fe-4S]⁺² cluster that suffers oxidation, changing its DNA binding affinity, thus sensing the redox state of the cell. Inactivation of anr decreased expression of phaC and phaR (coding for the transcriptional regulator of pha biosynthetic genes) and PHB production under aerobic and micro aerobic conditions. The presence of two putative binding sites in the regulatory region of these genes suggested a direct regulation of transcription of these genes by Anr (Tribelli et al., 2010).

In Herbaspirillum seropedicae SmR1, a nitrogen-fixing endophyte, where PHB has an important role in the maintenance of the intracellular redox balance, it was found that two of the three Fnr regulators encoded in this bacterium control expression of the phaC synthase gene. A triple Fnr mutant showed a 50% reduction on PHB accumulation under low oxygen tension, with respect to the wild-type strain (Batista et al., 2018).

A role for Anr in the control of mcl-PHA metabolism has also been shown in P. chlororaphis PA23. Inactivation of anr significantly reduced PHAs accumulation, although their monomeric composition was not drastically changed. Although expression of phaC1, phaZ, phaC2, phaD, phaF, and phaI genes related to PHAs metabolism was analyzed, only phaF showed a significant reduction in the mutant; nevertheless, Anr-recognition sequences were not found upstream of phaF (Mohanan et al., 2019).

The participation of an Fnr homolog (CydR) in the control of PHB metabolism in response to oxygen has also been reported in A. vinelandii, where this transcriptional regulator seems to be a repressor of the biosynthetic genes (Wu et al., 2001). This regulation is further described in section “Regulation of PHB Synthesis in Azotobacter.”

In Acinetobacter sp. PHB accumulation is induced by phosphate starvation and a regulatory mechanism has been proposed. Schembri et al. (1995) showed that phosphate limitation induces transcription of the pha biosynthetic locus. The presence of two potential Pho-box consensus sequences, upstream of two transcription start points, and the requirement of PhoB (the pho regulon activator) and phosphate limitation for the expression of these transcripts, showed direct positive regulation of these genes by the transcriptional regulator PhoB under phosphate limitation.

Control by the global regulator Crc has been demonstrated in P. putida KT2442 (La Rosa et al., 2014). Crc is a protein that in Pseudomonas has a key role in carbon catabolite repression, a process that allows to first assimilate a preferred carbon source among a mixture of several compounds. Crc acts post-transcriptionally by binding to its target mRNAs on motifs close to their ribosome binding site, affecting translation (Rojo, 2010). In P. putida, La Rosa et al. (2014) showed that Crc links PHAs metabolism with the regulatory networks controlling carbon utilization. They identified a Crc binding motif on the PHA synthase gene phaC1 mRNA, and demonstrated that Crc inhibits phaC1 mRNA translation and not transcription. They also showed that the activity of this regulator is affected both by the carbon source present and by the C/N ratio of the medium. This is carried out by the antagonist small RNAs CrcZ and CrcY, which sequester Crc and the expression of these RNAs is higher under conditions in which PHAs synthesis is high and Crc is strongly antagonized. They also reported that Crc negatively controls PHAs synthesis when growing in media with a balanced carbon/nitrogen ratio. Under unbalanced carbon/nitrogen conditions (where PHAs synthesis is allowed), Crc had no activity. Interestingly, in the P. putida crc negative mutant, the requirement for nitrogen limitation for polymer synthesis was lost, allowing to accumulate the same percentage of PHAs under nitrogen replete conditions. In LB medium with octanoate, the highest accumulation percentages and also the highest PHAs volumetric production were obtained with the crc mutant, eliminating the need to restrict growth by limiting nitrogen (La Rosa et al., 2014). Therefore, this is another example of the potential benefits of modifying the regulatory systems controlling PHAs metabolism for strain improvement.

NtrB and NtrC proteins are the sensor kinase (NRII) and the response regulator (NRI) of a two-component regulatory system that in diverse bacteria is a key player in the control of expression of many genes of nitrogen metabolism (van Heeswijk et al., 2013). These regulators are also involved in the regulation of PHB in some bacteria.

In Azospirillum brasilense SP7, PHB production is inhibited by the presence of high ammonia concentration. Inactivation of ntrB or ntrC considerably increased PHB accumulation (up to 9-fold) on different C/N ratios. The ntrB or ntrC mutants couple PHB production and growth, producing PHB both during the growth phase and stationary phase. The absence of these regulators eliminates the inhibitory effect of ammonia, and interestingly, greatly diminished the respiration of A. brasilense (Sun et al., 2000). All these phenotypes show that NtrB and NtrC somehow regulate PHB metabolism; therefore, the corresponding mutants are interesting starting strains for future improvements.

In H. seropedicae SmR1, an ntrC mutant was also shown to accumulate higher levels of PHB than the wild type (up to 2-fold increase). The inactivation was found to increase the activity of glucose-6-phosphate dehydrogenase (a reaction producing NADPH), leading to a 2.1-fold increase in the NADPH/NADP + ratio. This increase in NADPH was proposed to cause the increased synthesis of the polymer (Sacomboio et al., 2017).

The NtrB regulator of Paracoccus denitrificans PD1222 was also shown to somehow regulate PHB accumulation. In this bacterium inactivation of ntrB induced the overproduction of the polymer (5-fold), when growing on nitrate, and also increased expression of the PHA synthase gene phaC. An interesting metabolic change was also observed. The acetyl-CoA concentration increased 4-fold in the ntrB mutant. Therefore, the increased accumulation of PHB in this mutant was a combination of the effect on phb gene expression with the induction of a metabolic change caused by altering the regulatory network (Olaya-Abril et al., 2018).

Many other transcriptional regulatory systems have been reported to participate in the control of PHAs metabolism. One very interesting is the GacS-GacA two-component system, where GacS functions as a sensor histidine kinase protein that phosphorylates GacA, which is the response regulator that when phosphorylated activates transcription of its target genes. The GacS-GacA system is conserved in gram-negative bacteria. Its participation in the control of PHB metabolism was discovered in A. vinelandii while looking for regulators of alginate synthesis, because inactivation of gacS had a negative effect on the synthesis of this polysaccharide, but also on PHB accumulation (Castañeda et al., 2000). It was shown later that phosphorylated GacA is a positive regulator of PHB synthesis, rpoS expression (Castañeda et al., 2001), and for the expression of the small regulatory RNAs RsmZ/Y, that interact with the translational regulatory protein RsmA (Manzo et al., 2011). This regulation is further described in section “Regulation of PHB Synthesis in Azotobacter.”

In P. putida CA-3 the participation of the GacS-GacA on the control of mcl-PHAs metabolism was found by random mini-Tn5 mutagenesis. A mutant unable to accumulate PHAs had disrupted the gacS gene. Transcription of the regulatory small RNAs RsmY and RsmZ, known to be regulated by GacS-GacA in other Pseudomonas, was not affected in the gacS mutant. Transcription of the PHA synthase phaC1 was also unaffected, but evidence of a lower content of the synthase enzyme suggested a post-transcriptional control (Ryan et al., 2013). In P. putida KT2442, inactivation of the sensor kinase GacS also reduced PHAs accumulation, but in this case, a reduction in the transcription rate of the PHAs gene cluster was observed and this was restored by expressing PhaD, an activator of the pha genes discussed in section “PhaD” (Prieto et al., 2016).

As previously mentioned, in P. chlororaphis PA23 inactivation of gacS strongly decreased PHAs accumulation and this correlated with a considerably diminished expression of genes phaC1, phaC2, phaD, phaF, phaI and phaZ. Interestingly, the gacS mutation also changed the monomeric composition of the mcl-PHAs produced by this bacterium (Sharma et al., 2018).

Small non-coding RNAs (sRNAs) play important regulatory roles in bacteria. Many of them act as antisense RNAs on multiple target mRNAs, other sRNAs act by binding to proteins, some of them regulatory, affecting their activity (Gottesman and Storz, 2011).

In the regulation of PHB metabolism, some examples of regulation by small RNAs have been documented. Free-living Sinorhizobium meliloti cells accumulate PHB under nitrogen-limiting conditions. For that reason, Lagares et al. (2017) studied a small RNA called MmgR (Makes more granules Regulator), whose expression is regulated by the amount of nitrogen source. They found that MmgR negatively regulates PHB accumulation under conditions of N starvation and C surplus. Inactivation of mmgR gene increased PHB production by 20% with respect to the wild type, without affecting growth. The increase in PHB was even higher when using higher C/N ratios. The overproduction occurred during the stationary phase, but not during the exponential growth, so PHB production was still affected by nutrient limitation. The mmgR mutation increased the number of granules and also resulted in the overexpression of two phasin proteins. So, MmgR sRNA participates in the regulation of PHB accumulation in this bacterium, through an unknown mechanism. This mutation is another candidate for further studying its potential to improve PHB production.

A regulatory small RNA called ArrF has been implicated in the regulation of PHB accumulation in response to iron in A. vinelandii (Pyla et al., 2009; Muriel-Millán et al., 2014). Iron-limiting conditions increase PHB accumulation in this bacterium (Page, 1982). The ArrF RNA participates in this regulation by affecting translation of its target, the mRNA of phbR, through complementary binding, although the effects of this interaction probably differ between strains (Pyla et al., 2009; Muriel-Millán et al., 2014). This is further described in section “Regulation of PHB Synthesis in Azotobacter.”

In Pseudomonadaceae, the non-coding small RNAs called RsmZ/Y/X regulate secondary metabolism and carbon storage (Lapouge et al., 2008). These RNAs, together with the RsmA protein, constitute a system of post-transcriptional regulation where RsmA acts as a translational repressor by binding to its mRNA targets, and the small RNAs counteract this repression through binding to RsmA. In A. vinelandii, inactivation of rsmA results in increased PHB production, because RsmA represses translation of the PHB biosynthetic operon phbBAC and of the phbR gene that codes for its transcriptional activator. At least two of the eight small RNAs present in this bacterium (RmZ1 and RsmZ2) have been shown to specifically bind RsmA and counteract this repression (Manzo et al., 2011; Hernández-Eligio et al., 2012). This regulation and its connection with the GacS/GacA two-component system will be described in section” Regulation of PHB Synthesis in Azotobacter.”

Another regulatory mechanism controlling the synthesis of PHB is carried out by the nitrogen-related phosphotransferase system (PTS^Ntr). The participation of this system in the control of PHB synthesis was first reported in A. vinelandii (Segura and Espín, 1998). This system is homologous to the carbohydrate PTS used in many bacteria for uptake and phosphorylation of different sugar substrates. PTS^Ntr is present in most gram-negative bacteria and is composed of EI^Ntr, Npr and EIIA^Ntr proteins encoded by the ptsP, ptsO and ptsN genes, respectively. These proteins participate in a phosphorylation cascade where EIIA^Ntr seems to be the final phospho-acceptor. The PTS^Ntr seems to be exclusively involved in regulatory functions (Pflüger-Grau and Görke, 2010). In P. putida, inactivation of PTS^Ntr components reduced the synthesis of PHAs by increasing the level of unphosphorylated EIIA^Ntr (Velázquez et al., 2007). In a similar way, A. vinelandii mutants where EIIA^Ntr is present in an unphosphorylated form are unable to produce PHB (Segura and Espín, 1998; Noguez et al., 2008). The mechanism by which the PTS^Ntr system regulates the synthesis PHB has been widely investigated in A. vinelandii and is described in section “Regulation of PHB Synthesis in Azotobacter.”

An interesting connection between PTS^Ntr and the stringent response was found in R. eutropha (Karstens et al., 2014). In an attempt to clarify the role of PTS^Ntr in the control of PHB metabolism in this bacterium, these authors screened for proteins able to interact with EIIA^Ntr. The bifunctional ppGpp synthase/hydrolase SpoT1 was shown to interact, but only with the unphosphorylated form of EIIA^Ntr. This could, in turn, alter the ppGpp levels, thus affecting PHB formation and the expression of genes responsive to this alarmone. A role for PTS^Ntr in the control of (p)ppGpp levels has also been described in Caulobacter crescentus, although its relation with PHAs metabolism was not investigated (Ronneau et al., 2016).

The participation in the control of PHB metabolism, not of the PTS^Ntr system, but of proteins of the sugar-specific-PTS system, has been documented in R. eutropha (Pries et al., 1991). Inactivation of ptsH or ptsI genes, coding for the components of the PTS, Enzyme I (EI) and the histidine phosphocarrier protein (HPr), respectively, caused a lower accumulation of PHB. These mutants showed a higher rate of PHB degradation, so the authors proposed that the PTS could be controlling, by phosphorylation, the activity of the PHB mobilizing enzymes or the affinity of a regulatory element controlling them (Pries et al., 1991). Later, Kaddor et al. (2012) reported a proteome analysis of a ptsHI double mutant of R. eutropha and showed that these mutations caused a down-regulation of the PHB biosynthetic acetoacetyl-CoA reductase and of two phasin proteins; also enzymes of central catabolism (Entner-Doudoroff pathway, the tricarboxylic acid cycle) were down-regulated in the mutant, whereas enzymes of gluconeogenesis showed enhanced expression in the mutant. Together, these changes explained the PHB-leaky phenotype observed.

It is interesting to point out that the permease components (EIIB and EIIC proteins/domains) that together with EI and Hpr proteins constitute a complete PTS for the transport of carbohydrates are not present in R eutropha, so EI and Hpr are likely to be involved in regulatory functions of carbon and PHB metabolism (Kaddor and Steinbüchel, 2011). Strikingly, this bacterium also lacks ptsP and ptsO genes of the PTS^Ntr, but has a homologous to ptsN (IIA^Ntr, H16_A0384) (Kaddor and Steinbüchel, 2011; Kaddor et al., 2012). Inactivation of ptsN increased PHB accumulation, as in A vinelandii, and P. putida (Kaddor and Steinbüchel, 2011), and phosphorylation of EIIA^Ntr through EI and Hpr has been shown and it favors PHB accumulation (Krausse et al., 2009).

The genes encoding proteins involved in PHAs synthesis are often located in conserved PHAs gene clusters, depending if the bacteria are scl- or mcl-PHA producers. In some PHB-producing bacteria, like A. vinelandii (Peralta-Gil et al., 2002), Azotobacter FA8 (Pettinari et al., 2001), Pseudomonas sp. strain 61-3 (Matsusaki et al., 1998), and Pseudomonas sp. USM 4-55 (Tan et al., 2010), the phbA, phbB and phbC genes are clustered in the same locus and are organized in a single transcriptional unit. In all these cases, upstream of the phbBAC operon and in the opposite direction, the gene called phbR is located. This gene codes for a member of the AraC/XylS family of transcriptional regulators, with a typical DNA binding domain helix-turn-helix (HTH). It is important to note that Matsusaki et al. (1998) reported for the first time the similarity between PhbR and other proteins belonging to the AraC family, such as OruR of P. aeruginosa and the virulence-associated regulator of Mycobacterium tuberculosis; thus, the ORF was referred to as phbR, and it is not related to other regulators also called PhaR, that are discussed later. In A. vinelandii and Pseudomonas sp. strain 61-3 it was demonstrated that PHB biosynthesis is positively regulated by PhbR, which activates transcription of the phbBAC biosynthetic operon (Matsusaki et al., 1998; Peralta-Gil et al., 2002; Hernández-Eligio et al., 2011).

Although the majority of bacteria accumulate either scl-PHAs or mcl-PHAs, some of them, like Pseudomonas sp. strain 61-3, are able to synthesize and accumulate a blend of PHB homopolymer and a random P(3HB-co-3HA) copolymer. This is possible because these bacteria possess two types of PHAs biosynthetic gene loci. Interestingly, when the activator phbR was overexpressed in trans, Pseudomonas sp. strain 61-3 increased its polymer content (51%). Furthermore, the overexpression of phbR enhanced PHB biosynthesis in this bacterium, resulting in an increase of the polymer content and enrichment of the 3HB fraction in the whole polymer (94 mol%) (Matsusaki et al., 1998).

Most Pseudomonas species belonging to the ribosomal RNA-homology group I synthesize mcl-PHAs, whose synthesis requires the proteins encoded by the phaC1ZC2DFI cluster. This cluster contains two polymerases (PhaC1 and PhaC2), a depolymerase (PhaZ), and three proteins (PhaDFI) involved in granule formation and in the regulation of gene expression. PhaD is a TetR-family transcriptional regulator that in P. aeruginosa GPo1 (formerly Pseudomonas oleovorans GPo1; Klinke et al., 2000; Ramos et al., 2005), P. putida U (Sandoval et al., 2007) and P. putida KT2440 (de Eugenio et al., 2010b), positively regulates PHAs synthesis. Inactivation of phaD reduces not only the production of the polymers, but also affects the size and number of the PHAs granules (Klinke et al., 2000). This regulation is performed in a direct manner, since PhaD binds to the promoter region of synthase phaC1ZC2D and phaIF operons, auto activating its expression and that of the other pha genes. This phenomenon is carbon source dependent, with higher activation on the preferred carbon source octanoate and this, together with 3D modeling of PhaD, allowed the authors to suggest the participation of a β-oxidation metabolite as an effector (de Eugenio et al., 2010b).

The finding that PhaD simultaneously activates expression of the PhaZ depolymerase and the PHA synthases is in agreement with the conclusion that synthesis and mobilization in Pseudomonas is a cyclic process allowing PHAs turnover to achieve a metabolic balance (de Eugenio et al., 2010a). However, knowledge about how this potentially futile cycle is regulated or balanced is relatively limited. As discussed in section “Control of Enzymatic Activity,” this could be accomplished through allosteric control of the enzymes involved. PHA synthase and the fatty acid oxidation complex, the provider or consumer of 3-hydroxyacyl-CoAs (precursors or depolymerization products), are affected by metabolites like R-3-hydroxyacyl-CoA, free CoA, acetyl-CoA, NADH and NAD; thus, the balance of this cycle could be determined by the metabolic status of the cell (Ren et al., 2009). In bacteria like R. eutropha, where a similar situation of constitutive phb gene expression and PHB turnover also occur (Doi et al., 1990; Brigham et al., 2012), the control of this cycle seems to be achieved also by controlling enzyme activities, through allosteric regulation by the intracellular levels of acetyl-CoA and free CoA and its redox state (Oeding and Schlegel, 1973; Haywood et al., 1988). As mentioned in sections “Control of Enzymatic Activity” and “Control by the Stringent Response,” the activities of the PHB synthase or PhaZa1 depolymerase can also be controlled by (p)ppGpp levels or through phosphorylation (Juengert et al., 2017, 2018).

When PHAs are synthesized, they accumulate as granules, which are composed of 97.5% PHA, 2% proteins, and some lipids (Jendrossek and Pfeiffer, 2014). Four types of proteins are found associated to the granules: structural proteins called phasins, PHA metabolism enzymes, like PHA synthases and depolymerases, and regulatory proteins (Pötter and Steinbüchel, 2005).