INTRODUCTION
Amino sugars, including
N-acetylglucosamine (GlcNAc) and
N-acetylneuraminic acid (NANA) (
1), are present in many glycans in all organisms from bacteria to animals, and in mammals, for example, they are found in human milk (
2) and cell surface mucus (
3). The microbiome member and model organism for Gram-negative bacteria,
Escherichia coli, is an important human pathogen and is used in numerous biotechnological applications.
E. coli tightly controls the utilization of amino sugars, which are excellent sources of both carbon and ammonia.
NANA is essential for the synthesis of some polysaccharides and the glycosylation of certain proteins and lipids in both eukaryotes and prokaryotes. Exogenous NANA is utilized by
E. coli via the transporter, NanT, and further hydrolyzed by a lyase, NanA, to produce pyruvate and
N-acetylmannosamine (ManNAc). ManNAc and mannosamine (ManN), as well as glucosamine (GlcN), are taken up from the growth medium by the ManXYZ enzyme complex of the bacterial phosphoenolpyruvate:sugar phosphotransferase system (PTS) and concomitantly phosphorylated to ManNAc-6P, ManN-6P, and GlcN-6P, respectively, and
N-acetylglucosamine is taken up via a distinct sugar phosphorylating PTS permease, NagE (
4–6) (
Fig. 1). In contrast, ManNAc, produced from NANA hydrolysis, is phosphorylated in the cell by an ATP-dependent ManNAc kinase: NanK, regulated by the NanR regulator (
7).
The pathway for
N-acetylmannosamine utilization includes an epimerase, NanE, which converts ManNAc-6P to GlcNAc-6P as part of the NANA and ManNAc utilization pathways (
Fig. 1). NagA further deacetylates GlcNAc-6P to GlcN-6P. Glucosamine 6-phosphate isomerase/deaminase, NagB, provides the last step in the amino sugar-specific catabolic pathway, converting GlcN-6P to NH
3 and fructose 6-phosphate (Fru-6P), the first common metabolite of glycolysis. This enzyme is essential for the utilization of amino sugars in
E. coli and is known to be allosterically activated by an intermediate of the GlcNAc metabolic pathway, GlcNAc-6P (
8–10). Orthologs of this enzyme are present not only in bacteria but also in mammals and other organisms.
E. coli NagB is encoded in an operon with the
nagA gene and is regulated by NagC (
11), being induced when GlcN, GlcNAc, or another amino sugar is available in the medium.
The biosynthetic pathway producing UDP-GlcNAc for incorporation into cell wall components involves the
glmS,
glmM, and
glmU gene products, and GlmS generates the cytoplasmic GlcN-6P pool essential for peptidoglycan biosynthesis. NagB redirects GlcN-6P to the glycolytic pathway and is thus important and tightly regulated. In
E. coli,
nagA and
nagB occur in an operon, regulated by the transcriptional regulator, NagC, and these genes are expressed only when amino sugars are present in the medium. NagB activity is regulated by two previously recognized factors, GlcNAc-6P, as noted above (allosteric regulation), and a primary protein constituent of the PTS, HPr, a sensor of the availability of extracellular PTS sugar substrates (
12), including GlcNAc, ManNAc, and GlcN.
NagB interactome data (
13), presented in part in
Table 1, suggest that NagB interacts with several cellular proteins, including the nitrogen-related signal transduction PII protein (
14,
15), the NanE epimerase described above, proline aminopeptidase PepP, nitroreductase NfsB (capable of reducing nitrofurazone and quinones), and even the riboflavin biosynthetic enzymes RibA, RibB, and RibC. The work reported here shows that NagB is activated by NanE in the presence or absence of GlcNAc-6P but not by BglA (used as a negative control) and is activated in the presence of GlcNAc-6P by the PII protein covalently modified by uridylylation, an indicator of nitrogen availability. The other interactions, suggested by the data in
Table 1, have not been examined.
The uridylylated PII protein (U-PII) is generated by posttranslational modification under nitrogen-limiting conditions involving the glutamine/α-ketoglutarate ratio-sensing uridylyltransferase/uridylyl-removing enzyme GlnD (
16,
17). Adenylylation of glutamine synthetase, GlnA, is stimulated by the PII protein, GlnB, and deadenylylation is stimulated by U-PII, thus comprising a dual bicyclic cascade.
The regulatory interdependence between different metabolic pathways has been considered (
18). For example, carbon metabolism is known to be controlled not only by carbon-derived signals but also by the availability of nitrogen, sulfur, and iron (
19–21). Components of the PTS participate in regulatory interactions (
22), resulting in the control of carbon and nitrogen metabolism (
23–25), and recently, the histidine-phosphorylatable phosphocarrier protein, HPr, was shown to control the activities of glycolytic enzymes, including NagB, by direct protein-protein interactions (
12).
We show here that activation of NagB by U-PII but not by PII in the presence of GlcNAc-6P leads to an increase in activity of >10-fold. The synergistic effects of HPr/U-PII and HPr/NanE but not of U-PII/NanE on NagB activation have been demonstrated. The modeling of HPr/U-PII/NagB and HPr/NanE/NagB complex formation confirmed the possibility that the two proteins (HPr/U-PII or HPr/NanE) can simultaneously interact with NagB, although U-PII and NanE cannot. These observations are rationalized from both mechanistic and physiological standpoints.
DISCUSSION
The PII protein is known to be a regulator of both the activity and the synthesis of glutamine synthetase (GS; GlnA) in enteric bacteria, and of nitrogen metabolism in many other bacteria, archaea, and eukaryotes, in response to the availability of nitrogen sources (
Fig. 1) (
29–33). The pathways that regulate
glnA gene expression and GS enzymatic activity both involve the covalent modification of proteins (
Fig. 1). The regulation of GS activity involves deadenylylation for activation and adenylylation for inactivation with both reactions catalyzed by the same enzyme, adenylyltransferase/adenylylase (GlnE). The direction of GS modification is dictated by the PII protein, the state of which is also regulated by reversible covalent modification by uridylylation catalyzed by GlnD, another bifunctional enzyme regulated oppositely by αKG and glutamine. The modified form, U-PII, is essential for the deadenylylation reaction acting on GS.
We found that the modified form of PII, U-PII, activates NagB. Coordinate activation of both NagB and GS by U-PII makes teleological sense since activation of the former releases NH
3, while activation of GS facilitates its incorporation into glutamine for the synthesis of numerous other nitrogenous compounds. The effects and consequences of the NagB allosteric interactions can be summarized as follows. (i) The presence of amino sugars in the medium dephosphorylates HPr and activates NagB if and only if GlcNAc-6P is present. (ii) An increase in the cytoplasmic GlcNAc-6P concentration promotes high levels of
nagB expression and high NagB activity. (iii) The activation of NagB by U-PII promotes successful utilization of amino sugars, thereby increasing levels of both carbon and nitrogen in the cell (
Fig. 7). (iv) GS will be converted to the unmodified active form, allowing the incorporation of the NH
3 released from GlcN-6P into glutamine. (v) NanE activation of NagB only occurs when NANA is available, promoting high-level expression of the
nanE gene.
The pathways for the utilization of different amino sugars—NANA, ManNAc-6P, and GlcNAc-6P—converge with the production of GlcN-6P, the substrate of NagB (
Fig. 1). The protein-protein interactome data for HPr have led to the suggestion that NagB is a connecting point with the nitrogen regulatory module, with the U-PII protein playing a central role (
Fig. 7). If GlcN or GlcNAc is transported into the cell by the PTS, HPr, present in the nonphosphorylated form, will activate NagB in response to the availability of PTS sugar substrates if and only if cytoplasmic GlcNAc-6P is present (
12). However, what if NANA is utilized? During exogenous NANA utilization, the transporter is NanT, and no accumulation of GlcNAc-6P occurs; this means that the PTS protein, HPr, should be largely phosphorylated (HPr-P), and HPr-P has no effect on the activity of NagB (
12). Under these conditions,
nanE gene expression is induced in response to the availability of cytoplasmic NANA, so that even in the absence of GlcNAc-6P, NagB will be activated by NanE. The concentration of GlcNAc-6P can be low during NANA utilization, since NanE is the epimerase for ManNAc-6P, catalyzing a reversible reaction to GlcNAc-6P as product.
nagB expression is dependent on the GlcNAc-6P concentration, and during NANA utilization, when cytoplasmic concentrations of GlcNAc-6P are low, activation of NagB by NanE is physiologically relevant.
In the situation where the nitrogen source is limiting, the utilization of low concentrations of amino sugars produces the essential level of nitrogen, and the synergistic effect of U-PII- and HPr-dependent activation of NagB (as sensors of nitrogen limitation and extracellular amino sugar availability, respectively) is also physiologically relevant.
Thus, we propose that NanE activates NagB when cytoplasmic or exogenously derived NANA is available and metabolized. HPr activates NagB only when the PTS is used for amino sugar uptake, and U-PII activates NagB primarily under nitrogen-limiting conditions. Thus, NanE transmits a signal indicating the presence of cytoplasmic NANA, HPr signals the availability of an extracellular amino sugar substrate, and U-PII signals nitrogen deficiency since GlnD, which uridylylates PII, senses the ratio of cytoplasmic αKG to glutamine. These signal transducing systems (
Fig. 7) allow the bacteria to respond to at least three different signals, all converging to regulate the activity of glucosamine-6-phosphate deaminase in response to nitrogen and carbon source availability.
We have further shown that HPr and U-PII, as well as HPr and NanE, act synergistically under appropriate conditions, enhancing the activating effect of either one. In contrast, U-PII and NanE had no measurable synergistic effect, suggesting that they bind to the same site or overlapping sites on NagB or the NagB-HPr complex. One noteworthy aspect of the NagB interface (adjacent/overlapping binding sites for U-PII, NanE, and HPr) is that it is highly disordered. There is almost no secondary structure at all, although there are loops. Also noteworthy is that this disordered area is large, approximately one-fourth of the entire surface area of NagB. This suggests that the NagB interface is highly flexible. Analysis of the models suggests that HPr, U-PII, and NanE bind very close to each other on NagB. It is possible that the flexibility of the NagB interface allows for the simultaneous binding of HPr/U-PII, as well as HPr/NanE. However, an entire alpha helix of NanE overlaps with the core of the U-PII binding site, making simultaneous binding of these two NagB activators unlikely, no matter what extent of flexibility is allowed. Moreover, the modeling of the HPr/U-PII/NagB and HPr/NanE/NagB complexes suggested which residues are involved in the simultaneous interactions of the two proteins (HPr/U-PII or HPr/NanE) with NagB (data not shown). They explain why the possibility of simultaneous U-PII/NanE binding could be excluded.
MATERIALS AND METHODS
GlcNAc-6P, GlcN-6P, NADP, UTP, l-glutamine, α-ketoglutarate, DTT, and other chemicals were purchased from Sigma-Aldrich, USA.
Protein-protein interaction analysis for the NagB interactome.
The scores recorded in
Table 1 represent the log likelihood scoring (LLS), which was computed by integrating the HyperGeometric Spectral Counts score (HGSCore [
34]) and the Comparative Proteomic Analysis Software Suite (ComPASS)
S-score (
35) into a single combined score to define high-quality associations. The procedures for the LLS calculation and precise recombination and perfect in-frame fusion of the SPA-tag to the natural C terminus of the target proteins are described elsewhere (
13).
Cloning nagB into pMST3.
The nagB gene, encoding the GlcN-6P deaminase, NagB, was PCR amplified from the E. coli BW25113 chromosome using the oligonucleotides nagB-Bam-F (ATAGGATCCAGACTGATCCCCCTGACTACCGCTGAAC) and nagB-Sal-R (CTCGTCGACTTACAGACCTTTGATATTTTCTGCTTC). The product was gel purified, digested with BamHI and SalI, and then cloned into the pSMT3 vector digested with the same restriction endonucleases. Individual clones were confirmed by colony PCR and subsequently by DNA sequencing. The resultant recombinant plasmid, pMST3-nagB, carried the nagB structural gene (without the first codon) fused to the 3′ end of the SUMO gene (without its stop codon) encoding the SMT3-His tag. Expression of “SUMO:nagB” was under the control of the T7 promoter. The SMT3 tag, present in the fusion protein, was removed using the Ulp1 Sumo protease. The resultant NagB enzyme (Ser1-NagB) has a serine residue instead of the N-terminal methionine residue (Met1-NagB).
Protein purification.
Recombinant proteins NagB, NanE, PII (GlnB), GlnD, Zwf, Pgi, Tsf, and BglA, all containing an N-terminal His
6 tag, were overexpressed in
E. coli and purified using Ni
2+-chelating chromatography.
E. coli OE strains for NanE, PII, GlnD, Zwf, Pgi, Tsf, and BglA, all from the ASKA collection (
36), were used for protein purification. Strains were grown in Luria-Bertani medium (50 ml), induced by the addition of 0.6 mM IPTG (isopropyl-β-
d-thiogalactopyranoside), and harvested after 4 h of shaking. Rapid purification of recombinant proteins on Ni-nitrilotriacetic acid-agarose minicolumns was performed as described previously (
37). PII protein was refolded as described previously (
12).
NagB activity measurements.
The activity of the purified recombinant NagB protein was routinely assayed in a cuvette at 37°C using a standard enzymatic coupling assay involving phosphoglucose isomerase (Pgi) and glucose 6-phosphate dehydrogenase (Zwf) by measuring the increase in absorbance at 340 nm resulting from the reduction of NADP as described previously (
26). NagB kinetics as a function of the GlcN-6P concentration were measured using 0 to 0.4 mM GlcNAc-6P in a 0.1-ml assay mixture in the presence of 0.2 M Tris (pH 6.5 to 8.1), 5 mM phosphate, 10 mM MgSO
4, 3 mM NADP, 50 mM KCl, 1.2 U of Zwf, 1.2 U of Pgi, and 10 to 50 nM NagB.
Effect of different protein-protein interactions on NagB activity.
We examined the effects of His-tagged, purified, recombinant proteins, NanE, PII (GlnB), GlnD, Zwf, Tsf, and BglA on NagB activity. GlnD proved to have no effect on NagB activity under the conditions used in the assay mixture. The activity of NagB was measured after purification with a His tag, followed by proteolytic removal of the His tag resulting in the Ser-1 NagB derivative. Met-1 in NagB has been shown to play a role in activation by GlcNAc-6P (
10), possibly explaining the higher
Ka of 2.1 mM (
12) in the activation of NagB by GlcNAc-6P reported here compared to that reported previously (
9).
E. coli NagB activity was measured by monitoring the increase in absorbance at 340 nm resulting from the reduction of NADP in a coupled assay involving Pgi and Zwf. This assay is based on the conversion of GlcN-6P to fructose 6-phosphate by NagB, followed by isomerization to glucose 6-phosphate by Pgi and further oxidation of glucose 6-phosphate to gluconate 6-phosphate by Zwf. We showed that the
E. coli Zwf (NADPH producing), under the conditions used, had no effect on NagB activity (data not shown). The Zwf with a different cofactor specificity (NADH-producing) from
Leuconostoc mesenteroides was used as the coupling enzyme (
38) to test the effect of
E. coli Zwf, and the NagB kinetics were the same regardless of which coupling enzyme was used.
GlnD-dependent uridylylation of the PII protein.
GlnD was assayed in a 1-ml assay mixture containing the purified refolded PII recombinant protein at a concentration of 10 μM. The reaction for the covalent modification of PII included 0.2 M Tris (pH 7.5), 1 mM ATP, 3 mM UTP, 1 mM DTT, 200 nM GlnD, 50 mM KCl, and 0.5 mM α-ketoglutarate. The reaction mixture was incubated at 30°C for 20 min and for 4 h at 25°C. The level of PII uridylylation was measured by native gel electrophoresis (data not shown). The fully posttranslationally modified U-PII was not purified from the PII uridylylation reaction mixture, but no effect of this mixture on NagB activity was noticed. In this control, the PII uridylylation mixture without the PII protein was added to the NagB assay mixture. It should be noted that the former mixture was added to the latter mixture with a 10-fold dilution (data not shown). Loss of uridylylation occurred during incubation of the U-PII protein for 3 days at 4°C in the elution buffer.
Structural modeling of the U-PII/HPr and NanE/HPr protein interactions with NagB.
HPr, NanE, and U-PII were each docked to NagB individually using the HADDOCK webserver with CPORT-predicted interface residues as active and passive restraints. All structures from clusters with negative z-scores (below average energy scores among clusters of the top 200 structures) were considered in the modeling. Docked complexes for HPr/NagB, NanE/NagB, and U-PII/NagB were aligned by NagB in Pymol. The four, one, and three clusters for HPr/NagB, NanE/NagB, and U-PII/NagB, respectively, were selected for further analysis, since these allowed unobstructed orientations for HPr and NanE, as well as HPr and U-PII, in their bound states with NagB. The HPr/NagB and NanE/NagB clusters were both the largest (greatest number of docking models) clusters from their respective docking runs, while the U-PII/NagB cluster was the second largest. The PDB identification codes and chains used for NagB, HPr, and U-PII were IFS5 chain A (
1FS5 chain A), 3CCD chain A (
3CCD chain A), and 5L9N chain A (
5L9N chain A), respectively.
1FS5 chain A is a structure of the open, “R” conformation of NagB, and
5L9N chain A is a structure of uridylylated PII. For NanE, the full-length Swiss Model Repository model based on the template
3IGS chain A (79.7% sequence identity) was used. For fitting of the model into the NagB hexamer, PDB
1CD5 chains A to F were used.
1CD5 is the NagB hexamer in the closed T form. For studying the fitting of the model within the hexamer, six copies of the open R conformer,
1FS5 chain A, were aligned onto each of the six subunits of
1CD5. For conservation analyses ConSurf-DB (
39) was used, and only the most highly conserved (level 9/9) residues are highlighted in
Fig. 6.