Glucose- and glucokinase-controlled mal gene expression in Escherichia coli

Abstract

MalT is the central transcriptional activator of all mal genes in Escherichia coli. Its activity is controlled by the inducer maltotriose. It can be inhibited by the interaction with certain proteins, and its expression can be controlled. We report here a novel aspect of mal gene regulation: the effect of cytoplasmic glucose and glucokinase (Glk) on the activity and the expression of MalT. Amylomaltase (MalQ) is essential for the metabolism of maltose. It forms maltodextrins and glucose from maltose or maltodextrins. We found that glucose above a concentration of 0.1 mM blocked the activity of the enzyme. malQ mutants when grown in the absence of maltodextrins are endogenously induced by maltotriose that is derived from the degradation of glycogen. Therefore, the fact that glk malQ(+) mutants showed elevated mal gene expression finds its explanation in the reduced ability to remove glucose from MalQ-catalyzed maltodextrin formation and is caused by a metabolically induced MalQ(-) phenotype. However, even in mutants lacking glycogen, Glk controls endogenous induction. We found that overexpressed Glk due to its structural similarity with Mlc, the repressor of malT, binds to the glucose transporter (PtsG), releasing Mlc and thus increasing malT repression. In addition, even in mutants lacking Mlc (and glycogen), the overexpression of glk leads to a reduction in mal gene expression. We interpret this repression by a direct interaction of Glk with MalT concomitant with MalT inhibition. This repression was dependent on the presence of either maltodextrin phosphorylase or amylomaltase and led to the inactivation of MalT.

FIG. 1.

In vitro MalQ activity with maltose as substrate. (A) The reaction was performed with 4.4 μg of C-terminally His₆-tagged MalQ in a total volume of 16 μl containing 250 mM Tris-HCl (pH 7.5), 10 mM MgCl₂, and different concentrations of unlabeled glucose. After 10 min of incubation at room temperature, [¹⁴C]maltose (final concentration, 75 μM) was added to the mixture, followed by incubation for another 20 min at 30°C. Portions (5 μl) of the reaction were spotted. The figure represents the autoradiogram of the TLC plate. The standards were as follows: lane 1, [¹⁴C]glucose; lane 2, [¹⁴C]maltose; lane 3, [¹⁴C]maltotriose. Glucose additions to the enzymatic assays were as follows: lane 4, no glucose; lane 5, 0.0002 mM; lane 6, 0.001 mM; lane 7, 0.002 mM; lane 8, 0.01 mM; lane 9, 0.02 mM; lane 10, 0.1 mM; lane 11, 0.2 mM; lane 12, 1.0 mM; lane 13, 2.0 mM; lane 14, 10 mM. (B) The reaction was performed as in panel A with the additional 50 mM MgCl₂, 25 mM ATP, and 100 μM glucose without (lane 3 to 7) and with 1.5 μg of Glk (lanes 8 to 12). After 0, 1, 2, 5, and 20 min, 4 μl of the reaction was spotted. The standards were as follows: lane 1, [¹⁴C]glucose; lane 2, [¹⁴C]maltose; lane 13, [¹⁴C]glucose-6-phosphate.

FIG. 2.

Interaction of Glk with PtsG. Portions (50 μg) of crude extract of strain JMG77(ΔptsG) (lanes 1 to 4) and JMG77 harboring plasmid-encoded PtsG (lanes 5 to 8) were analyzed by SDS-PAGE followed by Western blotting with anti-His tag antibodies to detect Glk. Lanes: 1 and 5, first membrane-free supernatants (SN1); lanes 2 and 6, first pellets (P1); lanes 3 and 7, second supernatants (SN2); lanes 4 and 8, second pellets (P2).

FIG. 3.

Superimposition of GlK on interaction between Mlc and PtsG. (A) Partial view of Mlc (yellow) in its interaction with the EIIB domain of PtsG (green) (31). The interacting amino acids between the two proteins are shown in blue. Glk (in gray) (26) has been superimposed onto the Mlc structure (48) by using the Coot program (http://www.ysbl.york.ac.uk/∼emsley/coot/). The amino acids contacting the EIIB domain are in cyan. The C-terminal helix harboring the 16 amino acids lacking in Glk15 (Δ305-321) are shown in red. The helix-turn-helix domain of Mlc (lacking in Glk) is on the lower left-hand corner. (B) Detailed view of the interaction site seen from the side opposite that shown in panel A.

FIG. 4.

Size of the truncated version of Glk. Proteins were analyzed by SDS-PAGE, followed by Western blotting with anti-Glk antibodies. Lane 1, wild-type MG1655; lane 2, JW2385 (Δglk); lane 3, JW2385 harboring glk15 (Δ305-321); lane 4, JW2385 harboring plasmid-encoded N-terminal His₆-tagged wild-type Glk under IPTG induction.

FIG. 5.

Structure of Glk. The structure of the Glk dimer (gray) (26) is shown. The C-terminal helix harboring the 16 amino acids lacking in Glk15 (Δ305-321) are shown in red. They most likely constitute the interaction domain with MalT. The putative interaction site with EIIB of PtsG is indicated in cyan.

FIG. 6.

Effect of cytoplasmic glucose on the activity of MalQ and on glycogen-dependent endogenous induction. Shown is the pathway of glycogen degradation producing maltotriose, the inducer of MalT, the transcriptional activator of all mal genes. Glucosyl residues are indicated by small circles, horizontal lines between the circles indicate α(1-4) linkages, bent arrows indicate α(1-6) linkages. Unlinked arrows indicate the reducing end of the maltodextrins and glucose. Solid circles indicate the origin of maltotriose from glycogen and its further metabolism. (Lower left branch) In a glk mutant, glucose formed by the action of MalQ or by hydrolysis of glucose containing substrates (not shown) inhibit MalQ, preventing the removal of maltotriose, thus establishing high endogenous induction. (Lower right branch) Overexpression of Glk removes free glucose, allowing high MalQ activity. This leads to the removal of maltotriose by the formation of larger maltodextrins, thus causing the loss of MalT activation and a reduction in mal gene expression. Not shown here is the action of MalZ or the synthesis of glycogen.

FIG. 7.

Binding of Glk to MalT inhibits MalT activity. Two situations are shown. On the right-hand side, the “normal” situation is shown. MalT exists in an equilibrium of the inactive monomer and the active dimer (multimer). On the left-hand side, the overproduction of glucokinase is shown. In the proposed model, high concentrations of Glk result in the binding of Glk dimers to monomeric and inactive MalT, preventing mal gene expression. Not shown in this scheme is the essential role of MalP or MalQ in Glk-dependent MalT inhibition. Also not shown is the role of maltotriose in preventing the inhibition by Glk. The established inhibition of MalT by other proteins such as Aes, MalK, or MalY is not shown but was the basis for this model. Also novel is the proposal that even in the absence of the inducer maltotriose MalT can form a transcriptionally active species.

FIG. 8.

The competition between Mlc and glucokinase for PtsG binding affects the expression of malT. Two situations are shown. On the right-hand side, the “normal” situation is shown. A certain portion of Mlc, the repressor of malT, is bound by PtsG and is not available for the inhibition of malT transcription. On the left-hand side, the overexpression of Glk is shown. The high concentration of Glk replaces Mlc on PtsG. The increased concentration of Mlc leads to the inhibition of malT transcription and therefore to a reduced mal gene expression. Not reflected in this scheme is the dependence of malT expression on the cAMP/CAP complex and the state of PtsG phosphorylation on binding Mlc (or Glk).