A. Kornberg Tel. (415) 723 6167; Fax (415) 723 6783.
The regulatory protein AlgR2 in Pseudomonas aeruginosa positively regulates nucleoside diphosphate kinase (Ndk) and succinyl-CoA synthetase, enzymes critical in nucleoside triphosphate (NTP) formation. AlgR2 positively regulates the production of alginate, GTP, ppGpp and inorganic polyphosphate (poly P). An algR2 mutant with low levels of these metabolites has them restored by introducing and overexpressing either the algR2 or the ndk gene into the algR2 mutant. Thus, Ndk is involved in the formation of these compounds and largely prevents the death of the algR2 mutant, which occurs early in the stationary phase. We demonstrate that the 12 kDa Ndk–pyruvate kinase (Pk) complex, previously shown to generate predominantly GTP instead of all the NTPs, has a low affinity for the deoxynucleoside diphosphates and cannot generate the dNTPs needed for DNA replication and cell division; this complex may thus be involved in regulating the levels of both NTPs and dNTPs that modulate cell division and survival in the stationary phase.
Many bacteria produce exopolysaccharides under starvation conditions, particularly in response to prolonged nitrogen or phosphate limitation (Sutherland, 1990). Like many secondary metabolites, polysaccharide synthesis is usually triggered at the onset of the stationary phase of growth when nutrients in the growth media become limiting. An example is alginate synthesis by Pseudomonas aeruginosa, triggered in non-mucoid (non-alginate-producing) cells during growth on poor phosphate or nitrogen sources. With acetamide as a source of nitrogen, under phosphate limitation in chemostats or in the presence of inhibitors of energy metabolism that deplete the levels of nucleoside triphosphates (NTPs), non-mucoid cells become heavily mucoid because of secreted alginate (Speert et al., 1990; Woods et al., 1991; Terry et al., 1992). Another example of the transition of non-mucoid P. aeruginosa to mucoidy is during pulmonary infections of cystic fibrosis (CF) patients in whom mucoidy is believed to contribute to bacterial resistance to phagocytosis, to the actions of neutrophils and to allowing biofilm formation leading to resistance to antibiotics (May and Chakrabarty, 1994; Shankar et al., 1995).
Another polymer produced in the stationary phase and presumably in response to starvation is inorganic polyphosphate (poly P). Mutants of Escherichia coli lacking polyphosphate kinase (PPK), the enzyme that makes poly P, fail to survive in stationary phase and lack resistance to heat, oxidants and osmotic changes (Akiyama et al., 1992; Crooke et al., 1994; Rao and Kornberg, 1996). In Myxococcus xanthus, the levels of poly P as well as poly P-AMP phosphotransferase increase more than 10-fold after the vegetative phase of growth (T. Shiba and A. Kornberg, unpublished results). The increase in poly P level is preceded by an increase in the level of guanosine-3′-di-5′-phosphate (ppGpp); mutants that fail to produce ppGpp also fail to enhance this poly P level. Thus, ppGpp appears to have a regulatory role in poly P accumulation (Kornberg, 1995). More recently, Kuroda et al. (1997) have demonstrated in E. coli that high levels of ppGpp and guanosine pentaphosphate (pppGpp) lead to massive accumulations of poly P as a result of inhibition of the exopolyphosphatase (PPX) responsible for the hydrolytic breakdown of poly P.
In P. aeruginosa strain 8830, a mucoid CF isolate, the algR2 gene positively regulates alginate synthesis and the production of two key enzymes of energy metabolism, nucleoside diphosphate kinase (Ndk), responsible for the generation of NTPs or deoxy-NTPs (dNTPs), and succinyl-CoA synthetase (Scs), an ATP-generating enzyme of the tricarboxylic acid (TCA) cycle (Kavanaugh-Black et al., 1994; Schlictman et al., 1994). In this report, we demonstrate that a null mutation in the algR2 gene leads to a reduced level not only of alginate, but also of ppGpp and poly P as well. Membrane fractions of this mutant have been shown previously to allow reduced levels of GTP synthesis (Sundin et al., 1996a), and we demonstrate in this report that the intracellular levels of poly P, GTP, pppGpp and ppGpp are significantly reduced after growth in a low-phosphate medium in the presence of serine hydroxamate, an inducer of stringent response in E. coli (Cashel and Rudd, 1987; Cashel, 1994). Introduction of either the algR2 or the ndk gene and their hyperexpression from the tac promoter restores the levels of alginate, poly P, GTP and (p)ppGpp simultaneously, suggesting that the reduced levels of these metabolites in the algR2 mutant are primarily caused by extremely low levels of Ndk in this mutant. Additionally, we demonstrate that the promoter of the algR2 gene is responsive to phosphate starvation, so that an elevated level of Ndk under such starvation conditions enhances the efficiency of NTP and dNTP generation, compensating for the low levels of intracellular NTPs and dNTPs.
Poly P and alginate production in P. aeruginosa is co-regulated by algR2 and ndk genes
An alginate-negative (alg ) mutant of stably mucoid alginate-producing (Alg+) P. aeruginosa strain 8830 is deficient in Ndk and Scs formation (Schlictman et al., 1994). A double mutant, algR2 algH, has no detectable Ndk as measured by autophosphorylation or Western blotting (Schlictman et al., 1995). Alginate synthesis requires phosphorylated sugars, such as fructose 6-phosphate (F6P) or mannose 6-phosphate (M6P) as precursors of GDP-mannose (Fig. 1). Previous reports have suggested the presence of poly P in pseudomonads (Miguez et al., 1986; Takade et al., 1991), but there is little information on its relation to other metabolic events. We, therefore, compared the levels of poly P in the CF-isolate mucoid strain 8830 with those of its spontaneous non-mucoid segregant 8822 and the non-mucoid laboratory strain PAO1. During growth in L broth, the bulk of poly P accumulation in strain 8830 takes place at the onset of the stationary phase and continues well into the late stationary phase (Fig. 2). The poly P level in mucoid strain 8830 is much higher than is normally found in other microorganisms, such as E. coli (30–35 nmol mg−1 protein during growth in low-phosphate MOPS medium), Helicobacter pylori (180–190 nmol mg−1 protein) and compares favourably with that found in yeasts (850 nmol mg−1 protein). It is noteworthy that poly P levels are highly variable, depending upon media composition and the state of growth of the microorganisms. Inasmuch as poly P accumulation occurs predominantly under stringent conditions in E. coli (Kornberg, 1995), poly P levels in P. aeruginosa were determined in a low-phosphate (0.4 mM) MOPS medium containing serine hydroxamate, an inducer of the amino acid stringent response (Cashel and Rudd, 1987). Strain 8830 grew slowly under such conditions but accumulated large amounts of poly P up to 5 h, after which the level declined (Fig. 3A). In contrast, the non-mucoid PAO1 accumulated about 10% as much as the mucoid strain under similar starvation conditions (Fig. 3A). As for the spontaneous segregant non-mucoid 8822 subjected to serine hydroxamate, poly P accumulation was far less than in the mucoid strain (Fig. 3B), indicating that the loss of alginate production is accompanied by a lack of poly P accumulation. In each case, the level of poly P was much lower (10% or less) when serine hydroxamate was omitted or when the medium was supplemented with L broth, suggesting that conditions leading to poly P accumulation in P. aeruginosa parallel those in E. coli. Also, the levels were found to be variable in different experiments, as shown by the results in 3Fig. 3A and B, presumably because of differences in cell density used in individual experiments and the physiological state of the cells.
Synthesis of both alginate and poly P requires NTPs, the production of which is controlled by Ndk, which in turn is positively regulated by algR2. As with alginate synthesis, the poly P level in the algR2 mutant is extremely low, but is restored to a large extent in the presence of a functional algR2 gene (pJK662) (Fig. 3B). This synthesis continued until 4–5 h, as shown previously for the wild type, and reached values approaching 65–70% of the wild-type poly P level (data not shown). It is not clear why the poly P level in the algR2-complemented strain never reached a value similar to that of the wild-type strain 8830. It is, however, clear that both alginate synthesis and poly P accumulation are regulated by the algR2 gene.
For lack of positive regulation of ndk by algR2, the algR2 mutant has very low levels of Ndk. The loss of alginate in the mutant can be restored to 60% of the wild-type level by cloning ndk under the tac promoter and inducing enzyme synthesis with IPTG (Sundin et al., 1996a). Likewise, overexpression of the ndk gene restored the poly P level to about 60% of the wild-type level (Fig. 4), suggesting that the level of NTPs generated by Ndk determines the accumulation of both poly P and alginate.
GTP and ppGpp synthesis is co-regulated by algR2 and ndk genes
In order to understand how Ndk may promote both alginate and poly P synthesis, we determined the levels of GTP and (p)ppGpp in P. aeruginosa 8830. This is because GTP is required for the synthesis of GDP-mannose, a precursor of alginate (Fig. 1) (May and Chakrabarty, 1994), and GTP itself is a precursor of (p)ppGpp (Cashel, 1994), the accumulation of which leads to poly P accumulation (Kuroda et al., 1997). Nucleoside diphosphate kinase is known to be involved in preferential GTP synthesis in the cell, both in P. aeruginosa and in other microorganisms, such as Mycobacterium smegmatis, through complex formation with specific proteins, such as pyruvate kinase, a Ras-like protein Pra or the elongation factor Tu (Sundin et al., 1996a; Chopade et al., 1997; Mukhopadhyay et al., 1997; Shankar et al., 1997). Sundin et al. (1996a) previously reported that the membrane fractions of the algR2 ::Cm null mutant generate only 10–15% of the GTP generated by the wild type. To measure the level of (p)ppGpp, we grew wild-type strain 8830, its algR2 ::Cm null mutant and the complemented strains algR2 ::Cm with either the ndk gene under the tac promoter (pGWS95) or the algR2 gene under the tac promoter (pJK662) in a low-phosphate (0.4 mM) MOPS medium containing [32P]-orthophosphate (Pi) and induced with serine hydroxamate, as described previously (Kuroda et al., 1997). After 4 h of growth, the cells were harvested at 1 h and 4 h, suspended in cold 10 M formic acid, frozen in liquid nitrogen and thawed, centrifuged, and the levels of GTP, (p)ppGpp and poly P were determined in the supernatant after thin-layer chromatography (TLC) separation on polyethyleneimine-cellulose sheets. The results with poly P are shown in 3Figs 3B and 4, while the results with GTP and (p)ppGpp are shown in Fig. 5 and Table 1. The ppGpp levels, shown in Table 1, were determined by autoradiography followed by PhosphoImager scanning as described previously (Kuroda and Kornberg, 1997). As can be seen from Table 1, the ppGpp levels were significantly reduced in the algR2 mutant, but were restored to near the wild-type level by complementation with either the algR2 gene (pJK662) or the ndk gene (pGWS95). Thus, the level of ppGpp, like the alginate, poly P and GTP levels, is positively regulated by the algR2 or ndk gene, presumably through enhanced synthesis of the precursor GTP.
Table 1. . Level of ppGpp in 8830 and its algR2 mutant complemented by ndk or algR2. Cells were grown in low-phosphate medium and induced with IPTG for 60 min in the presence of serine hydroxamate before being extracted for ppGpp measurement, as described by Cashel (1994) and under Experimental procedures. PGWS95 is a plasmid with the ndk gene and pJK662 with the algR2 gene, both under the tac promoter.
algR2 expression responds to phosphate starvation
To determine whether algR2 is regulated by phosphate, an algR2–lacZ transcriptional fusion (in which the lacZ expression depends on the algR2 promoter) proved to be responsive to phosphate starvation (Fig. 6). During growth in a basal synthetic medium with low (0.1 mM) phosphate, the algR2 promoter was activated and remained active throughout the stationary phase of growth. This activation was absent in a phoB mutant (which is unresponsive to low Pi) or in the presence of high (2 mM) phosphate. It is notable that the increased activity of the algR2 promoter is not merely the result of cessation of growth, because then an increase would also be seen in the phoB mutant, which is not the case. Furthermore, we have tested algR2–lacZ activity under conditions of carbon and nitrogen starvation, in which only a small increase in activity is seen. Thus, under prolonged phosphate starvation, which results in low NTP levels, the algR2 promoter is activated to generate higher levels of AlgR2, which in turn stimulates Ndk and Scs formation; elevated levels of Ndk and Scs then raise NTP levels to overcome the effects of starvation. AlgR2, therefore, appears to play an additional role in restoring the energy balance during periods of phosphate starvation. A similar induction of ppGpp formation by phosphate starvation in E. coli has also been reported (Spira et al., 1995). It is interesting to note that the algR2 promoter activation under phosphate starvation is abolished in a phoB mutant. Visual examination does not reveal a Pho box in the algR2 promoter region, suggesting that the effect of PhoB on algR2 expression is likely to be indirect. Given the regulation of algR2, a key regulator of alginate synthesis, by PhoB, it would be of interest to construct a phoB mutant of the mucoid strain 8830 to identify directly any role of PhoB in alginate synthesis.
algR2 and ndk contribute to stationary phase survival
Inasmuch as algR2 promotes the expression of ndk, it may upregulate the levels of GTP and ppGpp and, hence, poly P. Furthermore, poly P accumulation is defective in the algR2 mutant, and this defect can be overcome by complementation by ndk (Fig. 4). The fact that poly P is essential for the survival of E. coli in the stationary phase (Rao and Kornberg, 1996) prompted the study of the viability of the algR2 mutant. As in E. coli, survival of the mutant in stationary phase is greatly diminished, a defect that in this case can be overcome by plasmids bearing either the algR2 or ndk genes (Fig. 7). Thus, the levels of GTP, ppGpp or poly P may singularly or collectively promote the survival of P. aeruginosa in the stationary phase.
Differential synthesis of GTP, NTPs and dNTPs by the 12 kDa form of Ndk
In P. aeruginosa 8830, the 16 kDa cytoplasmic Ndk produces all the NTPs and dNTPs during the exponential phase of growth, while the 12 kDa membrane form produces predominantly GTP at the onset of the stationary phase (Shankar et al., 1996). Both RNA and DNA syntheses slow down at the onset of stationary phase, caused perhaps in part by reduced levels of NTPs and dNTPs.
To determine whether the truncated 12 kDa form of Ndk, complexed with Pk, has an altered capacity to synthesize dNTPs, the specificities of the complex were compared with those of the membrane fraction (Fig. 8). The membrane fraction makes predominantly GTP, an activity inhibited by anti-Ndk antibody (Fig. 8). Anti-Pk antibody changes the specificity from GTP to all three NTPs. With regard to the synthesis of dNTPs, the membrane fraction has very little activity to begin with and is unaffected by anti-Ndk or anti-Pk antibodies (Fig. 8). The reconstituted complex between purified 12 kDa Ndk and Pk resembles activities found in the membrane fraction (Fig. 9). Pk and the 16 kDa Ndk can each generate NTPs and dNTPs from NDPs, dNDPs and [γ-32P]-ATP; however, Pk requires PEP and Mn2+ for this activity (Sundin et al., 1996a). Because the 16 kDa Ndk fails to form a complex with Pk, the addition of Pk had no effect (Fig. 9). However, with the 12 kDa form and adequate levels of Pk, the resulting complex was ineffective in the formation of dNTPs (Fig. 9); the complex has reduced specificity for dNDPs compared with either the 12 kDa Ndk or Pk separately.
AlgR2 positively regulates nucleoside diphosphate kinase (NdK) synthesis and, as a result, may modulate the levels of GTP, ppGpp, alginate and poly P. Alginate synthesis requires a supply of GTP, as does (p)ppGpp; the latter promotes the accumulation of poly P (Kuroda et al., 1997). The conversion of the 16 kDa cytoplasmic form of Ndk to the 12 kDa membrane-associated form at the onset of stationary phase (Shankar et al., 1996) promotes the synthesis of GTP and, in turn, ppGpp (Cashel, 1994). The latter induces expression of the stationary phase sigma factor, σS, in E. coli (Gentry et al., 1993), which controls over 40 genes essential for stationary phase resistance and survival (Loewen and Hengge-Aronis, 1994). However, the effect of ppGpp on the level of RpoS in P. aeruginosa is unknown at present. As the membrane-associated 12 kDa Ndk appears at the onset of the stationary phase (Shankar et al., 1996; Chopade et al., 1997) and makes GTP preferentially and very little dNTPs (Figs 8 and 9), the precursor pools of RNA and DNA are probably reduced with a consequent slowdown in RNA and DNA synthesis. In addition to Pk, a Ras-like protein, Pra, also appears in the membrane at the stationary phase and forms a complex with 12 kDa Ndk, directing its specificity towards GTP synthesis (Chopade et al., 1997). Apart from 12 kDa Ndk, Pra can also form a complex with Pk, modulating its specificity from NTP synthesis to GTP (Chopade et al., 1997). In the presence of 12 kDa Ndk, however, Pk has a higher affinity for 12 kDa Ndk, so that the Pk–Pra complex dissociates. In the algR2 mutant, which has very little Ndk, it is Pk that allows NTP synthesis, and it is likely that the Pk–Pra complex generates the GTP and very little dNTP for transition to stationary phase (Chopade et al., 1997).
The co-regulation of two polymers, alginate and poly P, by a common regulator, such as AlgR2, is interesting. The mucoid Alg+ cells are rich in poly P, while the non-mucoid segregant or the non-mucoid laboratory isolate produces much less poly P. Thus, the genetic switch that turns on alginate synthesis in P. aeruginosa also turns on poly P synthesis, implying a common mode of genetic regulation. As hyperexpression of the ndk gene largely restores alginate, GTP, ppGpp and poly P synthesis in the algR2 mutant (Table 1, Figs 3 and 4[link]), Ndk plays a critical role in the synthesis of these two polymers. As AlgR2 positively regulates Ndk and Scs formation and as hyperexpression of the ndk gene can compensate for AlgR2 deficiency, it is likely that the phenotypes demonstrated by the algR2 mutant are largely caused by the reduced formation of Ndk. Although P. aeruginosa is normally non-mucoid, it is converted to a mucoid form under several conditions: in the lungs of CF patients (Shankar et al., 1995), during growth under starvation conditions and in the presence of inhibitors of energy metabolism (Speert et al., 1990; Woods et al., 1991; Terry et al., 1992). Starvation conditions, particularly for phosphate, allow an activation of the algR2 promoter, leading to enhanced Ndk synthesis and, presumably, an enhanced formation of GTP, pppGpp and ppGpp. High concentrations of pppGpp or ppGpp are known to lead to the accumulation of poly P (Kuroda et al., 1997). Indeed, the poly P accumulation in P. aeruginosa is maximal in the stationary phase (Fig. 2) when the membrane-associated complexes of Pk and Pra with the 12 kDa Ndk generate large amounts of GTP. Alginate is also known to accumulate predominantly in the late exponential–early stationary phase cells (Tatnell et al., 1993; Hassett, 1996). As alginate synthesis requires both phosphorylated sugars and GTP (Fig. 1) and as poly P may substitute for ATP (Kornberg, 1995), it is possible that poly P can contribute to alginate synthesis. Recent efforts in our laboratory to isolate polyphosphate kinase-negative (ppk ) mutants of stably mucoid P. aeruginosa strain 8830 have produced presumptive mutants, many of which are non-mucoid with little PPK activity (A. Zago, unpublished observations). When the cloned ppk gene of P. aeruginosa becomes available, its capacity to restore alginate synthesis in the presumptive ppk non-mucoid mutants can be determined. Given the ubiquitous nature of poly P and its accumulation in the stationary phase (Kornberg, 1995), it can be imagined that poly P, among its many functions, can be used by bacteria to produce capsular polysaccharides for their protection in stressful situations, as for P. aeruginosa in the CF lung. Similarly, isolation of a relA mutant, incapable of converting GTP to pppGpp or ppGpp, in the mucoid strain 8830 may provide insights into the role of these regulatory nucleotides in alginate synthesis.
Bacterial strains and growth
P. aeruginosa 8830 is a stable alginate-producing strain. The organism was routinely propagated in L broth at 37°C. The algR2 ::Cm, which is a null mutant in the algR2 gene because of insertion of a chloramphenicol cassette, was propagated in L broth or in MOPS media for studies of poly P accumulation.
Assay for guanosine tri- and tetraphosphates
GTP and ppGpp production was analysed by one-dimensional chromatography (PEI-cellulose; E. Merck) as described by Cashel (1994). Non-labelled E. coli ppGpp and pppGpp were used as standards and detected by UV light. The wild-type strain 8830, the algR2 null mutant and the algR2 mutant complemented by either the algR2 or the ndk gene under the tac promoter were grown from an OD540 of 0.05 to an OD of 0.25 in a low-phosphate (0.4 mM) MOPS medium containing [32P]-orthophosphate. Aliquots were transferred to cold 10 M formic acid, frozen in liquid nitrogen, and the thawed lysates were examined for GTP and (p)ppGpp levels, as described by Cashel (1994) and shown in Fig. 5. The authenticity of the (p)ppGpp was determined by its susceptibility to purified yeast exopolyphosphatase (PPX), which converts it to GTP and Pi, all of which can be separated on polyethyleneimine TLC plates as described earlier (Kuroda and Kornberg, 1997).
Quantitative measurement of poly P
Assays according to S. Liu and A. Kornberg (unpublished) used overnight cultures grown with [32P]-Pi in a synthetic medium. Cells were harvested and lysed in a FUSE buffer (3 M formic acid, 2 M urea, 20 mM EDTA, 1% SDS, 1 μg μl−1 polyphosphate 65) and sonicated briefly. Poly P in the lysate was adsorbed on a DE81 filter disc, washed with a TKP buffer (10 mM Tris-HCl, pH 8, 100 mM KCl, 5 mM K3PO4) and eluted with TKP buffer containing 500 mM KCl. Poly P in the eluant was further enriched by removal of 32P contaminants with Norit (activated charcoal), then readsorbed to a DE81 filter disc, washed and treated with yeast PPX (Kuroda and Kornberg, 1997) in 40 mM Tris-HCl, pH 7.4, 100 mM ammonium acetate/4 mM MgCl2, 10 mg ml−1 BSA and 5 μg ml−1 PPX. The amount of poly P was determined by the loss of 32P from the filter. This loss was subsequently correlated with the appearance of [32P]-Pi in the supernatant, and the level of poly P was also occasionally determined by averaging the loss of 32P from the filter and the appearance of [32P]-Pi in the eluant.
Plasmid pDS25 (algR2–lacz ) was introduced into strain PAO1 and its phoB mutant (Shortridge et al., 1992). The strains were grown in MOPS minimal media (Neidhardt et al., 1974), containing either 2 mM or 0.1 mM Pi. Strains were grown in 1 l of media with 100 μg ml−1 carbenicillin in 2 l flasks at 37°C with shaking at 250 r.p.m. β-Galactosidase activity was determined according to Miller (1972). Protein was determined using the Bio-Rad Protein Assay kit with BSA as a standard.
Assay for NTP and dNTP synthesis by membrane fractions or by combinations of Ndk and Pk