|Home | About | Journals | Submit | Contact Us | Français|
The genome of Pseudomonas putida KT2440 encodes five proteins of the phosphoenolpyruvate-carbohydrate phosphotransferase system. Two of these (FruA and FruB) form a dedicated system for fructose intake, while enzyme INtr (EINtr; encoded by ptsP), NPr (ptsO), and EIINtr (ptsN) act in concert to control the intracellular accumulation of polyhydroxyalkanoates, a typical product of carbon overflow.
The phosphoenolpyruvate (PEP)-carbohydrate phosphotransferase transport system (PTS) (4, 10) mediates the phosphorylation and subsequent uptake of a large number of carbohydrates in both gram-positive and gram-negative bacteria through a mechanism that allows the sequential utilization of these sugars and the cross talk between their transport and other cellular utilities (25, 32, 36). Any canonical PTS involves a set of three major phosphotransfer catalytic activities (enzyme I [EI], HPr, and EII), which mediate the flow of a high-energy phosphate from PEP all the way to the sugar to be transported. EII components are virtually sugar specific and may consist of a single polypeptide carrying three subdomains (EIIA, EIIB, and EIIC) or any combination of the same moieties. With many variations, the same general scheme holds true for a large variety of microorganisms (4, 10).
In order to ascertain the functions of the PTS proteins borne by the soil bacterium Pseudomonas putida (8, 9), the genome of P. putida strain KT2440 (23) was inspected in silico for PTS domains (Fig. (Fig.1).1). Such a survey pinpointed as few as five open reading frames (ORFs) containing recognizable PTS modules. By similarity with what appear to be homologous proteins in the genome of Escherichia coli (36) and other gram-negative bacteria (4), we maintain the following nomenclature for each of them: EINtr (encoded by ptsP), NPr (ptsO), EIIANtr (ptsN), EI::HPr::EIIAFru (fruB), and EIIB::EIICFru (fruA). The chromosomal context of each of these genes (Fig. (Fig.2)2) suggested that fruA/fruB and ptsN/ptsO belong to discrete functional clusters, while ptsP did not offer any hint as to possible functional partners. The fruA- and fruB-encoded proteins appeared to form the only system in P. putida for intake of sugars (fructose) through the phosphorylation-linked transport scheme that is so distinctive of the PTS enzymes. In fact, FruB is a multiphosphoryl transfer protein, which consists of an EI domain fused to an equally standard HPr module and one EIIAFru module (Fig. (Fig.1).1). In contrast, PtsO (NPr) and PtsN (EIIANtr) were already known to cluster downstream of rpoN, encoding the nitrogen-related alternative sigma factor σ54, in P. putida and many other bacteria (10). Moreover, PtsP (EINtr) is believed to form, along with PtsO and PtsN, a nitrogen-related branch of the PTS (10, 27, 29).
The hints as to the functions of the PTS proteins of P. putida brought about by the available annotations (28) and the known properties of (as well as the many hypotheses about) similar genes and proteins found in other microorganisms (4, 10) set the stage for addressing experimentally the roles of the reduced complement of phosphotransfer proteins found in P. putida. To this end, we employed nonpolar mutations in each of the genes under scrutiny, the genomic context of which is shown in Fig. Fig.22.
As a first approach, we performed quantitative growth assays of the ptsP, ptsN, ptsO, and fruB strains on four C sources (succinate, glucose, fructose, and glycerol) (Fig. (Fig.3)3) which enter the C supply at distinct stages of the central metabolism (20, 37). The reference conditions were set for the wild-type strain on glucose, as this sugar is known to be taken up by P. putida altogether independently of any PTS (33). The most obvious result of these tests was the complete lack of growth of the fruB mutant on fructose (Fig. (Fig.3B).3B). This piece of information was not only compatible with the genomic predictions but also verified that there are no alternative mechanisms of entry of fructose in P. putida other than the one mediated by the multiphosphoryl transfer protein FruB. On the other hand, the fruB mutant did grow on the other C sources, thus suggesting that the complete PTS branch to which FruB belongs (along with FruA) is unique to fructose intake.
A closer inspection of the curves in Fig. Fig.33 also revealed a recurrent slower growth of the ptsN mutant (but not of the ptsO or ptsP counterpart) on every C source tested. Although the lag periods of the wild type and the ptsN mutant were roughly similar, the growth rates during exponential proliferation differed significantly in medium with succinate (μ = 0.87 ± 0.02 versus 0.60 ± 0.01 h−1). Furthermore, the lag period of the ptsN strain in glycerol was ≥65 h (Fig. (Fig.3C).3C). This reflected a long physiological adaptation phase, as instant growth in glycerol was not retained after transferring the cells back to LB (contamination was completely ruled out by use of proper controls). In addition, a noticeable slower growth of the ptsN mutant was observed in medium with glucose (μ = 0.31 h−1 ± 0.01 versus 0.27 ± 0.01 h−1) or fructose (μ = 0.15 ± 0.00 versus 0.12 ± 0.01 h−1). That the ptsN strain eventually grew, albeit slower or later, in all C sources tested suggested that the growth problem was associated with a defect in metabolism, in transport, or both. This was in contrast to the case for the fruB mutant, which did not grow at all in fructose (see above).
To examine whether the growth problems of the ptsN mutant (and some of the other strains as well) were related only to C or also involved nitrogen, we ran equivalent experiments in nitrogen-free glucose-M9 or succinate-M9 medium and various N compounds: 10 mM of isoleucine, glutamine, or sodium nitrate or 1.0 to 10 mM NH4+. In these conditions, all strains turned out to have approximately the same lag period, so we focused on growth rates. The data in Table Table11 indicate that the growth rate of each of the mutants did vary with respect to that of the wild-type strain according to the various N sources and their combination with the two C compounds tested. In the glucose medium, the only significant effect of varying the N source was the slower growth of the ptsN mutant with glutamine. However, in the succinate medium, varied N compounds caused different effects on the growth rates of the mutants. Specifically, Table Table11 shows that all PTS mutants cultured in succinate amended with a high ammonia concentration (10 mM NH4+) or with glutamine grew faster than the wild-type cells. This suggested that PTS proteins could affect the utilization of these N sources. Taken together, the data in Fig. Fig.33 and Table Table11 raised the prospect that the PTS products could be related to channeling the available carbon and nitrogen compounds to diverse metabolic destinations.
One helpful approach to test whether these PTS products could be connected to the gross balance of C and/or N was to examine the synthesis of polyhydroxyalkanoates (PHAs) in the various mutants. Although the complex regulation of PHA buildup in Pseudomonas and other gram-negative species is not yet understood in detail (14-16, 35), is seems that these insoluble compounds do accumulate intracellularly under conditions of C overflow as a sort of dependable sink for carbon and reducing equivalents (1). By the same token, the presence of PHA granules reflects an imbalance of the ability of cells to metabolize available C sources with respect to other limiting nutrients in the medium, specifically N. Figure Figure4A4A shows the intracellular levels of PHAs found in ptsN, ptsO, ptsP, fruB, and ptsN ptsO mutants of P. putida grown in minimal medium supplemented with 15 mM of sodium octanoate (conditions known to favor accumulation of PHAs ). The data in Fig. Fig.4A4A revealed not only that the ptsP and ptsO mutants were largely impaired in accumulation of PHAs but also that the ptsN strain contained a considerably larger fraction of the same polymer. That the ptsN ptsO double mutant behaved similarly to the single ptsO mutant suggested an epistatic relationship between ptsN and ptsO in their influence on the PHA synthesis phenotype. In contrast, the fruB mutant had PHA contents indistinguishable from those of the wild-type P. putida strain. Taken together, this set of data revealed a connection between the abridged PTS formed by ptsP, ptsO, and ptsN and accumulation of PHAs.
The next question was whether the PHA phenotypes associated with the various PTS genes could be traced to the traffic of high-energy phosphate through the PTS. To sort this out, we transformed the ptsN mutant with low-copy-number plasmids encoding the wild-type PtsN protein or variants in which the phosphorylatable His 68 residue had been replaced by either an Ala (pRK154HA) or an Asp (pRK154HD). We have argued before (8) that such PtsN variants fix the conformation of the polypeptide in a form that mimics, respectively, the nonphosphorylated (H68A) or the phosphorylated (H68D) protein. The results of the PHA accumulation assays with P. putida strains bearing each of the ptsN alleles are shown in Fig. Fig.4B.4B. As expected, complementation of the ptsN deletion by the intact ptsN gene brought down PHA contents to wild-type levels (in the range of 20% of the dried cell biomass). In contrast, expression of H68D and H68A in the same ptsN mutant kept the PHA hyperaccumulation phenotype high (and even exacerbated it). These results implied that traffic of high-energy phosphate through the PTS proteins participates in the regulatory process that leads to production of PHAs. Yet, the genetic data did not elucidate the direction of such traffic between the corresponding PTS products. One way or the other, the combined results in Fig. Fig.33 and and44 are consistent with the notion that these PTS genes could be related to signaling the N versus C balance (27, 29).
The intriguing side of what appears to be a separate branch of the PTS is that the protein encoded by ptsN (EIIANtr) lacks the membrane-associated permease moieties EIIB and EIIC that typically tie PTS proteins to sugar transport (4). The whole set of EINtr, NPr, and EIIANtr proteins of E. coli have been purified and shown in vitro to sustain a typical flow of high-energy phosphate from PEP to EINtr to NPr to EIIANtr (26, 27). Furthermore, E. coli strains lacking these proteins, as well as mutants of many other bacteria, display a plethora of phenotypes (2, 3, 5-9, 13, 17, 19, 21, 24, 26, 30, 31, 38-40). Interestingly, the loss of PtsP makes Azotobacter vinelandii unable to accumulate poly-β-hydroxybutyrate (34). None of the biological qualities affected by these PTS genes in diverse microorganisms appear to involve transport of nutrients, so it plausible that the PtsP/PtsO/PtsN partnership has evolved to fulfill a different regulatory function. While the question of the physiological signal(s) that is connected to PtsP/PtsO/PtsN remains open, our data argue that the loss of PtsN is sensed by the PHA synthesis machinery as a factual situation of C surplus with respect to other limiting nutrients, which channels much of the available octanoate to the synthesis of PHAs. In contrast, in cells lacking either PtsP or PtsO, the PHA polymerization process might detect a shortage of C, which makes octanoate be directed into other functions. These observations indicate that the PTS composed of EINtr, NPr, and EIIANtr could control a physiological balance of C versus N sources. The mechanisms involved in such control are currently under study in our laboratory.
Auxiliadora Prieto is kindly acknowledged for critical review of the manuscript.
This work was supported in part by EU grants of the 5th and the 6th Framework Programs.
Published ahead of print on 6 April 2007.