Instability of aquaglyceroporin (AQP) 2 contributes to drug resistance in Trypanosoma brucei

Abstract

Defining mode of action is vital for both developing new drugs and predicting potential resistance mechanisms. Sensitivity of African trypanosomes to pentamidine and melarsoprol is predominantly mediated by aquaglyceroporin 2 (TbAQP2), a channel associated with water/glycerol transport. TbAQP2 is expressed at the flagellar pocket membrane and chimerisation with TbAQP3 renders parasites resistant to both drugs. Two models for how TbAQP2 mediates pentamidine sensitivity have emerged; that TbAQP2 mediates pentamidine translocation across the plasma membrane or via binding to TbAQP2, with subsequent endocytosis and presumably transport across the endosomal/lysosomal membrane, but as trafficking and regulation of TbAQPs is uncharacterised this remains unresolved. We demonstrate that TbAQP2 is organised as a high order complex, is ubiquitylated and is transported to the lysosome. Unexpectedly, mutation of potential ubiquitin conjugation sites, i.e. cytoplasmic-oriented lysine residues, reduced folding and tetramerization efficiency and triggered ER retention. Moreover, TbAQP2/TbAQP3 chimerisation, as observed in pentamidine-resistant parasites, also leads to impaired oligomerisation, mislocalisation and increased turnover. These data suggest that TbAQP2 stability is highly sensitive to mutation and that instability contributes towards the emergence of drug resistance.

Fig 1. Schematic representation of constructs used in this study.

A) 3D structural predictions of the AQP2 harbouring three haemagglutinin tags at either terminus. Top panel; lateral and cytoplasmic face view of simulated model of T. brucei AQP2 tetramer embedded in a POPC lipid bilayer. Lipids are shown in surface and line representations in cyan. Each monomer of AQP2 is shown in cartoon representation. Bottom panel; lateral and cytoplasmic face view of T. brucei AQP2 showing key amino acids (in spheres) from NSA (cyan), NPS (orange) and IVLL (magenta) domains. B) N- and C-terminal tagged TbAQP2 variants with a tandem of three hemagglutinin (3xHA) epitopes. Positions of predicted trans-membrane domains (TMD) are indicated with numbers above solid blocks. Similarly, lysine residues that were manipulated in this study are highlighted. C) Wild type TbAQP1 (blue), TbAQP2 (grey), TbAQP3 (green), and chimeras used (40AT, AQP2^TMD4, and AQP2^TMD5). TMDs for AQP1, 2 and 3 are shown as blocks and in blue, grey and green, respectively.

Fig 2. Characterisation of tagged TbAQP2.

A) Fluorescence microscopy of T. brucei 2T1 cells expressing tetracycline-regulated N- or C-terminal tagged AQP2 (^3xHAAQP2 or AQP2^3xHA, respectively, in yellow). These proteins localise similar to ISG75 (magenta) at the flagellar pocket/endosomes. The triple aqp-null T. brucei 2T1 cells (ΔAQP) were also included as control. Scale bar 5 μm. B) Tet-regulated expression of N- or C-terminal HA-tagged AQP2. Both native-PAGE (upper panel) and SDS-PAGE (lower panel) αHA blots are shown. α–β tubulin was used as loading control. The presence of the different oligomeric species is indicated in the right-hand side of the panel. Note the presence of a high molecular weight form under SDS-PAGE in ^3xHAAQP2 but not AQP2^3xHA. The triple aqp-null T. brucei 2T1 cells (ΔAQP) were also included as control. C) EC₅₀ values for pentamidine (left panel) or salicylhydroxamic acid (SHAM; right panel) with or without 5 mM glycerol following expression of either ^3xHAAQP2 or AQP2^3xHA. For multiparametric ANOVA, we compared the average values (n = 4 independent replicates) from wild type T. brucei 2T1 cells as reference for pentamidine, or from aqp-null cells for SHAM. * p<0.01, ** p<0.001, *** p<0.0001 from four independent replicates. D) Left panel; Representative western blotting (n = 3 independent replicates) from protein turnover assay monitored by cycloheximide (CHX) treatment in T. brucei 2T1 cells expressing either ^3xHAAQP2 (upper panel) or AQP2^3xHA (lower panel). Right panel; Protein quantification from western blotting analysis in left panel for either ^3xHAAQP2 (black square) or AQP2^3xHA (grey circles). Results are the mean ± standard deviation of three independent experiments (n = 3 independent replicates). The estimated half-life (t_1/2) was calculated based on regression analysis using PRISM.

Fig 3. TbAQP2 is ubiquitylated in T. brucei.

A) Cells expressing ^3xHAAQP2 were treated with either NH₄⁺Cl (10 mM) or MG132 (25 μM) for 1h prior to harvesting. Cell lysates were resolved in a 4–12% acrylamide gel and detected with anti-HA antibody by western blotting. The intensity of anti-β tubulin was used as loading control. B) Immunoprecipitation of Δaqp or ^3xHAAQP2 cell lysates with anti-HA beads followed by anti-ubiquitin detection by western blotting. An anti-HA blot was also included to confirm protein expression upon induction with tetracycline. Anti-β tubulin was used as loading control. ‘*’ indicates the predicted migration poition of a monoubiquitylated ^3xHAAQP2. C) As in (B), but immunoprecipitation conducted using ubiquitin capture matrix and analysed by western blotting (left panel). The total (“T”), unbound (“Unb.”), wash (“W”), and elution (“E”) fractions were resolved by SDS-PAGE electrophoresis and analysed with anti-HA immunoblotting (right panel). ‘*’ is as in panel B.

Fig 4. TbAQP2 transits through the endosomal compartment and is efficiently delivered and degraded in the lysosome.

A) Cell lines expressing a tetracycline-regulated copy of ^3xHAAQP2 (Alexa Fluor 488; yellow) were co-stained with anti-TbRab5a and anti-TbRab5b (early endosomes), anti-TbRab11 (recycling endosomes), and anti-p67 (lysosome). All endosomal and lysosomal markers were labelled with secondary antibodies coupled to Alexa Fluor 568 (magenta). DAPI (cyan) was used to label the nucleus and kinetoplast. Scale bars 5 μm. A schematic depiction of the results from confocal microscopy is included in the right panel, generated with BioRender. B) Left panel; Protein turnover was monitored by cycloheximide (CHX) treatment. Cells were harvested at various times and the protein level monitored by immunoblotting. ISG75 was included as a control. Right panel; Quantification for ISG75 and ^3xHAAQP2. Results are the mean ± standard deviation of three independent experiments. C) Upper panel; As in (B), but cells were untreated or exposed to 100 nM of bafilomycin A1 (BafA1), or to 25 μM of MG132 for 1 h prior to harvesting. Cell lysates were resolved by SDS-PAGE followed by western immunoblotting using anti-HA antibody. Lower panel; Quantification from three independent experiments—dotted line represents 100% (signal at 0h). Data presented as mean ± standard deviation (n = 3 independent replicates). Statistical analysis was conducted using t test; * p<0.01 and the signal from untreated cells at 1 h as reference.

Fig 5. N-terminal lysine residues in the N-terminal cytoplasmic tail are important for protein stability, oligomerisation, and anterograde transport.

A) Left panel; Structural predictions of ^3xHAAQP2 generated with i-Tasser, indicating the three N-terminal lysine residues (magenta) mutated in AQP2^3K>R. The 3xHA tag has been omitted for simplicity. Right panel; Fluorescence microscopy of cells expressing N-terminal HA-tagged wild type AQP2 (AQP2^WT) or lysine mutant AQP2^3K>R. Both proteins are shown in yellow. DAPI (cyan) was used to label the nucleus (N) and the kinetoplast (K). Scale bars, 5 μm. Western blot of cell lysates upon induction with tetracycline are also included. B) EC₅₀ values for pentamidine (left panel) and salicylhydroxamic acid (SHAM) with or without 5 mM glycerol (right panel) following recombinant expression of either AQP2^WT or AQP2^3K>R with a tetracycline-regulated (Tet-on) copy in T. brucei 2T1 bloodstream forms. Multiparametric ANOVA calculated as for Fig 4 (N = 3 independent replicates). C) Cell lines expressing AQP2^WT or AQP2^3K>R (Alexa Fluor 488; yellow) were co-stained with anti-BiP (endoplasmic reticulum marker). All markers were labelled with secondary antibodies coupled to Alexa Fluor 568 (magenta). DAPI (cyan) was used to label the nucleus and the kinetoplast, as indicated in (A). Scale bars, 5 μm. D) Native-PAGE immunoblot of total cell lysates expressing either AQP2^WT or AQP2^3K>R. Coomassie blue staining of the same fractions was used as loading control. The triple aqp-null T. brucei 2T1 cells (ΔAQP) were also included as control. E) Left panel; Protein turnover monitored as in Fig 4 for AQP2^WT or AQP2^3K>R. Cells were either untreated or treated with 25 μM MG132 for 1 h prior to harvest. Cells were harvested at 0 hours and 2 h post-CHX treatment, and lysates analysed by immunoblotting. α–β tubulin was used as loading control. Right panel; Protein quantification representing the mean ± standard deviation (n = 3 independent replicates). Dotted line represents 100% (signal in untreated samples). Statistical analysis was conducted using the signal from untreated cells at 2 h as reference group. ** p<0.001, ns = not significant, using a t test.

Fig 6. Requirement for cytoplasmic-oriented lysine residues for AQP2 stability and trafficking.

A) Cell lines expressing a tetracycline-regulated copy of the constructs mentioned in (A) (Alexa Fluor 488; yellow) were co-stained with either αBiP (ER) or αISG75 (localises to flagellar pocket/endosome), both stained with secondary antibodies coupled to Alexa Fluor 568 (magenta). DAPI (cyan) was used to label the nucleus and the kinetoplast. Scale bars, 5 μm. B) Representative western blot (n = 3 independent replicates) of protein turnover monitored by cycloheximide (CHX) treatment followed by pulse-chase of cells expressing the constructs in (A). Cells were either untreated or treated with 25 μM MG132 for 1 h prior to harvest. Cells were harvested at 0 hours and 2 h post-CHX treatment and analysed by immunoblotting. Uninduced controls (“Un.”) were also included. C) EC₅₀ values (average ± standard deviation; n = 3 independent replicates) of pentamidine (upper panel) and salicylhydroxamic acid (SHAM) with or without 5 mM glycerol (lower panel) following recombinant expression of either AQP2^WT, AQP2^5K>R, or single arginine-to-lysine AQP2 mutants (AQP2^R19K, AQP2^R45K, and AQP2^R54K). Statistical test for significance was conducted using a pairwise t test comparison with uninduced cell lines. * p<0.01, ** p<0.001, *** p<0.0001.

Fig 7. Differential turnover rate of the repertoire of AQPs in the bloodstream form of T. brucei.

A) Cell lines expressing N-terminal HA-tagged TbAQP1, TbAQP2, TbAQP3, field-isolate chimeric AQP2/3 (40AT) (Alexa Fluor 488; yellow) co-stained with the endoplasmic reticulum marker anti-BiP (magenta). DAPI (cyan) was used to label the nucleus and the kinetoplast. Scale bars, 5 μm. B) Three different confocal planes are shown for 2T1 cells expression TbAQP1, TbAQP2, or TbAQP3. The planes are defined from “Bottom” (far from the flagellar pocket) to “Top” (close to the flagellar pocket). Note a change in DAPI intensity as the images progress through the different planes. TbAQPs are denoted in yellow, DAPI in cyan, and TbISG75 in magenta. Scale bar, 10 μm. C) Upper panel; Representative western blot (n = 3 independent replicates) of protein turnover monitored by cycloheximide (CHX) treatment followed by pulse-chase assay. Cells were either untreated or treated with 100 nM of Bafilomycin A1 (BafA1) or 25 μM of MG132 for 1 h prior to harvest. Cells were harvested at 0 hours and 2 h post-CHX treatment and analysed by immunoblotting. Lower panel; Protein quantification representing the mean ± standard deviation of three independent experiments (n = 3 independent replicates). Statistical analysis was conducted using the signal from untreated cells at 2 h post-CHX treatment as reference group. * p<0.01, ** p<0.001, ns = not significant, using a t-test.

Fig 8. Chimerisation of TbAQP2 leads to mislocalisation, reduction in glycerol transport activity and rapid turnover.

A) Cell lines expressing N-terminal HA-tagged TbAQP1, TbAQP2, TbAQP3, field-isolate chimeric AQP2/3 (40AT) or a single TMD mutant (AQP2^TMD4) (Alexa Fluor 488; yellow) co-stained with anti-ISG75 (magenta). DAPI (cyan) was used to label the nucleus and the kinetoplast. Scale bars 5 μm. Western immunoblotting analysis from lysates of cells expressing these constructs are also included. Anti-β tubulin was used as loading control. B) BN-PAGE immunoblot of total cell lysates expressing the constructs in (A). Coomassie blue staining of the same fractions was used as loading control. C) EC₅₀ values (average ± standard deviation; n = 3) for salicylhydroxamic acid (SHAM) with or without 5 mM glycerol following recombinant expression of the constructs in (A). D) Left panel; Representative western blotting (n = 3 independent replicates) analysis of protein turnover monitored by cycloheximide (CHX) treatment followed by pulse-chase of cells expressing the constructs in (A). Right panel; Protein quantification representing the mean ± standard deviation of three independent experiments. Dotted line represents 50% of protein abundance. Data presented as mean ± standard deviation (n = 3 independent replicates).