The IIANtr (PtsN) Protein of Pseudomonas putida Mediates the C Source Inhibition of the ς54-dependent Pu Promoter of the TOL Plasmid*

The gene cluster adjacent to the sequence of rpoN (encoding sigma factor ς54) ofPseudomonas putida has been studied with respect to the C source regulation of the Pu promoter of theupper TOL (toluene catabolism) operon. The region includes four open reading frames (ORFs), two of which (named ptsNand ptsO genes) encode proteins similar to components of the phosphoenolpyruvate:sugar phosphotransferase system. Each of the four genes was disrupted with a nonpolar insertion, and the effects in the inhibition caused by glucose on Pu activity were inspected with a lacZ reporter system. Although cells lacking ORF102, ORF284, and ptsO did not display any evident phenotype under the conditions tested, the loss ofptsN, which encodes the IIANtr protein, madePu unresponsive to repression by glucose. TheptsN mutant had rates of glucose/gluconate consumption identical to those of the wild type, thus ruling out indirect effects mediated by the transport of the carbohydrate. A site-directedptsN mutant in which the conserved phospho-acceptor site His68 of IIANtr was replaced by an aspartic acid residue made Pu blind to the presence or absence of glucose, thus supporting the notion that phosphorylation of IIANtr mediates the C source inhibition of the promoter. These data substantiate the existence of a molecular pathway for co-regulation of some ς54 promoters in which IIANtr is a key protein intermediate.

Pseudomonas putida cells harboring the catabolic TOL plasmid pWW0 degrade toluene, m-xylene, and p-xylene through a pathway encoded by two separate operons (1). The upper operon encodes enzymes for the oxidation of the methyl group of toluene, whereas the lower operon determines activities for the fission of the aromatic ring leading to pyruvate and acetaldehyde (2). The upper operon is transcribed from the 54 -dependent promoter Pu, which is activated at a distance by the pWW0-borne regulator called XylR, of the NtrC family of proteins, in response to pathway substrates (1). Like other enhancer binding proteins, XylR has the modular structure common to this type of activator (3). XylR variants devoid of its N-terminal module (called the A domain), activate transcrip-tion from Pu (4,5) in the absence of inducers. Besides XylR and the form of RNA polymerase containing 54 ( 54 -RNAP), Pu activation also requires the integration host factor, which both favors a DNA geometry optimal for the interplay of the different factors (6,7) and assists the binding of the RNA polymerase (8).
The comparison of Pu promoter activity in vitro and in vivo has revealed a paradox that determines the regulation of this promoter under real environmental conditions (9). Although Pu is active in vitro by just mixing purified and preactivated XylR (i.e. deleted of its N-terminal domain) with 54 -RNAP and integration host factor (5), promoter activity in vivo is downregulated by various growth conditions (10 -16). In particular, the presence in the medium of carbon sources like glucose, gluconate, and ␣-ketogluconate inhibit Pu activity, whereas other compounds (citrate and fructose) do not seem to have any influence (11).
The molecular basis of this phenomenon is unknown. In enteric bacteria, such as Escherichia coli, glucose repression is mediated mainly by the cAMP receptor protein (17) and the PTS 1 (18). The latter is a complex and very branched whole of phosphotransfer proteins that mediate the intake of certain carbohydrates (which vary widely among species) through a mechanism involving the simultaneous phosphorylation of the sugar and its transport (18). The availability of adequate carbohydrates in the external medium that act as a drain of high energy phosphate determines the accumulation or the depletion of phosphorylated protein intermediates that have the ability to interact with and modify the activity of many other cell proteins. The PTS proteins have been traditionally classified as enzymes type I, HPr, and enzymes type II, the latter frequently involving a multiprotein array (18). As opposed to the situation in E. coli, cAMP appears to play no role in carbon repression in pseudomonads (19). In fact, the known cAMP receptor protein analogue in Pseudomonas aeruginosa (called Vfr) is fully alien to carbon regulation (20,21). On the other hand glucose, which inhibits Pu activity (11), is not transported in Pseudomonas through the PTS (22), thus making unlikely an involvement of the housekeeping PTS intermediates in the effect. Finally fructose, which is transported through the PTS system in Pseudomonas (23) does not inhibit Pu activity. The mechanism should therefore be different of those known for other systems subjected to catabolite control.
An early clue on additional regulatory mechanisms for 54dependent promoters in vivo was revealed by the sequencing of the DNA region around the gene encoding the sigma factor (rpoN) in Klebsiella pneumoniae (24) and E. coli (25). In both species rpoN is followed by four open reading frames later named ORF95, ptsN, ORF284, and ptsO (see Fig. 1A). Homologues of most of these genes have also been found adjacent to rpoN in various other Gram-negative species (26 -30), including P. aeruginosa (31) and P. putida (32). Two of the most conserved ORFs, ptsN and ptsO, encode proteins with a considerable sequence similarity with known intermediates of the PTS sugar transport system. In particular, ptsN appears to encode a IIA-type enzyme, whereas ptsO seems to be a variant HPr protein (25). The corresponding proteins have been designated as IIA Ntr and NPr, respectively, on the basis of their hypothetical involvement in N metabolism (25). This was initially suggested by the observation that disruption of the ptsN gene of K. pneumoniae (originally called ORF152) or the adjacent ORF95 increased to various degrees the activity of the 54 -dependent promoter PnifH (24), whereas the loss of ptsO gene reduced it significantly (33). On the contrary, mutations in the equivalent ptsN (ORF2) gene P. aeruginosa did not affect the activity of the 54 -dependent promoters for pilae and flagellin genes (31). Furthermore, some (but not all) 54 -dependent promoters tested of Caulobacter and Rhizobium become less active upon the loss of ptsN (29). This protein may also be involved in more general metabolic activities, because the E. coli ptsN mutant displayed certain incompatibilities between C and N sources, typically glucose and alanine (25). In addition this mutation also suppressed a temperature-sensitive allele of the gene era, which encodes an essential GTPase of unknown function (25).
Taken together, the observations above suggest that ptsN and the accompanying genes of the cluster may play a role not only in expression of 54 -dependent promoters but also in their potential connection to N and C metabolism (34). Because such a connection with the central metabolism could account for the C source regulation observed in promoters such as Pu (11), we set out to explore the issue systematically in a well defined genetic assay system. The data presented in this work show that the ptsN gene of P. putida, formerly called ORF154 (32), participates in the inhibition of the Pu promoter of the TOL plasmid by glucose and that such inhibition involves a phosphorylation site of the ptsN-encoded IIA-type enzyme.

EXPERIMENTAL PROCEDURES
Strains, Plasmids, Media, and General Methods-P. putida MAD2 (4) was used in all cases to examine Pu activity. This strain is a derivative of P. putida KT2442 bearing a chromosomal Pu-lacZ fusion along with the xylR allele named xylR⌬A, all assembled within a tellurite resistance (Tel r ) mini-transposon vector. The loss of the N-terminal A domain of the protein makes XylR constitutively active in the absence of aromatic inducers (4). Bacteria were grown in either rich LB medium or in synthetic mineral M9 medium (35) supplemented with 0.2% casamino acids (CAA) to equal growth rates and to avoid effects related to the stringent response (36). To test carbon inhibition of promoter activity, the M9-CAA medium was amended with glucose, gluconate, or fructose at a final concentration of 10 mM. Overnight cultures of P. putida MAD2 or its derivatives were diluted to an A 600 of approximately 0.05 in fresh medium and regrown at 30°C until the end of the exponential phase (A 600 of approximately 1.0). Samples were collected 30 min after this point, and promoter activity was measured by assaying the accumulation of ␤-galactosidase in cells permeabilized with chloroform and SDS as described by Miller (37). Each enzymatic measurement was repeated at least twice, and deviations were less than 15%. The concentration of glucose and gluconate in culture supernatants were determined as described by Schleissner et al. (38). When required, the culture medium was supplemented with streptomycin (Sm; 50 g/ ml), kanamycin (Km; 50 g/ml), ampicillin (150 g/ml), and potassium tellurite (80 g/ml).
Plasmids were generally maintained in E. coli DH5␣ and E. coli CC118 strains, although those containing conditional R6K replication origins were hosted in E. coli strain CC118pir (39). DNA was manipulated following standard protocols (35). For constructing pJM154 (a ptsN ϩ broad host range plasmid), a 0.78-kb PstI fragment from pARG5.1 (spanning the corresponding chromosomal region of P. putida; Ref. 40) was cloned in pNot18. This vector is a variant of pUC18 in which the entire ␣-lac segment and the Plac promoter of the plasmid are flanked by NotI sites (41). The result of this cloning step was plasmid pNot18 -154, in which the promoterless ptsN sequence is transcribed through the Plac promoter of the ␣-lac segment. The corresponding 1.2-kb NotI portion of pNot18 -154 was then excised and cloned in the single NotI site of pJPS9, a mobilizable derivative of the broad host range pPS10 replicon bearing an Sm resistance gene (41). When required, the resulting construction (pJM154) was mobilized into P. putida recipients through triparental matings with helper strain E. coli HB101 (RK2013) as described by de Lorenzo and Timmis (39). For the polymerase chain reaction, separate colonies of the strains under scrutiny were resuspended in 10 l of H 2 O and boiled for 5 min. 1 l of this material was diluted 100-fold and directly subjected to 25 cycles (1 min at 92°C, 1 min at 55°C, and 2 min at 72°C) of amplification with Taq polymerase in the presence of 1.5 mM MgCl 2 , 1 mM dNTPs, and 50 pmol of the primers indicated in each case. DNA sequencing (kindly performed by E. Diaz) was carried out by applying a primer walking strategy along the insert of plasmid pDORF2 (see below) with fluorescent dNTPs. The sequence of the rpoN region has been deposited in the EMBL nucleotide sequence data bank under the reference PPU007699.
Construction of Insertion Mutants-Stepwise nonpolar knockouts of the four genes adjacent to rpoN were first generated in plasmids by directed insertions of a promoterless Km resistance cassette devoid of transcriptional termination signals. Such a cassette was obtained by polymerase chain reaction amplification of the sequence within coordinates 1920 to 2732 of pACYC177 (42). The resulting 0.8-kb DNA fragment was then recovered as a BamHI insert in pUC19 giving rise to plasmid pUCKm. Insertions in each of the ORFs were achieved as follows. A 3.2-kb SalI-ScaI fragment from plasmid pNTR1 (spanning rpoN and 1.5 kb downstream; Ref. 40) was cloned in vector pBlueScript KS ϩ digested with SmaI and SalI. To generate a Km insertion in ORF102, the resulting plasmid (pBSNtr) was digested with HindIII, and the overhanging ends were filled in with dNTPs using the Klenow fragment of DNA polymerase (35) and ligated to an equally filled in 0.8-kb segment of pUCKm bearing the promoterless kanamycin cassette mentioned above. The resulting plasmid (pBSN102) bears ORF102 truncated at its 46th codon. Similarly, disruption of ptsN required the digestion of pBSNtr with SmaI and its ligation to the blunt-ended kanamycin cassette of pUCKm. In the resulting plasmid (pBSN154) the Km resistance phenotype became apparent only when two cassettes were inserted in tandem, disrupting ptsN at its 53th structural codon. To knockout ORF284, the 1.2-kb XmaI fragment from pARG5.1 was ligated to the 2.3-kb XmaI fragment of the pHP45⍀ (43). The resulting plasmid pHP284 was digested with BamHI/BglII and ligated to the 0.8-kb BamHI fragment from pUCKm, giving rise to pHP284Km. In this plasmid, the 853 nucleotides following the 3rd structural codon of ORF283 have been replaced by the Km r cassette. The 1.2-kb fragment resulting from partial digestion of pHP284Km with XmaI was then used to replace the corresponding XmaI fragment in plasmid pDORF2. This is a pNot19 derivative containing a 3.4-kb EcoRI/HindIII fragment from pARG5.1, spanning ptsN, ORF284, ptsO and 1.5 kb further downstream. The exchange of XmaI segments between pHP284Km and pDORF2 rendered plasmid pDORFK. Disruption of ptsO was engineered by first cloning the 0.8-kb SalI fragment of pARG5.1 into pNot18 (41) thereby originating plasmid pNot18-90. This one was digested with SmaI and ligated to the blunt-ended 0.8-kb BamHI segment of pUCKm encoding the promoterless Km cassette, thus yielding the plasmid pNot18-90Km, which encodes a NPr protein truncated at position 16 of its predicted amino acid sequence.
To transfer the various insertions into the chromosome of P. putida MAD2, the SalI-BamHI fragments of each of the pBSNtr derivatives bearing Km insertions in ORF102 (pBSN102) and ptsN (pBSN154) as well as the NotI fragments from pDORFK (ORF284::Km) and pNot18-90Km (ptsO::Km) were cloned in pKNG101 (44), a mobilizable suicide vector bearing the positive counter selection marker sacB. The resulting plasmids pKNG102, pKNG154, pKNG284, and pKNG90 were transferred to P. putida MAD2 through triparental matings with helper strain E. coli HB101 (RK2013) as described above. Homologous recombination of the mutated ORFs in the Pseudomonas chromosome was made by first selecting cointegrates of the suicide plasmids on LB plates with Km, Sm, and potassium tellurite (potassium tellurite for counterselection of the donor and helper strains). Plates with approximately 1000 exconjugant colonies were kept for 1 week at 4°C, pooled in 1 ml of 1% NaCl and then replated in LB with Km and 5% sucrose for selection of a second recombination event. The colonies grown in such a medium were then scored for sensitivity to Sm. Those with a phenotype Sm s Km r were analyzed by polymerase chain reaction with oligonucleo-tides spanning the first and last codon of each gene as follows. Disruption of ORF102 was verified with primers ORF102D (5Ј-ATGCAGAAT-TCTTTCAGTCAATATCAGTGGACAG-3Ј) and ORF102R (5Ј-CGGGAT-CCTCAGCGGGCAGCTGCACCTTGCAG-3Ј). The ptsN::Km knockout was confirmed with primers ORF154D (5Ј-ATGCAGAATTCTTTCACG-ACTTGAAACCATCCTGACC-3Ј) and ORF154R (5Ј-CGGGATCCTCAG-TGCTCGTTCTGTACGTCCAG-3Ј). Disruption of ORF284 was verified with primers ORF284D (5Ј-GCGGATCCC GCCTGATCATCGTC-3Ј) and ORF284R (5Ј-ATCGCAAGCTTCTAGAGGTCGCGATG-3Ј). Finally, insertion of the Km gene in the ptsO coding sequence was authenticated with primers ORF90D (5Ј-GCGGATCCCGCCCGCGAAATC-3Ј) and ORF90R (5Ј-ATCGCAAGCTTCCTATTCGCCTTCGTCAAA-3Ј).
Site-directed Mutagenesis of ptsN and Expression of the Mutant Alleles-The method developed by Kunkel et al. (45) was used to generate the H68A and H68D variants of ptsN. To this end, the SalI-BamHI fragment of pBSN102 described above was cloned in vector pGC2 (46). Extension in vitro of the single-stranded, uracyl-containing DNA from the resulting plasmid (pGC154) was primed independently with mutagenic oligonucleotides PTSNH68A (5Ј-GGCATCGCCATCCCAGCAT-GCCGGCTTGAAGGA-3Ј) for H68A substitution and PTSNH68D (5Ј-GGCATCGCCATTCCGGACTGCCGGCTTGAAGGA-3Ј) for the H68D mutant (the changed nucleotides are shown in bold, and the triplets corresponding to the new amino acid residues are underlined). These changes introduced new SphI and BspEI restriction sites, respectively, to facilitate the screening of the mutants. The resulting plasmids, pGC154-H68A and pGC154-H68D were used as templates for amplification with polymerase chain reaction of the new alleles with primers ORF154D and ORF154R, which incorporate new EcoRI and BamHI sites in the 5Ј and 3Ј ends of the resulting amplification products. This amplification step modified the N terminus of the mutant proteins from MI to MGILS. The resulting 0.45-kb EcoRI-BamHI fragments were cloned in the pVTR-B expression vector (47), originating plasmids pVTR154-H68A and pVTR154-H68D. Their NotI inserts containing the mutant alleles under the control of an inducible Ptrc promoter and the lacI q gene were cloned into the single NotI site of pJPS9 vector (41), creating plasmids pJMT/154-H68A and pJMT/154-H68D. An equivalent plasmid (pJMT/154) containing the wild type ptsN allele was also constructed and used as a control. Each of the three resulting plasmids were transferred to the wild type P. putida MAD2 or P. putida MAD2 ptsN::Km strains by triparental mating as described above.
Protein Techniques-Western blot assays to detect the ptsN product were made as described in Cases et al. (16). To this end, equal amounts of whole Pseudomonas cells (typically 10 8 ) were lysed in a sample buffer with 2% SDS and 5% ␤-mercapthoethanol and run in denaturing 12% polyacrylamide gels. These were subsequently blotted and probed with a 1:1000 dilution of a preadsorbed rabbit serum raised against a purified MalE-IIA Ntr fusion protein expressed in E. coli. The bands in the blots corresponding to the product of ptsN were developed with protein A coupled to horseradish peroxidase and H 2 O 2 /diamino-benzidine as described in Cases et al. (16).

Organization of the rpoN Locus of P. putida and Its Encoded
Products-The rpoN gene of P. putida was reported by Köhler et al. (48) to be followed by two ORFs that were called at that time ORF102 and ORF154 (32). To visualize the complete organization of the region, we sequenced 1.35 kb of the DNA further downstream. The presence of two additional genes (ORF284 and ptsO) completed the overall picture of the locus in this bacterium (Fig. 1B). Besides rpoN itself, the region includes a gene cluster located 78 base pairs downstream and consisting of four adjacent cistrons. Distances from ORF102 to ptsN or from ORF284 to ptsO are as small as 13 and 16 base pairs, respectively. The ptsN sequence overlaps by 1 base pair with that of ORF284. Such an organization resembles with relatively minor differences that found in E. coli (49) and K. pneumoniae (33) as shown in Fig. 1.
The protein encoded by ORF284 in P. putida is well conserved among bacterial species, but its sequence does not manifest any predictable function. The only hint of a role for the product borne by ORF284 of P. putida (49% identity and Ͼ60% similarity with the E. coli and K. pneumoniae homologues) is the presence of an 8-amino acid sequence that matches the phosphate binding loop of many ATP-and GTP-binding pro-teins (Ref. 25 and Fig. 1B). The P. putida ptsO gene was also very similar (34% identity and 56% similarity) to the ptsO gene of E. coli, which encodes the HPr-like enzyme termed NPr (25) and also contains the two sequence motifs that are conserved among phosphorylation sites of HPr-type phosphotransferases ( Fig. 1B and Ref. 18).
Step XylR⌬A, in which the N-terminal domain has been deleted to yield an effector-independent constitutive activator (4). In this genetic background, we inserted a promoterless Km cassette devoid of transcription terminators in each of the four cistrons of the rpoN cluster, namely ORF102, ptsN, ORF284, and ptsO (see "Experimental Procedures"). The effect of the C source added was examined in P. putida MAD2 and each of the derived mutant strains by growing them in mineral medium supplemented not only with the sugar under study but also with casamino acids to equal growth rates and to avoid effects related to the stringent response (36). As shown in Table I, the MAD2 strain faithfully reproduced the C source-dependent inhibition of the Pu promoter reported by Holtel et al. (11) but without the need to add an aromatic inducer to the medium, hence reducing the number of variables to be considered. The results of these assays indicated that although disruption of ORF102, ORF284, or ptsO did not have an apparent effect on the inhibition of Pu activity by glucose (not shown), the ptsN::Km mutant gave rise to a clear phenotype of apparent insensitivity to the added carbohydrate (Table I). The behavior of the ptsN::Km strain with glucose could be extended to other repressive carbon sources such as gluconate, whereas Pu activity remained unaffected with fructose. That the release of Pu inhibition by glucose or gluconate was due to the loss of ptsN and not to any other indirect effect caused by the Km insertion was verified by transforming P. putida MAD2 ptsN::Km with the ptsN ϩ plasmid pJM154. As shown in Table I, the transformed strain reverted to the same phenotype of glucose inhibition of Pu activity as the wild type strain. These data suggested that ptsN was involved in the C source inhibition of Pu.
The Loss of ptsN Does Not Affect Glucose Transport or Metabolism-A trivial explanation to justify the phenotype of ptsN mutants regarding the inhibition of Pu by glucose could be that cells fail to transport and/or to metabolize this sugar. However, this is unlikely, because both the mutant and the wild type strain grew well in a minimal medium with glucose as the only carbon source (not shown). However, to rule out that glucose transport in the ptsN mutant could be inhibited in the richer M9-CAA medium, we measured directly the concentration of glucose in the supernatants of cultures of ptsN ϩ and ptsN Ϫ P. putida MAD2 cells along the growth curve. Because glucose is predominantly converted by P. putida cells to gluconate prior to its intake and subsequent metabolism (38), we also measured the extracellular levels of this carbohydrate. The results shown in Fig. 2 did not indicate any significant differences in either the growth rate or in consumption of glucose between the two strains. The sugar was, in fact, rapidly depleted from the medium, and the low levels of gluconate detected in the cultures suggested that glucose was normally used as a carbon source. The modest difference between the two strains regarding glucose consumption (Fig. 2) might simply reflect a small growth defect of the ptsN mutant.
A Genetic Approach to the Role of the His 68 Residue of IIA Ntr in C Source Control of Pu-The similarity of IIA Ntr with the PTS enzymes type II raised the possibility that the same phosphorylation switch that governs their activity (18) could also effect the action of the ptsN product on Pu. Fig. 3A shows that most members of the IIA Ntr family, with the one exception of IIA Ntr of Hemophilus influenzae, maintain almost perfectly the site of phosphorylation, including the conserved His residues that in various PTS enzymes type II is phosphorylated by the HPrϳP (18). To examine whether the equivalent site of P. putida IIA Ntr, located in position 68, played a role in the repression of Pu by glucose, we produced site-directed mutants of the protein in which His 68 was replaced by either an alanine or an aspartic acid residue. Because the H68A substitution prevents the site from being phosphorylated by a cognate kinase (50), the phenotype endowed by IIA Ntr H68A should correspond to either a total loss of function or that due to the protein locked in a nonphosphorylated state. On the other hand, the H68D substitution is also predicted to prevent phosphorylation while adding a negative charge at the former His residue, which may help the protein to maintain its native conformation and even mimic the phosphorylated state (50). The actual phenotypes raised by these mutants in connection with Pu activity are described separately below.
Phenotype of Mutant ptsNH68A-To assess the behavior conferred by the ptsNH68A allele we employed plasmid pJMT/154-H68A, which expresses a IIA Ntr variant not amenable to phosphorylation. This plasmid, in which ptsNH68A can be induced with IPTG, was then introduced into wild type P. putida MAD2 as well as in its ptsN::Km derivative, so we could assess both the phenotype endowed by the mutant allele and its dominance or recesiveness versus the native protein. As a control, an equivalent plasmid encoding the wild type ptsN gene expressed through the same system was also introduced in identical strains. Each of the P. putida exconjugants was then grown in the presence or absence of glucose, and the accumulation of ␤-galactosidase was measured as before with or without addition of IPTG. The results of such an experiment are shown in Fig. 3B. The Western blot of each strain confirmed that plasmid  pJMT/154-H68A results in a product without any indication of instability or proteolytic degradation, thus suggesting that the IIA Ntr -H68A mutant is produced as a full-size protein. The data of Fig. 3B indicated that the mutant protein failed to restore the phenotype of repression by glucose that is lost in the ptsN::Km mutant. In the same conditions, plasmid pJMT/154, which encodes wild type IIA Ntr , fully complemented the lack of ptsN in that respect. Finally, plasmid pJMT/154-H68A did not have any effect in the inhibition of Pu by glucose displayed by the wild type P. putida MAD2 host, even if the H68A variant was overproduced upon addition of IPTG, thereby indicating that the H68A allele was a recessive mutation and that it is unable to mediate any effect of glucose in Pu.
Phenotypes of Mutant ptsNH68D-The expression plasmid pJMT/154-H68D (ptsNH68D ϩ ) was conjugated into wild type P. putida MAD2 and the ptsN::Km derivative, so the resulting phenotypes with respect to the inhibition of Pu by glucose could be determined. As with the H68A counterpart, each of the P. putida exconjugants bearing pJMT/154-H68D were grown with or without glucose and with or without IPTG, with the results on Pu-lacZ activity shown in Fig. 3C. The Western blots corresponding to each strain confirmed the apparent integrity of the mutant protein produced. Interestingly, expression of the ptsNH68D allele not only restored the low levels of Pu activity in the presence of glucose but also kept the down-regulation of the promoter in the absence of the sugar (a sort of super repressed phenotype). Such a glucose-independent repression was exacerbated when IPTG was added to the medium, i.e. under higher intracellular concentrations of IIA Ntr /H68D. This behavior is consistent with that observed when plasmid pJMT/ 154-H68D (ptsNH68D ϩ ) was introduced into the P. putida wild type strain MAD2. In this case (Fig. 3C), Pu-lacZ activity was inhibited to a degree dependent on addition of IPTG but could be further down-regulated when glucose was present in the medium. This can be explained by the presence of a wild type copy of the ptsN gene whose product could still respond to C source inhibition. Because the IIA Ntr /H68D protein was produced to higher levels than the wild type product, even without IPTG, we tested whether the presence of an equivalent plasmid harboring the wild type allele had the same effect on Pu in media lacking a repressive carbon source. As shown in Fig. 3C, this was not the case. Overexpression of the ptsN gene made the promoter more sensitive to glucose but had no effect in the cultures not added with the sugar. This result rules out differences in intracellular concentrations of the wild type and the mutant ptsN products as the origin of the super repressed phenotype raised by ptsNH68D. DISCUSSION We have analyzed the role of rpoN gene cluster in the physiological control of the transcription of an operon encoding the early steps for biodegradation of toluene in P. putida (pWW0). The 1.35-kb fragment of new DNA sequence downstream of the already known ORF102 and ptsN revealed the presence of two additional genes (ORF284 and ptsO), the whole of which was similar to those found in E. coli and K. pneumoniae adjacent to rpoN (Fig. 1). The most salient features of the gene cluster was the virtual identity of the ORF154 product to the E. coli protein named IIA Ntr on the basis of its homology to EII-type proteins of the PTS system (32) as well as the identity of ORF90 to the E. coli protein named NPr on the basis of its homology to the housekeeping HPr phosphotransferases (25). These similarities are particularly true around the predicted phosphorylation sites of both proteins (Fig. 1B). That the array ORF102-ptsN-ORF284-ptsO forms a defined gene cluster is suggested by the fact that the only coding sequence found downstream of ptsO (similar to the gufA gene of Myxococcus xanthus) is expressed in an opposite orientation (Fig. 1B). The conservation of these genes and their genomic organization in E. coli, K. pneumoniae, and P. putida is in contrast with that found in other bacteria. The conserved His residue in position 68 was changed to either Ala or to Asp, the latter introducing a negative charge. B, phenotype endowed by the ptsNH68A allele. Plasmid pJMT/154-H68A, which expresses the ptsNH68A allele from an IPTG-inducible Ptrc promoter, was transferred to P. putida MAD2 and its ptsN::Km derivative. The resulting exconjugants were grown in M9-CAA with or without glucose and amended, where indicated, with IPTG. The expression levels of the ptsN gene and its variants in each of the strains and growth conditions was monitored with the blots probed with an anti-IIA Ntr serum shown below the bar diagrams. Equal amounts of total cell protein were loaded in the lanes, so that the intensities of the bands highlighted in the blots are representative of the relative intracellular concentrations of the ptsN products. Note that pJMT/154-H68A failed to restore the inhibition by glucose of Pu activity that is lost in the ptsN::Km strain, even when the mutant protein was overproduced with IPTG. Similarly, pJMT/154-H68A had no effect in Pu activity measured in the wild type strain P. putida MAD2. C, phenotype endowed by the ptsNH68D allele. Plasmid pJMT/154-H68D expressing the ptsNH68A allele was transferred to P. putida MAD2 and its ptsN::Km derivative. The exconjugants were grown as before, and the levels of the ptsN variants were examined with the same procedure described above. Expression of the ptsNH68D allele in both the wild type and the ptsN::Km strain produced a decrease of Pu activity regardless of the presence or the absence of glucose. Such an inhibition was exacerbated upon addition of IPTG (i.e. overproduction of the H68D protein variant). Note that Pu activity in the wild type strain bearing pJMT/154-H68D was still responsive to inhibition by glucose, perhaps owing to the presence of a chromosomal copy of the wild type ptsN gene.
For instance, Caulobacter crescentus has an additional ORF of unknown function between the rpoN gene and the ORF102 homologue (which has in fact 208 amino acid residues in this microorganism; Ref. 29). A similar gene is present downstream of rpoN2 of Bradyrhizobium japonicum. This organism harbors two different rpoN genes, and one of them, rpoN1, lacks these ORFs in its 3Ј end (26). On the contrary, H. influenzae, which has no rpoN gene as it has been revealed by the complete sequencing of its chromosome, presents an abbreviated cluster, consisting only in the ptsN gene and the ORF284 homologue (28). A short version of the cluster is also found in Acinetobacter calcoaceticus, in which only the ORF102 homologue is present, followed by two unrelated genes (27).
When each of the four genes of the rpoN cluster of P. putida were knocked out with a Km cassette, none but the ptsN::Km insertion resulted in a detectable effect on the activity of the 54 -dependent Pu promoter. Disruption of ptsN made Pu to lack inhibition by either glucose or gluconate but did not influence significantly the consumption of these sugars (Table I and Fig. 2). This suggested that the loss of ptsN did not interfere with glucose or gluconate metabolism but rather with the transduction pathway that translates the presence in the medium of a repressive carbon source into inhibition of the Pu promoter. The phenotypes endowed by the ptsNH68A and ptsNH68D alleles (Fig. 3) indicate that the phosphorylable H68 residue of IIA Ntr plays a key role in carbon inhibition of the upper TOL operon. Many proteins of the PTS system are known to interact directly with a variety of polypeptides of diverse functions not necessarily related in sequence or structure. Typical examples of this include the phosphorylation-dependent ability of IIA Glu to interact with a number of permeases (51) or the formation of the CcpA-HPr repressor complex in Gram-positive bacteria (52). Similarly, the interplay of phosphorylated/nonphosphorylated forms of IIA Ntr could cause the inhibition of Pu activity through the interaction of this protein with a thus far unknown target in the transcription machinery.
Although our data would fit well in a model in which the phosphorylated form of IIA Ntr (imitated by the H68D mutant) would cause repression and the nonphosphorylated form (simulated by the H68A mutant) would prevent such a repression, the mechanism by which the His 68 residue may be modified in vivo deserves further clarification. Whereas the IIA Ntr proteins of E. coli and K. pneumoniae can be phosphorylated by the PTS enzyme HPr in vitro (25,53), it is also possible that IIA Ntr of P. putida follows in vivo a route alternative to the housekeeping PTS pathway. A potential good candidate for being the phospho-donor to IIA Ntr could be NPr, the protein encoded by ptsO (Fig. 1), because its homologous protein HPr phosphorylates PTS enzymes IIA. In vitro, however, the flow of high energy phosphate between E. coli proteins seems to go from IIA Ntr to NPr and not the contrary (25). This issue will be the subject of future investigations.
The loss of ptsN has been reported to affect the activity of other 54 promoters in several species (24,29,31,54). However, a proper correlation of our data with the effects observed in other instances is flawed by the very different composition of the media and the physiology of the species used by each group of authors. For instance, Du et al. (54) reported that the inhibition by glucose of Ps, a second 54 promoter of the TOL plasmid, is not relieved when promoter activity is tested in a ptsN mutant of E. coli, a behavior that might be different if the same had been tested in P. putida. Interestingly, every function reported so far to be influenced by ptsN is connected to nitrogen or carbon metabolism. Even suppression of the lethality of the era ts allele described in E. coli is restricted to some culture conditions (25). These observations have led Reizer et al. (34) to propose a role for ptsN in the co-ordination of C and N metabolism, a hypothesis consistent with the results reported in this article for the P. putida gene. In this respect it could well happen that IIA Ntr has evolved to modulate at least some 54 promoters in response to the presence of specific carbon sources. This is the case for other members of the PTS system that have been recruited to a number of regulatory processes aside from their participation in transport and phosphorylation of sugars (18).