The Kdp system of Escherichia coli is composed of the high-affinity K+ transporter KdpFABC and the two regulatory proteins KdpD (sensor kinase) and KdpE (response regulator), which constitute a typical two-component system. The kdpFABC operon is induced under K+-limiting conditions and, to a lesser extent, under high osmolality in the medium. In search for the stimulus sensed by KdpD, we studied the inhibitory effect of extracellular K+ on the Kdp system at pH 6.0, which is masked by unspecific K+ transport at higher pH values. Based on KdpD derivatives carrying single aspartate replacements in the periplasmic loops which are part of the input domain, we concluded that the inhibition of the Kdp system at extracellular K+ concentrations above 5 mM is mediated via KdpD/KdpE and not due to inhibition of the K+-transporting KdpFABC complex. Furthermore, time-course analyses of kdpFABC expression revealed that a decline in the extracellular K+ concentration efficiently stimulates KdpD/KdpE-mediated signal transduction. In this report we provide evidence that the extracellular K+ concentration serves as one of the stimuli sensed by KdpD.
K+ is the predominant monovalent cation in the cytoplasm of most living cells, including bacteria (Epstein, 2003). In case of Escherichia coli, K+ serves several important functions: it is involved in the maintenance of cell turgor (Epstein, 1986), in the activation of cytoplasmic enzymes, in pH homeostasis (Booth, 1985; 1999) and, together with glutamate, it is also known to positively regulate the expression of certain osmoresponsive genes (Lee and Gralla, 2004; Gralla and Vargas, 2006). In order to fulfil this important role for cellular function, E. coli has established three transport systems for K+, which operate under a wide range of different environmental conditions: the constitutively expressed, low-affinity K+ transport systems Trk and Kup (Km = 1.5 mM) and the inducible KdpFABC complex with a high affinity for K+ (Km = 2 μM). The latter serves as an emergency system to scavenge K+ when the external K+ concentration is too low to satisfy the requirement of the cell via the Trk and Kup systems. Expression of the kdpFABC operon encoding the KdpFABC complex is regulated by the sensor kinase KdpD and the response regulator KdpE, constituting a typical two-component system (Walderhaug et al., 1992; Altendorf et al., 2009). ‘TrkF’ (unspecific K+ transport) represents the residual K+ transport activity in cells that have mutations in the three K+ uptake systems (Buurmann et al., 2004). This activity, completely uncharacterized on molecular level, is just the sum of all so far not identified ways of K+ entering the cell i.e. paths that, regardless of their physiological function, allow K+ uptake by a uniport process (Buurmann et al., 2004). The rate of unspecific K+ transport is linearly proportional to the external K+ concentration and increases with increasing external pH (Buurmann et al., 2004). Therefore, the authors proposed that the driving force for the unspecific K+ transport is the transmembrane electrical potential (ΔΨ). It is generally known that the transmembrane electrical potential together with the chemical proton gradient (ΔpH) constitute the proton motif force (pmf). At neutral external pH values the pmf is made up almost entirely by ΔΨ. During acidification of the environment ΔΨ is interconverted into ΔpH to maintain a constant pH of about 7.7 in the cytoplasm (Padan et al., 1976). This interconversion requires electrogenic K+ uptake to compensate for proton extrusion (Bakker and Mangerich, 1981). Therefore, cells lacking all saturable K+ uptake systems fail to maintain a constant cytoplasmic pH in an acidic environment and consequently their growth rate decreases (White et al., 1992).
The two-component system KdpD/KdpE induces kdpFABC expression under K+-limiting growth conditions and, to a lesser extent, in an environment of high osmolarity (Rhoads et al., 1976; Laimins et al., 1981). A variety of environmental conditions including the pH, growth temperature and the concentration of other cationic solutes are known to additionally modulate the strength of the stimulus perceived by KdpD (Asha and Gowrishankar, 1993). The hypothesis that KdpD senses changes in turgor (Laimins et al., 1981; Malli and Epstein, 1998) was challenged by the fact that iso-osmolal concentrations of non-ionic solutes did not induce the same kdpFABC expression levels as seen with NaCl and, additionally, KCl did not induce kdpFABC expression at all. Furthermore, the level of kdpFABC expression of cells subjected to K+-limiting conditions or to high concentrations of NaCl does not correlate with changes in the cytoplasmic volume of E. coli (Hamann et al., 2008). Recently, it was shown that not changes in turgor, but binding of UspC to the input domain of KdpD enhances kdpFABC expression under salt stress, but not under K+-limiting conditions (Heermann et al., 2009). However, the finding that kdpFABC expression is only induced by ionic solutes, with the exception of K+, remains enigmatic.
Although an inhibitory effect of high extracellular K+ concentrations on the Kdp system has clearly been shown for E. coli (Roe et al., 2000) and Salmonella typhimurium (Frymier et al., 1997), the extracellular K+ concentration was excluded as stimulus for KdpD, since a derivative lacking the four transmembrane helices still induces high levels of kdpFABC expression (Heermann et al., 2003a). Furthermore, the level of kdpFABC expression at a given K+ concentration depends on the K+ transport systems present, whereas the threshold of the extracellular K+ concentration at which kdpFABC expression is induced is generally higher in a strain inactivated in the Trk and Kup systems compared with a wild-type strain (Laimins et al., 1981). Therefore, the stimulus that is sensed by KdpD has often been described as the cells' need for K+ (Laimins et al., 1981; Malli and Epstein, 1998) implying that kdpFABC expression is induced when the cells' requirement for K+ is not fully satisfied by the K+ transport systems Trk and Kup. Accordingly, mainly intracellular parameters were suggested to be sensed by KdpD. Based on truncated versions of the sensor kinase KdpD it was postulated that intracellular K+ is sensed by the C-terminal domain of KdpD (Rothenbücher et al., 2006). In vitro assays based on right-side-out membrane vesicles have shown that KdpD kinase activity is inhibited by K+ concentrations higher than 1 mM (Jung et al., 2000). However, the hypothesis that changes in the cytoplasmic K+ concentration are the stimulus for KdpD were challenged by the fact that the cytoplasmic K+ concentration does not change within the first minutes upon subjecting E. coli to K+-limiting conditions, although kdpFABC expression is already very high (Hamann et al., 2008). In this report we provide evidence that the stimulus sensed by KdpD cannot solely be attributed to intracellular parameters but also to the extracellular K+ concentration.
It is generally known that the level of kdpFABC expression is higher in a strain lacking Trk and Kup compared with a wild-type strain, in particular at moderate extracellular K+ concentrations ranging from 10 to 50 mM, giving rise to the assumption that mainly intracellular parameters serve as stimulus for KdpD. Since the aim of this study was to analyse stimulus perception of KdpD at higher extracellular K+ concentrations, all strains used in this study are inactivated in the genes coding for the constitutively expressed K+ uptake systems Trk and Kup. However, it is important to notice that K+ uptake in such a strain cannot solely be attributed to the KdpFABC complex, but is also due to unspecific K+ uptake (‘TrkF activity’), the latter being largely dependent on the pH of the environment (Buurmann et al., 2004).
Unmasking the inhibition of the Kdp system of E. coli by extracellular K+
To distinguish between unspecific and KdpFABC-mediated K+ uptake, E. coli strains TK2281 (Δkdp) defective in the three saturable K+ uptake systems Trk, Kup and Kdp, and TK2240 (kdp+) only defective in the Trk and Kup uptake systems were used. Strains TK2281 and TK2240 were grown on minimal medium agar plates with defined K+ concentrations at pH 6.0 and pH 7.8 respectively. In parallel, we measured the cytoplasmic K+ concentration and pH of both strains cultivated in liquid minimal medium under the same growth conditions. The cytoplasmic pH of TK2240 and TK2281 was measured using the pH-sensitive derivative GFPmut3 variant. The cytoplasmic pH was calculated from fluorescence intensity using standard curves obtained from cell suspensions containing 30 mM benzoate, which equalizes the cytoplasmic pH with the external pH (Wilks and Slonczewski, 2007). To test the influence of the given growth conditions on the regulation of kdpFABC expression in TK2240, a transcriptional fusion of the region upstream of the kdpFABC operon to promoterless lacZ gene in plasmid pWPA1lacZ was used as quantitative reporter for kdpFABC expression.
At pH 7.8, TK2281 maintained a high cytoplasmic pH irrespective of the extracellular K+ concentration (Table 1). Under these conditions ΔpH is minimal and consequently Δψ and unspecific K+ uptake is maximal allowing maintenance of a sufficiently high cytoplasmic K+ concentration and growth of TK2281 at K+ concentration above 10 mM (Fig. 1A and B). Due to the fact that the rate of unspecific K+ uptake is linearly proportional to the extracellular K+ concentration, the cytoplasmic K+ concentration and growth of TK2281 were slightly reduced at 10 mM K+ and severely affected at 1 mM K+ (Fig. 1A and B). At an external pH of 6.0, ΔΨ is to a great extent converted to ΔpH through electrogenic influx of K+ (Bakker and Mangerich, 1981). For TK2281, the lowered ΔΨ caused a decrease in unspecific K+ uptake preventing further interconversion of ΔΨ and ΔpH. Therefore, the generation of ΔpH and subsequent net K+ uptake via TrkF are mutually exclusive in TK2281. As a consequence, the cytoplasmic K+ concentration at a given extracellular K+ concentration was significantly lower at pH 6.0 compared with pH 7.8 (Fig. 1B). Additionally, the cytoplasmic pH was almost equal to that of the minimal medium with 0.1 mM K+ and progressively increased with increasing unspecific K+ uptake (Table 1). Accordingly, no growth of TK2281 was observed at 30 mM K+ or less (Fig. 1A). Growth and the cytoplasmic K+ concentration slightly increased at 50 mM K+ and a high K+ concentration of 115 mM could compensate for the lowered ΔΨ restoring a sufficiently high rate of unspecific K+ uptake to allow pH homeostasis and, consequently, growth.
Table 1. Influence of K+ uptake on the ability to maintain pH homeostasis
E. coli TK2240 (kdp+ Δtrk Δkup) and E. coli TK2281 (Δkdp Δtrk Δkup) transformed with the GFPmut3b reporter-plasmid pMMB1311 were pre-cultured in K115 at pH 7.2 to exponential growth phase (optical density, OD600 = 0.4). Cells were washed and resuspended in minimal medium with defined K+ concentrations adjusted to pHex of 6.0 and 7.8 respectively. After 1 h of incubation, cell suspensions were diluted to OD600 = 0.4 and the cytoplasmic pH (pHin) was determined as described in Experimental procedures. The data represent average values and standard deviations obtained in three independent experiments.
6.27 ± 0.10
6.72 ± 0.13
7.13 ± 0.15
7.36 ± 0.06
7.71 ± 0.08
7.05 ± 0.10
7.11 ± 0.10
7.28 ± 0.03
7.63 ± 0.08
7.77 ± 0.06
7.77 ± 0.10
7.71 ± 0.03
7.85 ± 0.10
7.75 ± 0.09
7.80 ± 0.09
7.78 ± 0.03
Comparable to TK2281, TK2240 maintained a high cytoplasmic pH irrespective of the extracellular K+ concentration (Table 1) at an external pH of 7.8. However, in contrast to TK2281, growth of TK2240 was not affected at the whole range of K+ concentrations tested (Fig. 1A). At the lowest K+ concentrations of 1 mM and 10 mM, TK2240/pWPA1lacZ exhibited high β-galactosidase (β-gal) activities (Fig. 1C) and the high kdpFABC expression compensated the low unspecific K+ uptake (Fig. 1B). With increasing K+ concentrations β-gal activities decreased and unspecific K+ uptake increased. Consequently, the cytoplasmic K+ concentration of TK2240 remained nearly constant irrespective of the extracellular K+ concentration at pH 7.8 (Fig. 1B). At pH 6.0, TK2240/pWPA1lacZ showed maximal β-gal activity at 1 mM K+ (Fig. 1C). The high level of kdpFABC expression caused a high cytoplasmic K+ concentration of about 450 nmol mg−1 dry weight (Fig. 1B). Accordingly, TK2240 exhibited normal growth and a high cytoplasmic pH under this condition (Fig. 1A and Table 1). The data clearly show that the KdpFABC complex as the only active K+ transporter has generally the ability to supply the cell with sufficient K+. However, β-gal activities of TK2240/pWPA1lacZ gradually decreased with increasing K+ concentrations (Fig. 1C). Due to the fact that unspecific K+ uptake is low at pH 6.0, the decreased levels of kdpFABC expression were accompanied with a decline in the cytoplasmic K+ concentrations and pH values at 10–50 mM K+ (Fig. 1B and Table 1) causing a significant growth inhibition under these conditions (Fig. 1A). Again, a further increase up to 115 mM K+ caused an increase of unspecific K+ uptake, restoring a sufficiently high cytoplasmic K+ concentration and pH and, in consequence, growth (Fig. 1A and B and Table 1). It is worth mentioning that a strain lacking only the Kup system was able to grow at the whole range of conditions tested in this study (data not shown), corroborating the assumption that growth inhibition of TK2240 at 10–50 mM K+ at pH 6.0 is based on a failure of K+ and pH homeostasis, in accord with decreased levels of kdpFABC expression. This indicates that an insufficient amount of KdpFABC complex is the cause of the given phenotype. It is important to note that the inhibitory effect of K+ on kdpFABC expression is not dependent on the extracellular pH, but masked by unspecific K+ uptake (TrkF) at higher pH values. However, besides the fact that expression of kdpFABC was reduced at increasing extracellular K+ concentrations we cannot rule out that additionally, the activity of the KdpFABC complex itself was affected by high amounts of K+.
It should be noted that β-gal activities at a given K+ concentration were generally higher at pH 6.0 compared with pH 7.8 (Fig. 1C). It is known that the pH of the environment modulates the strength of the stimulus perceived by KdpD (Asha and Gowrishankar, 1993). However, kdpFABC expression is still regulated in a K+-dependent manner at pH 7.8 (Fig. 1C), although the cytoplasmic and extracellular pH remain constant. Therefore, it seems unlikely that KdpD senses pH per se.
Nature of the inhibition of the Kdp system
As shown above, kdpFABC expression gradually decreased with increasing extracellular K+ concentration, although the cells need for K+ is constantly high at K+ concentrations ranging from 10 to 50 mM at pH 6.0. This implies that induction of kdpFABC expression is repressed by an extracellular signal, e.g. the extracellular K+ concentration. Consequently, KdpD must be able to sense the extracellular K+ concentration and, therefore, the corresponding sensor domain must be located in the periplasmic space.
KdpD, active as a dimer, has two short periplasmic loops consisting of 9 and 13 amino acids respectively. In order to search for a potential K+ sensor domain, we focused on two aspartate residues, each located in one periplasmic loop. The carboxylate side-chain of aspartate residues often serves as donor group for the octahedral co-ordination of K+ in respective binding sites (Harding, 2002). Therefore, site-directed mutagenesis was performed on both aspartate residues (D424, D474). Two different groups of mutants were obtained exhibiting an altered phenotype concerning the growth inhibition at higher extracellular K+ concentrations at pH 6.0. The first group of mutants exhibited an increased sensitivity towards the inhibitory effect of extracellular K+ on the Kdp system compared with the wild-type and these mutations were exclusively located in the first periplasmic loop (D424Y, D424C, D424N). In contrast, the second group of mutants showed a decrease in sensitivity towards the K+-mediated inhibition, mainly comprising aspartate substitutions in the second periplasmic loop (D474A, D474C, D474Y) with the exception of D424A. It is worth mentioning that a third group of mutants retained the ability to respond to increasing K+ concentrations like the KdpD wild-type (KdpD-WT) protein (D424E, D424G, D474E, D474N). In the following, the results typical for one KdpD derivative of each group were presented. At pH 6.0, cells expressing either the KdpD-WT or the derivatives KdpD-D474A and KdpD-D424Y exhibited normal growth on agar plates containing a low K+ concentration of 1 mM (Fig. 2), indicating that the mutations in KdpD do not affect KdpD/KdpE signal transduction or stimulus perception in general. With increasing K+ concentrations up to 50 mM growth of cells synthesizing KdpD-D424Y was more strongly affected than the growth of the KdpD-WT cells. At 115 mM K+ growth was restored by increasing unspecific K+ uptake. The cytoplasmic K+ concentration was measured in parallel to the doubling time of cells expressing KdpD-WT or the derivatives in liquid minimal medium under the same conditions as described above (Fig. 3). The general pattern of cells expressing KdpD-WT or derivative KdpD-D424Y showed a negative correlation between the doubling times and cytoplasmic K+ concentrations at extracellular K+ concentrations ranging from 1 mM to 115 mM. At 1 mM K+ and 115 mM K+, the doubling times were low and the cytoplasmic K+ concentrations were in a physiological range of about 400 nmol mg−1 dry weight. The increasing doubling times of the KdpD-WT and the derivative KdpD-D424Y at increasing K+ concentrations up to 50 mM were accompanied by a severe decline of the intracellular K+ concentration (Fig. 3). It should be noted that the doubling times as well as the cytoplasmic K+ concentration were more strongly affected in cells expressing KdpD-D424Y than in the KdpD-WT cells.
Cells expressing the derivative KdpD-D474A showed no growth inhibition and the cytoplasmic K+ concentration hardly changed at all over the whole range of K+ concentrations tested (Fig. 3). Since a strain lacking the KdpFABC complex as well as the Trk and Kup systems was not able to grow at 1–50 mM K+ at pH 6.0 (Fig. 1A and B), a prerequisite for growth of cells expressing KdpD-D474A is active K+ uptake via the KdpFABC complex, demonstrating that the activity of the KdpFABC complex is not affected by high amounts of K+.
Influence of aspartate substitutions in KdpD on the regulation of kdpFABC expression
To test the influence of the aspartate substitutions in the periplasmic loops of KdpD on the K+ dependence of kdpFABC expression, again plasmid pWPA1lacZ was used as quantitative reporter for kdpFABC expression. The β-gal assays were performed in E. coli strain TKV2209 defective in trk and kup and deleted in the kdpDE operon. The kdpDE deletion was complemented using plasmid pPV-2 and its derivatives pPV-2 D424Y and pPV-2 D474A coding for KdpD-WT and KdpE-WT, KdpD-D424Y and KdpE-WT or KdpD-D474A and KdpE-WT respectively. E. coli TKV2209 was co-transformed with plasmids pWPA1lacZ and pPV-2 or its pPV-2 derivatives. The phenotypes of the transformants regarding the growth inhibition at high extracellular K+ concentrations were in agreement with the data presented in Fig. 2. Cells were grown under the same conditions as described above, e.g. extracellular K+ concentrations ranging from 1 to 115 mM at pH 6.0. To confirm that the mutations did not alter induction of kdpFABC expression at low K+ concentrations, an extracellular K+ concentration of 0.1 mM was also included in the β-gal assay.
For KdpD-WT or derivatives KdpD-D424Y and KdpD-D474A maximal and comparable kdpFABC promoter (PkdpFABC)-mediated β-gal activities were obtained at the lowest K+ concentration of 0.1 mM (Fig. 4). The general pattern for KdpD-WT and the derivatives showed continuously decreasing PkdpFABC-mediated β-gal activities with increasing extracellular K+ concentrations, whereas β-gal activities were in the same range at 0.1 mM K+ and significantly diverged at K+ concentrations > 1 mM. KdpD-WT induced PkdpFABC-promoted β-gal activities at K+ concentrations lower than 50 mM. Half-maximal β-gal activity was observed at 5–10 mM K+, the same range at which the growth inhibition became evident in Fig. 2. Compared with KdpD-WT, PkdpFABC-governed β-gal activities of the derivative KdpD-D424Y were much lower at K+ concentrations above 1 mM. Furthermore, β-gal activity was reduced to approximately 50% at 5 mM K+, again the same concentration at which growth inhibition of cells expressing KdpD-D424Y became apparent in Fig. 2. Compared with KdpD-WT, PkdpFABC-mediated β-gal activities of KdpD-D474A remained on a high level at K+ concentrations up to 30 mM. β-gal activity was reduced to approximately 50% at 50 mM K+ and only at a high K+ concentration of 115 mM total inhibition of β-gal activity was observed. The results obtained for cells expressing the derivative KdpD-D474A showed that an adequately high induction of kdpFABC expression prevented cellular K+ starvation and growth inhibition at K+ concentrations of 5–50 mM at pH 6.0 (Fig. 3). These results confirm that the growth inhibition of cells expressing the KdpD-WT protein is based on insufficient induction, or in other words, progressive repression of kdpFABC expression via the sensor kinase KdpD.
Time-course of kdpFABC expression at different extracellular K+ concentrations
Based on the knowledge of the direct influence of the extracellular K+ concentration on KdpD kinase activity, we reexamined kdpFABC expression at different K+ concentrations in more detail. Since β-gal activities allow only rough estimations of kdpFABC expression, Q-RT-PCR was used as a direct approach to quantify kdpFABC expression. E. coli strain TK2240 was grown under kdp non-inducing conditions (K60 minimal medium) to exponential growth phase. An immediate change of the extracellular K+ concentration was obtained by filtration, resuspension and subcultivation of the cells in fresh medium with defined K+ concentrations (K0.02, K0.1, K1, K5 and K10). As a control, cells were also shifted into growth medium containing the same K+ concentration in order to identify disturbances due to the filtration procedure, which were negligible. To measure the time-course of kdpFABC expression following the shift in the extracellular K+ concentration, samples were taken at the times indicated (Fig. 5A) and normalized kdpA mRNA levels were quantified by Q-RT-PCR as representative for expression of the kdpFABC operon.
The time-resolved induction profile of cells exposed to a shift in the extracellular K+ concentration revealed unexpected dynamics of kdpFABC expression. The signal transduction of the KdpD/KdpE system responds very fast to a sudden decrease of the extracellular K+ concentration. Thirty seconds after the shift, each of the five subcultures already exhibited a high level of kdpFABC expression (Fig. 5A). In each case, after reaching maximal expression levels a few minutes after the K+ downshifts, the amount of the kdpFABC transcripts decreased and finally remained on a basal level. It should be noted that the basal level of kdpFABC expression in the five subcultures was still higher compared with the expression level before the shift. However, the time period observed between high kdpFABC expression following the K+ downshift and the basal level reached, differed in the five subcultures in dependence on the differences in the extracellular K+ concentration before and after the shift; the lower the K+ concentration of the subculture, the longer the period of high kdpFABC expression. Cells grown in K0.02 and K0.1 medium showed a second increase in kdpFABC expression after 60 min and 80 min, respectively, whereas the amount of kdpFABC transcript in the cultures supplemented with higher K+ concentrations remained on the basal level (Fig. 5A). Growth of E. coli in K0.02 and K0.1 medium was restricted by the limited amount of K+. Cells grown in K+-limited medium continuously decreased the extracellular K+ concentration, until growth ceased due to K+ starvation. Accordingly, the determination of the extracellular K+ concentration at the times indicated revealed that cells subcultured in K0.02 medium and K0.1 medium completely depleted the extracellular K+ pool within the first 120 min after the shift (Fig. 5B, closed symbols), whereas only a slight reduction in the K+ concentration in the K1 subculture and no significant reduction in the K+ concentration in the K5 and K10 subculture could be observed (Fig. 5B, open symbols). Therefore, in contrast to cells subcultured in K1, K5 and K10, those subcultured in K+-limited media (K0.02, K0.1) are continuously exposed to a decline in K+ caused by the accumulation of the ion during growth. These results clearly demonstrate that a decline in the extracellular K+ concentration efficiently stimulates the expression of kdpFABC.
There is evidence that intracellular parameters serve as stimulus for KdpD leading to the assumption that KdpD senses the cells need for K+ (Laimins et al., 1981; Malli and Epstein, 1998). However, the data presented in this work revealed that in a kdp wild-type strain lacking trk and kup, the kdpFABC expression levels observed at K+ concentrations ranging from 10 to 50 mM continuously decrease with increasing extracellular K+ concentrations, although the cells' need for K+ is high under these growth conditions at pH 6.0. From the results obtained, we conclude that the extracellular K+ concentration is sensed by KdpD antagonizing the activating effect of the postulated intracellular stimulus. The sensitivity of the sensor kinase KdpD towards higher extracellular K+ concentrations can be altered by single amino acid substitutions in the periplasmic loops. It is worth mentioning that the extent of kdpFABC expression in cells expressing either the KdpD-WT or its derivatives is quite comparable in K+-limiting growth medium. Therefore, signal transduction of the mutants is not affected in general and the mutations in KdpD only influence the response of KdpD towards K+ supplementation. It is surprising that replacement of KdpD-D424 and KdpD-D474, respectively, causes opposite phenotypes concerning the influence of the extracellular K+ concentration on kdpFABC expression. Based on the observation that removal of the negative charge does not necessarily cause an altered sensing property of KdpD (D424G, D474N) it seems unlikely that the negative charges of the aspartate residues are involved in the co-ordination of K+ in a potential binding site. However, the backbone carbonyl group of the aspartate residues might be directly part of a K+ binding site, or conformational changes introduced by the mutations might affect the property of a potential proximate K+ binding site. To date it is impossible to identify a putative K+ binding site in the periplasmic loops of the sensor kinase KdpD due to limited structural data. Moreover, K+ binding sites in deposited structures (http://www.rcsb.org/pdb) reveal a strong heterogeneity in distances and number of also non-protein co-ordination partners like water molecules impeding a solid bioinformatical prediction from the primary structure. However, the information obtained by the mutants clearly supports the notion that this region is of structural importance in conferring sensitivity towards extracellular K+. Replacement of aspartate 474 against alanine in the KdpD protein abolishes repression of kdpFABC expression at higher K+ concentrations without affecting the postulated intracellular sensing mechanism. Consequently, the levels of kdpFABC expression in TKV2209/pPV-2-D474A at a given K+ concentration and pH 6.0 reflect perfectly the cells' need for K+: PkdpFABC-mediated β-gal activities are high at K+ concentrations up to 30 mM when unspecific K+ transport is low. The increase in unspecific K+ transport at 50 mM K+ is accompanied with a decrease of PkdpFABC-governed β-gal activity and high unspecific K+ uptake at 115 mM completely represses β-gal activity. This regulation of kdpFABC expression ensures optimal K+ supply of cells expressing KdpD-D474A and the cytoplasmic K+ concentration is in a physiological range of about 400 nmol mg−1 dry weight under all conditions tested in this study.
The assumption that the extracellular K+ concentration influences sensing by KdpD is corroborated by the finding that an increase in the level of kdpFABC expression correlates very well with a decline in extracellular K+ concentration. Stimulus perception and signal transduction of cells that underwent a shift in the extracellular K+ concentration is remarkable fast. Therefore, KdpD must perceive the stimulus immediately after the shift. Similar observations were made by Hamann et al. (2008) for E. coli wild-type strain K-12, in which the fast response of KdpD/KdpE is neither correlated with a loss of cytoplasmic water nor correlated with significant fluctuations in the cytoplasmic concentrations of K+ or other compounds (i.e. ATP, glutamate, proline, trehalose). Therefore, we propose that the shift in the extracellular K+ concentration itself stimulates kdpFABC expression. First, magnitude and duration of kdpFABC expression depends on the extent of the K+ decrease applied to the cells by the shift, whereas a shift into minimal medium containing the lowest K+ concentration of 0.02 mM provokes the longest period of kdpFABC expression. Second, cells shifted into K+-limiting growth medium (0.02 mM and 0.1 mM K+) showed a second increase in kdpFABC expression correlating with a decline in the extracellular K+ concentration caused by K+ accumulation of the cells during growth. The share in K+ accumulated by the cells until reaching stationary phase depends on the initial K+ concentration of the growth medium: cells grown in media containing a K+ concentration of 0.1 mM and 0.02 mM accumulate K+ completely and enter stationary phase early due to cellular K+ starvation. In contrast, cells grown in medium with 1 mM K+ accumulate about 25% K+ and cells grown in media containing 5 mM K+ or higher do not significantly alter the K+ concentration of the growth medium until reaching stationary phase. It is well established that kdpFABC expression is highly induced under K+-limiting growth conditions, but the question remained how the cells are able to sense the K+-limited state of the environment; we suggest that the low amount of external K+ is sensed by KdpD. Induction of kdpFABC expression as an anticipatory response to a decline in the extracellular K+ concentration enables the cell to adapt to an impending K+ starvation before the stress enters the cell. However, due to the fact that growth of E. coli in media with a low K+ concentration of 0.02 mM or 0.1 mM completely depletes the extracellular K+ pool, preventing further K+ uptake, the second increase in kdpFABC expression of cells grown under these conditions is conform with sensing extracellular K+ as well as intracellular parameters changing under cellular K+ starvation. This is in accord with the finding that long-term starvation for K+ leads to a decrease in the cytoplasmic K+ concentration (Weiden et al., 1967). Analysis of the K+ pool of K+ starving cells is under way in our laboratory.
The observed inhibitory effect of extracellular K+ on the kinase activity of KdpD might also help in explaining the expression of kdpFABC under osmotic stress, whereby NaCl, but not KCl is effective and sucrose shows no effect at all. We propose that high concentrations of Na+ might interfere with the postulated K+ sensor domain in the periplasm, thereby mimicking a decline in the extracellular K+ concentration and preventing K+ mediated inhibition of kdpFABC expression. This is in line with the proposal of Asha and Gowrishankar (1993) that Na+-mediated induction of kdpFABC expression does not solely depend on the osmotic property of this solute.
In contrast to the results presented in this study, Heermann et al. (2003a) reported that the transmembrane domains including the periplasmic loops are not essential for stimulus perception of KdpD, since a derivative lacking the four transmembrane domains still exhibits high levels of kdpFABC expression under K+-limiting conditions. However, the truncations abolished repression of kdpFABC expression at higher extracellular K+ concentrations causing constitutive induction of kdpFABC expression at extracellular K+ concentrations above 5 mM. This fits very well with our hypothesis that the transmembrane domains confer sensitivity towards the inhibitory effect of extracellular K+. Based on truncated versions of KdpD, different domains of KdpD were proposed to be responsible for sensing an intracellular stimulus. Rothenbücher et al. (2006) reported that the C-terminal domain of KdpD (amino acids 499–894) senses intracellular K+. In contrast, Heermann et al. (2003b) proposed that the N-terminal domain of KdpD (amino acids 1–395) enhances kdpFABC-expression under K+-limiting conditions. However, both truncated KdpD derivatives exhibited high levels of kdpFABC expression under K+-limiting growth conditions, but constitutive kdpFABC expression at higher extracellular K+ concentrations, similar to the KdpD derivative lacking the transmembrane domains. Therefore, we propose that an intracellular parameter changing under cellular K+ starvation (i.e. low intracellular K+ concentration) is sensed by the cytoplasmic domains of KdpD stimulating KdpD kinase activity, whereas the periplasmic loops sense the extracellular K+ concentration, triggering KdpD phosphatase activity at K+ concentrations above 5 mM.
Media and growth conditions
Escherichia coli cells were grown aerobically at 37°C in minimal phosphate-buffered medium adjusted to the desired pH and identified by the K+ content in millimolars as described previously (Epstein and Kim, 1971). Briefly, K115 contains 115 mM K+, and K0 is identical to K115 except that K+ is replaced by Na+. Every mixture of these two media is iso-osmotic and allows adjustment of any K+ concentration in the range of 0.02 mM to 115 mM. Solid media were prepared by the addition of 19 g l−1 agar. In general, glucose (0.4%) served as carbon and energy source and thiamine was added in final concentration of 1 mg l−1. Carbenicillin (100 μg ml−1) and Chloramphenicol (34 μg ml−1) for plasmid selection were added as needed.
Bacterial strains, plasmids and oligonucleotide-directed site-specific mutagenesis
All strains are derivatives of E. coli K-12 and are inactivated in the genes coding for the constitutively expressed K+ uptake systems Trk and Kup (formerly TrkD). Strain TK2240 [kdp+ trkA405 trkD1 nagA thi rha lacZ] (Epstein et al., 1978) encodes the wild-type kdp-system, whereas TK2281 [ΔkdpFABCDE trkA405 trkD1 nagA thi rha lacZ] (Polarek et al., 1988) carries a deletion of the kdpFABCDE genes and is therefore inactive in all saturable K+ uptake systems. Strain TKV2208 [ΔkdpD trkA405 trkD1 nagA thi rha lacZ] (Puppe et al., 1996) transformed with plasmid pBD5-9 or its derivatives was used to characterize the phenotype of the mutated KdpD proteins at different extracellular K+ concentrations. Strain TKV2209 [ΔkdpDE trkA405 trkD1 nagA thi rha lacZ] (Zimmann et al., 1995) co-transformed with plasmids pWPA1lacZ and pPV-2 or its derivatives was used for determination of kdpFABC expression in vivo. For simplicity, the mutated genes trk and kup are referred to as Δtrk and Δkup. In plasmid pBD5-9 (Zimmann et al., 2007), kdpD was cloned into pBAD18, under control of the inducible and repressible arabinose promoter araBAD (Guzman et al., 1995). Growing cells in the absence of the inducer arabinose but in the presence of the repressor glucose, the small amounts of KdpD synthesized are sufficient for complementation of the kdpD deletion strain (Heermann et al., 2003a). In plasmid pPV2 (CmR), the kdpDE operon was cloned into pSU19 (Walderhaug et al., 1992), carrying a p15A origin compatible with the common ColE1 origin. Plasmid pWPA1lacZ (AmpR) (Puppe et al., 1996) contains an in frame kdpA–lacZ fusion under control of the kdpFABC promoter and was used as a quantitative reporter for kdpFABC expression.
A codon change corresponding to the replacement of Asp424 or Asp474 against the indicated amino acids was introduced by synthetic mutagenic oligonucleotide primers using the PCR overlap extension method (Ho et al., 1989). The PCR products were cloned into vector pBD5-9 digested with restriction endonucleases NotI and StuI resulting in plasmids pBD5-9/D424* and pBD5-9/D474* respectively (the asterisk stands for the corresponding amino acid replacement). The mutations were verified by sequencing through the ligation junctions. All kdpD constructs were cloned into pPV-2 using XcmI and StuI restriction sites, resulting in plasmid pPV-2/D424* and pPV-2/D474* respectively.
An immediate change of the extracellular K+ concentration was applied to growing cells as described previously (Hamann et al., 2008). Briefly, cells of strain TK2240 were grown to exponential growth phase in K60 minimal medium and, subsequently, the culture was vacuum filtered through a 0.45 μm nitrocellulose membrane. Cells were washed and resuspended in the same volume of pre-warmed (37°C) minimal medium containing a lower K+ concentration, and growth of the subcultures was continued under shaking. The moment when filter-collected cells were placed in the new culture flask was considered time zero and the first sample was taken 30 s thereafter.
Determination of kdpFABC expression in vivo
In vivo signal transduction was probed with TKV2209 transformed with the plasmids described. Cells were grown in minimal medium containing the indicated amounts of K+ (see Fig. 5) to exponential growth phase and harvested by centrifugation. β-Galactosidase activity was determined as described previously (Miller, 1972) and is given in Miller units.
Determination of extracellular and cytoplasmic K+ concentrations
Cellular K+ contents were determined by flame photometry of cell-extracts that were collected by centrifugation through silicon oil (Hamann et al., 2008). Cells of 1 ml cell culture were collected by centrifugation through a 200 μl silicon oil layer and extracted by using 5% trichloroacetic acid, freeze–thawing at −20°C and heating to 100°C for 10 min. Cell debris were pelleted and K+ in the supernatant was measured by using flame photometry (Eppendorf ELEX 6361 flame photometer). The K+ content was corrected for the amounts of K+ in the periplasmic and extracellular water space (Hamann et al., 2008). K+ concentrations of minimal media were directly measured after appropriate dilution.
Determination of kdpFABC expression by Q-RT-PCR
Total RNA was extracted from cells using the RNeasy minikit (Qiagen, Hilden, Germany) according to the manufacturer's instructions. Residual DNA contaminations were eliminated by DNase I digestion and cDNA was synthesized via the RevertAid First Strand cDNA synthesis kit (Fermentas). Quantitative reverse transcription-PCR (Q-RT-PCR) was performed as described previously (Hamann et al., 2008) by the use of primer pair kdpAfor2/kdpArev2 annealing in kdpA (GCC GCC AGC GGG ATT GCG G and CTT CAA CGG TAT TCA CAG CCT G). As an internal standard, the primer pair gapAfor1/gapArev1 (CTC CAC TCA CGG CCG TTT CG/CTT CGC ACC AGC GGT GAT GTG) annealing in the E. coli housekeeping gene gap (glyceraldehyde-3-phosphate dehydrogenase A) was used. To correct for sampling errors, the levels of expression of kdpA as determined from the cycle threshold (CT) values were normalized to the expression level of the housekeeping gene.
Cytoplasmic pH measurements
Escherichia. coli strains TK2281 (Δkdp Δtrk Δkup) and TK2240 (kdp+ Δtrk Δkup) were transformed with the pH-dependent GFPmut3b reporter-plasmid pMMB1311 (Kitko et al., 2010), kindly provided by Dr Ryan Kitko (Ohio, USA). The transformed strains were grown in K115 minimal medium at pH 7.0 to exponential growth phase (OD600 = 0.4). The cultures were harvested, washed and resuspended in the same volume of pre-warmed minimal medium containing the indicated K+ concentration adjusted to pH 6.0 or pH 7.8 and incubation was continued for 60 min. The cytoplasmic pH of cultures diluted to OD600 = 0.4 was calculated from fluorescence measurements as described previously (Kitko et al., 2010). Briefly, excitation spectra of the cell suspensions were measured from 480 to 510 nm with an emission wavelength of 545 nm using a Jasco FP-6500 spectrofluorometer. After recording the excitation spectra, 30 mM sodium benzoate were added to the cell suspensions to deplete the ΔpH, thereby equalizing the internal pH with the external pH, and pH and fluorescence intensity were measured. The extracellular pH was then raised with NaOH or lowered with HCl for a second measurement of pH and fluorescence intensity. Fluorescence intensities of the cell suspensions were converted to cytoplasmic pH values using the standard curves generated from the known pH/fluorescence intensity points of the cell suspensions including 30 mM sodium benzoate.
This work was financially supported by grants from the Deutsche Forschungsgemeinschaft, the Friedel & Gisela Bohnenkamp-Stiftung, and the Ministry for Research and Culture of the State of Lower Saxony. We thank Ryan Kitko for the gift of plasmid pMMB1311, Wolf Epstein for providing mutant strains of E. coli and Gabriele Deckers-Hebestreit for critically reading the manuscript.