Suramin exposure alters cellular metabolism and mitochondrial energy production in African trypanosomes

Abstract

Introduced about a century ago, suramin remains a frontline drug for the management of early-stage East African trypanosomiasis (sleeping sickness). Cellular entry into the causative agent, the protozoan parasite Trypanosoma brucei, occurs through receptor-mediated endocytosis involving the parasite's invariant surface glycoprotein 75 (ISG75), followed by transport into the cytosol via a lysosomal transporter. The molecular basis of the trypanocidal activity of suramin remains unclear, but some evidence suggests broad, but specific, impacts on trypanosome metabolism (i.e. polypharmacology). Here we observed that suramin is rapidly accumulated in trypanosome cells proportionally to ISG75 abundance. Although we found little evidence that suramin disrupts glycolytic or glycosomal pathways, we noted increased mitochondrial ATP production, but a net decrease in cellular ATP levels. Metabolomics highlighted additional impacts on mitochondrial metabolism, including partial Krebs' cycle activation and significant accumulation of pyruvate, corroborated by increased expression of mitochondrial enzymes and transporters. Significantly, the vast majority of suramin-induced proteins were normally more abundant in the insect forms compared with the blood stage of the parasite, including several proteins associated with differentiation. We conclude that suramin has multiple and complex effects on trypanosomes, but unexpectedly partially activates mitochondrial ATP-generating activity. We propose that despite apparent compensatory mechanisms in drug-challenged cells, the suramin-induced collapse of cellular ATP ultimately leads to trypanosome cell death.

Keywords: Trypanosoma brucei; differentiation; drug action; drug mechanisms; energy homeostasis; glycosomes; metabolomics; parasite metabolism; polypharmacology; proteomics; sleeping sickness; suramin; trypanosome.

Figure 1.

Suramin is taken up rapidly and accumulates to high intracellular concentrations. Shown are scintillation counts from 1 × 10⁷ cells incubated in 200 nm [³H]suramin over 15 min and surface binding cold-chased. Bars, mean of three independent experiments for WT cells (wt), ISG75-overexpressing cells (ISG75O.E.), and ISG75- and ISG65-silenced cells (ISG75RNAi and ISG65RNAi), respectively, with the S.E. (error bars) and significance intervals from Student's unpaired t test (*, p < 0.05; ***, p < 0.001) indicated. The intracellular suramin concentration was calculated using a volume of 29 μm³ for T. brucei BSF (33). Inset, suramin uptake in the presence of 250 nm suramin monitored over 1 h (red line). Suramin is taken up in two phases, a rapid initial phase over 20 min with 19.0 pmol/min (linear regression shown as a black dashed line; r² = 0.97; p = 0.002, nonzero slope) and a slower phase (20–60 min) with 4.5 pmol/min (linear regression shown as a dark blue dashed line; r² = 0.99; p = 0.048, nonzero slope). Saturability of uptake was determined by incubation of cells with 250 nm [³H]suramin in the presence of 100 μm unlabeled suramin (p = 0.48, zero slope) (green line).

Figure 2.

Suramin has no significant impact on abundance or location of selected enzymes of the glycosomal pathway. A and B, steady-state levels of PFK, PGK, and PYK were analyzed by immunoblotting at 0 (CON), 3, or 4 days post-suramin treatment (1× EC₅₀). β-tubulin was used as loading control. Graphs represent the mean of three independent experiments, with the S.E. indicated. C, cells were treated with 1× EC₅₀ suramin for 0 or 4 days, and the subcellular localizations of PFK, PGK, and PYK were detected using antibodies specific to each protein. Cells were stained with 4′,6-diamidino-2-phenylindole to visualize nuclear and mitochondrial DNA. Scale bar, 2 μm.

Figure 3.

Suramin has minimal impact on glycosome cellular frequency. Cells were cultured in the presence of suramin for either 0 h (top) or 72 h (bottom). The respective distribution of glycosome volumes and for cells treated for 72 h with suramin is shown as a bar graph. Upon treatment with suramin, median individual glycosome volume increased from 0.0113 to 0.0127 μm³ (p < 0.037). Example segmentations of whole T. brucei cell are shown for both cases. Scale bar, 2 μm.

Figure 4.

Suramin leads to an enlarged flagellar pocket. Cells were grown in the presence or absence of suramin (35 nm) over 4 days and analyzed by thin-section transmission EM. Images show ultrastructure in untreated cells or suramin-treated cells after 3 or 4 days. Enlarged flagellar pockets are marked by an asterisk. The day 3 micrograph is representative of 7 of 15 thin sections across the flagellar pocket, whereas day 4 is representative of 14 of 22 flagellar pocket cross-sections. In untreated cells, a large pocket is observed in 2 of 63.

Figure 5.

Suramin impacts both cellular ATP production and mitochondrial membrane potential. Cells were exposed to 35 nm (1× EC₅₀) and 105 nm (3× EC₅₀) suramin and analyzed at the indicated time points. A, the intracellular ATP concentration was measured by a bioluminescent assay with oligomycin (2 μg/ml) as positive control. Values are presented as percentage versus untreated control. B, Ψm changes were determined using the accumulation of the fluorescent indicator dye tetramethylrhodamine ethyl ester. Valinomycin (100 nm) was used as a depolarization control. For both panels, bars show the average and S.E. of three independent determinations, and statistical differences were determined using Student's unpaired two-tailed t test: *, p < 0.05; **, p < 0.01; ***, p < 0.001 relative to untreated control.

Figure 6.

F_oF₁-ATPase activity is not affected by suramin. In situ generation of the mitochondrial membrane potential in the presence of ATP in digitonin-permeabilized cells untreated or pretreated with suramin for 8 h prior to permeabilization. OLM, oligomycin (2.5 μg/ml). SF6847 (250 nm) is an uncoupler. The displayed results represent the average activities obtained from three independent measurements.

Figure 7.

Suramin affects levels of cytosolic and mitochondrial ATP. A and B, subcellular localization of luciferase_v5 (A) and MLS_luciferase_v5 (with N-terminal mitochondrial localization signal) (B) was determined in BSF whole-cell lysates (WCL) and the corresponding organellar (ORG) and cytosolic (CYT) fraction separated by digitonin extraction. Purified fractions were analyzed by Western blotting with the following antibodies: anti-luciferase, anti-mt Hsp70 (mitochondrial marker), and anti-adenosine phosphoribosyltransferase (APRT) (cytosolic marker). The relevant sizes of the protein marker are indicated on the left. C and D, the cytosolic and mitochondrial ATP content was measured by a TECAN M200 plate reader upon the addition of luciferin to living cells. The y axis represents the percentage of the measured light units of cells treated with suramin to BSF cells, relative to those that were not treated with suramin (0-h treatment). The displayed results represent the average maximal signals obtained from three independent measurements. Error bars, S.E.; ****, p < 0.001.

Figure 8.

Metabolite changes upon suramin treatment. Heat map of selected metabolites quantified at the indicated time points in cells after suramin treatment and nontreated controls. Colors in the scale bar (right) correspond to intensities normalized to the 0-h point.

Figure 9.

Suramin impact on glycosomal and mitochondrial pathways. Enzymatic steps and transport processes are represented by black arrows with different thickness dependent on the observed abundance change of the respective protein as a result of 48-h suramin treatment (for ratio changes, see the key on the left). The respective metabolites changes upon 29 h of suramin treatment are color-coded (white circles, not determined). Dashed lines indicate transport processes, where the transport proteins are unknown.

Figure 10.

Global proteome changes upon 48 h of suramin exposure. Shown is a volcano plot of normalized SILAC ratios, averaged from triplicate experiments, plotted against the respective −log₁₀-transformed p values. Selected data points are annotated. For complete annotation, see Table S4. DH, dehydrogenase; AQP, aquaporin; AA, amino acid; AATP1, amino acid transporter 1 Tb927.8.7610; AATP10, amino acid transporter 10 Tb927.4.4820; NT4, adenosine transporter Tb927.2.6220; AAT, amino acid transporter Tb927.4.4830; JBP1, J-binding protein 1 Tb927.11.13640; PAD, protein associated with differentiation; PIP39, PTP1-interacting protein, 39 kDa.

Figure 11.

Correlation of proteome changes upon suramin treatment with proteome differences between BSF and PCF and BSF-LS and BS-SS. −log₂-transformed abundance shifts after 48 h of suramin (35 nm) exposure were plotted against corresponding −log₂-transformed abundance differences between BSF and PCF (A–C) (50) and between BSF-LS and BSF-SS (D–F) (65). Selected GO terms are indicated by color: red, mitochondrion; blue, glycosome; green, transmembrane transporter activity. Infinite changes were omitted for clarity. The BSF-SS data are derived from the pleomorphic strain EATRO1125 (65). DH, dehydrogenase; PPDK, pyruvate phosphate dikinase; PYK, cytosolic pyruvate kinase; gGAPDH, glycosomal glyceraldehyde 3-phosphate dehydrogenase; PGI, phosphoglucose isomerase; HK, hexokinase; IDH, isocitrate dehydrogenase; PGKC, phosphoglycerate kinase; PFK, phosphofructokinase; ALD, aldolase; G3PDH, mitochondrial glycerol-3-phosphate dehydrogenase; AOX, alternative oxidase; NT4, adenosine transporter Tb927.2.6220; AATP, amino acid transporter; AQP, aquaporin.

Figure 12.

Correlation of proteome changes upon suramin treatment with proteome differences between BSF and PCF and between BSF-LS and BS-SS. Shown is a bar graph of Pearson coefficients from correlations of proteome changes upon suramin treatment between BSF and PCF (gray) (50) and between BSF-LS and BSF-SS (black) for all shared, finite quantified protein groups and selected GO terms. The BSF-SS data are derived from the pleomorphic strain EATRO1125 (65). Corresponding correlation plots are shown in Fig. 11 (BSF-SS) and Fig. S5 (PCF).

Figure 13.

Schematic view of potential impact of suramin on glycolytic complexes. There are several lines of evidence that suramin is a promiscuous binder, and multiple diverse low-affinity targets have been described. It is possible that suramin binding disrupts glycolytic protein complexes even at low concentration through conformational impact (A). Due to rapid accumulation of suramin by endocytic uptake, higher concentrations could be reached, potentially resulting in complete inhibition of single (B) or multiple protein targets (C) within a complex. All of these scenarios would ultimately lead to a decrease in glycolytic flux and hence ATP production.