What is interesting to note, however, is that when both the ΔnagA mutants were grown on Aga, the induced levels of nagB fell drastically to about 10% of that in glycerol grown ΔnagA mutants (Table 1). A very likely reason why this happens is that upon induction of agaA in ΔnagA mutants by Aga, the induced AgaA deacetylates the accumulated GlcNAc-6-P to GlcN-6-P thereby lowering the intracellular concentration of GlcNAc-6-P which results in turning down the expression of the nag regulon. This strongly suggests that AgaA can deacetylate GlcNAc-6-P

in addition to Aga-6-P just like NagA can substitute for the absence of AgaA. Finally, in Aga grown EDL933 ΔnagA the induced levels of agaA and agaS were about 220% and 180%, respectively, of that in Aga grown EDL933 and likewise, in E. coli C ΔnagA grown on Aga, the induced levels of agaA and agaS were about 550% and 150%, respectively, of Batimastat ic50 that in E. coli C grown on Aga. Why this happens remains to be investigated. Constitutive expression of the aga/gam regulon enables a ΔnagA mutant to grow on GlcNAc The induction of nagB in ΔnagA mutants grown on glycerol and its repression when grown on Aga (Table 1) indicated that AgaA deacetylated GlcNAc-6-P. Unlike ΔagaA mutants which grew on Aga (Figure 2) because nagA was expressed in these mutants by Aga (Table 1), check details ΔnagA mutants did not grow on GlcNAc most likely

Erastin in vivo because agaA is not expressed with GlcNAc (Figure 1). If this is true, then deleting the agaR gene, that codes for the repressor of the aga/gam regulon, in a ΔnagA mutant would result in the constitutive expression of the aga/gam regulon and thereby of agaA that would allow its growth on GlcNAc. Therefore, agaR deletion mutants in E. coli C and in E. coli C ΔnagA were constructed and examined for growth on GlcNAc. As shown in Figure 3, E. coli C and E. coli C ΔagaR grew on GlcNAc and the ΔnagA mutant

did not grow but the double knockout strain, E. coli C ΔnagA ΔagaR, did indeed grow on GlcNAc. Phenotypic microarray [13] done with E. coli C ΔnagA ΔagaR also revealed that it regained the ability to utilize ManNAc and N-acetylneuraminic acid in addition to that of GlcNAc (data not shown) as their utilization is nagA dependent [5]. Analysis by qRT-PCR was done to confirm that agaA and agaS were constitutively expressed in E. coli C ΔagaR and in E. coli C ΔnagA ΔagaR. As shown in Table 2, agaA and agaS were expressed in E. coli C ΔagaR and E. coli C ΔnagA ΔagaR irrespective of the carbon source used for growth but nagA and nagB were induced only by GlcNAc and, as expected, nagA expression was not detected in E. coli C ΔnagA ΔagaR. In fact, agaA and agaS were induced higher in these ΔagaR mutants than that in Aga grown E. coli C, the only exception being that of agaA whose induction was slightly lower in GlcNAc grown E. coli ΔagaR.