Evidence for direct interaction between enzyme I(Ntr) and aspartokinase to regulate bacterial oligopeptide transport.

Bradyrhizobium japonicum transports oligopeptides and the heme precursor delta-aminolevulinic acid (ALA) by a common mechanism. Two Tn5-induced mutants disrupted in the lysC and ptsP genes were identified based on the inability to use prolyl-glycyl-glycine as a proline source and were defective in [(14)C]ALA uptake activity. lysC and ptsP were shown to be proximal genes in the B. japonicum genome. However, RNase protection and in trans complementation analysis showed that lysC and ptsP are transcribed separately, and that both genes are involved in oligopeptide transport. Aspartokinase, encoded by lysC, catalyzes the phosphorylation of aspartate for synthesis of three amino acids, but the lysC strain is not an amino acid auxotroph. The ptsP gene encodes Enzyme I(Ntr) (EI(Ntr)), a paralogue of Enzyme I of the phosphoenolpyruvate:sugar phosphotransferase (PTS) system. In vitro pull-down experiments indicated that purified recombinant aspartokinase and EI(Ntr) interact directly with each other. Expression of ptsP in trans from a multicopy plasmid complemented the lysC mutant, suggesting that aspartokinase normally affects Enzyme I(Ntr) in a manner that can be compensated for by increasing the copy number of the ptsP gene. ATP was not a phosphoryl donor to purified EI(Ntr), but it was phosphorylated by ATP in the presence of cell extracts. This phosphorylation was inhibited in the presence of aspartokinase. The findings demonstrate a role for a PTS protein in the transport of a non-sugar solute and suggest an unusual regulatory function for aspartokinase in regulating the phosphorylation state of EI(Ntr).

Bacteria utilize small peptides as nutrients, chemoattractants, and quorum sensing signals, and their metabolism is a target for antibiotics (1)(2)(3). Escherichia coli and Salmonella typhimurium contain distinct dipeptide (Dpp) 1 and oligopeptide permease (Opp) systems with some overlap in substrate specificity. The two permease systems are structurally homologous, each one containing five proteins, including a periplasmic peptide-binding protein (4,5). The Opp system binds peptides two to five peptides in length, with the highest affinity for tripeptides (6,7). For both permease systems, the amino acid side chain appears not to be important for specificity, and therefore these systems transport peptides independent of sequence.
Our interest in oligopeptide transport in the bacterium Bradyrhizobium japonicum is founded on studies of heme biosynthesis, where it has been demonstrated that the heme precursor ␦-aminolevulinic acid (ALA) is taken up by a system that also transports oligopeptides (8). ALA is structurally similar to glycyl-glycine. It is taken up by the Dpp system in E. coli (9) and S. typhimurium (10), but ALA is taken up in a dpp mutant if the opp system is activated (8). B. japonicum lives as a free-living soil bacterium or as an endosymbiont of soybeans within root nodules. In symbiosis, B. japonicum may utilize ALA synthesized by the plant host for heme formation (11,12).
The phosphoenolpyruvate (PEP):sugar phosphotransferase system (PTS) couples the transfer of a high energy phosphoryl group from PEP to a sugar with the concomitant transport into the cell (13)(14)(15). Paralogues of the PTS enzymes EI, Hpr and IIA, called EI Ntr , Npr, and IIA Ntr , respectively, have been identified in several organisms (16 -21), and evidence for a parallel phosphoryl transfer chain has been presented (22). Phenotypes of mutants in the parallel PTS system (herein called the PTS Ntr system), gene organization and the structure of EI Ntr , suggest a role for this system coordinating nitrogen and carbon metabolism. However, the physiological functions of the PTS Ntr system is unclear. Although the conventional PTS system transports sugars, no solute transport activity has been linked to the PTS Ntr proteins. Herein, we demonstrate a role for EI Ntr , encoded by the ptsP gene, in oligopepide transport in B. japonicum. Furthermore, a novel function for the amino acid biosynthesis enzyme aspartokinase was also implicated, and evidence for interaction between EI Ntr and aspartokinase is presented.

MATERIALS AND METHODS
Chemicals and Reagents-All chemicals were reagent grade and were purchased from Sigma Chemical Co., St Bacterial Strains, Plasmids, Media, and Growth-B. japonicum strain I110proC is a proC mutant derivative of strain I110 and is a proline auxotroph (23). All B. japonicum strains were grown at 29°C in GSY media or minimal media as described previously (24). Tetracycline (75 g/ml) was added to the media for growth of strains bearing the broad host range plasmid pVK102 (25) and its derivatives. Strains I20, Q5, I110I20, and I110Q5 were grown in the presence of kanamycin (50 g/ml) and streptomycin (50 g/ml). Strains I110proC, I20, and Q5 were grown in the presence of spectinomycin (50 g/ml) and streptomycin. Strains I110proC, I20, and Q5 required the addition of proline (500 g/ml) for growth. E. coli strain DH5␣ was used for propagation of plasmids and was grown at 37°C on Luria-Bertani (LB) medium with appropriate antibiotics. E. coli strains HB101(pDS4101) and HB101(pRK2013) were grown in media containing ampicillin (200 g/ ml) and kanamycin (50 g/ml), respectively, for tri-parental matings. Plasmid pSUP1011 is a pACYC184 derivative carrying the transposon Tn5 (26).
Tn5 Mutagenesis and Construction of B. japonicum Mutant Strains I110I20 and I110Q5-A random transposon Tn5 mutagenesis of B. japonicum strain I110proC was carried out as described previously (26). Approximately 3500 mutant colonies were streaked onto media containing either proline or prolyl-glycyl-glycine (50 g/ml) to screen for mutants that are unable to use the latter as a proline source. Two mutants displaying this phenotype were identified and designated strains I20 and Q5. Southern blot analysis revealed that Tn5 from each mutant were contained on EcoRI fragments ϳ12 kb in length, which were then isolated as described previously (26). These fragments, cloned into pBR322, were used to generate strains for the corresponding mutants in parent strain I110 by homologous recombination as described previously (26). The nucleotide sequence of both strands of the HpaI-XhoI region bearing the B. japonicum lysC and ptsP genes was determined.
ALA Uptake Activity by Cells-B. japonicum cells grown to mid-log phase in 100 ml of minimal media were washed and assayed for [ 14 C] ALA (150 M; 2.67 Ci/mol) as described previously (8). Strains I110, I110I20, and I10Q5 were used in the studies for monitoring uptake in wild type, ptsP, and lysC strains, respectively.
Isolation of RNA and mRNA Analysis-Total RNA was prepared and analyzed as described previously (27,28). Analysis of RNA levels of specific genes was carried out by a quantitative RNase protection assay. Plasmid pAspK(RV) is a derivative of pBluescriptSK ϩ that carries a 347-bp EcoRV fragment containing a portion of the lysC gene. Plasmid pPtsP(XhoI) is a derivative of pBluescriptSK ϩ that carries a 402-bp XhoI fragment containing a portion of the ptsP gene. Each plasmid was used as a template to synthesize antisense RNA as described previously (28,29).
Expression and Purification of c-myc/HisX6 Fusion Proteins-AspK and EI Ntr were overexpressed as c-myc/HisX6 fusion proteins from the pTrcHis2 expression vector (Invitrogen). PCR reactions were performed using 5Ј-GGCCATGGCCCGCCTCGTGATG-3Ј and 5Ј-CCTAAGCTTG-ATCGAGGCCGTAGAGCG-3Ј as the forward and reverse primers to amplify the lysC coding region and introduce NcoI and HindIII sites, respectively. The pTrcHis2B-ptsP construct was generated by introducing a Csp45I site at the 5Ј-end and a HindIII site at the 3Ј-end of the ptsP open reading frame by PCR using the primers 5Ј-GGTTCGAA-GCGCGTCGGGAGGTCC-3Ј and 5Ј-GCCAAGCTTAAGGCCAGGCCT-TCGGC-3Ј, respectively. One liter of LB, containing 200 g of ampicillin per ml of media, was inoculated with 10 ml of an overnight culture of E. coli XL-1 Blue cells containing either pTrcHis2B-ptsP or pTrcHis2C-lysC, grown and induced as described previously (28). The cleared lysates were applied to columns each containing 1 ml of nickel-nitrilotriacetic acid settled resin charged with 400 mM NiSO 4 according to the manufacturer's instructions (Novagen). The resin was washed with 15 volumes of PO 4 binding buffer followed by 25 volumes of PO 4 wash buffer (50 mM potassium phosphate, 10% glycerol, 300 mM NaCl, and 30 mM imidazole, pH 8.0). Proteins retained by the resin were eluted with 3 volumes of PO 4 elute buffer (50 mM potassium phosphate, 10% glycerol, 300 mM NaCl, and 250 mM imidazole, pH 8.0) and dialyzed overnight at 4°C against 1 liter of 50 mM potassium phosphate, 10% glycerol, 300 mM NaCl, and 1 mM ␤-mercaptoethanol, pH 8.0. A precipitate formed in the eluted AspK-c-myc/HisX6 fraction after dialysis and was removed by centrifugation. Concentration of proteins was deter-mined using the Bradford protein assay (Bio-Rad).
Expression of GST Fusion Proteins-AspK and EI Ntr were overexpressed as GST fusion proteins from the pGEX-6P-2 expression vector (Amersham Pharmacia Biotech). The pGEX-6P2-lysC construct was generated by introducing a BamHI site at the 5Ј-end and a SmaI site at the 3Ј-end of the lysC open reading frame by PCR using the primers 5Ј-CTCGGATCCATGAGCCGCCTCGTGAT-3Ј and 5Ј-GAGCCCGGGC-TAAGCCTGATCGAGGC-3Ј, respectively. The pGEX-6P2-ptsP construct was generated by introducing a BamHI site at the 5Ј-end and a SmaI site at the 3Ј-end of the ptsP open reading frame by PCR using the primers 5Ј-CTCGGATCCATGCGGAGCGCGTCGGG-3Ј and 5Ј-GAGC-CCGGGCCCGCTACAAGGCCA-3Ј, respectively. Expression was induced by adding 1 mM isopropyl-␤-D-thiogalactopyranoside to the media, and cultures were grown overnight at 15°C. Cells were lysed and cleared as described above, and protein was purified with prewashed glutathione-Sepharose beads (Amersham Pharmacia Biotech) according to the manufacturer's instructions.
GST Pull-down Assay-Crude lysates of E. coli BL21(DE3)pLysS containing the overexpressed proteins GST, GST-AspK, or GST-EI Ntr were prepared as described. 1 ml of each cleared lysate preparation was added to a separate Eppendorf tube containing 50 l of prewashed glutathione-Sepharose beads (Amersham Pharmacia Biotech) and rotated on an Orbitron for 30 min at 4°C. Beads were washed five times with PBS ϩ 0.1% Triton X-100 by centrifugation at 2000 ϫ g for 5 min at 4°C followed by aspiration of supernatant. Beads were resuspended in 250 l of PBS ϩ 0.1% Triton X-100, and the amount of protein bound to the beads was determined using the Bradford protein assay (Bio-Rad). The results of this assay were normalized against a control containing unbound glutathione-Sepharose beads. 5 g of purified B. japonicum EI Ntr -c-myc/Hisx6 fusion protein was added to a volume of beads that contained 20 g of GST or GST-AspK, and 5 g of purified B. japonicum AspK-c-myc/Hisx6 fusion protein was added to a volume of beads that contained 20 g of GST or GST-EI Ntr , in a total volume of 0.5 ml. Reactions were rotated on an Orbitron for 60 min at 4°C. Beads were washed five times with PBS plus 0.1% Triton X-100 by centrifugation at 2000 ϫ g for 5 min at 4°C followed by aspiration removing any trace of supernatant. Beads were resuspended in 25 l of 2ϫ SDSsample buffer, boiled for 5 min, and resolved by SDS-PAGE. After transferring to Immobilon-P (0.45 m polyvinylidene difluoride, Millipore), proteins were detected with anti-c-myc/peroxidase conjugate (Roche Molecular Biochemicals) by chemiluminescence (Renaissance, PerkinElmer Life Sciences).
Assay for Aspartokinase Activity-Aspartokinase activity was determined by measuring asparthydroxamate produced when aspartate is incubated with enzyme in the presence of ATP and hydroxylamine (30). The assay mixture contained 10.4 mM Mg-ATP, 94 mM Tris, HCl buffer (pH 8.0), 1.6 mM MgSO 4 , 10 mM ␤-mercaptoethanol, 10 mM L-aspartate, 800 mM NH 2 OH, 800 mM KCl, and purified B. japonicum AspK-c-myc/ Hisx6 fusion protein in a total volume of 0.5 ml. After incubation at room temperature for 15 min, the reaction was stopped by the addition of 0.5 ml of a 1.7% solution of FeCl 3 in 1 N HCl. After centrifugation, the optical density of the asparthydroxamate-iron complex was measured at 540 nm using a Beckman DU spectrophotometer. Enzyme activity is expressed as the optical density units ϫ 1000. A blank reaction mixture that contained all components except for enzyme served as a control. Each reaction was performed in triplicate in the presence or absence of Mg-ATP.
PEP-dependent Phosphorylation of EI Ntr -B. japonicum EI Ntr -c-myc/ HisX6 fusion was overexpressed and purified as described above. The mixture was incubated for 90 min at 30°C. Because this is an exchange reaction, the concentrations of pyruvate and PEP do not change; therefore, the specific activity of the [ 32 P]PEP can be calculated from the known specific activity of the [␥-32 P]ATP and the initial ATP and PEP concentrations to be ϳ2 ϫ 10 3 Ci/mol. The reaction mixture was used as a [ 32 P]PEP source without further purification. The EI Ntr phosphorylation assay contained, in a 20-l volume, 94 mM Tris, HCl buffer (pH 8.0), 1.6 mM MgSO 4 , 10 mM ␤-mercaptoethanol, 4 g of purified B. japonicum EI Ntr -c-myc/HisX6 fusion protein, and either 125 M [ 32 P]PEP (220 Ci/mol) or 10 mM [␥-32 P]ATP (50 Ci/mol). The reactions were incubated for 30 min at 37°C. Reactions were stopped by adding an equal volume of 2ϫ SDS-sample buffer and incubated at room temperature before resolving on a 7.5% SDS-PAGE. Gels were stained to visualize protein standards, and labeled proteins were detected by autoradiography.

RESULTS
Isolation of Oligopeptide Uptake Mutants of B. japonicum-B. japonicum strain I110proC is disrupted in the proline biosynthesis gene proC and requires an exogenous source of proline for growth (23). We established that strain I110proC could use the proline-containing peptides Pro-Gly or Pro-Gly-Gly, to satisfy its auxotrophy as discerned by growth in liquid or solid media supplemented with 50 g/ml of either compound (Table I). Thus, the strategy for obtaining oligopeptide transport mutants was to screen for mutants of strain I110proC that could not use prolyl-glycyl-glycine (Pro-Gly-Gly) as a proline source. Strain I110proC was mutagenized with Tn5, and kanamycin-resistant colonies were screened for those that could no longer use Pro-Gly-Gly as a proline source, but still grew on proline. Two mutants, strains I20 and Q5, exhibited this phenotype on plates and in liquid cultures and were also unable to use Pro-Gly as a proline source (Table I). Both mutant strains retained the ability to grow on proline as well as strain I110proC.
Mutations in Strains I20 and Q5 Affect ALA Uptake-Radiolabeled tripeptides were not commercially available, but [ 14 C]ALA was available to carry out uptake experiments. The mutations in strains I20 and Q5 were reconstructed in a wild type background, and ALA uptake activity was measured. The mutants were constructed by isolation of the Tn5-containing EcoRI fragment from strains I20 and Q5 followed by introduction into the genome of strain I110 by homologous recombination to generate strains I110I20 and I110Q5 (see "Materials and Methods"). The uptake of [ 14 C]ALA by the mutant and wild type strains was assessed using cells cultured in minimal media. Both mutants had severely reduced ALA uptake activities compared with the parent strain I110 (Fig. 1). These data show that the mutations in strain I20 or Q5, which affect their ability to use proline-containing peptides, also severely inhibit ALA uptake activity. The defect in uptake of ALA, a dipeptide analogue, indicates that the inability of Pro-Gly or Pro-Gly-Gly to satisfy the proline auxotrophy in strains I20 and Q5 is the result of a defect in transport rather than another step of oligopeptide metabolism.
Identification and Characterization of the lysC Gene and Its Product Aspartokinase-Initial subcloning and analysis of Tn5containing genomic fragments from strains I20 and Q5 revealed that the transposon from each strain was inserted into one of two different regions of the same EcoRI fragment. Consequently, the EcoRI fragment isolated from strain I20 contained the wild type copy of the DNA mutated in strain Q5, and vice versa. Therefore, from these EcoRI fragments, the wild type genes were cloned and sequenced, and DNA fragments containing one or the other gene were constructed (see "Materials and Methods"). These two open reading frames have the same orientation, and are separated by a 288-bp intergenic region ( Fig. 2A).
The upstream gene corresponding to that mutated in strain Q5 is 1257 bp in length and encodes a 418-amino acid polypeptide with extensive similarity to aspartokinase from numerous organisms, with the greatest identity (44.6%) to that from Corynebacterium glutamicum (32) (Fig. 3). Aspartokinase, encoded by the lysC gene, catalyzes the ATP-dependent phosphorylation of aspartate yielding aspartyl ␤-phosphate. The putative lysC gene was analyzed further by overexpression of the gene in E. coli (Fig. 2B). The purified recombinant had aspartokinase activity as determined by as the production asparthydroxamate from aspartate, ATP, and hydroxylamine (30). Thus, a bona fide lysC gene was identified.
Aspartyl ␤-phosphate formed by aspartokinase is the first intermediate in the lysine, threonine, and methionine biosynthetic pathways (33). Although only one gene was detected using the lysC open reading frame as a probe against B. japonicum wild type DNA in a Southern blot (data not shown), the lysC mutant strain I110Q5 is not an amino acid auxotroph. Thus, it is likely that B. japonicum contains an additional aspartokinase gene. Indeed, E. coli contains three aspartokinase isozymes, all of which must be mutated to obtain an amino acid auxotrophic phenotype (34,35). We note, however, that the phenotypes of the B. japonicum lysC mutants indicate that a proline-containing peptides with ptsP or lysC expressed in trans Cells were plated on solid media containing: no proline, 500 g of proline ml Ϫ1 media, 50 g of prolyl-glycine ml Ϫ1 media, or 50 g of prolyl-glycyl-glycine ml Ϫ1 media. ϩ, growth; Ϫ, no growth. The complementing genes were harbored in pVK102.

Strain (genotype) and gene on plasmid Growth on medium with
None Proline Pro-Gly-Gly Pro-Gly Enzyme I Ntr and Aspartokinase Interaction putative second aspartokinase gene cannot compensate for the mutated gene described in this study. Finally, addition of lysine, threonine, or methionine, or combinations of them, to growth media did not complement lysC strain Q5 for oligopeptidedependent growth in the absence of proline. We suggest that lysC has a role other than, or in addition to, amino acid biosynthesis in B. japonicum.
Identification and Characterization of the ptsP Gene and Its Product Enzyme I Ntr -The downstream open reading frame corresponding to the gene mutated in strain I20 is 2268 bp in length and encodes a 755-amino acid polypeptide that is homologous to an unusual protein called Enzyme I Ntr (EI Ntr ), which has been identified in several organisms. B. japonicum EI Ntr shows greatest identity (36.4%) to EI Ntr from Azotobacter vinelandii (21) (Fig. 4). EI Ntr is a paralogue of Enzyme I (EI) of the PTS system except that it contains an additional domain at the N terminus that is homologous to the N-terminal sensory domain of NifA from A. vinelandii (36). NifA is a regulatory protein of the 54 -dependent family of transcriptional activators responsible for the activation of genes related to nitrogen fixation in a wide variety of diazotrophs (37), including B. japonicum (38).
We overexpressed the ptsP gene in E. coli, and found that the recombinant EI Ntr can be phosphorylated by PEP, but not by ATP (Fig. 2C), similar to what is observed for other EI proteins, including the EI Ntr from E. coli (22). The current work indicates that like EI, EI Ntr is involved in solute transport into cells. However, EI Ntr is involved in transport of oligopeptides, which is not a known PTS substrate. Furthermore, the ptsP strain, as well as the lysC mutant, grew as well as the parent strain on glycerol, succinate, or glucose as sole carbon sources. Therefore, it is unlikely that the phenotypes of those mutants is an indirect consequence of a defect in the ability to metabolize a carbon source. Data base searches reveal that the gene organization of lysC and ptsP in B. japonicum is not found in other bacteria where either gene, along with flanking sequence, has been identified.
Evidence that lysC and ptsP Are Expressed as Separate Transcriptional Units and Are Both Required for Oligopeptide Transport-The genetic organization of lysC and ptsP led us to ask whether the phenotype of the mutants was due to disruption of the respective gene, or whether it was due to a polar effect of the Tn5 on a downstream gene. To evaluate the necessity for each gene in the utilization of oligopeptides, we tested for complementation of lysC strain Q5 and ptsP strain I20 in trans using wild type copies of lysC or ptsP, respectively, harbored in the broad host range vector pVK102. The results show that each mutant strain can be complemented in trans for growth on Pro-Gly or Pro-Gly-Gly by a wild type copy of the respective gene (Table I) and indicates that the phenotypes exhibited by strains Q5 and I20 are due to the loss-of-function of lysC and ptsP, respectively.
Expression of lysC and ptsP was examined at the RNA level to determine the effect of the Tn5 on the transcription of both genes in the mutant strains. RNase protection analyses of total RNA isolated from strains Q5 and I20 and the parent strain revealed that lysC and ptsP mRNA accumulated in the parent strain, confirming that they are expressed genes (Fig. 5). lysC mRNA was not detected in the lysC mutant, but that strain did accumulate normal levels of ptsP transcript. Therefore, the Tn5 inserted in lysC did not have a polar affect on the downstream ptsP gene, which is consistent with the complementation data. Finally, lysC mRNA was found in the ptsP mutant; the levels were somewhat higher than found in the parent strain, for which we offer no explanation. These results support the conclusion that lysC and ptsP each have an effect on oligopeptide and ALA uptake in B. japonicum, and disruption of either gene results in a loss of those activities.
Overexpression of ptsP in Trans Compensates for the lysC Mutation-lysC and ptsP are each involved in the same cellular process, but their relationship to each other does not appear to be at the level of gene activation. Nevertheless, we found that ptsP complemented the lysC mutant in trans for growth on oligopeptides, suggesting that ptsP can compensate for the lysC mutation when expressed on a low copy plasmid (Table I). This is an interesting result given that transcription of the endoge-nous ptsP gene is normal in lysC strain Q5. However, lysC did not complement ptsP strain I20 when expressed in trans, indicating aspartokinase is insufficient to promote oligopeptide utilization in the absence of EI Ntr . The data indicate that aspartokinase normally affects EI Ntr activity in some manner that can be compensated for by expression of ptsP from a multicopy plasmid. This idea is addressed further as described below.
Evidence for Direct Interaction between Aspartokinase and Enzyme I Ntr -Complementation analysis suggests that aspartokinase affects EI Ntr activity. Therefore, we examined whether the two proteins interact with each other by "pulldown" experiments using purified, recombinant proteins. In the first experiment, we tested for the ability of a GST-aspartokinase fusion, immobilized on glutathione-Sepharose, to interact with an myc-tagged EI Ntr protein. The results show that the myc-tagged EI Ntr was pulled down by the GST-aspartokinase fusion protein but not when GST alone was used as bait. The reverse experiment was carried out using a GST-EI Ntr fusion and a myc-tagged aspartokinase protein, and the results show an interaction between these two proteins as well (Fig. 6). The findings suggest that the mechanism by which aspartokinase affects EI Ntr activity involves direct protein-protein interaction.
Recombinant EI Ntr Is Phosphorylated by ATP in the Presence of B. japonicum Cell Extract, Which Is Negatively Affected in the Presence of Aspartokinase-Aspartokinase catalyzes the transfer of phosphate from ATP to aspartate, whereas EI Ntr is autophosphorylated by PEP. Initial experiments were carried out to determine whether PEP and ATP were interchangeable in the respective reactions, either alone or in combination with the other protein. Using purified recombinant proteins, we did not find evidence supporting phosphoryl transfer between aspartokinase and EI Ntr , nor did one protein affect the activity of the other in vitro. It seemed plausible that a cellular factor was required for a functional interaction that was not present in these preliminary experiments. Therefore, we carried out a series of experiments where B. japonicum cell extract was included in the reactions. Purified GST-EI Ntr fusion protein was used so that it could be separated from the other components after the reaction was completed using glutathione-Sepharose beads, and subsequently analyzed by autoradiography of SDS-PAGE gels. GST-EI Ntr was not phosphorylated when incubated by [␥-32 P]ATP alone, but it was strongly radiolabeled when cell extracts from the lysC strain I110Q5 were included in the reaction (Fig. 7). Extracts from the lysC strain were used so that the only aspartokinase present was that added as purified protein. GST alone was not phosphorylated under those conditions. These observations indicate a factor in FIG. 5. RNase protection analysis of ptsP and lysC expression in B. japonicum strains I110proC, I20, and Q5. Cells were grown in a yeast extract-based media. 8 g of total RNA was analyzed per reaction.
FIG . 6. Protein-protein interactions between B. japonicum EI Ntr and AspK. A, purified, recombinant EI Ntr (His-myc-tagged) was incubated with either GST or GST-AspK, and protein bound to glutathione-Sepharose beads was analyzed by Western blots using anti-myc antibodies. B, purified, recombinant aspartokinase (His-myc-tagged) was incubated with either GST or GST-EI Ntr , and protein bound to glutathione-Sepharose beads was analyzed by Western blots using antimyc antibodies.
B. japonicum extracts that allow GST-EI Ntr to be phosphorylated by ATP. However, when recombinant aspartokinase was included in the reaction, GST-EI Ntr was substantially underphosphorylated (Fig. 7). Addition of aspartate or dialysis of cell extracts did not affect the phosphorylation of GST-EI Ntr , nor did it affect the inhibition by aspartokinase (data not shown). Thus, the underphosphorylated GST-EI Ntr in the presence of aspartokinase could not be explained by consumption of ATP from the enzyme activity of aspartokinase. We suggest that aspartokinase controls EI Ntr function by regulating the phosphorylation state of EI Ntr . The complementaton data (Table I), indicating that aspartokinase exerts an affect on EI Ntr , are consistent with the in vitro phosphorylation experiments. DISCUSSION The intriguing primary structure of Enzyme I Ntr suggested a role in transport, but previous studies had not demonstrated a physiological role for it or for the other PTS Ntr . In this study, we show that EI Ntr is involved in oligopeptide transport, and therefore it, and perhaps the PTS Ntr system as a whole, has a role in transport of a non-sugar solute. The N-terminal domain of EI Ntr is similar to the N-terminal sensory domain of NifA, and therefore it is probably significant that this protein is involved in transport of a nitrogen-containing compound. The present study identified a novel role for aspartokinase, the lysC gene product, in oligopeptide transport. The B. japonicum lysC strain is not an amino acid auxotroph, and thus its role in transport differs from its amino acid biosynthetic function. Collectively, the data indicate that aspartokinase interacts directly with EI Ntr to control its activity.
EI Ntr (GST-EI Ntr ) was phosphorylated by ATP in the presence of cell extract in vitro, indicating a cellular factor that acts as an EI Ntr kinase or that allows EI Ntr to phosphorylate itself.
Roseman's group identified two ATP-dependent EI kinase activities in E. coli (39,40), and therefore the input signal to EI and EI Ntr is not limited to PEP. ATP-dependent phosphorylation of EI Ntr was severely inhibited in the presence of aspartokinase, which provides a plausible regulatory function that the complementation data suggested. These findings, along with the fact that both aspartokinase and EI Ntr are required for oligopeptide uptake, strongly suggests that EI Ntr positively affects transport activity in the unphosphorylated state. Thus, it is likely that the lysC mutant was complemented in trans by the ptsP gene from a multicopy plasmid, because the overexpressed protein was primarily in the under phosphorylated state, thereby obviating the need for aspartokinase. An activity for unphosphorylated EI Ntr to promote oligopeptide transport contrasts sharply with conventional EI enzymes, which trans-fer a phosphoryl group to the next protein in the cascade, ultimately resulting in phosphorylation and concomitant uptake of a sugar. In that case, the high energy phosphate drives the transport process. However, oligopeptide transport complexes bind ATP (4), and therefore EI Ntr need not directly couple transport with energy. The ability of EI Ntr to function differently depending on the phosphorylation status may allow it to serve as a branch point in different signal transduction systems. Recent studies involving a ptsN mutant of Pseudomonas putida indicate that the PTS Ntr protein IIA Ntr is a general regulator in that organism (17). Bacterial proteome and transcriptosome analyses should help reveal the extent of control that these PTS paralogues exert on cellular activities.