Validation of aminodeoxychorismate synthase and anthranilate synthase as novel targets for bispecific antibiotics inhibiting conserved protein-protein interactions

Abstract

Multi-resistant bacteria are a rapidly emerging threat to modern medicine. It is thus essential to identify and validate novel antibacterial targets that promise high robustness against resistance-mediating mutations. This can be achieved by simultaneously targeting several conserved function-determining protein-protein interactions in enzyme complexes from prokaryotic primary metabolism. Here, we selected two evolutionary related glutamine amidotransferase complexes, aminodeoxychorismate synthase and anthranilate synthase, that are required for the biosynthesis of folate and tryptophan in most prokaryotic organisms. Both enzymes rely on the interplay of a glutaminase and a synthase subunit that is conferred by a highly conserved subunit interface. Consequently, inhibiting subunit association in both enzymes by one competing bispecific inhibitor has the potential to suppress bacterial proliferation. We comprehensively verified two conserved interface hot-spot residues as potential inhibitor-binding sites in vitro by demonstrating their crucial role in subunit association and enzymatic activity. For in vivo target validation, we generated genomically modified Escherichia coli strains in which subunit association was disrupted by modifying these central interface residues. The growth of such strains was drastically retarded on liquid and solid minimal medium due to a lack of folate and tryptophan. Remarkably, the bacteriostatic effect was observed even in the presence of heat-inactivated human plasma, demonstrating that accessible host metabolite concentrations do not compensate for the lack of folate and tryptophan within the tested bacterial cells. We conclude that a potential inhibitor targeting both enzyme complexes will be effective against a broad spectrum of pathogens and offer increased resilience against antibiotic resistance.

Importance: Antibiotics are indispensable for the treatment of bacterial infections in human and veterinary medicine and are thus a major pillar of modern medicine. However, the exposure of bacteria to antibiotics generates an unintentional selective pressure on bacterial assemblies that over time promotes the development or acquisition of resistance mechanisms, allowing pathogens to escape the treatment. In that manner, humanity is in an ever-lasting race with pathogens to come up with new treatment options before resistances emerge. In general, antibiotics with novel modes of action require more complex pathogen adaptations as compared to chemical derivates of existing entities, thus delaying the emergence of resistance. In this contribution, we use modified Escherichia coli strains to validate two novel targets required for folate and tryptophan biosynthesis that can potentially be targeted by one and the same bispecific protein-protein interaction inhibitor and promise increased robustness against bacterial resistances.

Keywords: antibiotic resistance; antibiotic target validation; enzyme complex; glutamine amidotransferases; protein-protein interactions.

Fig 1

Reaction scheme and hotspot interface residues in ADCS and AS. (A) ADCS [homology model of the enzyme from E. coli (15)] consists of a glutaminase subunit PabA (light blue) and a synthase subunit PabB (blue). AS [crystal structure of the enzyme from Serratia marcescens, PDB-ID: 1I7Q (16)] is organized as a dimer of heterodimers (one heterodimer shown here) comprising the glutaminase subunit TrpG (light green) and the synthase subunit TrpE (green). The homologous complexes share the same fold and have r.m.s.d. values of 1.5 Å and 1.1 Å when superimposing the synthase or glutaminase subunits, respectively. The glutaminases hydrolyze glutamine to produce nascent ammonia, which is channeled to the respective synthase active site. In PabB, ammonia is incorporated at the C₄ position of chorismate to produce 4-amino-4-deoxy-chorismate (ADC), which is further processed by the aminodeoxychorismate lyase PabC to pyruvate and para-amino-benzoate (pABA), a precursor of folate. TrpE catalyzes the incorporation of ammonia at the C₂ position of chorismate and a subsequent lyase reaction to yield pyruvate and anthranilate, a precursor of tryptophan. (B) The positions of putative hotspot residues in the PPI of ADCS (left panel) and AS (right panel) are shown. (C) Schematic representation of the functional ADCS/AS complex with putative hotspot interface residues (left panel) and disruption of complex formation by the introduction of hotspot mutations (middle panel) or a putative PPI inhibitor (right panel).

Fig 2

Growth in liquid M9 minimal medium of E. coli strains with genomic manipulations of ecpabA or/and ectrpG. Growth curves of indicated strains were observed at 37°C in the absence (orange circles) and presence (black circles) of 100 nM pABA and 10 µM tryptophan (Trp), and SDs are shown as orange or gray frames. “Wild-type” represents the E. coli strain with unmodified pabA and trpG genes as positive control, “M9 control” indicates non-inoculated medium, “ΔpabA” and “ΔtrpG” denote genomic deletions of the respective gene, and “pabA Y127A, S171A” and “trpG Y132A, S173A” denote the introduction of hotspot mutations into the respective gene. Hotspot mutations resulted in a significant reduction of bacterial growth, which could be fully restored by the addition of pABA and Trp.

Fig 3

Growth on solid M9 minimal medium of E. coli strains with genomic manipulations of ecpabA or/and ectrpG. Colony formation velocity at 37°C of an E. coli wild-type control strain (wild-type, line 1) is compared with strains containing genomic manipulations of ecpabA (lines 2 and 3), ectrpG (lines 4 and 5), and the combined manipulations (lines 6 and 7). The addition of pABA and tryptophan (Trp) is indicated. Shown are growth plates after incubation for 1, 2, 3, and 4 days. Hotspot mutations resulted in a significant delay of bacterial growth, which could be fully restored by the addition of pABA or Trp.

Fig 4

Growth of E. coli strains with genomic manipulations of ecpabA or/and ectrpG on solid M9 minimal medium supplemented with heat-inactivated human plasma. Colony formation velocity at 37°C of an E. coli wild-type control strain (wild-type) is compared with a strain containing genomic manipulations of ecpabA and ectrpG (pabA Y127A, S171A and trpG Y132A, S171A) and the double deletion mutant (ΔpabA ΔtrpG). The addition of pABA and tryptophan (Trp) is indicated. Shown are growth plates after incubation for 1, 2, 3, and 6 days. Hotspot mutations resulted in a significant delay of bacterial growth, which could be fully restored by the addition of 100 nM pABA and 10 µM Trp.