Coupling chemical mutagenesis to next generation sequencing for the identification of drug resistance mutations in Leishmania

Abstract

Current genome-wide screens allow system-wide study of drug resistance but detecting small nucleotide variants (SNVs) is challenging. Here, we use chemical mutagenesis, drug selection and next generation sequencing to characterize miltefosine and paromomycin resistant clones of the parasite Leishmania. We highlight several genes involved in drug resistance by sequencing the genomes of 41 resistant clones and by concentrating on recurrent SNVs. We associate genes linked to lipid metabolism or to ribosome/translation functions with miltefosine or paromomycin resistance, respectively. We prove by allelic replacement and CRISPR-Cas9 gene-editing that the essential protein kinase CDPK1 is crucial for paromomycin resistance. We have linked CDPK1 in translation by functional interactome analysis, and provide evidence that CDPK1 contributes to antimonial resistance in the parasite. This screen is powerful in exploring networks of drug resistance in an organism with diploid to mosaic aneuploid genome, hence widening the scope of its applicability.

Fig. 1. Drug susceptibility and mutations in Leishmania selected for resistance.

a Susceptibility to miltefosine (MIL; left panel) and paromomycin (PMM; right panel) were performed on individual clones. The wild-type L. infantum (WT) is shown for both drugs. The MIL resistant mutants were selected after mutagenesis with either EMS or HMPA (Supplementary Table 1) while the PMM-resistant mutants were selected after mutagenesis with EMS, ENU, or MMS (Supplementary Table 1). Data are mean ± SEM. For the MIL susceptibility assay, n = 5 biologically independent replicates for the wild-type and n = 3 biologically independent replicates for the mutants. For the PMM susceptibility assay, n = 3 biologically independent replicates. Statistical analyses were performed using unpaired two-tailed t-tests. ***P < 0.001. b Genome-wide distribution of SNVs in mutants selected against MIL (left panel) and PMM (right panel). Bars represent the genes on each chromosome. Colored bars represent genes mutated in defined numbers of mutant clones; gray bars represent non-mutated genes. Source data are provided as a Source Data file.

Fig. 2. Kinase activity of the L. infantum CDPK1.

a Phosphorylation of proteins immunoprecipitated with CDPK1-HA as determined by autoradiography after in vitro kinase assay followed by electrophoresis in 12% SDS–PAGE. Assays were performed without γ-P³²-ATP (lane 1), with 2.5 μCi γ-P³²-ATP (2) and with γ-P³²-ATP in presence of excess of non-radioactive ATP (3). In an independent set of experiments, assays were performed with immunoprecipitated CDPK1-HA (4) and CDPK1^K61G-HA (5) both using 2.5 µCi of γ-P³²-ATP. Lower panels depict the amount of immunoprecipitated CDPK1 involved in each reaction as observed by Western blotting of the immunoprecipitated complex using mouse anti-HA IgG. b Impact of the CDPK1 K61G mutation on PMM resistance. EC₅₀ values were determined by dose responsive curves against PMM. The wild-type and K61G CDPK1 versions were cloned in PURO or ZEO vectors respectively, thus explaining two PMM25 controls transfected with the empty PURO and ZEO vectors. Data are mean ± SEM for n = 6 (PMM25(PURO); PMM25 + CDKP1) or n = 3 (PMM25(ZEO); PMM25 + CDPK1^K61G) biologically independent replicates. Statistical analyses were performed using unpaired two-tailed t-tests. ***P < 0.001. ns, not significant. Source data are provided as a Source Data file.

Fig. 3. The essential L. infantum CDPK1 and its role in paromomycin resistance.

a Schematic representation of the CDPK1 locus in L. infantum before and after integration of the inactivation cassettes hygromycin phosphotransferase B (5′-HYG-3′) and puromycin phosphotransferase (5′-PURO-3′) and expected sizes after digestion with SalI. PCR primers (P_f and P_r) and expected size are shown below the map. b Southern blot of DNAs digested with SalI and hybridized with a probe covering 300 bp of the 5′-UTR of CDPK1. L. infantum wild-type strain (1); CDPK1^+/H (2); CDPK1^+/P (3); CDPK1^H/P (4) and CDPK1^H/P + CDPK1 at passage 3 (5). c EC₅₀ values of L. infantum cells with various versions of CDPK1 were determined by dose responsive curves against PMM. Data are mean ± SEM for n = 8 (CDPK1^+/H), n = 6 (WT(HYG); CDPK1^+/H + CDPK1) or n = 5 (all other samples) biologically independent replicates. Statistical analyses were performed using unpaired two-tailed t-tests. ***P < 0.001. d Enumeration of activated macrophage (THP1) infectivity and dose responsiveness of wild-type (white) and CDPK1^H/+ recombinant parasites supplemented (light gray) or not (dark gray) with an episomal wild-type copy of CDPK1, as determined by calculating P_index after 96 h of PMM treatment. Superscript H and Z refer to the HYG and ZEO selectable markers, respectively. Superscript + refers to the wild-type allele. Data represents mean ± SEM for n = 4 independent biological replicates. Statistical analyses were performed using unpaired two-tailed t-tests. ***P < 0.001; **P < 0.01; ns, not significant. Source data are provided as a Source Data file.

Fig. 4. CDPK1 is linked to paromomycin and antimonial resistance.

a Dose responsiveness to PMM of the single knockout line CDPK1^+/H complemented with an episomal rescue coding for the wild-type, K61G or E629K versions of CDPK1. Superscript H and + refer to the HYG-inactivated and wild-type alleles, respectively. Data are mean ± SEM for n = 3 biological replicates. Statistical analyses were performed using unpaired two-tailed t-tests.***P < 0.001; **P < 0.01. Dose responsiveness against SbIII (b) and miltefosine (c) for the single knockout line CDPK^+/H complemented or not with an episomal rescue coding for the wild-type protein. Data are mean ± SEM for n = 3 biological replicates. Statistical analyses were performed using unpaired two-tailed t-tests. ***P < 0.001. d Mutations in CDPK1 detected in PMM (red) and SbIII (blue) resistant mutants. The mutation in black (E629K) was detected in one mutant selected by step wise exposure to SbIII in an independent study. An asterisk represents a stop codon. Source data are provided as a Source Data file.

Fig. 5. CDPK1 interactome and candidate targets.

a Nodes inferring specific gene-IDs as retrieved from SAINT analysis (AvP > 0.8) of IP data for CDPK1-HA, CDPK1^K61G-HA and CDPK1^V366E-HA. Visualization of predicted interacting partners by Cytoscape. Independent IP experiments, carried out using HA-DHFR-TS and HA-PTR1 expressed episomally in L. infantum, were treated as controls. Test and control experiments were set in three biological replicates. SAINT analysis was performed using total spectral counts for peptides identified with <1% FDR. The networks were merged using DyNet analyzer. Unique nodes are presented in orange, green and blue respectively for CDPK1^V366E-HA, CDPK1^K61G-HA and CDPK1-HA. Common nodes are presented in shades of red depending on the level of overlap. An enlarged version of the three interactomes is shown in Supplementary Fig. 8. b Number of overlapped proteins identified by immunoprecipitation of the three CDPK1-HA versions is shown by Venn diagram. Lists of unique proteins associated with CDPK1 and proteins present in all samples are found in Supplementary Data 3. c Cytoscape visualization of network depicting significantly (P < 0.05) over-represented GO-molecular function terms for the gene list generated by SAINT analysis of possible interactome for CDPK1. The network file was generated using REVIGO following analysis of the gene list for GO-enrichment using existing tools in TriTrypDB. Nodes represent specific GO-MF terms and intensity of shades signify extent of enrichment for the term.

Fig. 6. CDPK1 modulates translation efficiency.

Functional GO-annotation of the 206 proteins identified by LC-MS/MS from phosphorylated bands derived from SDS–PAGE of in vitro IP kinase reaction of CDPK1-HA (a); Functional GO-annotation of the 40 proteins (out of the 206) reported to be phosphorylated by Tsigankov et al. (2013) (b). Proteins were grouped functionally based on GO-slim terms. Fold enrichment of particular GO-slim terms describes ratio of representation of the terms in the identified pool with that in the genomic repertoire. Enriched GO-terms for biological process (BP), molecular function (MF), and cellular component (CC) are shown. c A competitive kinase assay was performed in the presence of increasing concentrations of AMARA peptide, a CDPK1 substrate (Supplementary Fig. 9d), using immunoprecipitated lysates (anti-HA antibody) prepared from L. infantum expressing HA-CDPK1 alone or in combination with HA-L23a. The left panel shows the phosphotransfer reaction (top panel) and the amount of immunoprecipitated HA-CDPK1 (middle panel) and HA-L23a (bottom panel) involved in each reaction, as observed by western blotting (WB) of the immunoprecipitated complex using mouse anti-HA (α-HA) IgG. The normalized phosphotransfer signal intensities for the different AMARA peptide concentrations are shown on the right panel. d Active translation was monitored by S³⁵-Methionine incorporation assay in log-phase promastigotes from the strains mentioned. The fold change of S³⁵-Methionine incorporation in TCA precipitated protein fractions were compared with wild-type cells. Superscript H and +refer to the HYG-inactivated and wild-type alleles, respectively. Data are mean ± SEM for n = 5 biologically independent experiments. Statistical analyses were performed using paired two-tailed t-tests. **P < 0.01; *P < 0.05. e Polysome profile of wild-type, CDPK1^+/H and CDPK^1+/H + CDPK1 parasites in exponential growth phase. The polysome/monosome ratio was determined by measuring the area of the polysome peaks and of the 80S monosome peak using the ImageJ software (developed and maintained by the National Institutes of Health, Bethesda, MD). Shown are representative profiles from two independent experiments. Source data are provided as a Source Data file.