Résumé : The two human pathogenic sub-species T.b.gambiense and T.b.rhodesiense can be distinguished from the morphologically identical T.b.brucei by their ability to infect humans, enabling them to cause sleeping sickness. This is because they are resistant to lysis by the lytic factor (APOL-I) present in normal human serum (NHS). In T.b.rhodesiense resistance to this lytic factor is due to a truncated VSG gene termed SRA which blocks lysis by interacting with APOL-I in the lysosome. SRA does not exist in T.b.gambiense. The search for a similar truncated VSG gene lead to the identification of a T.b.gambiense specific glycoprotein termed TGSGP. TGSGP transfected alone into the sensitive T.b.brucei is unable to confer resistance to this sub-species. This is either due to incorrect processing of this gene is this sub-species or because TGSGP requires a partner to confer resistance.

In the search for a partner, the genomic locus of TGSGP was cloned and sequenced. We found that TGSGP is linked to a truncated gene homologous to the S.cerevisiae AUT1 gene, a gene implicated in autophagy and more specifically in membrane expansion. Southern blot hybridization and PCR analysis on genomic DNA from several isolates demonstrated that this feature was a specific to T.b.gambiense. In addition, we observed a correlation between the aut1 allele size and the geographical origin of the isolate.

Since in trypanosomes lysis by NHS is due to an uncontrolled expansion of the lysosome, we speculated that the truncation of the aut1 allele could be implication in the resistance to human serum. We characterized the genomic organisation of the AUT1 locus. T.b.brucei possesses two native AUT1 alleles whilst T.b.gambiense possesses a truncated aut1 allele, as well as a native AUT1 allele. We showed that in the T.b.gambiense LiTAR isolate (aut1/AUT1), despite the presence of a wild-type allele this gene is no longer expressed at the mRNA and protein level. Our complimentary results by run-on transcription assay showed that the AUT1 region is transcribed but that the messenger is unstable. LiTAR is a functional knock-out for AUT1, but Northern blot analysis on several T.b.gambiense isolates showed that this is not a generalised T.b.gambiense characteristic.

We explored the role of AUT1 in trypanosomes by invalidation of the AUT1 gene in T.b.brucei and by the over-expression of the AUT1 and aut1 alleles in T.b.brucei. By functional analysis of AUT1 knocked-down cells we showed that AUT1 is not essential in trypanosomes. By recreating in T.b.brucei the T.b.gambiense AUT1/aut1 genotype we were able to show that the expression of the aut1 UTR down-regulated the expression of the wild-type AUT1 allele. We speculated that this may be due to a natural RNAi mechanism. Par northern blot, using probes covering the potential target region of AUT1, we detected a 50nt small RNA specific to T.b.gambiense. In addition, we showed that in a LiTAR strain in which the RNAi pathway was abolished AUT1 expression is restored.

We continued to investigate TGSGP’s role in the resistance to human serum by invalidation of TGSGP in T.b.gambiense and by expressing TGSGP in the NHS-sensitive T.b.brucei. Because T.b.gambiense cannot be cultured in vitro we established a new in vivo transfection technique and as the knock-out of TGSGP is most probably lethal, we created an inducible RNAi T.b.gambiense cell strain. These indispensable tools will be used to test whether invalidation TGSGP is sufficient to confer resistance to NHS. Many strategies were tested in order to correctly expressing TGSGP in T.b.brucei; in none of these transfectants was TGSGP correctly located in the flagellar pocket as is the case in T.b.gambiense and only partial resistance was ever obtained. In order to identify the factors in human serum that could interacts with TGSGP, we subjected NHS to affinity chromatography using TGSGP as bait. We showed that TGSGP interacts with APOA-I, a major component of HDLs.