The T. brucei genome encodes three proteins containing SPFH domains, TbPhb1, TbPhb2 and a putative stomatin-like protein (Tb927.5.520). Blast searches with the putative T. brucei stomatin-like protein against the human proteome revealed the most significant alignment with SLP-2/STOML2. Pairwise sequence alignment of the deduced full-length T. brucei stomatin-like protein (TbSlp2) with human SLP-2 revealed an overall sequence identity of 30% (41% sequence similarity). The stomatin domain alone revealed a 57% sequence identity and a 74% sequence similarity with human SLP-2. The deduced TbSlp2 protein has a calculated molecular mass of 56 kDa, contains a conserved SPFH domain and a C-terminal domain, but lacks transmembrane domains or membrane hairpins (Lapatsina et al., 2012) (Figure 1A).

To study TbSlp2, we generated T. brucei procyclic forms expressing C-terminally HA-tagged TbSlp2 from their genomic locus (henceforth called in situ tagged TbSlp2-HA) and analyzed total parasite protein by SDS-PAGE and immunoblotting. The results show that TbSlp2-HA migrated as a single ∼55 kDa band (Figure 1B). Subsequently, the subcellular localization of TbSlp2-HA was studied using immunofluorescence microscopy. Co-staining of parasites expressing TbSlp2-HA using antibodies against HA and the mitochondrial ADP/ATP carrier (TbAAC) revealed that TbSlp2-HA localizes to the mitochondrion (Figure 1C).

Members of the SPFH family are membrane-associated proteins that form ring-like complexes involved in compartmentalization of the inner mitochondrial membrane (Langhorst et al., 2005; Browman et al., 2007). TbSlp2, however, has no predicted transmembrane domain or predicted lipid modification that would indicate membrane association. To study if TbSlp2-HA is membrane-associated, we used carbonate extraction at different pH to test its affinity to mitochondrial membranes. The results show that TbSlp2-HA is firmly membrane-associated at pH 10.5, but soluble at pH 12.5 (Figure 1D). At pH 11.5, TbSlp2-HA is present in both fractions. A similar behavior at pH 11.5 and pH 12.5 was also observed for the mitochondrial matrix chaperone TbHsp60 (Figure 1D). In contrast, the two membrane-integral proteins TbPhb1 and TbAAC are completely membrane-associated at pH 10.5 and 11.5 and partly membrane-associated even at pH 12.5 (Figure 1D).

We next studied a possible binding affinity of TbSlp2 towards different glycerophospholipids using a lipid overlay experiment. TbSlp2 was either purified by immunoprecipitation from T. brucei expressing in situ tagged TbSlp2-HA or by affinity chromatography from E. coli expressing recombinant glutathione-S-transferase (GST)-conjugated TbSlp2 (GST-TbSlp2). Incubation of glycerophospholipid-coated membranes with purified TbSlp2-HA or GST-TbSlp2 and subsequent detection with anti-HA or anti-GST antibodies revealed that TbSlp2 interacts strongly with PA, and weakly with phosphoinositides, while no interactions with CL or other phospholipids were observed (Figure 1E).

To study the role of TbSlp2 in T. brucei procyclic forms, we used stably Cas9-expressing SmOx P9 parasites (Beneke et al., 2017) to knock-out gene Tb927.5.520 by sgDNA-targeted ssDNA nicking near the 5′- and 3′-ends of the open reading frame (ORF) (Supplementary Figure S1A). Providing a repair template containing a G418 resistance cassette flanked by 30 nucleotide homology regions matching the untranslated regions adjacent to the open reading frame, we were able to generate TbSlp2 knock-out (TbSlp2−/−) parasites using a single resistance cassette. Analyzing individual G418-resistant clones by PCR with different combinations of gene-specific primers, we identified clones that integrated the resistance gene into one allele (TbSlp2^+/−) or two alleles simultaneously (TbSlp−/−) (Supplementary Figure S1B). When we analyzed cell proliferation, we found a slightly reduced growth rate of TbSlp2−/− parasites (doubling time 12.2 ± 1.2 h) as compared to the wild-type cells (doubling time 11.5 ± 0.8 h). These results show that TbSlp2 is dispensable for survival in culture under ideal growth conditions.

In SLP2-deficient mouse T cells, the mitochondrial membrane potential was shown to be decreased and formation of respiratory chain supercomplexes was affected (Mitsopoulos et al., 2015). To study possible effects of TbSlp2 depletion on mitochondrial function, we examined the mitochondrial membrane potential ΔΨm in T. brucei wild-type and TbSlp2−/− parasites grown in high- or low-glucose medium. In low-glucose medium, T. brucei procyclic forms depend primarily on mitochondrial amino acid metabolism for ATP production as compared to glycolysis in high-glucose conditions (Lamour et al., 2005). Our results show that ΔΨm-dependent uptake of tetramethylrhodamine ethylester (TMRE) in vivo was not affected by depletion of TbSlp2, irrespective of the culture medium used (Supplementary Figure S1C). In line with these results, we detected no differences between wild-type and TbSlp2−/− parasites in the size and stability of the respiratory complexes IV (detected using anti-Cox4 antibody) and III (detected using anti-Cyc c1 antibody), and the migration of TbAAC after native gel electrophoresis (native PAGE) (Supplementary Figure S1D). As additional readout for mitochondrial health, we measured oxygen consumption rates using Seahorse flux analyser. We found that basal respiration, maximal respiration, spare respiratory capacity and non-mitochondrial respiration were not significantly different between wild-type and TbSlp2−/− parasites (Supplementary Figure S1E). Finally, using immunofluorescence microscopy we observed similar reticulate mitochondrial staining patterns of TbAAC in wild-type and TbSlp2−/− parasites (Supplementary Figure S1F). Together, these results show that under standard culture conditions, TbSlp2 is not required to maintain mitochondrial integrity and function in T. brucei procyclic forms.

To study possible interactions between TbSlp2 and proteins involved in CL biosynthesis and binding, we co-expressed in situ tagged TbSlp2 with in situ tagged TbPgs or TbPhb1 in T. brucei procyclic forms and performed reciprocal immunoprecipitation experiments. The results show that TbSlp2 co-precipitated with both TbPgs-HA (Figure 2A) and TbPhb1-Myc (Figure 2B). The reciprocal experiments show that TbSlp2 was able to co-precipitate TbPgs-HA (Figure 2A), while it failed to co-precipitate TbPhb1-Myc (Figure 2B). Since TbPgs (Serricchio and Bütikofer, 2013) and TbPhb1 (Tyc et al., 2010) have been shown to form mitochondrial high molecular mass complexes, we wondered if TbSlp2 plays a role in the formation and/or stabilization of these complexes. We therefore in situ tagged TbPgs and TbPhb1 in wild-type and TbSlp2−/− parasites and analyzed mitochondrial protein complexes by native PAGE and immunoblotting. Our results show that in wild type cells TbPgs-HA migrated as a doublet at ∼300 kDa, whereas TbPhb1-Myc was detected as a broad band migrating at ∼1 MDa (Figure 2C). The sizes and intensities of these complexes were unchanged in TbSlp2−/− parasites (Figure 2C), indicating that TbSlp2 is not involved in their assembly or stability. In addition, using metabolic labeling experiments with [³H]-glycerol and analysis of [³H]-labeled lipid extracts by thin-layer chromatography and radioisotope scanning, we found no differences in de novo CL or phosphatidylglycerol formation between wild-type and TbSlp2−/− parasites (Supplementary Figure S2), indicating that TbSlp2 is not involved in CL biosynthesis.

To identify additional interaction partners of TbSlp2, we used an unbiased proteomic approach by immunoprecipitating in situ tagged TbSlp2-HA and analyzing co-precipitated proteins by mass spectrometry. Two different extracts were prepared for analysis: 1) total protein from parasites lysed with Triton X-100 and NP-40, and 2) crude mitochondrial membranes solubilized with n-dodecyl-maltoside (DDM). Our analyses revealed three proteins that were precipitated in TbSlp2-HA expressing parasites but absent in control untagged cells, independent of sample preparation: 1) the bait TbSlp2, 2) a protein annotated as putative mitochondrial ATP-dependent zinc metallopeptidase (Tb927.10.7620) and 3) mitochondria-localized TbHsp60 (Tb927.10.6510). Sequence comparisons indicated that Tb927.10.7620 is a member of the FtsH protease family with a high degree of homology to human YME1L (30% sequence identity, 42% similarity). The deduced protein contains one, or possibly two, transmembrane domains (Figure 2D) that anchor it to the mitochondrial inner membrane, facing the mitochondrial matrix (Kovalinka et al., 2020) (see below); we have re-named this protein TbYme1. To confirm the interaction between TbSlp2 and TbYme1, we generated double in situ tagged T. brucei procyclic forms and performed reciprocal co-immunoprecipitation experiments. The results show that TbSlp2-HA co-precipitated TbYme1-Myc, and vice versa (Figure 2E), indicating their tight interaction. To prevent rapid degradation of TbYme1-Myc during the immunoprecipitation experiments, the metalloprotease inhibitor 1,10-phenanthroline was added to the lysis buffer. Immunofluorescence microscopy of in situ tagged TbYme1-Myc and co-staining with TbAAC confirmed that TbYme1 is a mitochondrial protein (Figure 2F).

To study the role of TbYme1 in T. brucei, we used CRISPR/Cas9 to replace the TbYme1 ORF sequentially with two resistance cassettes (Supplementary Figure S3A). PCR was performed to confirm the absence of the TbYme1 ORF from clones obtained after first-allele and second-allele replacement (Supplementary Figure S3B). TbYme1−/− parasites were viable in culture but grew slower than parental parasites, with cell doubling times of 14.3 ± 0.6 h as compared to 9.8 ± 0.2 h. Analysis of respiratory chain complexes by native PAGE and oxygen consumption by Seahorse flux analysis revealed no significant differences between TbYme1−/− and parental parasites (Supplementary Figure S3C–E). Loss of TbYme1 had no effect on mitochondrial morphology, as immunofluorescence staining of TbAAC in wild-type and TbYme1−/− parasites appeared similar (Figure 3A).

It has been shown that human SLP-2 is part of a mitochondrial complex migrating at ∼2 MDa (Wai et al., 2016). To test if TbSlp2 migrates as a complex in T. brucei, we separated DDM-solubilized mitochondrial membranes by native PAGE. Detection of in situ tagged TbSlp2-HA by immunoblotting revealed a large complex of ∼1.5 MDa (Figure 3B). A complex of similar size was detected when analyzing the migration of TbYme1-Myc (Figure 3B), providing additional evidence that the two proteins interact with each other (see also Figure 2E). We then studied the formation and/or stability of the TbSlp2-TbYme1 complex in parasites lacking either protein. For this we expressed TbSlp2-HA in TbYme1−/− parasites and, conversely, TbYme1-Myc in TbSlp2−/− parasites. Analysis by native PAGE and immunoblotting revealed that the ∼1.5 MDa complex containing TbSlp2-HA in parental cells was shifted downwards to ∼1.1 MDa in parasites lacking TbYme1 (Figure 3C). Unexpectedly, analysis of TbYme1-Myc revealed that the ∼1.5 MDa complex containing TbYme1-Myc in parental cells was absent in TbSlp2−/− parasites (Figure 3C). Instead TbYme1-Myc migrated as a broad band in the molecular mass range 250 kDa - 1 MDa (Figure 3C).

To examine a possible role of TbYme1 on TbSlp2 expression, we then compared TbSlp2 protein levels between wild-type and TbYme1−/− parasites. Analysis by SDS-PAGE followed by protein quantification showed an increase in TbSlp2-HA levels by ∼50% in TbYme1−/− parasites as compared to wild-type cells (Figure 3D). In control cells expressing TbYme1-Myc in the TbYme1−/− background, TbSlp2-HA levels were reduced to ∼80% of wild-type levels, or to 55% of the levels in parental TbYme1−/− parasites (Figure 3D). We next expressed a mutant form of TbYme1, TbYme1-Myc E286Q, containing a point mutation in the predicted catalytic site rendering it catalytically inactive (Hartmann et al., 2016; Shi et al., 2016). TbSlp2-HA levels again increased to 135% compared to wild-type cells, or to 90% compared to parental TbYme1−/− parasites (Figure 3D), however, it should be noted that expression levels of TbYme1-E286Q were consistently lower when compared to wild-type TbYme1-Myc. Together, these results indicate that TbSlp2 expression is regulated by the presence of TbYme1.

To study if TbYme1 and TbSlp2 affect each other`s stability, we followed the turnover of in situ tagged TbSlp2-HA expressed in parental or TbYme1−/− parasites after cycloheximide treatment. In parental cells, TbSlp2-HA had a turnover rate t_1/2 of ∼15 h, while in TbYme1−/− cells, TbSlp2-HA was stable for at least 24 h (Figures 4A,B). These results are in line with our findings (see Figure 3D) that TbSlp2-HA protein levels are approximately two-fold up-regulated in TbYme1−/− parasites and the turnover of TbSlp2-HA is attenuated in TbYme1−/− cells. Next, we analyzed the stability of in situ tagged TbYme1-Myc in wild-type and TbSlp2−/− parasites and found that the level of TbYme1-Myc in TbSlp2−/− parasites was 2.5-fold higher than in wild-type cells (Figure 4C). In addition, our results show that the turnover of TbYme1-Myc in TbSlp2−/− parasites (t_1/2 ≈ 6.3 h) was significantly decreased compared to wild-type cells (t_1/2 ≈ 1.8 h) (Figure 4D).

Since TbYme1 is a putative metalloprotease and is unstable during parasite lysis (as mentioned above), we studied if it may undergo autocatalytic degradation. We expressed wild-type and the catalytically inactive TbYme1-E286Q variant in TbYme1−/− parasites and quantified protein levels after cycloheximide treatment. We noted a small downward size-shift in TbYme1-E286Q compared to the wild-type protein, possibly due to the change in acidity (Figure 4E). The results show that wild-type TbYme1-Myc was degraded quickly, while the mutant TbYme1-E286Q remained stable (Figures 4E,F), indicating that TbYme1 undergoes autocatalytic processing. In addition, we found that the turnover of TbSlp2-HA was completely blocked in TbYme1−/− parasites expressing the catalytically inactive TbYme1-E286Q variant (Figures 4G,H), indicating that the turnover/degradation of TbSlp2-HA is regulated by TbYme1-Myc.

Mitochondrial proteases are involved in many processes including the unfolded protein response (Pickles et al., 2018; Vogtle, 2021). To investigate a possible role for TbSlp2 and TbYme1 during stress, we determined parasite proliferation at elevated culture temperature. At the standard culture temperature of 27°C, growth of TbSlp2−/− and TbYme1−/− was only slightly reduced (Figure 5A; see above). In contrast, at 37°C only wild-type parasites were able to proliferate, while the density of TbSlp2−/− parasites remained constant and that of TbYme1−/− parasites dropped substantially (Figure 5B). Examination by light microscopy revealed that TbSlp2−/− parasites were mobile and morphologically unaffected, while TbYme1−/− parasites were immobile and appeared dead after 24 h of culture at 37°C. To further substantiate the role of TbYme1 in heat tolerance, we compared growth at 37°C of TbYme1−/− parasites complemented with wild-type or the catalytically inactive TbYme1-Myc variant. The results show that expression of wild-type TbYme1 in TbYme1−/− parasites restored heat tolerance, while the inactive TbYme1-E286Q variant was unable to support growth at 37°C (Figure 5C).

To assess mitochondrial integrity at elevated temperature, parasites were cultured at 27°C and after a heat-pulse (4 h at 37°C) stained with antibodies against the mitochondrial marker proteins TbAAC and TbHsp70. Examination by fluorescence microscopy revealed that the mitochondria of wild-type parasites were not visibly affected by heat stress (<0.5% of mitochondria fragmented at 37°C) (Figure 5D). In contrast, the majority of the mitochondria of TbYme1−/− parasites appeared fragmented (>76%), while the mitochondria of TbSlp2−/− parasites were still reticulated (<4% of mitochondria fragmented) (Figure 5D).

Finally, since heat stress is known to induce oxidative damage (Slimen et al., 2014), we studied if the absence TbYme1 may increase oxidative stress. Quantification of the amount of reactive oxygen species (ROS; hydroxyl, peroxyl and other reactive oxygen species) using dichlorodihydrofluorescein diacetate (Kalyanaraman et al., 2012) revealed that although ROS levels were increased after a heat-shock for 2 h at 37°C (Supplementary Figure S4), as expected, no differences between wild-type and TbYme1−/− parasites were observed at normal or elevated temperatures (Supplementary Figure S4).

To identify additional potential substrates of TbYme1 we immunoprecipitated TbYme1-Myc expressed in TbYme1−/− parasites and identified co-purified proteins by mass spectrometry. In a first attempt, using wild-type TbYme1-Myc as bait, TbSlp2 was the only protein that was specifically precipitated (Figure 6A), confirming the previous immunoprecipitation using TbYme1-Myc as bait and the results showing that TbSlp2 and TbYme1 interact with each other (see Figure 2E). In a second attempt, we took advantage of a previous observation showing that proteolytically inactive human Yme1-E286Q was able to interact with its substrates (Stiburek et al., 2012). We therefore used TbYme1-E286Q expressed in TbYme1−/− parasites as bait to co-precipitate and identify potential substrates (Figure 6B). The results show that the catalytically inactive TbYme1-E286Q variant pulled down a large number of proteins, including TbSlp2, TbPhb1 and TbPhb2, TbMSP-B, multiple known mitochondrial proteins (elongation factor Tu; TbCOX4, TbCOX10, TbMCP13, among others) and several hypothetical, i.e. uncharacterized, proteins (Supplementary Table S1). Interestingly, when we repeated the experiment after a 2 h heat pulse at 37°C, two proteins, TbPOMP24 (present in the outer mitochondrial membrane proteome 24) and succinate dehydrogenase flavoprotein subunit 5 (TbShd5), were specifically precipitated compared to the pull-down done at 27°C (Figure 6C). We then in situ tagged a subset of the proteins pulled down using TbYme1-E286Q to compare their expression levels between wild-type and TbYme1−/− parasites. The results show that, in addition to TbSlp2 (see also Figure 3D), TbSdh5 was clearly more abundant in TbYme1−/− parasites compared to wild-type cells (Figure 6D). Quantification of protein levels revealed that TbSdh5 was increased to 184 ± 16% in TbYme1 −/− parasites compared to wild-type cells. Analysis of protein turnover using cycloheximide treatment demonstrated that TbSdh5-HA, similar to TbSlp2-HA (see Figure 4A), was more stable in TbYme1−/− parasites compared to wild-type cells (Figures 6E,F). In contrast, TbPOMP24-HA showed a faster turnover in the absence of TbYme1 (Figures 6E,F). Together these results indicate that TbYme1, or the TbYme1/TbSlp2 complex, may have different effects on mitochondrial proteins.

T. brucei procyclic form SmOx P9 parasites (Poon et al., 2012) expressing hSpCas9 (Beneke et al., 2017) were cultured at 27°C in SDM79 containing 10% (v/v) heat-inactivated fetal bovine serum, 5 μg/ml blasticidin (InvivoGen), 2 μg/ml puromycin. Knock-out parasites were cultured in presence of additional 15 μg/ml G418 (geneticin; Santa Cruz Biotechnology, Heidelberg, Germany) and 25 μg/ml hygromycin (InvivoGen, Nunningen, Switzerland). TbYme1 addback cultures were selected and cultured in the presence of 150 μg/ml nourseothricin (Jena Bioscience, Jena, Germany).

Genome editing was achieved following the genome editing toolkit for kinetoplastids (Beneke et al., 2017). Donor resistance cassettes for knock-outs were amplified by PCR from pPOTv6 plasmids (Dean et al., 2015) with hygromycin and geneticin resistance cassettes using primers as described (Beneke et al., 2017). 5 and 3′ short-guide DNAs comprising a T7 promoter sequence, a Cas9 binding site, and a stretch of 20 nucleotide long 5 and 3′ target sequences were PCR-assembled as described (Beneke et al., 2017). Donor cassettes for tagging were amplified from pMOTag vectors (Oberholzer et al., 2006). PCR reactions were performed with Expand High Fidelity PCR System (Roche, Basel, Switzerland) with primers generated by the online tool www.leishgedit.net. PCR reactions were pooled and transfected using the 4D-Nucleofector system (Lonza, Basel, Switzerland). After 4 h recovery, selection antibiotics were added. Control PCRs to verify gene knockout were done with primers TbSlp2_UTR_fwd: GTT​GTT​GGT​GCT​ATT​GTT​GCT​ATG; TbSlp2_UTR_rev: CGCATCCGCCTATGCAA; TbSlp2_ORF_fwd: GTA​GTT​TGG​TGC​ACT​CGT​CTC; TbSlp2_ORF_rev: CCA​ACG​AAG​TGG​CAC​TAA​CC; TbYme1_UTR_fwd: GAG​GGA​TAA​TAC​GAA​GAG​GAG​AAC; TbYme1_UTR_rev: CGT​GTG​CAT​GCT​GCT​TAG; TbYme1_ORF_fwd: CGC​TGT​CCC​ATC​ATA​CCA​G; TbYme1_ORF_rev: CAA​CAA​TAA​CCA​CAG​TCG​GG.

Crude mitochondrial preparations were obtained by digitonin extraction as described (Charriere et al., 2006). Briefly, trypanosomes were washed in TBS (10 mM Tris·HCl pH 7.5, 144 mM NaCl) and suspended in 0.5 ml SoTE (20 mM Tris·HCl, pH 7.5, 0.6 M sorbitol, 0.2 mM EDTA) followed by the addition of 0.5 ml SoTE containing 0.05% (w/v) digitonin. After 5 min on ice, crude membranes were collected by centrifugation (6000 x g, 5 min, 4°C).

Crude membrane preparations from 2 × 10⁸ parasites were re-suspended in water containing protease inhibitors (Roche), separated into three individual tubes, and equal volumes of 0.2 M sodium carbonate at pH 10.5, pH 11.5 or pH 12.5 were added. After 1 h of incubation on ice, the membrane fractions were collected by ultra-centrifugation (100’000 x g, 30 min, 4°C), dissolved in 100 µl SDS sample buffer and heated for 5 min at 65°C. Alkaline-solubilized protein fractions were precipitated (Wessel and Flugge, 1984), dissolved in 100 µl SDS sample buffer and heated for 5 min at 65°C.

Immunoprecipitations were performed as described in (Serricchio et al., 2018) with minor modifications. Log-phase parasites (approx. 10⁸) were collected, washed in TBS and lysed in 900 µl lysis buffer [10 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1 mM EDTA, 1% (v/v) Triton X-100, 0.5% (v/v) NP-40, protease inhibitor cocktail (Roche)]. Alternatively, some experiments were done with lysis buffer containing 2% (w/v) DDM. After clearing the lysate for 10 min at 17′000 x g, an input sample was removed and the lysate distributed into three tubes, to which 1 µg of the following antibodies were added: mouse anti-c-Myc antibody (9E30, Santa Cruz Biotechnology, Heidelberg, Germany), or mouse anti-HA antibody (HA.11, 16B12, Enzo Life Sciences, Lausen, Switzerland). After 2 h rotation at 4°C, Protein G Dynabeads (Invitrogen, Reinach, Switzerland) were added for 16 h. Beads were washed in cold lysis buffer, proteins eluted with SDS sample buffer and heated at 62°C for 10 min.

To isolate TbSlp2-HA from T. brucei, parasites (3.5 × 10⁸) were lysed in 140 µl RIPA buffer [10 mM Tris-HCl, pH 8.0, 140 mM NaCl, 1 mM EDTA, 0.5 mM EGTA, 1% (v/v) Triton X-100, 0.1% (w/v) sodium deoxycholate, 0.1% (w/v) SDS, protease inhibitor cocktail] for 5 min at 65°C, then diluted with 1,260 µl IP buffer [10 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1 mM EDTA, 1% (v/v) Triton X-100, 0.5% (v/v) NP-40, protease inhibitor cocktail (Roche)]. After clearing the lysate at 17’000 x g for 30 min at 4°C, anti-HA agarose (Roche) was added for 16 h at 4°C. After extensive washing with IP buffer, proteins were eluted with 0.1 M glycine (pH 2.5) and dialyzed against PBS (137 mM NaCl, 2.7 mM KCl, 10 mM Na₂HPO₄, 1.8 mM KH₂PO₄).

To purify GST-TbSlp2, TbSlp2 was cloned into plasmid pGEX-6P-1 and transformed into Escherichia coli BL21. Exponentially growing cultures were induced with 0.1 mM IPTG for 3 h at 25°C. Pelleted bacteria were suspended in 40 ml lysis buffer (PBS, pH 7.4, 0.5 mg/ml lysozyme, 1 mM DTT, 2 mM MgCl₂) and kept on ice for 30 min. After three freeze-thaw cycles, lysates were centrifuged at 12,000 x g for 30 min at 4°C. Glutathione-sepharose 4B (1 ml; GE Healthcare) was added to the cleared lysates and incubated under rotation at 4°C for 60 min. After extensive washing with PBS, beads were equilibrated with equilibration buffer (50 mM Tris-HCl, 2 mM MgCl₂, pH 8.0). The GST fusion proteins were eluted with freshly prepared equilibration buffer supplemented with 20 mM reduced glutathione. Eluted fractions were analyzed by SDS-PAGE and immunoblotting, pooled and dialyzed against PBS. Lipid overlay experiments were done using Membrane Lipid Strips (Echelon Biosciences, Salt Lake City, United States) according to the manufacturer’s instructions.

Parasites were incubated with [³H]-glycerol and phospholipids were extracted and analyzed exactly as described before (Serricchio and Bütikofer, 2012).

Parasites were washed in PBS, counted and diluted in PBS to 5 × 10⁶/ml. Aliquots of 100 µl were added per well into 96-well plates. H₂O₂ (1 ml) was added to control wells for 5 min, followed by addition of 20 µM dichlorodihydrofluorescein diacetate to all wells. After incubation for 20 min at 27°C, the samples were analyzed using an excitation wavelength of 485 nm and emission wavelength of 530 nm using a Spark Microplate reader (TECAN, Männedorf, Switzerland).

The TbYme1 ORF was amplified from gDNA using primers SalI_fwd ATA​AGT​CGA​CAT​GCA​CCG​GCG​CTG​TC and Xho_rev AGA​CTC​GAG​CGT​TAT​GGA​AAC​GGG​GCG​TTG (restriction sites underlined), digested and ligated into pGS plasmids (Gottier et al., 2017) containing a C-terminal Myc-tag and a SAT1 resistance cassette under a constitutive procyclin promoter. For mutagenesis, inverse PCR was performed using primers fwd CCA​ATG​CGT​CGA​TTT​GGT​CTA​TAA​AAA​TAA​GCG and rev CGC​TTA​TTT​TTA​TAG​ACC​AAA​TCG​ACG​CAT​TGG.

Cells were diluted to 0.6 × 10⁷/ml in 5 ml medium and treated with 100 μg/ml cycloheximide. Aliquots (0.9 ml) were removed at each time point, washed in TBS and lysed in 50 µl lysis buffer [10 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1 mM EDTA, 1% (v/v) Triton X-100, 0.5% (v/v) NP-40, protease inhibitor cocktail]. After addition of SDS sample buffer, extracts were heated at 62°C for 5 min.

Native PAGE was performed with DDM-solubilized crude mitochondrial membrane fractions and separated on 3–12% native PAGE gradient gels (Invitrogen Reinach, Switzerland) at 4°C (Wittig et al., 2006). Proteins were transferred onto nitrocellulose membranes (Thermo Scientific) using a semi-dry protein blotting system (BioRad, Cressier, Switzerland). After blocking in TBS containing 5% (w/v) milk powder, membranes were exposed to primary antibodies rabbit anti-Cox4, rabbit anti-Cyt c1, rabbit anti-ATP synthase subunit β, or mouse anti-Hsp60 antibody (kindly provided by André Schneider, University of Bern, Bern, Switzerland), diluted 1:1,000 in TBS containing 5% (w/v) milk powder. Horseradish peroxidase-conjugated (HRP) anti-mouse, anti-rabbit (Dako, Glostrup, Denmark), anti-HA (HA.11, 16B12, Enzo Life Sciences, Lausen, Switzerland) or anti-cMyc (Santa Cruz Biotechnology) were used at dilutions of 1:5,000, and detected using an enhanced chemiluminescence detection kit (Thermo Scientific). Protein sizes were determined using NativeMark™ Unstained Protein Standard (Invitrogen, Reinach, Switzerland). Bands on blots were quantified using the gel analyzer function of Fiji (Schindelin et al., 2012). Signal intensities of multiple blots and different exposure times were analyzed with the multiple t-test function in GraphPad Prism software (Version 6.0 g).

Parasites at a cell density of 10⁶ in mid-log growth phase were washed and suspended in PBS, allowed to adhere on a microscope slide (Thermo Scientific) for 15 min and fixed in PBS containing 4% (w/v) paraformaldehyde for 10 min. After washing and permeabilization for 5 min with PBS containing 0.1% (w/v) Triton X-100, cells were blocked for 30 min with 2% (w/v) bovine serum albumin in PBS. The following antibodies were diluted 1:250 in blocking solution and added to cells for 30 min: mouse anti-c-Myc IgG (c-Myc 9E30; Santa Cruz Biotechnology, Heidelberg Germany), mouse anti-HA (HA.11, 16B12; Enzo Life Sciences, Lausen, Switzerland), rabbit anti-TbACC (provided by Alena Ziková, Biology Centre of the Czech Academy of Sciences) (Pena-Diaz et al., 2012), or mouse anti-Hsp70 (provided by André Schneider, University of Bern, Bern, Switzerland). After washing, cells were incubated for 30 min with goat anti-mouse Alexa Fluor 488 or goat anti-rabbit Alexa Fluor 594 (1:500 in blocking solution), washed again and air-dried before mounting with Vectashield (Vector Laboratories, Burlingame, CA) containing 1.5 μg/ml 4’,6-diamidino-2-phenylindole (DAPI). The images were acquired using a Leica DMI6000 B microscope with ×60 oil objective.

Parasites were counted, washed twice in fetal bovine serum-free SDM79 and diluted to a density of 5.5 × 10⁶/ml. Aliquots (180 μl; 10⁶ cells) were added to the wells of an 8-well microplate pre-treated with Cell-Tak (Corning), as described elsewhere (Shah-Simpson et al., 2016). The plate was centrifuged for 2 min at 300 x g and immediately analyzed using a Seahorse XFp Flux analyzer (Agilent Technologies). Parasites were still adhered to the bottom and motile after the experiments. Data were analyzed using Wave software (version 2.6.1, Agilent Technologies).