Dynamic Compartmentation of Vacuolar Amino Acids in Penicillium cyclopium

The regulation of amino acid transport from the vacuolar reservoir into the cytoplasm has been studied in hyphal cells of Penicillium cyclopium. To avoid artifacts caused by the isolation of vacuoles, efflux was examined “in situ,”i.e. in cells whose plasma membranes were permeabilized for micromolecules by a treatment with nystatin. The ATP-dependent proton gradient and amino acid transport activities at the vacuolar membrane remained intact under these conditions. Accumulation of amino acids in the vacuole proved to be the result of a dynamic equilibrium of active, ATP-dependent uptake and energy-independent efflux. The latter was strongly accelerated after the vacuolar amino acid content had surpassed a threshold level. Efflux of vacuolar amino acids was specifically controlled by extravacuolar adenylates: ATP, 5′-adenylyl imidodiphosphate (an ATPase-resistant ATP-analogue), ADP, or AMP caused a strong inhibition in the concentration range around 200 μmol/liter, whereas both lower and higher concentrations allowed significant efflux rates. Estimates of the cytosolic adenylates (which consisted mainly of ATP) were close to 2 mmol/liter in glucose-metabolizing cells, which concentration allowed maximum rates of both vacuolar uptake and efflux. During 24 h of carbon and nitrogen starvation, the adenylate level decreased toward the efflux-inhibiting region around 200 μmol/liter, whereas 3–4 d of carbon and nitrogen starvation caused a further decline of the adenylate content, leading again to efflux-permitting concentrations. Thus, the cytosolic adenylate pool appears to effectively control the availability of vacuolar amino acids for the cellular metabolism.

The metabolism of amino acids in fungal and plant cells is strongly influenced by transport activities of the vacuole. The majority of free amino acids is usually concentrated in this organelle, the relation between vacuolar and cytosolic pools being different for individual amino acid species (1)(2)(3)(4)(5). Vacuolar amino acids are used as a reservoir to maintain the concentrations of cytosolic amino acids within distinct limits, thus compensating for changes of the external or biosynthetic supply and metabolic needs. As amino acid precursors are required for many essential anabolic and catabolic activities, the control over their vacuolar pool is assumed to have a high significance for metabolic regulation. Classical examples are the controlled exchange of vacuolar and cytoplasmic arginine and ornithine during the transition between anabolic and catabolic steady states in Neurospora crassa (6 -8) or the response of the vacuolar amino acid pool to nutrient-induced changes of the rate of protein synthesis in leaves of barley or spinach (9).
In eukaryotic microorganisms various tonoplast transporters have been identified that catalyze the uptake of amino acids into the vacuole at the expense of proton-motive force generated by the tonoplast H ϩ -ATPase and tonoplast H ϩ -pyrophosphatase (10 -15). In plant cells similar evidence is less abundant (16,17), and proton-motive force-independent transport (stimulated by ATP and its ATPase-resistant analogue) has been reported as well (4,9,18).
The cellular control over the vacuolar amino acid pools suggests that tonoplast transport systems are involved in regulatory circuits that sense cytosolic conditions and respond by switching between accumulation into the organelle and efflux into the cytosol. Whereas some information about the regulation of vacuolar uptake can be deduced from kinetic and regulatory properties of amino acid transport systems at the tonoplast (3,4,10,15,17,18) and the energizing proton pumps (19 -21), the mechanism(s) governing the efflux of accumulated amino acids into the cytosol are largely unknown. Such knowledge would promote the understanding of the regulatory properties of the vacuolar transporters and of their impact on enzymic processes in the cytoplasm, whose conversion rates might be controlled by the efflux of precursors from their vacuolar pools (e.g. see Ref. 22).
In N. crassa, Weiss and co-workers (8, 23) described a stimulating effect of glutamine limitation on the energy-requiring efflux of vacuolar arginine. In vacuoles isolated from barley mesophyll protoplasts Dietz et al. (9) found the efflux of amino acids to be stimulated by ATP and its ATPase-resistant analogue AMP-PNP 1 and to be inhibited by neutral cytosolic amino acids. In the present report we describe two novel control mechanisms of vacuolar efflux that became apparent in fungal cells whose plasma membrane was selectively permeabilized for micromolecules: (a) stimulating as well as inhibitory effects exerted by distinct concentrations of extravacuolar adenylates and (b) a sigmoidal concentration dependence of amino acid efflux. * This work was supported by the Deutsche Forschungsgemeinschaft. 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  After 2 d of growth, a 100-ml submerged culture was diluted with 500 ml of maleate buffer, pH 4.5, 50 mmol/liter, and filtered through a nylon mesh (pore diameter ϳ500 m) to remove hyphal pellets. The fine, hairy grown mycelium was washed twice with 250 ml of the same buffer by suction filtration. The cells were then resuspended (4 mg fwt/ml) in half-concentrated culture medium without ammonium tartrate and shaken for another 12 h in order to allow the derepression of nitrogenregulated transport activities (25).
Permeabilization of the Plasma Membrane of Hyphal Cells-After 2 d of growth, submerged hyphae (approximately 100 mg fwt) were filtered and washed by suction filtration with 20 ml of sodium citrate buffer, 5 mmol/liter, pH 4.5, containing 1 mol/liter sorbitol, resuspended in 10 ml of the same buffer, and gently shaken for 5 min at 24°C. Then 5 l of a nystatin solution (10 mg/ml in Me 2 SO) were added, and the cells were incubated with shaking for another 6 min (or times indicated otherwise). The effect of nystatin was stopped by adding 20 ml of HEPES buffer, pH 7.0, 50 mmol/liter containing 150 mmol/liter KCl, 50 mmol/ liter NaCl, 600 mmol/liter sorbitol, 5 mmol/liter CaCl 2 , and 5 mmol/liter MgCl 2 . Finally, the cells were washed by suction filtration and resuspended in the last buffer containing 1 mmol/liter ATP.
The success of permeabilization and the intactness of the vacuolar accumulation capacity was checked prior to each transport experiment by following the vacuolar accumulation of neutral red from 2 mmol/liter solutions in the presence and absence of 1 mmol/liter ATP (see Fig. 2). This was done both microscopically and by following the decrease of A 530 of the supernatant. The test is based on the observation that ATP does not support dye accumulation in the presence of bafilomycin and that intact cells stain much slower and do not react to ATP, which is analogous to the uptake of Phe (see Table I).
Transport Assays-Hyphal cells were resuspended (4 mg fwt/ml) in 60 mmol/liter sodium maleate buffer, pH 3.5 (intact cells), or in 60 mmol/liter sodium MES buffer, pH 7.0, with 0.4 mol/liter sorbitol, 0.15 mol/liter KCl, and 1 mmol/liter MgATP (permeabilized cells). These suspensions were gently aerated through a glass sinter funnel. After 1 min the unlabeled amino acid (final concentration 50 mol/liter), together with the tracer amino acid ( 14 C-or 3 H-labeled, final concentration Ͻ1 mol/liter), and effectors in the appropriate buffer were added. Further details for uptake and efflux protocols are given in Figs. 3 and 4. The content of labeled amino acids given in the figures refers to the amount of free amino acids extractable from the cells with 80% ethanol. It was determined in 5-ml samples, which were rapidly injected into 10 ml of the same buffer at 0°C, sucked onto cellulose nitrate filters (1.2 m cutoff), and washed with 15 ml of this buffer (intact cells) or with 0.7 mol/liter sorbitol (permeabilized cells) at 0°C. The cell pellet was then immersed in 2 ml of ethanol 80% (v/v), shaken for 2 h, and centrifuged at 5000 ϫ g for 10 min, and radioactivity in 1 ml of the supernatant was counted by liquid scintillation. A Hewlett Packard LS counter was used in a dual channel mode that allowed the simultaneous detection of 14 C and 3 H in discrete energy windows by automatically compensating for overlapping radiation via internal calibration curves. The separate detection of the isotopes was confirmed using test mixtures.
Initial rates of uptake and efflux were computed from the experimental curves fitted to a hyperbolic equation (uptake curves) or the equation of exponential decay, y ϭ a(e ϪKx ) ϩ b, (efflux curves).
At least 88% of the radioactivity extracted from the cells after 80 min of incubation with [ 14 C]Phe was chromatographically identical with phenylalanin as confirmed by reversed phase high performance liquid chromatography amino acid analysis after derivatization with Fmoc-chloride (9-fluorenylmethyl chloroformate) (26). After two extraction steps with ethanol 80% (v/v), the total protein solubilized with 10% SDS (see below) contained Ͻ10% of the radioactivity found in the ethanol-soluble fraction.
Determination of Nucleotide Content-Hyphal suspensions were filtered by suction, the cell pellet (approximately 10 mg fwt) was rapidly immersed in boiling glycine buffer (100 mmol/liter, pH 11.0) and extracted for 10 min at 100°C. After cooling to 20°C, the extract was centrifuged at 6000 ϫ g for 15 min, and the adenylates in the supernatant were assayed luminometrically.
For ATP, 10-l aliquots were injected into 300 l of a luciferin/ luciferase mixture reconstituted from lyophilized preparations (Boehringer Mannheim) in glycine buffer 100 mmol/liter, pH 7.4, and the luminescence counts integrated over 10 s in a luminometer (Clinilumat, Fa. Berthold). The method was calibrated with ATP solutions in the extraction buffer and validated via internal standards or addition of ATPase. S.D. ϭ 8%, n ϭ 5. ADP was assayed as the amount of ATP formed from phosphoenolpyruvate via pyruvate kinase as outlined by Hampp (27). AMP was assayed as the amount of ATP used by the adenylate kinase reaction. The reaction mixture contained adenylate kinase (0.9 unit/10 l) and 10 mmol/liter MgS0 4 in 25 mmol/liter HEPES-K ϩ , pH 7.5. ATP present in the native sample was used as the phosphate donor as its concentration always enabled a sufficient excess over AMP (see Fig. 5B). ADP and AMP could be determined in the presence of a Յ20-fold (ADP) or Յ7-fold (AMP) excess of ATP as estimated from calibration experiments at ATP ϭ 10 mol/liter.
The intracellular water volume was determined as the difference between the total water of a wet hyphal pellet (dry loss) and the extracellular water. The latter was estimated as the partial volume of the wet cell pellet that was accessible to the non-membrane-penetrating solute bovine serum albumin.
Indicator Enzyme Activities-Glucose-6-phosphate dehydrogenase (G6PDH) activities were assayed in hyphal suspensions after exposure to nystatin for various times (see permeabilization procedure above). 2.5 mg of cells (fwt) in 1 ml were mixed with the same volume of triethanolamine buffer, 100 mmol/liter, pH 7.6, containing 0.7 mol/liter sorbitol. After centrifugation at 16,000 ϫ g for 10 min, both the cell pellet (after resuspension in the original volume of the last mentioned buffers) and the supernatant were used to determine conversion rates of glucose 6-phosphate: 800 l of cell suspension or supernatant were mixed with 150 l of MgCl 2 solution, 100 mmol/liter, and 60 l of NADP ϩ solution, 15 mmol/liter. After reading the initial extinction the conversion was started by adding 60 l of glucose 6-phosphate solution, 30 mmol/liter, and reciprocal shaking at 24°C. After 30 and 60 s, the reaction was stopped by rapid cooling on ice, and the mixture was centrifuged at 10,000 ϫ g for 10 min at 0°C. The OD of the supernatants was read at 340 nm after warming to room temperature.
Total cellular protein was extracted from cell pellets by shaking with 10% SDS for 4 h. After centrifugation at 6000 ϫ g for 10 min, the supernatant was assayed by the Lowry et al. (28) procedure.

Accumulation of Amino Acids in Intact Cells
Hyphal cells of P. cyclopium are able to concentrate amino acids to a considerable extent. For example, nitrogen-starved cells accumulate externally added 14 C-labeled amino acids from 5 mol/liter solutions by the following concentration factors (C in /C out , calculated on the basis of total intracellular water): Arg, 2600-fold; Glu, 2000-fold; Leu, 1700-fold; Phe, 1300-fold; even from 5 mmol/liter solutions concentration factors of 38 (Arg), 34 (Glu), 32 (Leu), and 24 (Phe) are reached. As to be expected, the majority of absorbed amino acids accumulated in the vacuolar compartment; after selective lysis of the plasma membrane (see next section) the vacuoles retained 89% of total [ 14 C]Arg, 65% of total [ 14 C]Glu, 93% of total [ 14 C]Leu, and 81% of total [ 14 C]Phe, each preaccumulated from a 5 mol/ liter solution.
Vacuoles liberated from intact hyphal cells via protoplasts contained approximately 65% of the total ninhydrin-positive compounds but showed very low accumulation capacities for added amino acids compared with the protoplasts of their origin (e.g. less than 20% in the case of [ 14 C]Phe or [ 14 C]Arg). This situation was only marginally influenced by the isolation procedure, i.e. lysis with DEAE-dextran (29), EGTA-induced lysis (according to Kringstad et al. (30)), lysis by shearing forces via ultracentrifugation (31), all followed by flotation in sorbitol/ Ficoll density gradients. Addition to isolated vacuoles of a 20,000 ϫ g supernatant (obtained after disintegration of protoplasts in a glass homogenizer at 0°C, centrifugation and 3 h flow dialysis against 10 mmol/liter MES, pH 6.5) increased their accumulation capacity for [ 14 C]Phe and [ 14 C]Arg by 30 -50% as well as their lifetime detectable by microscopic obser-vation. 2 Thus, contact with cytosolic macromolecules appears to be critical for maintenance of tonoplast transport capacities that can therefore not reliably be assayed with isolated vacuoles. Similar objections have been raised against the use of isolated vacuoles in an artificial environment for the elucidation of regulatory functions of vacuolar ion channels (32). In situ systems can serve as a verifiable compromise between the reduced complexity of the isolated organelle and the functional stability and variability of the intact cell. Therefore we decided to probe the dynamics of vacuolar accumulation and efflux with an in situ system to be established by selective permeabilization of the plasma membrane for micromolecules under conditions that keep the vacuole functionally intact within its macromolecular intracellular environment.

Selective Permeabilization of the Plasma Membrane for Micromolecules
Selective lysis of the fungal plasmalemma without damaging the tonoplast has been reported hitherto only for yeast cells that had been treated with either DEAE-dextran (33) or CuCl 2 (34). While experiences with the CuCl 2 procedure argue for a damaging side effect on the tonoplast H ϩ -ATPase (uptake of arginine or Ca 2ϩ was not only reduced compared with intact cells but also insensitive to uncouplers, cf. Ohsumi et al. (34)), treatment with DEAE-dextran permeabilized hyphal cells of our strain of P. cyclopium only partially. In this fungus, treatment with the polyene antibiotic nystatin in a buffer of low ionic strength proved to permeabilize the plasma membrane with sufficient efficiency and selectivity. Polyenes, such as nystatin or amphotericin B, are known for their pore-forming effect, which is based on their binding to membrane sterols of fungal cells (35)(36)(37). Selectivity toward the plasma membrane can be expected because the sterol (mainly ergosterol) content of this membrane is more than tenfold higher than that of other cellular membranes. This is not only true for P. cyclopium (38), but for other fungal species as well.
The breakdown of the plasma membrane permeability barrier was indicated by the increasing availability of extracellular substrates for intracellular ("latent") enzymes. As seen in Fig.  1, a cytosolically localized indicator enzyme, glucose-6-phosphate dehydrogenase (G6PDH) reports an increasing availability of its external substrates NADP ϩ and glucose 6-phosphate during a 6-min incubation with nystatin. Similar findings were obtained with the conversion of added 4-methylumbelliferyl phosphate (4-MUP) by cellular phosphatases (data not shown).
The leakage of bulk protein, measured as A 280 in the supernatant, remained very low for up to 8 min of nystatin treatment (less than 2% compared with detergent treated cells) and correlated with the appearance of marker enzyme activities in the outer medium.
Thus it appears very likely that a 6-min treatment with 5 g/ml nystatin permeabilizes the plasma membrane for micromolecules as NADP ϩ , glucose 6-phosphate, and 4-MUP 3 with- out causing a substantial loss of protein. On the other hand, this treatment did not significantly impair the functional integrity of the vacuoles as indicated by the following data. (a) The vacuoles were able to maintain an ATP-dependent proton gradient. Microscopic observations on the vacuolar trapping of neutral red indicated that this dye accumulated much faster in permeabilized than in intact cells (Fig. 2, a and b). The diagnostic value of neutral red accumulation to probe the intactness and energization of the vacuole is supported by showing that the dye accumulates neither in the absence of ATP nor after prolonged nystatin treatment or addition of the detergent CTAB (Fig. 2, c-e). Furthermore, the nucleic acid-based viability probe Fungo-light® (Molecular Probes) accumulated in vacuoles of permeabilized cells supplied with ATP much like in those of intact cells (data not shown). ATP-dependent accumulation of neutral red was used as a test criterion in routine checks of successful and selective permeabilization. (b) The uptake of 14 C-labeled amino acids meets various criteria for vacuolar transport energized by a H ϩ -ATPase (Table I). The most evident properties diverging from intact cells are a pH optimum near 7.0, strong dependence on ATP supply and osmotic stabilization, sensitivity toward bafilomycin and nitrate (inhibitors of tonoplast H ϩ -ATPases), but not toward vanadate and azide (inhibitors of plasma membrane-and mitochondrial H ϩ -ATPases, respectively).
Quantitative data suggest that the transport capacities of the permeabilized cells are close to that of vacuoles in intact cells; the accumulation ratio (C in /C out at an external concentration of 50 mol/liter) reached 80% for Phe or Arg and 75% for Leu of that observed with intact cells. The high degree of inhibition by bafilomycin and the absence of inhibition by azide implies that other organelles of the permeabilized cells (e.g. mitochondria) did not significantly contribute to the uptake of the labeled amino acids.
Summarizing, it seems justified to assume that the established procedure permeabilizes the plasma membrane for micromolecules without significantly impairing the ability of the vacuole to maintain a proton gradient and to accumulate external amino acids. Thus, reliable studies of the vacuolar amino acid transport can be performed in situ, i.e. with the vacuolar transport agency kept in place within its macromolecular environment.

Dynamic Compartmentation of Amino Acids Probed by Efflux Analysis
Initial rates of efflux of labeled amino acids were determined after the cells had accumulated these compounds over different periods of time and hence had reached different vacuolar concentrations. Most of the data were obtained by a double tracer technique (''indirect method'') that allowed the simultaneous monitoring of unidirectional uptake and net accumulation. Efflux was then calculated as the difference of these both parameters (Fig. 3A). This method, which does not require the exchange of the external medium, is more sensitive and yields more precise results than the direct approach, i.e. following the   B). B, threshold values of the intracellular amino acid content that is required for minimum efflux (abscissa) versus accumulation capacity (ordinate), linear regression line. Data were determined with 14 C-labeled Phe, Leu, Arg, and Glu in experiments like those shown in A. Accumulation capacity is given as the maximum content of the labeled amino acid, which was calculated via hyperbolic fitting of the accumulation curve. Data for Phe were determined not only in nitrogen-starved cells (as in all other experiments) but also in nitrogen-sufficient cells as they express lower transport capacities (25). Means from five independent experiments are given, S.D. ϭ 9 -12% for threshold concentrations, 11-14% for accumulation capacities. Fit to linear regression r 2 ϭ 0.99. The endogenous contents of the used amino acids, present at t ϭ 0, and their vacuolar portions are as follows (nmol/mg protein, % in vacuole): Phe: 8, 80%; Arg: 20, 15%; Glu: 140, 31%, Phe in nitrogen-sufficient cells: 30, 50% (W. Roos and K. Starke, unpublished results). As these amounts are well below the labeled amino acids accumulated in vacuoles during the experiment, they could be neglected for the determination of accumulation ratios and efflux rates.
decrease of the cellular content of preaccumulated labeled amino acids after resuspension of the cells in tracer-free medium (Fig. 4). Both methods reported similar efflux properties.
Efflux from Intact Cells-As shown in Fig. 4, A, the amino acids did not significantly leak out of intact cells until their total content had reached a distinct threshold value. Above this level efflux tended to increase with the cellular amino acid content. The threshold content required for efflux was linearly related to the cell's actual accumulation capacity (Fig. 3B). Under the conditions used this relation was true for the four investigated amino acids and for hyphal cells that express different accumulation capacities for a given amino acid (Phe), e.g. hyphae from nitrogen-sufficient compared with nitrogenstarved cell suspensions (25). No threshold level was required for efflux to occur in the presence of uncouplers or respiratory inhibitors, indicating that the threshold phenomenon involves active transport processes (Fig. 4A).
Efflux from Permeabilized Cells-Vacuoles in permeabilized cells allowed an efflux of preaccumulated amino acids from any initial content, i.e. no detectable threshold concentration had to be surpassed. As an example, Fig. 4B shows efflux of [ 14 C]Phe from in situ vacuoles with Phe contents below 200 nmol/mg protein. In the corresponding experiment with intact cells, similar Phe contents do not allow a measurable efflux. The release of efflux from this restriction after permeabilizing the plasma membrane indicates that the threshold phenomenon requires the plasma membrane to function as a permeability barrier. Efflux from the in situ vacuoles is not an energyrequiring process as it proceeded in the presence of uncouplers (cf. Fig. 4B) or in the absence of ATP. It was not influenced by the actual redox state of the medium as it occurred unchanged in the presence of dithiothreitol (Fig. 4B, efflux after a 10-min  accumulation).
The efflux of vacuolar amino acids shows a biphasic concentration dependence (Fig. 4B, inset). The ratio of efflux rate versus actual Phe content (given as the percentage of lost tracer content/min) increased after the Phe content surpassed a critical level around 200 nmol/mg of protein. This content comes close to the above threshold value that was required to trigger efflux from intact cells (cf. Figs. 3B and 4A, inset). The switch to a steeper concentration dependence of efflux is not an artifact that might be caused by instability of the in situ vacuoles but a reversible event (cf. Fig. 4B). Vacuolar Phe contents above the critical level were reached only transiently (e.g. within 20 and 60 min of accumulation), and after the return to lower levels (e.g. after 60 min), the enhanced efflux/content ratio was readjusted to the original values. Thus, allosteric acceleration of vacuolar efflux might also be involved in the regulation of the vacuolar pool size of amino acids in intact cells.

Effects of Adenylates on Initial Efflux Rates
The initial rate of amino acid efflux from in situ vacuoles is controlled by the concentration of ATP and other adenylates in the surrounding solution (Fig. 5A). In the absence of external ATP (its endogenous concentration detectable under these conditions was Ͻ1 mol/liter), a fairly high rate of Phe efflux was measurable, which decreased with increasing concentrations of added ATP. A strong inhibition was observed at ATP levels around 200 mol/liter. Further increase up to 1 mmol/liter of ATP relieved the efflux from inhibition and led to the same or even higher rates than in cells that were not supplied with ATP. The ATP-dependent release of vacuolar Phe is also demonstrated in Fig. 4A (curves starting at t ϭ 15 min). The ATPase resistant ATP analogue AMP-PNP influenced the efflux of Phe in a similar fashion as did ATP (Fig. 5A). AMP and ADP could replace ATP with similar efficiency and concentration dependence. The same is true for mixtures of the adenylates (Table II). The observed effects of adenylates hold true not only for the efflux of vacuolar Phe but also of Leu, Arg, and Glu (data not shown).
Obviously, the adenylate level exerts regulatory effects on the efflux of vacuolar amino acids which are not mediated via an ATPase dependent energy supply but rather by the carrier(s) responsible for amino acid efflux.
The uptake of [ 14 C]Phe by in situ vacuoles required ATP at a minimum concentration of approximately 100 mol/liter and increased to maximum rates when ATP was raised to 1 mmol/ liter (Fig. 5A). Expectedly, AMP-PNP, ADP, or AMP did not support Phe uptake into in situ vacuoles (data not shown), which reflects the ATPase-dependent energization of this process. ATP is the main component of the cellular adenylate pool (cf. Fig. 5B). As this nucleotide acts both as an energy source for uptake and a regulator of efflux of vacuolar amino acids, it appears very likely that its actual concentration determines distinct relations between amino acid uptake and efflux: (a) at ATP concentrations around 1 mmol/liter, both uptake and efflux reach maximum velocities, i.e. the exchange of vacuolar and cytosolic amino acids occurs at maximum rates; (b) at ATP concentrations around 0.2 mmol/liter, efflux is reduced to a minimum, and hence unidirectional uptake is the dominating process, even though it occurs with only 40% of its maximum rate; (c) at ATP concentrations below 0.1 mmol/liter, no uptake occurs due to insufficient energy input; unidirectional efflux is activated and lasts up to the depletion of the vacuolar pool.
Concentrations of ATP and other adenylates that exert the observed regulatory effects in permeabilized hyphae can be assumed to exist in intact cells under distinct physiological conditions. Fig. 5B displays adenylate concentrations (calculated on the basis of intracellular water) that arise during 4 days of starvation for carbon and nitrogen sources of submerged cultures. It can be seen that adenylate levels drop from 2 mmol/liter to 30 mol/liter, which indicates that the cells run through all the above mentioned relations of vacuolar uptake and efflux. At any time ATP was found to be the dominant constituent (around 70%) of the total adenylates. Obviously, within the period of starvation the adenylate pool as a whole was gradually degraded without severely changing the relation defined as the free energy charge (see "Discussion").
The effects of adenylates on the vacuolar efflux were not substantially modified by extravacuolar amino acids. External amino acids at concentrations Ͼ100 mol/liter accelerated the efflux of the same, preaccumulated amino acid especially in the presence of millimolar concentrations of ATP. However, efflux was not relieved from the inhibition caused by ϳ 200 mol/liter ATP or AMP-PNP, indicating that the trans-stimulation of efflux did not interfere with its control by adenylates.

Dynamics of Vacuolar Uptake and Efflux-
The characteristics of amino acid efflux from in situ vacuoles and the comparison with efflux data from intact cells indicate a dynamic equilibrium between active vacuolar accumulation and passive efflux into the cytosol. Most probably, in nutrient-sufficient cells efflux of vacuolar amino acids is a steady process, which is compensated for by active reuptake, thus keeping their cytosolic concentration low enough to prevent an efflux out of the cell. When the vacuolar concentration approaches a critical level, allosteric acceleration causes the rate of efflux to surpass the capacity of active uptake across the tonoplast. The resulting rise of the cytosolic amino acid concentration may then cause the observed efflux out of the cell. Thus, the threshold phenomenon observed with intact cells mainly reflects the functional balance between different transport steps at the tonoplast and the plasma membrane rather than massive binding of amino acids to protein components. The actual size of the vacuolar pool and its share of the cellular amino acid content is controlled by at least four factors. (a) The accumulation capacity of the tonoplasts active amino acid transport systems, which is supported by the observed linear relation between accumulation capacity and threshold content; (b) the affinity of the efflux carriers at the plasma membrane that prevents significant leakage of amino acids at low cytosolic concentrations, which is supported by our finding that efflux required no threshold content after the plasma membrane had been permeabilized; (c) the actual concentrations of vacuolar amino acids that drive the efflux with a biphasic concentration dependence (the involvement of carriers rather than channels in the efflux process is indicated by the trans-stimulation of efflux exerted by extravacuolar amino acids); and (d) by the actual ATP and adenylate content of the cytosol.
The Cellular Adenylate Pool as a Regulatory Signal for Vacuolar Efflux-To our knowledge, a direct regulatory effect of ATP and other adenylates on vacuolar efflux has not yet been reported from fungal cells. In part, our results might be related to earlier data of amino acid transport in plant vacuoles. Dietz et al. (9) found the efflux of various amino acids from isolated vacuoles of barley protoplasts was stimulated by 10 mmol/liter ATP or AMP-PNP. This finding is analogous to what occurs in the fungal system in the upper range of ATP concentrations. However, our data reveal an additional band of regulatory  active adenylate levels around 100 -200 mol/liter, which cause a drastic reduction of vacuolar efflux. When comparing barley vacuoles and our object, it must be realized that the uptake of amino acids shows significant differences. Whereas in barley vacuoles the transport of [ 14 C]Ala was independent of proton-motive force and hence not coupled to energization via the tonoplast H ϩ -ATPase, in P. cyclopium uptake of [ 14 C]Phe requires ATP and is sensitive to uncouplers. It is further possible that the effect of low ATP concentrations on vacuolar transport as observed with Penicillium vacuoles in situ cannot be detected by in vitro experiments (like that performed with barley vacuoles) as it requires cytosolic components that are present under in situ conditions but not in vitro. Extended work with different plant and fungal objects is required to determine whether the observed regulatory mechanism is a common phenomenon.
The biological significance of an ATP-mediated control over the efflux from the vacuole into the cytosol appears obvious if one considers the dynamics of cellular adenylates during the adaptive response to a long term carbon limitation (cf. Fig. 5B). Whereas in nutrient-sufficient cells the ATP and adenylate content is high enough to allow an exchange of vacuolar and cytosolic amino acids at maximum rates, cells respond to a decrease of the ATP and adenylate level in two steps. First, a drop toward ϳ10% triggers an amino acid-and ATP-saving strategy, i.e. stop of vacuolar efflux and reduction of vacuolar uptake. Second, a severe loss of energy (adenylate levels far below 10%) leads to increasing rates of efflux and metabolization of stored amino acids thus supporting the survival of the cell.
The mechanism behind the control of efflux exerted by the adenylate concentration remains to be elucidated. The present data argue for a direct influence of cytoplasmic adenylates on the vacuolar amino acid carriers. Some alternate possibilities appear less or not supported by our data, especially because of the similarity of effects caused by AMP, ADP, ATP, or AMP-PNP, which is also consistent with the passive, uncouplerinsensitive nature of the efflux step. Thus, proton-motive forcemediated effects on the amino acid transporters, ATPasedependent processes, or regulatory influences via the adenylate energy charge are rendered unlikely. In fact, the adenylate energy charge was maintained in the range between 0.75 and 0.85 (calculated from Fig. 5B) during a 4-day starvation period irrespective of whether the adenylate content reached an efflux-permitting or efflux-inhibiting range. This behavior is in contrast with the decline of adenylate free energy in yeast cells during prolonged starvation periods (39). Thus, it appears that in our object the total adenylate level rather than any relationship between the adenylate compounds exerts regulatory functions. Redox-mediated activities of vacuolar ion channels (as found with cation channels in the yeast vacuole) (40) are not consistent with the observed lack of influence of dithiothreitol on efflux rates. Furthermore, the involvement of cation channels cannot explain the similarities of efflux between the neutral (Phe), acidic (Glu), and basic (Arg) amino acids used in this study as well as the observed stimulation of efflux by extravacuolar amino acids. In plant cells, some vacuolar ion chan-nels (SV-type) have been found to be influenced by ATP (41,42), but in contrast to our object nonhydrolizable ATP analogues were ineffective. Thus, an adenylate-mediated control of the vacuolar pool might add a new facet to our knowledge of intracellular metabolic control and thus deserves further experimental efforts, e.g. to identify the site(s) of efflux control.