Functional analysis of drug resistance-associated mutations in the Trypanosoma brucei adenosine transporter 1 (TbAT1) and the proposal of a structural model for the protein

Abstract

The Trypanosoma brucei aminopurine transporter P2/TbAT1 has long been implicated in the transport of, and resistance to, the diamidine and melaminophenyl arsenical classes of drugs that form the backbone of the pharmacopoeia against African trypanosomiasis. Genetic alterations including deletions and single nucleotide polymorphisms (SNPs) have been observed in numerous strains and clinical isolates. Here, we systematically investigate each reported mutation and assess their effects on transporter function after expression in a tbat1(-/-) T. brucei line. Out of a set of six reported SNPs from a reported 'resistance allele', none significantly impaired sensitivity to pentamidine, diminazene or melarsoprol, relative to the TbAT1-WT allele, although several combinations, and the deletion of the codon for residue F316, resulted in highly significant impairment. These combinations of SNPs, and ΔF316, also strongly impaired the uptake of [(3)H]-adenosine and [(3)H]-diminazene, identical to the tbat1(-/-) control. The TbAT1 protein model predicted that residues F19, D140 and F316 interact with the substrate of the transporter. Mutation of D140 to alanine resulted in an inactive transporter, whereas the mutation F19A produced a transporter with a slightly increased affinity for [(3)H]-diminazene but reduced the uptake rate. The results presented here validate earlier hypotheses of drug binding motifs for TbAT1.

Figure 1

Effective concentration 50% (EC ₅₀) values for four trypanocidal drugs as determined using the Alamar Blue fluorescence assay. The strains tested were all derived from T . b. brucei s427 wild‐type (WT); B48 is a multidrug‐resistant strain lacking the genes for the TbAT1 and TbAQP2 drug transporters (Munday et al., 2014). B48 + EV is the resistant control of B48 transfected with the empty vector (EV) pHD1336 (Biebinger et al., 1997). All the variants of TbAT1 including the wild‐type copy (B48 + TbAT1) were expressed in B48 using this vector (Munday et al., 2013). Mutations (Mut) 1–6 are as listed in the box; Δ(F316) is the TbAT 1 allele from which the codon for Phe316 was deleted. The dotted line running across each panel denotes the EC50 value of the B48 + EV control for each drug. Statistical significance was determined using a one‐way ANOVA relative to B48 + EV, in GraphPad Prism 5.0. **, P < 0.01; ***, P < 0.001. Data shown are the average and SEM of at least five independent determinations.

Figure 2

Transport of (A) 0.1 μM [³ H]‐adenosine or (B) 0.1 μM [³ H]‐diminazene aceturate by T . b. brucei  B48 transfected with empty vector pHD1336 (B48 + EV) expressing either the wild‐type TbAT1 allele or a mutant TbAT1 allele as listed in Fig. 1. Incubation times were 30 s or 60 s for adenosine or diminazene transport respectively. All bars are the average of three independent experiments and error bars represent SEM; statistical significance was calculated relative to TbAT1WT (unless specifically indicated) using a paired Student's t‐test: *, P < 0.05; **, P < 0.02; ***, P < 0.01.

Figure 3

Transport of 0.1 μM [³ H]‐adenosine by B48 cells expressing various TbAT1 alleles as indicated (numbering as in Fig. 1), in the presence or absence of various concentrations of unlabelled adenosine. Main figure: experiment performed in triplicate with unlabelled adenosine as inhibitor, resulting in sigmoid competition curves. Inset: conversion of the same experiment to Michaelis‐Menten saturation plots displaying total adenosine uptake against adenosine concentration. The experiment depicted is representative of four independent experiments, which are summarised in Table 1.

Figure 4

TbAT1 model created using ROBETTA. Residues whose mutation is described in this study are displayed in space‐filling models. The ligand adenosine is shown in green carbon atoms as docked in the TbAT1 protein model. Extracellular and intracellular intra TM‐helical loops were not modelled and are therefore not shown. Helices are indicated by rigid cylinders. Both top view (A; extracellular side) and side view (B) are shown. TM‐domains are labelled with roman numerals. The image was created using PyMol version 1.50.04, Schrödinger.

Figure 5

Structural overlay between the LdNT1.1 model published by Valdes et al. (2009; 2012) – in brown; and the TbAT1 model obtained created by ROBETTA – in cyan. The extracellular side is shown. The ligand adenosine docked in the TbAT1 protein model is shown in green carbon atoms. Helices are indicated by rigid cylinders. The extracellular and intracellular TM‐helical loops were not modelled and are therefore not shown. The TbAT1 residue numbering is used, with the LdNT1.1 residues in brackets. The image was created using PyMol version 1.50.04, Schrödinger.

Figure 6

Overlay between the predicted binding poses of adenosine (green carbon atoms) and diminazene (yellow carbon atoms) in the TbAT1 model obtained created by ROBETTA (colored in cyan). Residues F19, D140 and F316 are displayed in space‐filling models. Helices are indicated by rigid cylinders. The hydrogen bonds between the side‐chain carboxylate group of Asp140 and (i) the amino group and (ii) the protonated nitrogen‐1 of the adenosine ring are shown in dashed blue lines. The image was created using PyMol version 1.50.04, Schrödinger.

Figure 7

Schematic model for the interactions between TbAT1 ligands adenosine (A), diminazene (B) and melarsoprol (C) with the critical amino acid residues in the TbAT1 binding pocket; Frame (D) shows the putative interaction of phenylarsine oxide with the binding pocket. Hydrogen bonds are depicted as dotted lines. The image was created using ChemDraw Pro 10.0, CambridgeSoft.

Figure 8

EC₅₀ values obtained by the Alamar Blue test for strains expressing mutant alleles of TbAT1. Bars represent the average and SEM of six independent experiments. Statistical significance was determined by a one‐way ANOVA, compared to B48 + EV except where specifically indicated.

Figure 9

Uptake of 0.1 μM [³ H]‐diminazene is significantly reduced following reintroduction of mutated TbAT1 relative to the wild type TbAT1. A. Transport of diminazene was assessed in each strain. Bars show mean rate of transport ± SEM at 1 min when normalised for background, n = 4. * = P < 0.05, ** = P < 0.01, *** = P < 0.001 (one‐way ANOVA and Dunnett's test), compared against B48 + EV, unless otherwise indicated. B. Inhibition (main figure) and saturation plot (inset) of 0.1 μM [³ H]‐diminazene over 60 s in the presence or absence of various concentrations of unlabelled diminazene as indicated. The experiment show was performed in triplicate and representative of four independent experiments, which are summarised in Table 2.