Generation of a Proton Motive Force by the Excretion of Metal-Phosphate in the Polyphosphate-accumulating Acinetobacter johnsonii Strain 210A*

The strictly aerobic, polyphosphate-accumulating Acinetobacter johnsonii strain 210A degrades its polyphosphate when oxidative phosphorylation is im-paired. The endproducts of this degradation, divalent metal ions and inorganic phosphate, are excreted as a neutral metal-phosphate (MeHPO,) chelate via the elec- trogenic MeHPO,/H+ symport system of the organism. The coupled excretion of MeHPO, and H+ in A. johnsonii 21OA can generate a proton motive force. In membrane vesicles and deenergized cells, a membrane potential of about -70 mV and transmembrane pH gradient of about -8 mV were formed in response to an imposed outwardly directed MeHPO, concentration gradient of 120 mV (ini- tial value). The MeHPO, efflux-induced proton motive force could drive energy-requiring processes, such as the accumulation of L-proline

Activated sludge in wastewater treatment plants is enriched with polyphosphate-accumulating bacteria, e.g. from the strictly aerobic genus Acinetobacter, when alternating aerobic and anaerobic conditions are applied (Fuhs and Chen, 1975;Deinema et al., 1985). The polyphosphate metabolism of one of these strains, Acinetobacter johnsonii 210A, has been studied in detail (van Groenestijn et al., 1989a, 198913;Bonting et al., 1992aBonting et al., , 1992bBonting et al., , 1993a. In the presence of excess energy and substrates, A. johnsonii 210A accumulates large amounts of Pi and metal ions as metal-polyphosphate in granules in the cy-Scientific Research (NWO) and the Dutch Technology Foundation *This was work supported by the Netherlands Organization for (STW). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
8 To whom correspondence and reprint requests should be addressed.
tosol. Under these conditions, the organism is able to take up the predominant Pi species from its aquatic environment by the concerted operation of two Pi transport systems (van Veen, 1994;van Veen et al., 1993avan Veen et al., , 1993b. An ATP-driven periplasmic binding protein-dependent system catalyzes the unidirectional uptake of H,PO; and HPOi-(Kt for Pi of 0.7 PM). The synthesis of this system is repressed by Pi a t medium concentrations above 10 VM. In addition, the organism possesses a constitutive secondary transport system mediating the uptake of a neutral metal-phosphate (MeHPO,) chelate via an electrogenic proton symport mechanism (Kt for MeHPO, of 7.7 VM). This chelate is formed by complexation of HPOi-and divalent cations like M e , Ca2+, Co2+, or Mn2+ (van Veen et al., 1994~).
Two enzymes are involved in the degradation of polyphosphate in A. johnsonii 210A: (i) polyphosphatase and (ii) polyphosphate:AMP phosphotransferase (van Groenestijn et al., 1989a). Polyphosphatase catalyzes the hydrolysis of polyphosphate to Pi (Bonting et al., 1993b). The activity of this enzyme enables the organism to use metal-polyphosphate as a source of Pi and divalent cations when the environmental concentrations of these nutrients are limiting (van Groenestijn and Deinema, 1985; van Groenestijn et al., 1988). Po1yphosphate:AMP phosphotransferase catalyzes the phosphorylation of AMP to ADP with polyphosphate as phosphoryl donor (van Groenestijn et al., 1989a;Bonting et al., 1991). The subsequent conversion by adenylate kinase of two molecules of ADP into one molecule of AMP and ATP enables A. johnsonii to regenerate AMP for the phosphotransferase reaction and may allow the organism to use its polyphosphate as a source of ATP when oxidative phosphorylation is impaired, e.g. under anaerobic conditions (van Groenestijn et al., 1987;van Groenestijn, 1988). During the degradation of metal-polyphosphate in A. johnsonii 210A, Pi and metal divalent ions are excreted into the environment via the secondary MeHPO, transport system (van Groenestijn et al., 1988;van Veen et al., 1993b.
Besides the direct synthesis of ATP via the polyphosphate: AMP phosphotransferaseladenylate kinase pathway, A. johnsonii 210A may conserve metabolic energy from polyphosphate degradation by the reversed process of MeHPO, uptake via the secondary MeHPO, transport system (van Veen et al., 1993a(van Veen et al., , 1993b. During MeHPO, uptake, the energy of the electrochemical proton gradient is converted into the energy of a chemical MeHPO, gradient, whereas during MeHPO, efllux the energy of a chemical MeHPO,, gradient may be converted back into the energy of an electrochemical proton gradient. In this paper, experimental support is given for this energy recycling mechanism. Energy transduction to electrogenic MeHPO,/H+ eMux was studied in deenergized cells and mem-

Energy
Recycling by brane vesicles of A. johnsonii 210A. In addition, in vivo 31P NMR was used to examine polyphosphate degradation and MeHPO, efflux in cells under physiological conditions. EXPERIMENTAL PROCEDURES Cells Growth and Preparation of Membrane Vesicles-A, johnsonii 210A was grown at 30 "C in a Tris-buffered medium (pH 7.0) supplemented with 20 mM sodium butyrate and 5 mM or 20 p~ sodium phosphate for cultivation of high and low PI-grown cells, respectively (van Veen et al., 1993a). Cells were harvested in midexponential phase by centrifugation (7,000 x g , 10 rnin). Membrane vesicles were prepared by osmotic lysis of high Pi-grown cells exposed to lithium chloride, a high concentration of lysozyme, and a temperature shock (van Veen et al., 1993b).
Polyphosphate Degradation in Cells-High Pi-grown cells containing polyphosphate granules and low Pi-grown control cells, in which polyphosphate granules were absent, were washed twice in ice-cold 50 mM Tris-HC1 (pH 7.8). To permeabilize the outer membrane, cells were given a n EDTA treatment as described previously (van Veen et al., 1993a). Subsequently, cells were washed in ice-cold 50 mM Tris-HC1 (pH 7.8) and resuspended in this buffer to about 5.5 mg of proteidml. The cell suspensions were transferred to Hungate tubes, flushed for 10 min with oxygen-free N,, and incubated anaerobically at 30 "C.
MeHPO, Efflux from Preloaded Cells and Membrane Vesicles-High Pi-grown cells were deenergized by incubation for 12 h at 30 "C in 20 m~ potassium Pipes' (pH 7.0) supplemented with 10 mM MgSO,, 2.5 m~ a-dinitrophenol and 50 pg of chloramphenicollml (van Veen et al., 1993a). Depletion of endogenous energy reserves was followed in time by measuring the endogenous respiration rate of the cells. Cells were treated with EDTA (van Veen et al., 1993a), washed extensively, and resuspended to a concentration of 0.5 mg of proteidml in 20 mM potassium Pipes (pH 7.5) containing 10 m~ MgSO, and 50 pg of chloramphenicol/ml (Buffer A). Cells were equilibrated overnight at 4 "C in BufferA supplemented with 5 mM potassium phosphate (pH 7.5) (about 3 mM MgHPO, (van Veen et al., 1994~)). The cell suspensions were concentrated to about 14 mg of proteidml. An outwardly directed MeHPO, concentration gradient of 120 mV was imposed by diluting preloaded cells 100-fold into 20 mM potassium Pipes (pH 7.5) containing 50 pg of chloramphenicoL'm1 (Buffer B). Deenergized cells equilibrated in Buffer A without added Pi served as a control. Membrane vesicles i n 10 m~ potassium Pipes (pH 7.5) supplemented with 10 mM MgSO, (Buffer C) were equilibrated in the presence and absence of 5 mM potassium phosphate (pH 7.5), and diluted into 10 mM potassium Pipes (pH 7.5) (Buffer D) as described for deenergized cells.
Determination of Membrane Potential-The membrane potential (A+, interior negative) in cells was determined from the distribution of the lipophilic tetraphenylphosphonium ion (TPP'), using a TPP+-selective electrode (Shinbo et al., 1978). For the determination of the A+ in cells during polyphosphate degradation, 800 pl of 50 mM Tris-HC1 (pH 7.8) supplemented with 5 mM potassium cyanide and 5 p~ TPP' was added to the TPP' electrode vessel. The vessel was sealed with a rubber septum. The buffer inside the vessel was flushed for 10 min with oxygen-free N,. Anaerobic cell samples of 200 pl were injected into the vessel after which TPP' accumulation was monitored. The induction of a A+ by MeHPO, efflux in deenergized cells was monitored with a TPP+-electrode by diluting MeHP0,-loaded cells 100-fold in Buffer B supplemented with 4 p~ TPP'. The ability of deenergized cell to generate a proton motive force (Ap) by the oxidation of glucose under aerobic conditions was checked by measuring the A+ (interior negative) as described previously (van Veen et al., 1993a). For calculations, an internal cell volume of 3 pllmg of protein was used (Kaback and Barnes, 1971;Bakker and Mangerich, 1981). Measurements were corrected for concentration-dependent nonspecific probe binding according to the model of Lolkema et al. (1982). For qualitative measurements of a MeHPO, efflux-induced A+ in membrane vesicles, MeHP0,-loaded membrane vesicles were diluted 100-fold in Buffer D supplemented with 9 p~ of the membrane potential indicator 3,3'-dipropylthiacarbocyanine iodide (DiSC3(5)). The A+-dependent fluorescence quenching was measured at excitation and emission wavelengths of 637 and 667 nm, respectively, with slit widths of 10 nm. The fluorescence signal was fonic acid); A+, transmembrane electrical potential difference; BCECF, ' The abbreviations used are: Pipes, piperazine-N,N"bis(2-ethanesul-2',7'-bis(2-carboxyethyl)-5(6)-carboxyfluorescein; DiSC3(5), 3,3'-diprobrane pH gradient (in mV); ZApMeHPO,, transmembrane MeHPO, con-pylthiadicarbocyanine iodide; Ap, proton motive force; ZApH, transmemcentration gradient (in mV); TPP', tetraphenylphosphonium ion.

MeHPO, I H+ Emux
averaged over time intervals of 0.3 s and calibrated in a A$ range of -44 mV to -90 mV by measuring the fluorescence as a function of an artificially imposed potassium diffusion potential (van Veen et al., 1993b).
Determination of Internal pH-The internal pH in cells and membrane vesicles was estimated from the fluorescence of the pH indicator BCECF entrapped in the intracellular or intravesicular space. MeHP0,-loaded and control cells were loaded with BCECF by a n acid shock treatment as described by Molenaar et al. (1991), washed five times with 1 ml of Buffer A with or without 5 mM potassium phosphate, respectively, and resuspended in these buffers to about 5 mg of proteid ml. The cells containing BCECF (as judged from the yellow color of cell pellets) after treatment with the smallest amount of HCl, were used for further experiments. Membrane vesicles were loaded with BCECF and MeHPO, by freeze-thaw-sonication (Driessen et al., 1985). An outwardly directed MeHPO, gradient was imposed in MeHPO, and BCECF-loaded cells and membrane vesicles as described above. Changes in the internal pH during MeHPO, efflux were monitored by continuous recording of BCECF fluorescence as described (van Veen et al., 1994a).
Amino Acid nunsport Assays-The uptake of 1.62 p~ L-lysine and 1.95 p~ L-proline in membrane vesicles was measured via the filtration method (Kaback, 1974) as described previously (van Veen et al., 1994a).
31P NMR Spectroscopy of Intact Cells-Low Pi-grown cells were washed and suspended to about 7 mg of proteidml in 150 mM potassium phosphate (pH 7.0) supplemented with 5 mM MgSO,. To allow the synthesis of polyphosphate, cells were kept under continuous aeration with oxygen at 21 "C. After 60 min of incubation, cells were washed twice in 50 mM Tris-C1 (pH 7.0), and resuspended to about 15 mg of proteidml in 50 mM Tris-C1 (pH 7.8) supplemented with 10 m~ potassium chloride and 2 mM MgSO,. Subsequently, the cell suspension was kept in a 10-mm NMR tube and gassed with argon using a n air-lift system. 31P NMR experiments were recorded using a 10-mm broadband probehead in a Bruker AMX500 spectrometer operating at 202.45 MHz for phosphorus. Spectra were acquired at 25 "C, without proton decoupling, using a 45" flip angle and 5.8-s repetition delay in 16,000 data points. Phosphorus resonances were referenced with respect to external 85% H3P0,.
Miscellaneous-Pi and total phosphorus, degraded to Pi by preliminary persulfate digestion, were determined colorimetrically by the ascorbic acid method (Rand et al., 1976). The presence of polyphosphate granules in cells was evaluated by light microscopy after staining according to Neisser (Gurr, 1965). For the determination of intracellular ATP concentrations, cells were extracted with perchloric acid by the method of Otto et al. (1984). ATP in neutralized extracts was determined with a firefly bioluminescence assay (LUMIT). The respiration rate of cells was measured with a Clark-type oxygen electrode. Protein was determined by the method of Lowry et al. (1951) using bovine serum albumin as the standard.
Chemicals-DiSC3(5) and BCECF were obtained from Sigma. The firefly bioluminescence NRBLUMIT-PM kit was purchased from Lumac, Omnilabo, Breda, The Netherlands. ~-[U-'~Cllysine (monohydrochloride (11.0 TBq/mol) and ~-[U-'~Clproline (9.5 TBq/mol) were supplied by the Radiochemical Centre, Amersham, Buckinghamshire, UK. Other chemicals were reagent grade and obtained from commercial sources. Fig. 1 demonstrate the ability of A. johnsonii 210A to conserve metabolic energy from the degradation of polyphosphate. In this study, high and low Pi-grown cells were used. High Pi-grown cells showed a relatively high phosphorus accumulation level of 4.1 pmol of phosphorus/mg of protein due to the presence of one or two large metal-polyphosphate granules in the cytoplasm. In contrast, low Pi-grown control cells were devoid of these granules and contained only 0.8 pmol of phosphorudmg of protein. During aerobic incubation for 2 h in the absence of an exogenous carbon and energy source, high Pi-grown cells hardly excreted MeHPO,. Their cellular ATP level and A+ were similar to those observed in low Pi-grown cells. The aerobic incubation period was followed by an anaerobic one. Under the latter condition, high Pi-grown cells rapidly degraded polyphosphate, resulting in the excretion of MeHPO, at an initial rate of about 3 nmoY midmg of protein (Fig. lA). High Pi-grown cells were able t o maintain a significant intracellular ATP concentration and A+ for at least 8 h under these conditions, whereas in low Pi-grown cells the levels of these parameters strongly decreased within 1 h (Fig. 1, B and C).

Conservation of Metabolic Energy from Polyphosphate Degradation-The experiments shown in
Pi Gradient during Degradation of Polyphosphate-Polyphosphate degradation was studied in A. johnsonii 210A using i n uivo 31P NMR (Fig. 2). During the first 8 h of anaerobiosis about 65% of soluble polyphosphates was degraded (Fig.  2 A ) . Strikingly, intracellular Pi accumulated up to 150 m~ in the course of polyphosphate degradation (Fig. 2 B ) . Since the external Pi concentration remained below 11 mM, an outwardly directed Pi gradient (ZA&J of 100-160 mV was maintained for 4 h (Fig. 2C). In view of (i) the important role of polyphosphate in the production of metabolic energy during anaerobiosis, (ii) the large outwardly directed Pi gradient that is maintained during polyphosphate degradation, and (iii) the presence of a phosphate carrier that mediates the translocation of MeHPO, via an electrogenic H+ symport mechanism, it was of interest to study energy transduction coupled t o MeHPO, efflux in A. johnsonii 210A.
MeHPO, Efflux i n Deenergized Cells-In order to study the recycling of metabolic energy by MeHP0,M' efflux in the absence of polyphosphate metabolism, cells of A. johnsonii 210A were depleted of polyphosphate and other endogenous energy reserves by aerobic incubation in the presence of the uncoupler a-dinitrophenol. As was shown previously (van Veen et al., 1993a), the cells retain the secondary MeHPO, transport system in an active form during the deenergization procedure and remain readily energizable after removal of a-dinitrophenol.
In deenergized cells, MeHPO,/H' efflux was coupled to the generation of a Ap. Thus, the lipophilic cation TPP' was accumulated almost 15-fold when an outwardly directed MeHPO, gradient (initial value of -120 mV) was imposed artificially by dilution of deenergized, MeHP0,-loaded cells into MeHP0,free buffer (Fig. 3 A ) . The maximum TPP' accumulation level suggested the generation of a A+ of about -70 mV. No significant accumulation of TPP' was observed in unloaded cells or in loaded cells in which the Ap was dissipated by valinomycin plus nigericin (each 2 nmoVmg of protein) (Fig. 3 A ) . To monitor the changes in the intracellular pH by MeHPO,/H+ efflux, deenergized cells were loaded with the fluorescent pH indicator BCECF. A rapid alkalinization of the internal milieu was observed when an outwardly directed MeHPO, gradient (initial value of 120 mV) was imposed (Fig. 3B). The addition of valinomycin plus nigericin (each 2 nmoVmg of protein) resulted in the decrease of the internal pH to base-line levels observed in unloaded cells. With a constant external pH during MeHPO, efflux, the degree of alkalinization of the cytoplasmic pH equalled a transmembrane pH gradient (-2ApH) of about -7 mV under these conditions. The Ap induced by MeHPO,/H+ efflux in deenergized cells could drive the synthesis of ATP from endogenous ADP and Pi via the membrane-bound H+-ATPase (Fig. 4). ATP levels remained very low in control cells in which a MeHPO, gradient was absent, e.g. through dilution of MeHP0,-loaded cells into buffers containing MeHPO, at a concentration equimolar to the internal one, or of unloaded cells into MeHP0,-free buffer. A significant synthesis of ATP was observed when MeHP0,loaded cells were diluted 100-fold into MeHP0,-free buffer. This synthesis is transient due to the rapid decrease of the MeHPO, gradient.
MeHPO, Efflux in Membrane Vesicles-In membrane vesicles, the mechanism of energy coupling to secondary transport of solutes can be studied in the absence of their metabolism by cytoplasmic enzymes. Membrane vesicles therefore offer an excellent model system to study energy transduction coupled to MeHPO,/H+ efflux in A. johnsonii 210A. The generation of a MeHPO, efflux-induced A+ was monitored in membrane vesicles using the fluorescent A+-indicator DiSC,(5) (Fig. 5A). Upon imposition of an outwardly directed MeHPO, gradient (initial value of 120 mV) in membrane vesicles, a rapid fluorescence quenching of DiSC,(5) was observed corresponding to a A+ of about -63 mV (Fig. 5A) membrane vesicles after dissipation of the Ap by valinomycin plus nigericin (each 1 nmoVmg of protein). The formation of a ApH by MeHPO,/H+ efflux was demonstrated by continuous recording of the fluorescence intensity of BCECF trapped within the membrane vesicles (Fig. 5 B ) . Imposition of an outwardly directed MeHPO, gradient (initial value of 120 mV) resulted in an alkalinization of the intravesicular pH and the generation of a -ZApH of -8 mV. A decrease of the BCECF fluorescence intensity down to the baseline level of unloaded membrane vesicles was observed upon addition of valinomycin plus nigericin (each 1 nmoVmg of protein). These results clearly demonstrate the generation of both components of the Ap by the electrogenic excretion of MeHPO, and H+. Membrane vesicles of this organism contain several cationamino acid transport systems that couple amino acid translocation to the Ap (van Veen et al., 1994a). Significant levels of L-lysine and ,L-proline accumulation were observed in MeHP0,loaded membrane vesicles that were diluted 100-fold in Me- HP0,-free buffer (Fig. 6). Accumulation of amino acids was not observed (i) upon dissipation of the MeHPO, efflux-induced Ap by valinomycin plus nigericin (each 1 nmoVmg of protein) and (ii) in the absence of an outwardly directed MeHPO, gradient, e.g. when MeHP0,-loaded vesicles were diluted into buffers containing MeHPO, at a concentration equimolar to the internal one or when unloaded membrane vesicles were diluted into MeHP0,-free buffer. These experiments show that a MeHPO, efflux-induced Ap can drive the uptake and accumulation of solutes in A. johnsonii 210A. DISCUSSION A. johnsonii 210A is a strictly aerobic nonfermentative bacterium. When oxidative phosphorylation is impaired, it degrades the metal-polyphosphate that was accumulated under aerobic conditions. The results of this investigation demonstrate that during this degradation, the organism is able to maintain its Ap and intracellular ATP at levels comparable with those observed under aerobic conditions. Two mechanisms for the conservation of metabolic energy from polyphosphate degradation have been suggested in A. johnsonii 210A. (i) the direct synthesis of ATP from polyphosphate via the polyphosphate:AMP phosphotransferaseladenylate kinase pathway (van Groenestijn et al., 1987;Van Groenestijn, 1988) and (ii) the generation of a Ap by the excretion of MeHPO,, a major endproduct of metal-polyphosphate degradation, together with H' (van Veen et al., 1993a, 199313).
Recently, the energetics and mechanism of the secondary MeHPO, transport system ofA. johnsonii 210A were examined in detail. The MeHPO, carrier catalyzes the coupled movement of a neutral MeHPO, chelate and a proton via an electrogenic symport mechanism (van Veen et al., 1993a. Thus, the driving force for MeHPO, translocation via this transport system is the sum of forces supplied by the Ap (A+ -ZApH) and the MeHPO, concentration gradient (ZApMeHPO,): Ap + ZApMe-HPO,. A steady state is reached when Ap = "ZApMeHPO,. During MeHPO, uptake, the Ap will exceed the -2ApMeHP0, whereas MeHPO, efflux occurs when the -2ApMeHP0, exceeds the Ap. In previous work, the generation of a Ap by electrogenic MeHPO,/H+ efflux was indirectly indicated by (i) the stimulation of MeHPO, efflux from cells by the uncoupler a-dinitrophenol and by N,N"dicyclohexylcarbodiimide, an inhibitor of the F,F, H+-ATPase ( van Groenestijn, 1988), (ii) the enhancement by protonophore carbonyl cyanide(3-chlorophen-y1)hydrazone of MeHPO, efflux from proteoliposomes containing reconstituted MeHPO, carrier protein (van Veen et al., 1993b), and (iii) the retardation of MeHPO, efflux from proteoliposomes by an artificially imposed ApH andor A+ (van Veen et al., 1993b).
The results presented in this study further corroborate the generation of a Ap by MeHPO,/H' efflux. Thus, the generation of a A+ by the efflux of MeHPO, was demonstrated directly by the fluorescence quenching of the A+ probe DiSC,(5) in membrane vesicles and by the accumulation of TPP' in deenergized cells under these conditions. In both systems, imposition of an outwardly directed MeHPO, gradient of 120 mV (initial value) resulted in the generation of a A+ of about -60 to -70 mV. The generation of a transmembrane pH gradient by MeHPO,/H+ efflux was shown by the fluorescence enhancement of the pHprobe BCECF, which was entrapped in the lumen of the membrane vesicles and deenergized cells. With the external pH remaining fairly constant during MeHPO, efflux, a -ZApH of about -8 mV was build up upon imposition of a MeHPO, gradient of 120 mV (initial value).
The MeHPO, efflux-induced Ap could be coupled to different metabolic energy requiring processes as summarized in Fig. 7. In membrane vesicles of A. johnsonii 210A, MeHPO,/H+ efflux could drive the accumulation of L-proline and L-lysine. Recent transport studies in membrane vesicles of this organism revealed the presence of single, high-affinity lysine and proline carriers that mediate the electrogenic symport of lysine and a proton and of proline and a sodium ion, respectively (van Veen et al., 1994a). The Na+ contamination in the uptake buffers (up to 150 w) greatly exceeds the low Kt of the proline carrier for Na+ (Kt = 26 p~ Na'). Na' is therefore present in sufficient amounts to allow Na+/proline symport. In addition, A. johnsonii 210A possesses a Na+/H+ antiporter, which converts the Ap into a sodium motive force (van Veen et al., 1994a). In view of the driving forces for lysine and proline uptake, 2A+ -ZApH and A+ -ZApNa, respectively, the almost 2-fold higher accumulation level of lysine compared with that of proline points to an effective generation of a A+ by MeHPO,/H+ efflux under the experimental conditions. This conclusion is consistent with the direct measurements of the composition and magnitude of the MeHPO, efflux-induced Ap in membrane vesicles and deenergized cells. Besides solute accumulation, the MeHPO, effluxinduced Ap could drive the synthesis ofATP via the membrane-bound H+-ATPase in deenergized cells ofA. johnsonii 210A. The results from these in vitro studies demonstrate the potential of MeHPO,/H+ efflux as an energy recycling mechanism in A. johnsonii 210A (Fig. 7).
31P NMR was used for in vivo studies of Pi gradients formed by the degradation of polyphosphate in anaerobic cell suspensions of A. johnsonii 210A. The concentration variations in internal Pi were much greater in magnitude and range than those in the external medium. The cells were able to maintain an outwardly directed Pi gradient of 100-160 mV for 4 h. Due to the high internal concentrations of Mg2 ' (up to 40 m M (Bontinget al., 1993a)) and Pi (up to 150 mM) during the degradation of magnesium-polyphosphate, the intracellular MgHPO, concentration will have reached saturating levels of about 20-30 m M (van Veen et al., 1994~). Thus, a substantial outwardly directed MgHPO, concentration gradient (of at least 100 mV) was present during the first 3 h of polyphosphate degradation, allowing MeHPO,/H+ efflux to be an effective energy conserving mechanism.
Under conditions of polyphosphate synthesis, the MeHPO, carrier and the ATP and binding protein-dependent Pi uptake system of A. johnsonii 210A enable the organism to efficiently acquire Pi from its habitat through uptake of the predominant Pi species (van Veen et al. , 1994~). However, the latter transport system has to be inactivated during the degradation of polyphosphate to prevent the reaccumulation of Pi with a concomitant waste ofATP. The inactivation of the primary Pi uptake system ofA. johnsonii 2lOAmay be exerted through trans-inhibition by the high internal Pi concentration that is established during the degradation of polyphosphate. Duns-inhibition by Pi has been described for a number of phosphate bond energy-driven uptake systems, including the Pi transporter of Lactococcus lactis (Poolman et al., 19871, and the sn-glycerol-3-P permease (Ugp) and phosphate-specific transport system (Pst) of Escherichia coli (Medveczky and Rosenberg, 1971;Brzoska et al., 1994).
Fermentative bacteria continuously excrete relatively large quantities of lactic acid and other organic acids into the environment. In their "energy recycling model," Michels et al. (1979) proposed that electrogenic efflux of these organic endproducts via ion symport systems may lead to the generation of an electrochemical ion gradient across the cytoplasmic membrane, thus providing metabolic energy to the cell. In recent years evidence has been obtained for H+-linked carrier-mediated excretion of lactate in lactic acid bacteria (Otto et al., 1980(Otto et al., , 1982Simpson et al., 1983a, 198313) and enteric bacteria (ten Brink and Konings, 1980), and Na+-linked excretion of succinate in Selenomonas ruminantium (Michel and Macy, 1990). Other organic acids, like acetate, may also be excreted via carrier-mediated processes (Driessen and Konings, 1990). The observations in A. johnsonii 210A (this work) extend the energy recycling model to the excretion of inorganic endproducts. Thus, the results show that a Ap is generated by the electrogenic excretion of MeHPO, and H+ via the secondary MeHPO, transport system of this organism. In view of the evidence for cotransport of Pi and divalent cations in other biological systems (van Veen et al., 199413, 19948, energy recycling by the excretion of metal-phosphate chelates may be a more general mechanism for the conservation of metabolic energy in polyphosphate-accumulating microorganisms. Recycling of metabolic energy by the excretion of inorganic endproducts may also have interesting implications for the efflux of NH; by Ureoplasma species (Masover et al., 1977;Smith et al., 1993) and the excretion of sulfate by Thiobacillus species and other sulfur-oxidizing bacteria (Kelly, 1988).

Energy Recycling by MeHP0,IH' Effzux
In conclusion, A. johnsonii 21OAis able to use metal-polyphosphate as a source of metabolic energy during anaerobiosis by (i) the direct synthesis of ATP via the po1yphosphate:AMP phosphotransferasefadenylate kinase pathway and (ii) the generation of a Ap via the coupled excretion of MeHPO, and H' . Polyphosphatase may enhance the latter energy recycling mechanism by providing the MeHPO, efflux process with a continuous supply of Pi and divalent metal ions. As a consequence of energy recycling by MeHPO, excretion, less ATP has to be hydrolyzed via the H+-ATPase t o generate a Ap when oxidative phosphorylation is impaired. Conservation of metabolic energy from metal-polyphosphate degradation may enable A. johnsonii 210A to survive alternating aerobidanaerobic conditions as encountered in certain natural habitats and in wastewater treatment plants.