Unbalance of L-Lysine Flux in Corynebacterium glutamicum and Its Use for the Isolation of Excretion-Defective Mutants

Transcription

1 JOURNAL OF BACTERIOLOGY, July 1995, p Vol. 177, No /95/$ Copyright 1995, American Society for Microbiology Unbalance of L-Lysine Flux in Corynebacterium glutamicum and Its Use for the Isolation of Excretion-Defective Mutants MARINA VRLJIC, WOLFGANG KRONEMEYER, HERMANN SAHM, AND LOTHAR EGGELING* Biotechnologie 1, Forschungszentrum Jülich GmbH, D Jülich, Germany Received 28 November 1994/Accepted 5 May 1995 We found that the simple addition of L-methionine to the wild type of Corynebacterium glutamicum results in excretion of the cellular building block L-lysine up to rates of 2.5 nmol/min/mg (dry weight). Biochemical analyses revealed that L-methionine represses the homoserine dehydrogenase activity and reduces the intracellular L-threonine level from 7 to less than 2 mm. Since L-lysine synthesis is regulated mainly by L-threonine (plus L-lysine) availability, the result is enhanced flux towards L-lysine. This indicates a delicate and not well controlled type of flux control at the branch point of aspartate semialdehyde conversion to either L-lysine or L-threonine, probably due to the absence of isoenzymes in C. glutamicum. The inducible system of L-lysine excretion discovered was used to isolate mutants defective in the excretion of this amino acid. One such mutant characterized in detail accumulated 174 mm L-lysine in its cytosol without extracellular excretion of L-lysine, whereas the wild type accumulated 53 mm L-lysine in the cytosol and 5.9 mm L-lysine in the medium. The mutant was unaffected in L-lysine uptake or L-isoleucine or L-glutamate excretion, and also the membrane potential was unaltered. This mutant therefore represents a strain with a defect in an excretion system for the primary metabolite L-lysine. The gram-positive organism Corynebacterium glutamicum has rather simple types of flux control within the biosynthetic pathway of the aspartate family of amino acids. Thus, C. glutamicum has only one aspartate kinase (14), with no indication that its synthesis is regulated (6), but which is inhibited in its activity by the common presence of L-lysine plus L-threonine (19, 28). Escherichia coli has three different aspartate kinase isoenzymes (30), which are independently regulated in activity or synthesis by one of the four possible end products of the highly branched pathway of aspartate-derived amino acid synthesis. Similarly, Bacillus subtilis has three differently regulated aspartate kinases (13). Another example of simple control structure within the synthesis of the aspartate-derived amino acids is the acetohydroxy acid synthase, where again C. glutamicum has only one enzyme (15), but E. coli is equipped with at least three isoenzymes (31). Moreover, E. coli has bifunctional proteins, having, for instance, kinase and homoserine dehydrogenase activities (30), which is not the case in C. glutamicum. Most of the structures of the genes of the aspartate family are analyzed in C. glutamicum (8, 10), as is the flux through the peculiar split pathway of L-lysine synthesis in this organism (Fig. 1) (29). Flux control towards L-lysine is exerted at the level of aspartate kinase and dihydrodipicolinate synthase (6). This type of control as evolved in C. glutamicum appears to be efficient for a balanced flux towards L-lysine and prevents wasting of this valuable metabolic building block. Only if the regulation of the kinase or synthase is artificially overcome is L-lysine synthesis in the cell increased (5). This results in an intracellular accumulation of L-lysine (2), accompanied by efflux of this amino acid. It is therefore initially surprising that C. glutamicum is supposed to have defined mechanisms at its disposal to excrete the primary metabolite L-lysine. Instead, specific efflux systems are known in detail for the removal of fermentation products (1) * Corresponding author. Mailing address: Institut für Biotechnologie 1, Forschungszentrum Jülich, D Jülich, Germany. Phone: Fax: or the removal of heavy metal ions (21), or they are present to expel antibiotics (25) and are thus directly required for the survival or metabolism of the cell. However, as recently shown, C. glutamicum can take up peptides, which become hydrolyzed to their amino acid constituents within the cell (12). As a consequence, the organism then has to deal with the surplus amino acids which cannot be metabolized and remove them in this particular situation. Formation of amino acid constituents derived from peptides has also been noted with E. coli (23) and Streptococcus faecalis (22). In the case of C. glutamicum, detailed biochemical investigations of L-lysine efflux characteristics have shown that a carrier-mediated L-lysine export system must be present in this organism (4). Transport of L-lysine from the cell to the medium is characterized as a secondary active process, which is modulated by the membrane potential, the L-lysine gradient, and the proton gradient (4). In the present work, we describe our finding that the flux of L-lysine biosynthesis in wild-type C. glutamicum is increased by L-methionine addition. We characterize this effect with respect to the flux disturbance within the aspartate family of amino acids. This system of inducible L-lysine excretion enabled us to isolate an L-lysine export-deficient mutant and to characterize the consequences of this mutation for the cell. MATERIALS AND METHODS Strains and media. The wild type of C. glutamicum was used (ATCC 13032), as well as the lysine-secreting strain 52-5, with feedback-resistant aspartate kinase (18), and the restriction-deficient strain R127 (17). The lysine auxotrophic strain 49/3 is a lysa mutant of the wild type, devoid of diaminopimelate decarboxylase activity. Strains were precultivated on complex medium CGIII (18). Growth on minimal medium (liquid cultures) was done on CGXII (15) or (as solidified medium) on CGIX (15). Isolation of mutants. N-Methyl-N -nitro-n-nitrosoguanidine was used as a mutagen as described before (18). Mutants were spread on CGIX containing strain 49/3 at a density of 10 7 cells per ml. Mutant colonies were inspected for L-lysine excretion as halo formation by growth of the lysine-dependent indicator strain 49/3. Those mutants showing reduced halo formation on CGIX (plus strain 49/3) plus 5 mm L-methionine but normal growth on the control (CGIX plus 49/3) were analyzed further. Determination of amino acid concentrations. Cells were separated from the medium and inactivated by silica oil centrifugation with the further working up procedure described before (7), yielding cellular and extracellular fractions. The 4021

2 4022 VRLJIC ET AL. J. BACTERIOL. TABLE 1. Excretion of L-lysine by the wild type of C. glutamicum (ATCC 13032) in response to amino acid addition a Amino acid added (5 mm) Lysine accumulated (mm) Growth rate (h 1 ) None (control) (n 3) L-Methionine (n 4) L-Isoleucine L-Valine L-Proline L-Tryptophan a Lysine accumulation was determined at the end of the log growth phase (12 h). Data from one experiment are given except for two growth rate determinations. FIG. 1. Simplified biosynthetic pathway of the aspartate family of amino acids in C. glutamicum, including the relevant steps for L-methionine-dependent L-lysine excretion and the known regulation by either transcriptional control of structural genes (shaded lines with square ends) and feedback inhibition of enzyme activities (shaded lines with arrowhead ends). amino acids were separated as their ortho-phthdialdehyde derivatives by automatic precolumn derivatization and reversed-phase chromatography, with fluorometric detection on the ChemStation LC1090A (Hewlett-Packard, Avondale, Calif.). Determination of enzyme activities, membrane potential, and excretion rates. Homoserine dehydrogenase was assayed in the forward direction by monitoring the NADPH-dependent reduction of aspartate semialdehyde (6). The feedback resistance of the aspartate kinase of mutants derived from strain 52-5 was assayed by picking the clones on minimal medium (CGIX) including 5 mm L-threonine plus 5 mm L-lysine analog S-(2-aminoethyl)-L-cysteine. Since the presence of threonine plus the analog inhibits the single aspartate kinase activity of C. glutamicum (6), growth of strains with feedback-sensitive enzyme is inhibited, whereas strains with feedback-resistant enzyme can grow. The membrane potential was determined by the use of tetraphenylphosphonium bromide (7). Excretion rates were calculated from at least six points during cultivation, obtained from the differential increase in L-lysine accumulated divided by the differential increase in cell mass formation (taking an absorbance of 1 at an optical density of 600 nm [A 600 ] as 0.3 g [dry weight] per liter). RESULTS The single aspartate kinase of C. glutamicum is subject to concerted feedback inhibition by L-lysine plus L-threonine (Fig. 1). Consequently, in strains auxotrophic for L-threonine, the feedback inhibition of the kinase is abolished, which results in increased L-lysine synthesis (19). We wanted to use this effect for the triggering of L-lysine synthesis and therefore performed experiments with a hom mutant (devoid of homoserine dehydrogenase activity) of C. glutamicum, which was assayed for L-lysine synthesis in response to a limiting supply of L-threonine and L-methionine (not shown). As a control, we also performed experiments with the wild type of C. glutamicum.to our surprise, we noticed that even the addition of L-methionine alone resulted in L-lysine excretion. Specificity of the methionine effect. We therefore assayed several amino acids for their possible effect on triggering of L-lysine formation and quantified the amounts of L-lysine excreted. As shown in Table 1, the addition of L-methionine resulted in a rise in the external L-lysine concentration from 0.04 to 1.2 mm, which corresponds to an approximately 30-fold increase. The addition of L-isoleucine also resulted in a considerably increased accumulation of L-lysine, whereas the other amino acids assayed had no positive effect on L-lysine formation. This shows that specifically the two amino acids of the homoserine branch of the aspartate family of amino acids influence L-lysine excretion. The excretion of amino acids other than L-lysine, such as L-alanine and L-valine, was not found to be influenced by L-methionine addition (not shown). Interestingly, the L-lysine excretion caused by L-methionine was always accompanied by a slight reduction in the growth rate. Intracellular L-lysine concentration and excretion rates. Since L-lysine excretion by C. glutamicum is so far only known to occur at an increased cytoplasmic L-lysine concentration, we assayed directly for this. The wild type was grown on minimal medium with and without L-methionine, growth was monitored, samples were prepared by the silica oil method, and amino acid analyses were performed to finally yield the intraand extracellular concentrations of L-lysine. As shown in Fig. 2A, the intracellular L-lysine concentration increased to a concentration of about 20 mm without the addition of L-methionine, and it was excreted at only a very low concentration ( 0.2 mm). In contrast to this, the addition of 5 mm L-methionine at the end of the exponential growth phase resulted in a substantially increased cytosolic L-lysine concentration, amounting to 70 mm (Fig. 2B). This demonstrates clearly that the addition of L-methionine results in a cellular flux alteration in L-lysine synthesis, as a consequence of which this amino acid is excreted. The compiled data were used to calculate the L-lysine export rates, since so far only data for nongrowing cultures are available. At the beginning of growth (4 to 8 h) and at an intracellular concentration of 15 to 30 mm L-lysine, the export rate was 1.5 nmol/min/mg (dry weight). However, in the second growth stage (8 to 42 h), it was only 0.5 nmol/min/mg in spite of the high cellular L-lysine concentration present (Table 2). This means that the L-lysine export characteristics are determined not only by the cytoplasmic L-lysine concentration but also by the physiological state of the cells. The known modulation of the export carrier activity by the lysine gradient (4) is unlikely because of the rather increased gradient in the second growth phase, whereas an influence by the ph gradient is possible, as

3 VOL. 177, 1995 LYSINE INCREASE AND EXPORT IN C. GLUTAMICUM 4023 TABLE 3. Homoserine dehydrogenase activities and intracellular L-threonine concentrations in two strains of C. glutamicum a Strain Methionine added Homoserine dehydrogenase sp act ( mol/min/mg) Intracellular L-threonine concn (mm) Wild type No Yes R127 No Yes a The enzyme data are the means from two separate experiments of extracts obtained from cells harvested at the exponential growth phase. Intracellular threonine concentrations were obtained from at least three data points during the log phase of growth at an A 600 of8to12. FIG. 2. Intracellular and extracellular L-lysine concentrations during growth of the wild type of C. glutamicum (A) without further addition and (B) with L-methionine addition. F, intracellular L-lysine;, extracellular L-lysine; å, growth. Data from one typical experiment are given. is regulation of the total activity at the genetic level. In addition to the analysis of the wild-type strain, C. glutamicum R127 was also analyzed. This strain is restriction deficient (17) and therefore of particular use for future cloning experiments. The methionine effect on L-lysine synthesis is also present with this strain (Table 2). However, the intracellular L-lysine concentrations were always higher than in the wild type, as were the L-lysine export rates. For comparison, strain 52-5, which excretes L-lysine because of a feedback-resistant aspartate kinase, was included in the analyses (18). This strain always has high internal L-lysine concentrations, and it has a high export TABLE 2. Intracellular and extracellular L-lysine concentrations and export rates in response to L-methionine addition in different C. glutamicum strains a Strain Methionine added Growth phase Lysine concn (mm) Intracellular Extracellular Export rate (nmol/min/mg) Yes Log Yes Stationary No Log ND b No Stationary ND R127 Yes Log No Log ND 52-5 No Log a Concentrations and rates were determined for at least four time points during cultivation. b ND, not determined. rate of 10 nmol/min/mg, for which altered carrier activity is probably responsible (2, 26). Homoserine dehydrogenase as the link between L-methionine and L-lysine flux. The following facts point to homoserine dehydrogenase as being involved in the methionine effect on L-lysine synthesis. (i) There is no known effect of L-methionine on any enzyme of L-lysine synthesis (6, 26). (ii) Besides L- methionine, L-isoleucine also has weak effects (Table 1), and both amino acids are known to repress the expression of the hom thrb operon in C. glutamicum by a still unknown mechanism (20, 24). (iii) Furthermore, the growth rate is reduced, which could indicate a limited availability of the homoserinederived amino acids. Therefore, a particular regulation of L- lysine flux exerted at the level of the homoserine dehydrogenase appeared obvious. This enzyme is located at the branch point of aspartate semialdehyde conversion to either D,L-diaminopimelate and L-lysine or L-threonine, L-methionine, and L-isoleucine (Fig. 1). Consequently, dehydrogenase activities were determined. As shown in Table 3, the specific enzyme activities were reduced to about 30% of the control levels in the wild type and in strain R127 by L-methionine addition. The basic level in both strains was different but found consistently, probably because the repeated mutagenesis to which strain R127 was subjected (17). Since, however, the repression of hom thrb alone should not interact with the known regulation of L-lysine biosynthesis, the intracellular L-threonine concentrations were also determined. L-Threonine is an allosteric effector of the aspartate kinase (19). The cytoplasmic concentration was 6 to 9 mm in the wild type (Table 3), but in cells grown in the presence of L-methionine, the concentration was only 0.5 to 2 mm. In strain R127, with reduced homoserine dehydrogenase activity, the intracellular L-threonine concentration was reduced from the beginning and was reduced even more in response to L-methionine addition. This shows clearly that the exogenous addition of L-methionine results in repression of the homoserine dehydrogenase, with a concomitant low concentration of L-threonine. The low concentration is surprising, since this implies that in the in vivo situation, the total activity of the methionine-repressed enzyme level cannot be counteracted by the less effective allosteric inhibition of the enzyme by L-threonine. With respect to L-lysine synthesis, we conclude that the decreased L-threonine concentration results in decreased allosteric inhibition of the aspartate kinase, which ultimately results in the observed increased L-lysine flux. This has been genetically proven for strains with constantly altered allosteric control of the kinase (2, 5, 26). Isolation of excretion-defective mutants. The discovered methionine effect was used to isolate mutants deficient in L- lysine export. For this purpose, mutagenized cells of strain R127 were plated on complex medium, and 12,000 clones were

4 4024 VRLJIC ET AL. J. BACTERIOL. TABLE 4. Intracellular lysine concentrations and extracellular lysine accumulation in mutants derived from R127 and 52-5 as identified on plates showing low L-lysine excretion a Strain Methionine added (5 mm) Growth (A 600 ) Lysine concn (mm) Intracellular Extracellular R127 Yes No NA5 Yes No NA6 Yes No NA7 Yes No NA8 Yes No NA9 Yes No No No No No No No No a Data are from one typical experiment. Cells were cultivated in liquid culture, and samples were taken after 24 h, when no further growth occurred. picked on two types of minimal medium plates, containing either 5 mm L-methionine or no L-methionine. In addition, a lysa strain of C. glutamicum was seeded into the minimal medium to monitor L-lysine excretion by the lysine-dependent growth of this strain. After two rounds of retesting, 11 colonies which did not enable growth of the indicator strain although L-methionine was added were finally identified. These clones were subsequently analyzed in liquid culture. In parallel to the work with R127, we also tried to isolate export-deficient mutants of strain 52-5, which has constantly high L-lysine synthesis because of the mutated aspartate kinase (18). Again, mutagenized clones were picked on minimal medium with incorporated indicator strain. Of 12,000 clones investigated, 148 no longer displayed any L-lysine formation. However, an examination of the aspartate kinase regulation indicated that in 137 clones, the enzyme inhibition by L-lysine was restored, and for this reason, no L-lysine could be excreted (26). Eleven potential candidates therefore remained from this approach, which, together with the 11 R127 derivatives, had to be characterized in more detail in liquid culture, since there is no simple test available for export deficiency. For this purpose, clones were grown in shake flask cultures (R127 derivatives with and without methionine) to determine growth and accumulated L-lysine and intracellular lysine concentrations. Data for selected strains are shown in Table 4. Three classes of mutants can be discerned. (i) In one case, growth was strongly reduced (NA5, derived from R127). (ii) In strain NA6 (or 60-4 derived from 52-5), the intracellular L- lysine concentration was severely reduced, and therefore no L-lysine was available for export. (iii) Further false-positive mutants were strains NA7, NA9, and 35-84, which excreted some L-lysine but in amounts too low to allow visible growth of the indicator strain seeded in the plates. (iv) Two strains appeared to be true export-deficient strains. Both mutants NA8 and excreted only very low amounts of L-lysine. Strain NA8 exhibited a threefold increase in the intracellular level of L-lysine compared with R127. The concentration can be increased by L-methionine to the very high concentration of 174 FIG. 3. Comparison of intracellular and extracellular L-lysine concentrations during growth of strain R127 and the export-deficient mutant NA8. Open symbols, strain R127; solid symbols, strain NA8. E, F, intracellular L-lysine;,, extracellular L-lysine. Methionine was added to the cultures at a concentration of 5 mm. Data from one typical experiment are given. mm, with the amount of L-lysine in the culture medium always being below the level of that excreted by R127 or the wild type. The second strain, 35-48, also shows virtually no extracellular accumulation of L-lysine and had a very high cytoplasmic concentration, comparable to that of NA8, when the intracellular L-lysine flux was increased by L-methionine addition. Characterization of the excretion-defective mutant NA8. For a detailed characterization, we focused on strain NA8, since this strain is easier to handle for genetic experiments. First, internal and external L-lysine concentrations in the presence of L-methionine were monitored during growth (Fig. 3). In the parent strain R127, the intracellular L-lysine concentration was about 30 mm between 12 and 18 h of cultivation and increased to about 100 mm. L-Lysine accumulated with time up to 8 mm. Strain NA8 exhibited an intracellular L-lysine concentration profile similar to that of R127, but the total concentration was increased by roughly about 100 mm. Very small amounts of L-lysine were excreted into the culture medium. In addition, export activity was determined in short time experiments (2, 3). As expected, an export activity of NA8 was not detectable ( 0.1 nmol/min/mg [dry weight]). The decrease in the cytoplasmic L-lysine concentration in the growing culture (Fig. 3) from 200 to 80 mm at the second stage of growth is surprising. This discrepancy was solved by calculating the amounts of L-lysine required for incorporation into cell material at this stage and the weak amounts excreted (see Discussion). As a second important characteristic of NA8, we assayed for the specificity of the export mutation. Although, in general, export is more difficult to assay than import (where substances can be simply added to the medium), there are fortunately two methods established for C. glutamicum to trigger specific amino acid excretion. Thus, C. glutamicum excretes L-isoleucine when the precursor 2-ketobutyrate is added to a culture (9). We therefore compared the L-isoleucine export of NA8 and R127 in the system outlined by Ebbighausen et al. (7) (Fig. 4). Upon addition of 2-ketobutyrate, there was an immediate onset of L-isoleucine excretion. Export was indistinguishable between the two strains, having a rate of 2.7 nmol/min/mg. Measurements of the intracellular L-isoleucine level show (Fig. 4) that L-isoleucine was exported at very low concentrations, excluding any efflux due to diffusion of the hydrophobic L- isoleucine. The other amino acid export system assayed was that of L-glutamate. This system is active when C. glutamicum

5 VOL. 177, 1995 LYSINE INCREASE AND EXPORT IN C. GLUTAMICUM 4025 uptake system was altered, since an inversion of the uptake system has also been discussed as an export mechanism (16). For the measurements, cells of NA8 and R127 were grown overnight on LB plus 5% glucose, and the uptake rate within a 2-min time interval was determined with radiolabeled L-lysine (4). The rates were 0.42 and 0.35 nmol/min/mg (dry weight) for NA8 and R127, respectively. We therefore conclude that in NA8, a gene involved in lysine export is specifically inactivated by an undirected mutation. FIG. 4. L-Isoleucine export in mutant NA8 and strain R127 in response to 2-ketobutyrate addition. Ketobutyrate was added during cultivation at time zero. Open symbols, strain R127; solid symbols, strain NA8. E, F, intracellular L- isoleucine;,, extracellular L-isoleucine. Data from one typical experiment are given. becomes limited for its essential growth factor biotin (28). Therefore, cells of NA8 and R127 were grown on minimal medium supplemented with 0.1 g of biotin per liter, and when growth was arrested after 15 h, glutamate accumulation was determined. Figure 5 shows that substantial amounts of glutamate were excreted, which accumulated to more than 14 mm in both cultures. The difference between the two cultures might be due to the slightly different growth characteristics of the two strains. The intracellular concentration of L-glutamate was not influenced, nor was the absence of L-glutamate excretion when biotin supply was high (not shown). As a third characteristic of NA8, we determined the membrane potential, since this is of major importance for determining the activity of the L-lysine export carrier (4). For this purpose, the distribution of the permeant cation tetraphenylphosphonium bromide over the membrane was measured, and the membrane potential was calculated from these data by the Nernst equation. It was 156 mv for NA8 and 149 mv for R127. These values show that the membrane potential, as the basic prerequisite for active lysine export, is not influenced in NA8. Additionally, we also determined whether the L-lysine FIG. 5. L-Glutamate export in mutant NA8 and strain R127 in response to biotin limitation. Open symbols, strain R127; solid symbols, strain NA8. E, F, intracellular L-glutamate;,, extracellular L-glutamate. Data from one typical experiment are given. DISCUSSION Although bacterial export carriers for a variety of purposes are known, only recently has increasing evidence emerged that such export carriers exist for amino acids (16). So far, no single mutant has been described as being deficient in excretion of a primary metabolic building block, and the consequences of such a mutation are also unknown. It was therefore a favorable precondition to discover in the course of our work the methionine effect on L-lysine export. This also gives us clues for important regulatory features of control within the aspartate family of amino acids in C. glutamicum. As we have shown, the cascade of regulatory reactions involved is (i) repression of homoserine dehydrogenase by L- methionine, with the consequence (ii) that the intracellular L-threonine level is reduced. This in turn (iii) reduces the known allosteric inhibition of the aspartate kinase (5), which (iv) leads to an increased flux of aspartate towards L-lysine, which (v) accumulates, and then in the last step (vi) becomes exported. The L-methionine-dependent repression of homoserine dehydrogenase was shown for C. glutamicum and its subspecies C. glutamicum subsp. flavum (19, 20) to result in a reduced enzyme level of at most 20% in synthetic medium. As we have directly proven, the intracellular concentration of L- threonine was reduced dramatically by about 6 mm, to 2 mm or less. The L-isoleucine concentration was always very low ( 1 mm). One astonishing problem is that the rather weak reduction of the dehydrogenase level to 26% of the original activity cannot be counteracted by a reduced allosteric inhibition of the enzyme by L-threonine. The specific activity of the dehydrogenase after L-methionine addition was 0.32 mol/min/mg of protein. At a growth rate of 0.4 h and a combined L-isoleucine and L-threonine demand of 414 mol/g (dry weight) (calculated from the cellular composition of C. glutamicum), and assuming 0.6 g of protein per g (dry weight), the flux through the homoserine dehydrogenase reaction in vivo is only mol/min/mg of protein. We must therefore conclude that the homoserine dehydrogenase operates in vivo at very low substrate concentrations (compared with in vitro), where the effects exerted by L-threonine are more conspicuous, since homoserine dehydrogenase is an allosteric enzyme. The K i for L-threonine inhibition is about mm (19). Therefore, the intracellular concentration determined would strongly influence the activity of the dehydrogenase. As genetic analysis has shown, the homoserine dehydrogenase-encoding gene hom forms a transcriptional unit together with the homoserine kinase-encoding gene thrb of C. glutamicum (24). Thus, upon L-methionine addition, the kinase level is also reduced by repression (20). This might additionally contribute to reduced flux towards L-threonine. In fact, in the opposite situation, when flux through homoserine dehydrogenase was increased, an appropriate level of thrb expression was shown to be required (11). It is obvious from these investigations that a delicate balance of hom thrb expression and enzyme activity control operates in C. glutamicum. Since E. coli is equipped with two isoenzymes with homoserine dehydrogenase activity, of

6 4026 VRLJIC ET AL. J. BACTERIOL. which only the metl-encoded enzyme is repressed by L-methionine (30), this organism is possibly better prepared for flux unbalances in the presence of L-methionine. The systems at present available to trigger L-lysine excretion in C. glutamicum are (i) mutations in the aspartate kinase preventing allosteric control, (ii) overexpression of dihydrodipicolinate synthase, (iii) use of peptides, and (iv) addition of L-methionine to cultures. The advantages and disadvantages are as follows. The mutation in the aspartate kinase always results in increased intracellular flux, as does overexpression of the synthase, and no regulation is possible. The use of L-lysinecontaining peptides results in only a transient increase in the intracellular L-lysine pool (12), and peptides are usually expensive. The methionine effect, however, represents an extremely simple on-off switch for flux increase and export and can be used in any strain (when the basic excretion is low). The disadvantage of using a strain with feedback-resistant aspartate kinase is clearly perceptible from the present work with strain Thus, at a very high frequency, mutants in which the feedback-resistant phenotype was lost have been found. Apparently, at a high intracellular L-lysine concentration, an inactive export carrier is disadvantageous, and probably a second mutation might be required to tolerate it, which might be the case for strain The false-positive mutants with low intracellular L-lysine concentrations might carry leaky mutations within genes of the biosynthetic pathway of L-lysine synthesis, as shown in Fig. 1. But mutations influencing the availability of aspartate (or pyruvate) should also result in reduced L-lysine synthesis. Because of the high reversion rate of strain 52-5 when used for isolating for export-deficient mutants (the strain is otherwise very stable), we focused on the inducible system with strain NA8, in which a selective pressure would be markedly present only when supplied with L-methionine. Detailed observation of the intracellular L-lysine pool of this strain with growth shows that the pool concentration varies with the growth phase. This is also the case for other strains investigated. In the wild type, the intracellular pool could be depleted by excretion, which should not be the case for strain NA8. To solve this discrepancy, we have to consider the metabolic activities filling or depleting the lysine pool, as well as the activities regarding exchange of the pool with the environment. Although we have a unique split pathway of L-lysine synthesis in C. glutamicum of still unknown function (27, 29), a regulation of synthesis except at the level of aspartate kinase and dihydropicolinate synthase is not known. The decrease by about 100 mm from 22 to 27.5 h in strain NA8 (Fig. 3) is accompanied by growth. Calculation of the dry weight increase, the lysine content of C. glutamicum of 191 mol/g (dry weight), and the cytoplasmic volume of the culture show that this decrease matches the sum of the weak increase in L-lysine in the culture (0.29 mm) plus that required for cell synthesis (0.4 mm). The calculated L-lysine efflux rate, however, is with the culture (Fig. 3) below 0.2 nmol/min/mg, and in a short time experiment almost no efflux ( 0.1 nmol/min/mg) was measurable. In contrast, other C. glutamicum strains at considerably lower intracellular L-lysine contrations have activities of at least 3.5 nmol/min/mg (4, 12), with the highest activity of 9.5 nmol/min/mg obtained for an L-lysine-producing C. glutamicum strain (26). Thus, NA8 is severely hampered in L-lysine export. Since L-isoleucine and L-glutamate export is not influenced by the mutation, we conclude that strain NA8 is the first mutant known to be specifically inactivated in a gene required for L-lysine export. This might open up entirely new perspectives on the investigation of the overproduction of this important metabolic building block. ACKNOWLEDGMENTS This work is part of a joint project with Degussa AG and was supported by grant from the Federal Ministry for Research and Technology. REFERENCES 1. Benthin, S., J. Nielsen, and J. Villadsen Galactose expulsion during lactose metabolism in Lactococcus lactis subsp. cremoris FD1 due to dephosphorylation of intracellular galactose 6-phosphate. Appl. Environ. Microbiol. 60: Bröer, S., L. Eggeling, and R. Krämer Strains of Corynebacterium glutamicum with different lysine productivities may have different lysine excretion systems. Appl. Environ. Microbiol. 59: Bröer, S., and R. Krämer Lysine uptake and exchange in Corynebacterium glutamicum. J. Bacteriol. 172: Bröer, S., and R. Krämer Lysine excretion by Corynebacterium glutamicum. 2. Energetics and mechanism of the transport system. Eur. J. Biochem. 202: Cremer, J., L. Eggeling, and H. Sahm Control of the lysine biosynthetic sequence in Corynebacterium glutamicum as analyzed by overexpression of the individual corresponding genes. Appl. Environ. Microbiol. 57: Cremer, J., C. Treptow, L. Eggeling, and H. Sahm Regulation of enzymes of lysine biosynthesis in Corynebacterium glutamicum. J. Gen. Microbiol. 134: Ebbighausen, H., B. Weil, and R. Krämer Isoleucine excretion in Corynebacterium glutamicum: evidence for a specific efflux carrier system. Appl. Microbiol. Biotechnol. 31: Eggeling, L Biology of L-lysine overproduction by Corynebacterium glutamicum. Amino Acids 6: Eggeling, I., C. Cordes, L. Eggeling, and H. Sahm Regulation of acetohydroxy acid synthase in Corynebacterium glutamicum during fermentation of -ketobutyrate to L-isoleucine. Appl. Microbiol. Biotechnol. 25: Eikmanns, B. J., L. Eggeling, and H. Sahm Molecular aspects of lysine, threonine, and isoleucine biosynthesis in Corynebacterium glutamicum. Antonie von Leeuwenhoek J. Microbiol. Serol. 64: Eikmanns, B., M. Metzger, D. Reinscheid, M. Kircher, and H. Sahm Amplification of three biosynthesis genes in Corynebacterium glutamicum and its influence on carbon flux in different strains. Appl. Microbiol. Biotechnol. 34: Erdmann, A., B. Weil, and R. Krämer Lysine secretion by wild-type Corynebacterium glutamicum triggered by dipeptide uptake. J. Gen. Microbiol. 139: Graves, L. M., and R. L. Switzer Aspartokinase III, a new isozyme in Bacillus subtilis 168. J. Bacteriol. 172: Kalinowski, J., J. Cremer, B. Bachmann, L. Eggeling, H. Sahm, and A. Pühler Genetic and biochemical analysis of the aspartokinase from Corynebacterium glutamicum. Mol. Microbiol. 5: Keilhauer, C., L. Eggeling, and H. Sahm Isoleucine in Corynebacterium glutamicum: molecular analysis of the ilvb-ilvn-ilvc operon. J. Bacteriol. 175: Krämer, R Secretion of amino acids by bacteria: physiology and mechanism. FEMS Microbiol. Rev. 13: Liebl, W., A. Bayerl, U. Stillner, and K. H. Schleifer High efficiency electroporation of intact Corynebacterium glutamicum cells. FEMS Microbiol. Lett. 65: Menkel, E., G. Thierbach, L. Eggeling, and H. Sahm Influence of increased aspartate availability on lysine formation by a recombinant strain of Corynebacterium glutamicum and utilization of fumarate. Appl. Environ. Microbiol. 55: Miyajima, R., and I. Shiio Regulation of aspartate family amino acid biosynthesis in Brevibacterium flavum. J. Biochem. (Tokyo) 68: Miyajima, R., and I. Shiio Regulation of aspartate family amino acid biosynthesis in Brevibacterium flavum: repression of the enzymes in threonine biosynthesis. Agric. Biol. Chem. 35: Nies, D. H., and S. Silver Plasmid-determined inducible efflux is responsible for resistance to cadmium, zinc, and cobalt in Alcaligenes eutrophus. J. Bacteriol. 171: Nisbet, T. M., and J. W. Payne The characteristics of peptide uptake in Streptococcus faecalis: studies on the transport of natural peptides and antibacterial phosphonopeptides. J. Gen. Microbiol. 128: Payne, J. W., and G. Bell Direct determination of the properties of peptide transport systems in Escherichia coli, using fluorescent-labeling procedure. J. Bacteriol. 137: Peoples, O. P., W. Liebl, M. Bodis, P. J. Maeng, M. T. Follettie, J. A. Archer, and A. J. Sinskey Nucleotide sequence and fine structural analysis of the Corynebacterium glutamicum hom-thrb operon. Mol. Microbiol. 2: Poole, K., K. Krebes, C. McNally, and S. Neshat Multiple antibiotic resistance in Pseudomonas aeruginosa: evidence for involvement of an efflux operon. J. Bacteriol. 175:

7 VOL. 177, 1995 LYSINE INCREASE AND EXPORT IN C. GLUTAMICUM Schrumpf, B., L. Eggeling, and H. Sahm Isolation and prominent characteristics of an L-lysine hyperproducing strain of Corynebacterium glutamicum. Appl. Microbiol. Biotechnol. 37: Schrumpf, B., A. Schwarzer, J. Kalinowski, A. Pühler, L. Eggeling, and H. Sahm A functionally split pathway for lysine synthesis in Corynebacterium glutamicum. J. Bacteriol. 173: Shiio, I., S. I. Otsuka, and M. Takahashi Effect of biotin on the bacterial formation of glutamic acid. J. Biochem. (Tokyo) 52: Sonntag, K., L. Eggeling, A. A. De Graaf, and H. Sahm Flux partitioning in the split pathway of lysine synthesis in Corynebacterium glutamicum: quantification by 13 C- and 1 H-NMR spectroscopy. Eur. J. Biochem. 213: Theze, J., D. Margarita, G. N. Cohen, F. Borne, and J. C. Patte Mapping of the structural genes of the three aspartokinases and of the two homoserine dehydrogenases of Escherichia coli K-12. J. Bacteriol. 117: Umbarger, H. E Biosynthesis of the branched-chain amino acids, p In J. L. Ingraham, K. B. Low, B. Magasanik, M. Schaechter, and H. E. Umbarger (ed.), Escherichia coli and Salmonella typhimurium: cellular and molecular biology. American Society for Microbiology, Washington, D.C.