Mechanism-based inhibition of gut microbial tryptophanases reduces serum indoxyl sulfate

Abstract

Indoxyl sulfate is a microbially derived uremic toxin that accumulates in late-stage chronic kidney disease and contributes to both renal and cardiovascular toxicity. Indoxyl sulfate is generated by the metabolism of indole, a compound created solely by gut microbial tryptophanases. Here, we characterize the landscape of tryptophanase enzymes in the human gut microbiome and find remarkable structural and functional similarities across diverse taxa. We leverage this homology through a medicinal chemistry campaign to create a potent pan-inhibitor, (3S) ALG-05, and validate its action as a transition-state analog. (3S) ALG-05 successfully reduces indole production in microbial culture and displays minimal toxicity against microbial and mammalian cells. Mice treated with (3S) ALG-05 show reduced cecal indole and serum indoxyl sulfate levels with minimal changes in other tryptophan-metabolizing pathways. These studies present a non-bactericidal pan-inhibitor of gut microbial tryptophanases with potential promise for reducing indoxyl sulfate in chronic kidney disease.

Keywords: Gut-derived uremic toxins; indoxyl sulfate; mechanism-based inhibitor; microbiome-targeted inhibition; tryptophanase.

Figure 1.. The gut-microbial tryptophanase-ome exhibits high functional homology

(A) Derivation pathway of indoxyl sulfate: tryptophan is broken down into indole by gut microbial tryptophanase (tryptophan-indole-lyase, TIL) enzymes and free indole can permeate the gut epithelial barrier and progress into the liver, where it is further metabolized into indoxyl sulfate by host cytochrome P450 (CYP) and sulfotransferase (SULT) enzymes. (B) Sequence similarity network of the 183 TIL sequences identified from the Integrated Gene Catalog (IGC). The sequences were organized using the EFI-EST tool with an E value of 1 x 10⁻¹⁶⁰ and a sequence identity threshold of 95%. (C) Percent occurrence frequency values of TILs identified within the IGC reveal several sequences that are present in >70% of samples. Sequences were organized by taxonomic class. See also Figure S1B. (D) Kinetic variables determined from Michaelis-Menten plots using the substrate L-Tryptophan for TIL enzymes from diverse microbes. Error bars represent SEM from n = 3 independent experiments. For K_M and K_cat/K_M, an ordinary one-way ANOVA with Tukey’s multiple comparison was conducted **p<0.005, ***p<0.001, ns= no significance. See also Figure S1C. Bt, Bacteroides thetaiotaomicron; Ap, Alistipes putredinis; Ec, Escherichia coli; Fv, Fusobacterium varium; Ff, Flavonifractor sp. An52; Bs, Butyricicoccus sp. BIOML-A1.

Figure 2.. TILs from diverse taxa display remarkable structural homology

(A) Dimers of three resolved TIL structures (Bt, pink; Fv, gray; Bs, blue) and one extant structure (Ec, yellow) reveal conserved core folds and secondary structural features. (B) TIL quaternary structure assembly from a monomer, showing the small and large domains (SD, LD), to a dimer and finally to the biological tetramer. The Bt TIL structure is shown and PLP is rendered as spheres.

Figure 3.. The active site architecture of TILs across the human gut microbiome is well-conserved

(A) The active site architecture of all resolved structures displays analagous placement of key residues involved in cofactor stabilization (left panel) and catalysis (right panel). (B) Crystal structure of Bs TIL reveals the same core fold and dimer formation as all other TILs but a 45 amino acid insert (midnight blue) near the dimerization and tetrameric interfaces. See also Figure S1. (C) Root mean square deviation (RMSD) and percent identity (% ID) comparison between TIL sequences and their structures reveal highly similar Cα position alignment while maintaining low percent sequence identity. Bs TIL 45 amino acid insert was removed for RMSD calculations. PLP was not shown for Fv or Bs TIL structures for clarity. Bt, Bacteroides thetaiotaomicron; Ec, Escherichia coli; Fv, Fusobacterium varium; Bs, Butyricicoccus sp. BIOML-A1.

Figure 4.. (3S) ALG-05 displays improved pan-activity and acts as a mechanism-based inhibitor of TILs

(A)Inhibitory assays against diverse TILs show improved activity of (3S) ALG-05 against Bt, Ap, and Ff compared to OxA. K_i values were calculated using IC50 values and the Cheng-Prusoff Equation. Error bars represent SEM from n = 3 independent experiments. Two-way ANOVA with multiple comparisons **p<0.01, ***p<0.001, ****p<0.0001, ns= no significance. See also Figure S2. (B) Ultraviolet-visible spectroscopy of Ec TIL alone, with tryptophan, and with (3S) ALG-05 confirms that the inhibitor stalls the quinonoid complex. See also Figure S3. (C) Resolved co-crystal structure of Bs TIL with (3S) ALG-05 reveals the PLP-inhibitor complex binding pose and key electrostatic, aromatic, and halogen bonding interactions. Structure was resolved at 2.07 Å resolution, [2Fo-Fc] maps were contoured at σ=1. Bt, Bacteroides thetaiotaomicron; Ap, Alistipes putredinis; Ec, Escherichia coli; Fv, Fusobacterium varium; Ff, Flavonifractor sp. An52; Bs, Butyricicoccus sp. BIOML-A1.

Figure 5.. (3S) ALG-05 non-lethally inhibits TIL in microbes, is well-tolerated by mammalian cells, and reduces IS levels in vivo

(A) Indole production over time by E. coli treated with (3S) ALG-05. Error bars represent SEM from n = 2 independent experiments. Two-way ANOVA with multiple comparison *p<0.05, **p<0.01, ***p<0.001, ns= no significance. (B) Microbial growth kinetics of E. coli exposed to different concentrations of (3S) ALG-05. Error bars represent SEM from n = 3 independent experiments. (C) Mammalian cell line HEK293T treated with a range of (3S) ALG-05 concentrations for 24 hours. Error bars represent SEM from n = 3 independent experiments. Ordinary one-way ANOVA with Tukey’s multiple comparison ***p<0.001, ****p<0.0001. (D) Mammalian cell line Caco-2 treated with a range of (3S) ALG-05 concentrations for 24 hours. Error bars represent SEM from n = 3 independent experiments. Ordinary one-way ANOVA with Tukey’s multiple comparison *p<0.05. (E) Mouse study design, n = 10 mice were oral gavaged with saline or (3S) ALG-05 and 15 hours post-treatment, mice were sacrificed and their blood and organs were harvested. (F) Targeted metabolomics performed on the serum using LC/MS-MS. Student’s T-test, *p<0.05, **p<0.01, ns= not significant. ND= not detected. (G) Targeted metabolomics of cecal contents using LC/MS-MS and GC/MS. Student’s T-test, *p<0.05, **p<0.01, ns= not significant. ND= not detected.