Fine-tuning of amino sugar homeostasis by EIIA(Ntr) in Salmonella Typhimurium

Abstract

The nitrogen-metabolic phosphotransferase system, PTS(Ntr), consists of the enzymes I(Ntr), NPr and IIA(Ntr) that are encoded by ptsP, ptsO, and ptsN, respectively. Due to the proximity of ptsO and ptsN to rpoN, the PTS(Ntr) system has been postulated to be closely related with nitrogen metabolism. To define the correlation between PTS(Ntr) and nitrogen metabolism, we performed ligand fishing with EIIA(Ntr) as a bait and revealed that D-glucosamine-6-phosphate synthase (GlmS) directly interacted with EIIA(Ntr). GlmS, which converts D-fructose-6-phosphate (Fru6P) into D-glucosamine-6-phosphate (GlcN6P), is a key enzyme producing amino sugars through glutamine hydrolysis. Amino sugar is an essential structural building block for bacterial peptidoglycan and LPS. We further verified that EIIA(Ntr) inhibited GlmS activity by direct interaction in a phosphorylation-state-dependent manner. EIIA(Ntr) was dephosphorylated in response to excessive nitrogen sources and was rapidly degraded by Lon protease upon amino sugar depletion. The regulation of GlmS activity by EIIA(Ntr) and the modulation of glmS translation by RapZ suggest that the genes comprising the rpoN operon play a key role in maintaining amino sugar homeostasis in response to nitrogen availability and the amino sugar concentration in the bacterial cytoplasm.

Figure 1. EIIA^Ntr specifically binds to glucosamine-6-phosphate synthase (GlmS).

(a) Ligand-fishing using EIIA^Ntr-His₆ as a bait. Total cell extract from S. Typhimurium SL1344 culture was incubated with EIIA^Ntr-His₆ or not. The proteins eluted from Ni-NTA resin (lane 1, whole cell extract only; lane 2, whole cell extract mixed with EIIA^Ntr-His₆) and purified EIIA^Ntr-His₆ (lane 3) were analyzed by SDS-PAGE. The protein band indicated by the arrow in lane 2 was identified as GlmS by LC-MS/MS analysis. Size markers (M) in kDa are aligned at left. (b) Co-purification of EIIA^Ntr and GlmS in vivo. E. coli BL21 (DE3) harboring pETDuet-glmS·ptsN was geneticallymanipulated to produce EIIA^Ntr and His₆-tagged GlmS simultaneously by isopropyl β-D-1-thiogalactoside (IPTG) addition. EIIA^Ntr bound to His₆-GlmS was co-purified using Ni-NTA metal-affinity resin. Proteins analyzed in SDS-PAGE are aliquots from total cell extracts (T), supernatant after centrifuging the cell extracts (S), column flow-through (F), wash (W), and the five elution fractions (E1–E5). Molecular masses of standards are presented in kDa on the left.

Figure 2. EIIA^Ntr phosphorylation status is influenced by nitrogen abundance.

(a) EIIA^Ntr modification strategy. The number of SDS molecules bound to a protein is affected by the charge of the amino acids around the phosphorylation site, which can change the mobility of the protein on SDS-PAGE. A histidine (73) amino acid of EIIA^Ntr was changed to alanine (73) to construct the unphosphorylated form of EIIA^Ntr (H73A), and a lysine (75) was substituted for aspartic acid (75) to provide negative charge effects on EIIA^Ntr. (b) Phosphorylation-dependent upshift of EIIA^Ntr (K75D). The intact form of EIIA^Ntr and its EIIA^Ntr (K75D) mutant derivative were incubated with or without 1 mM PEP under phosphorylating conditions and then analyzed by SDS-PAGE. EIIA^Ntr (K75D) showed excellent phosphorylation-dependent mobility shift (PDMS), whereas the intact form of EIIA^Ntr showed comparable mobility independent of its phosphorylation. (c) Differential phosphorylation status of EIIA^Ntr between metabolites. The phosphorylation-dependent mobility shift of EIIA^Ntr-His₆ (K75D) was assessed using different metabolites. EIIA^Ntr-His₆ (K75D) was incubated with PEP, His₆-EI^Ntr, and His₆-NPr in the presence of glutamine (Gln) or α-ketoglutarate (α-KG). The phosphorylation levels of EIIA^Ntr-His₆ (K75D) were compared by SDS-PAGE.

Figure 3. Phosphorylation status of EIIA^Ntr influences the binding affinity between EIIA^Ntr and GlmS.

(a) Increased binding affinities of EIIA^Ntr-His₆ (K75D) to GlmS after phosphorylation. Phosphorylated and unphosphorylated forms of EIIA^Ntr-His₆ (K75D) were incubated with equivalent amounts of GlmS, and the levels of GlmS bound to phosphorylated (P) or unphosphorylated (U) EIIA^Ntr-His₆ (K75D) were compared (top). PEP-dependent phosphorylation of EIIA^Ntr-His₆ (K75D) was verified in parallel (bottom inlet). Line (N) does not contain either the EIIA^Ntr-His₆ (K75D) or the GlmS protein as a control. (b) Differential interaction between EIIA^Ntr and GlmS depending on EIIA^Ntr phosphorylation status in vivo. Protein-protein interactions between EIIA^Ntr and GlmS were verified using a bacterial two-hybrid system. Plasmid pKT25 containing ptsN or ptsN (H73A) and plasmid pUT18 harboring glmS were introduced into a reporter strain respectively or in combination. The reporter strains were cultivated in LB broth supplemented with IPTG, and β-galactosidase activity was determined to examine the strength of the protein-protein interactions. This experiment was performed in triplicate.

Figure 4. EIIA^Ntr inhibits GlmS-mediated GlcN6P production.

(a) Negative effects of EIIA^Ntr on GlcN6P production. GlmS was incubated with different amounts of phosphorylated or unphosphorylated EIIA^Ntr-His₆ (K75D) (0–16 pmol) and GlcN6P production was measured by HPLC. The results from triplicates are plotted. (b) No effect of EIIA^Ntr on the expression of genes involved in amino sugar metabolism. Wild-type and ΔptsN mutant strains were transformed with pWJ04 containing ptsN and its presumable promoter, and qRT-PCR was conducted to compare mRNA levels of glmS, glmU, glmM, and ptsN. All mRNA levels were normalized using gyrB, and the relative expression ratios were averaged from three independent measurements.

Figure 5. Salmonella growth is influenced by the interaction between EIIA^Ntr and GlmS.

(a) The absence of glmS is lethal to bacteria. Effects of GlmS and EIIA^Ntr on bacterial growth were assessed individually or in combination. Introduction of pWJ10 (a pUHE21-2lacI^q::glmS) providing GlmS in trans partially restored the growth defect in ΔglmS mutant. GlmS production level was titrated using 10 μM IPTG (Supplementary Fig. S5). (b) Modulation of growth rate by the interaction between EIIA^Ntr and GlmS. Using the ΔglmS mutant complemented with pWJ10 as a parent strain, the effect of the EIIA^Ntr interaction with GlmS on bacterial growth was evaluated by deleting the chromosomal ptsN gene and introducing pWJ04 and pWJ05, which provided EIIA^Ntr and EIIA^Ntr (H73A), respectively. All growth measurements in (a,b) were performed in triplicate, and the average optical densities at 600 nm are plotted.

Figure 6. Trimeric complex is formed between EIIA^Ntr and GlmS.

(a) Stoichiometric analysis of the EIIA^Ntr and GlmS complex. A solution of the EIIA^Ntr and GlmS complex was analyzed by SEC-MALS. The molecular mass of GlmS alone in solution was 128 kDa, indicating the active dimer, and that of EIIA^Ntr was 20 kDa, indicating the monomeric form. A macromolecule with a peak of 148 kDa appeared in the solution composed of EIIA^Ntr and GlmS, and the size was accordant with a EIIA^Ntr-GlmS₂ heterotrimer composed of two GlmS molecules (128 kDa) and one EIIA^Ntr (20 kDa). (b) Identifying the eluted fractions by SDS-PAGE. The SEC-MALS eluted fractions from the 148, 128, and 20 kDa proteins, which were presumably the EIIA^Ntr-GlmS₂ complex, GlmS dimer, and the EIIA^Ntr monomer, respectively, were analyzed by SDS-PAGE and Coomassie Blue staining. Green, red, and blue lines indicate the 148, 128, and 20 kDa fractions, respectively.

Figure 7. EIIA^Ntr degradation is controlled by Lon protease in response to amino sugar availability.

(a) Comparison of EIIA^Ntr-FLAG protein levels between wild-type and ΔglmS mutant strains. Protein samples were isolated from wild-type and ΔglmS mutant strains in the presence or absence of GlcNAc 3 h post inoculation as described in Supplementary Figs S8A and S10, and subjected to Western blot analysis to compare the EIIA^Ntr-FLAG levels between conditions. (b) Effect of amino sugar availability on stability of EIIA^Ntr-FLAG. Chloramphenicol was added to the cultures of the wild-type and ΔglmS mutant strains at 3 h as described above, and total proteins harvested every 30 min were used to assess the EIIA^Ntr and DnaK decay rates. (c) Comparison of the stability of EIIA^Ntr-FLAG in the absence of Lon or ClpXP protease in the ΔglmS mutant strain. The ΔglmS, ΔglmS Δlon, and ΔglmS ΔclpXP Salmonella strains were cultured under conditions identical to those used above, and the level of EIIA^Ntr-FLAG and DnaK was assessed in each strain using Western blotting every 30 min after adding chloramphenicol. In all experiments, equivalent amounts of total protein were loaded in each lane for SDS-PAGE and DnaK levels were measured in parallel to verify equivalent total protein amounts between lanes.

Figure 8. Proposed regulatory model for the control of GlmS activity by EIIA^Ntr in response to nitrogen and amino sugar availability.

At high cellular glutamine (Gln) concentrations, EIIA^Ntr tends to be unphosphorylated and liberates the active form of GlmS that supplies GlcN6P and accelerates synthesis of the bacterial cell envelope (①). When GlcN6P is present in excess (②) or when Gln availability is restricted (③), phosphorylated EIIA^Ntr binds GlmS with increased affinity and inhibits its activity. The depletion of pre-existing GlcN6P leads to EIIA^Ntr degradation by Lon protease, freeing the active form of GlmS to supplement the lack of GlcN6P and maintain amino sugar homeostasis (④).