Cathepsin-L can resist lysis by human serum in Trypanosoma brucei brucei

Abstract

Closely related African trypanosomes cause lethal diseases but display distinct host ranges. Specifically, Trypanosoma brucei brucei causes nagana in livestock but fails to infect humans, while Trypanosoma brucei gambiense and Trypanosoma brucei rhodesiense cause sleeping sickness in humans. T. b. brucei fails to infect humans because it is sensitive to innate immune complexes found in normal human serum known as trypanolytic factor (TLF) 1 and 2; the lytic component is apolipoprotein-L1 in both TLFs. TLF resistance mechanisms of T. b. gambiense and T. b. rhodesiense are now known to arise through either gain or loss-of-function, but our understanding of factors that render T. b. brucei susceptible to lysis by human serum remains incomplete. We conducted a genome-scale RNA interference (RNAi) library screen for reduced sensitivity to human serum. Among only four high-confidence 'hits' were all three genes previously shown to sensitize T. b. brucei to human serum, the haptoglobin-haemoglobin receptor (HpHbR), inhibitor of cysteine peptidase (ICP) and the lysosomal protein, p67, thereby demonstrating the pivotal roles these factors play. The fourth gene identified encodes a predicted protein with eleven trans-membrane domains. Using chemical and genetic approaches, we show that ICP sensitizes T. b. brucei to human serum by modulating the essential cathepsin, CATL, a lysosomal cysteine peptidase. A second cathepsin, CATB, likely to be dispensable for growth in in vitro culture, has little or no impact on human-serum sensitivity. Our findings reveal major and novel determinants of human-serum sensitivity in T. b. brucei. They also shed light on the lysosomal protein-protein interactions that render T. b. brucei exquisitely sensitive to lytic factors in human serum, and indicate that CATL, an important potential drug target, has the capacity to resist these factors.

Figure 1. A genome-scale T. b. brucei RNAi library screening strategy for human serum resistance.

(A) T. b. brucei (MITat 1.2; 2T1) have an EC₅₀ of less than 0.00025% normal human serum (NHS) when incubated for 72 hours at 37°C. (B) Schematic showing selection of NHS-resistant parasites from the RNAi library. (C) A population resistant to NHS was selected. RNAi was induced in 1 µg/ml tetracycline (Tet) for 24 hours prior to selection initiated at day 0; red arrows indicate culture dilution and addition of fresh NHS and tetracycline at 0.0005% and 1 µg/ml, respectively. (D) The NHS resistance phenotype of the selected library was tetracycline-inducible; error bars represent standard deviation.

Figure 2. An RNAi screen reveals major determinants of human serum sensitivity in T. b. brucei.

(A) Genome-wide NHS RIT-seq profile representing 7,433 non-redundant protein-coding sequences, showing those represented by more than 100 reads per kbp; three known human serum sensitivity determinants, HpHbR (Tb927.6.440), p67 (Tb927.5.1810/1830) and ICP (Tb927.8.6450), and the only other high-confidence and novel ‘hit’ (Tb927.8.5240) are highlighted in red. (B) Profiles of mapped RIT-seq reads for the four high-confidence hits identified in (A); boxes represent protein coding sequences; total reads (red) and tagged reads (blue; containing the 14-bp RNAi construct signature sequence) mapped to each region are shown in the top right corner of each panel; see Table 1 for further details.

Figure 3. Specific RNAi-mediated depletion of Tb927.8.5240 renders cells less sensitive to NHS.

(A) Tb927.8.5240 depletion, as demonstrated by reverse transcriptase qPCR (see also Tables S1 and S2 in Text S1), has no significant effect on parasite population growth (B); bars represent standard deviation from three independent strains. (C, D) RNAi against Tb927.8.5240 increases NHS EC₅₀; bars represent standard error (C) or standard deviation (D); p values derived from extra sum-of-squares F-test (C) or paired t-test (D).

Figure 4. Chemical inhibition of T. b. brucei cathepsins in the presence of ICP has no effect on human serum trypanolytic activity.

(A) FMK024 fails to synergise with NHS against wild-type T. b. brucei; isobologram analysis showing 50% fractional inhibitory concentrations (FIC); dashed line indicates the expected output for no interaction. (B) Individual NHS EC₅₀ curves generated in the presence of 5 to 160 nM FMK024; bars represent standard error.

Figure 5. Specific depletion of T. b. brucei cathepsins in the presence of ICP has no effect on human serum trypanolytic activity.

(A) CATB depletion has no effect on cell growth, while (B) induction of CATL RNAi in 1 µg/ml tetracycline (Tet) results in a significant growth defect; bars represent standard deviation from four and two independent strains, respectively. CATB^12MYC and anti-CATL western blots confirm depletion of the individual cathepsins; coomassie-stained gels show loading. Representative NHS EC₅₀ analysis following (C) CATB and (D) CATL depletion in wild-type T. b. brucei; RNAi induced in 1 µg/ml and 2 ng/ml tetracycline, respectively; bars represent standard error.

Figure 6. Disruption of ICP in T. b. brucei renders cells less sensitive to NHS.

(A) Schematic showing segment deleted from the ICP locus (highlighted with an open triangle), and positions of restriction sites (arrow heads), and ICP and downstream (DS) flanking probes. (B) Southern blots confirming ICP deletion from the 2T1 cell line; BSD, blasticidin-S-deaminase; NPT, neomycin phosphotransferase. (C, D) ICP loss has a minor but significant effect on FMK024 sensitivity (mean fold EC₅₀ change, 1.31; five independent strains). (E) NHS EC₅₀ analysis of a representative icp null strain; bars represent standard error. (F) Reduced NHS sensitivity requires deletion of both ICP alleles (five independent strains); loss of one ICP allele has no significant effect on sensitivity to NHS (two independent strains). Bars represent standard error (C, E) and standard deviation (D, F); p values derived from extra sum-of-squares F-test (C, E) or paired t-test (D, F).

Figure 7. CATL resists human serum trypanolytic activity in the absence of ICP.

(A) FMK024 and NHS act synergistically against icp null T. b. brucei; isobologram analysis showing 50% fractional inhibitory concentrations (FIC) where the dashed line indicates the expected output for no interaction. (B) Treatment with 5–160 nM FMK024 reverses the icp null T. b. brucei NHS resistance phenotype. (C) Representative NHS EC₅₀ analysis following CATB depletion in icp null compared with wild-type and icp null T. b. brucei parental cell lines. (D) NHS EC₅₀ analysis following CATB depletion in four independent strains. (E) Representative EC₅₀ analysis following CATL depletion in icp null cells compared with wild-type and icp null T. b. brucei parental cells. (F) NHS EC₅₀ analysis of two independent CATL RNAi icp null strains. Bars represent standard error (C, E) and standard deviation (D, F); p value derived from paired t-test; CATB and CATL knockdowns were induced in 1 µg or 2 ng/ml tetracycline, respectively.

Figure 8. Model showing the proposed interactions among ICP, CATL and TLF.

The four proteins identified in our loss-of-function RNAi library screen are highlighted in green. TLF, including APOL1 (red square), is taken up via HpHbR in the parasite's flagellar pocket, the primary site of endocytosis in T. b. brucei. The lytic factor transits the endosomal system, eventually reaching the lysosome. APOL1 is thought to form pores in the lysosomal membrane, ultimately leading to cellular lysis. CATL activity is normally tightly regulated by ICP; however, we speculate that loss of ICP increases resistance to human serum due to increased proteolysis of TLF, and possibly APOL1, by CATL. The sub-cellular localisation of the putative channel protein, Tb927.8.5240, and its role in determining sensitivity to human serum are unknown. The lysosomal membrane protein p67 is also shown.