Genome-wide RNAi screens in African trypanosomes identify the nifurtimox activator NTR and the eflornithine transporter AAT6

Abstract

To be effective, therapeutic compounds must typically enter target cells and, in some cases, must be concentrated or modified. Thus, uptake and activation mechanisms often form the basis of selectivity against infectious agents. Loss-of-function screens can be used to identify proteins involved in drug uptake and metabolism and may also identify clinically relevant potential resistance mechanisms. We used a genome-scale RNA interference (RNAi) library to identify loss-of-function resistance mechanisms in bloodstream-form Trypanosoma brucei. Nifurtimox-Eflornithine Combination Therapy (NECT) was recently introduced for Human African Trypanosomiasis and we focus on these drugs here. Screens for resistance to nifurtimox and a related drug, benznidazole, identified loss of nitroreductase (NTR) pro-drug activator function. A screen for resistance to the amino-acid analogue, eflornithine, identified loss of amino-acid transporter (AAT6) function. Our results confirm recent findings and suggest that NTR or AAT6 loss-of-function represent major potential mechanisms of resistance to these drugs. Thus, bloodstream-form T. brucei RNAi libraries present a versatile tool for selective genetic screening and for the rapid identification of drug-activation, uptake and potential resistance mechanisms.

Fig. 1

RNAi library drug resistance screens. (A) The schematic representation illustrates the RNAi library selective screening protocol. (B) Cumulative growth curve during nifurtimox selection. Bloodstream-form T. brucei, MiTat 1.2, clone 221a-derived libraries were maintained as described ; >5 × 10⁶ cells per culture at a density of <2 × 10⁶ cells ml⁻¹. Tetracycline (Tet; 1 μg ml⁻¹) was added at −1 day to induce RNAi, and nifurtimox (Bayer, 3 μM; 1 × EC₅₀) was added at time 0. Cultures were maintained thereafter in medium containing Tet and selective drug. Genomic DNA was recovered on day 8 and increased resistance was confirmed to be Tet-dependent (data not shown). The inset shows the drug structure. (C) Cumulative growth curve during eflornithine selection. Eflornithine (Sanofi-Aventis) was applied at 70 μM (2.5 × EC₅₀). Other details as in B.

Fig. 2

Nifurtimox and eflornithine resistance mechanisms. (A) Amplified products from the nifurtimox (N) screen. PCR amplification was carried out over 30 cycles using the primers, LIB2f (TAGCCCCTCGAGGGCCAGT) and LIB2r (GGAATTCGATATCAAGCTTGGC); 95 °C for 30 s, 57 °C for 30 s and 72 °C for 130 s. The products were separated in 1% agarose gels and then sequenced directly and also cloned to derive sequence from both ends. This confirmed that each product did indeed represent an RNAi vector-derived target fragment and also allowed precise mapping to the reference genome. M, molecular weight markers; ST, sequences that mapped to repetitive, sub-telomeric domains. (B) Amplified products from the eflornithine (E) screen. Other details as in A. (C) Genetic map (black boxes represent protein-coding sequences) indicating the location of the RNAi target fragments recovered from the library following the nifurtimox and benznidazole screens (grey boxes). (D) Genetic map indicating the location of the RNAi target fragments recovered from the library following the eflornithine screen. The RNAi target fragment used in E is represented by an open box. The locations of predicted AAT6 transmembrane-coding sequences are indicated, as well as a schematic showing the predicted transmembrane (TM) topology (open TM regions are putative). (E) AAT6 knockdown confers resistance to eflornithine. The RNAit program was used to design primers and the pRPa^iSL construct was modified to target AAT6 for RNAi. The resulting hairpin RNAi constructs were transferred to 2T1 cells . For EC₅₀ determination, cells were seeded at 2 × 10³ ml⁻¹ in 96-well plates in an eflornithine 2-fold dilution series. After 66 h growth, 20 μl of Alamar blue (AbD serotec) was added to each well and the plates incubated for a further 6 h. Fluorescence was determined using a Gemini Fluorescent Plate reader (Molecular Devices) at an excitation wavelength of 530 nm, an emission wavelength of 585 nm and a filter cut-off of 570 nm . Error bars indicate sd from triplicate assays. These results were confirmed using three additional independent clones (data not shown).