Procyclic stage Trypanosoma brucei PT1 (Trenaman et al., 2019) cells were grown in SDM-79 medium at 27°C. Cell density was determined using a haemocytometer. For transformation, 2 × 10⁷ cells were spun for 10 min at 1,000 g at room temperature the supernatant discarded and washed in 2 ml of prewarmed cytomix and spun as before. The cell pellet was resuspended in prewarmed 100 µL cytomix solution (van den Hoff et al., 1992) with 10 µg linearised DNA and placed in a 0.2 cm gap cuvette, and nucleofected (Lonza) using the X-014 program. The transfected cells were placed into 10 ml SDM-79 medium only and placed in an incubator to allow the cells to recover overnight. Serial dilutions plated out into 96 well plates at a 1:25, 1:50, and neat. 1 × 10⁶/ml wild type or untransformed cells were added to the dilutions to condition the medium. Antat 1.1E (EATRO1125) cells were grown in HMI-11 medium at 37.4°C with 5% CO₂ and the density of cell cultures measured using a haemocytometer keeping cells bellow 1 × 10⁵ cells/ml. Transformation of cell lines was carried out by centrifuging 2.5 × 10⁷ cells at 1,000 g for 10 min at room temperature. The cell pellet was resuspended with 10 μg linearized DNA in 100 μL of warm cytomix solution (van den Hoff et al., 1992), placed in a cuvette (0.2 cm gap) and transformed using a Nucleofector™ (Lonza) (X-001 function). Transfected cells were recovered in 36 ml of warm HMI-11 at 37°C for 4–6 h, after which cells were plated out in 48 well plates at 1:5, 1:10, 1:50 and neat serial dilutions with the required drug selection. For metacyclic differentiation induction, the PT1 RBP6 overexpression cell line was grown in SDM-80 medium with 10% heat inactivated FBS, without glucose and 50 mM N-acetyl glucose-amine to block uptake of residual glucose molecules from the FBS (Dolezelova et al., 2020). RBP6 overexpression was induced with 10 μg/ml of tetracycline and the cell line maintained between 2–5 × 10⁶ cells/mL—exponential mid-log growth phase. Hygromycin and Blasticidin were selected at 2 μg ml⁻¹, 2.5 μg ml⁻¹, and 10 μg ml⁻¹ respectively. Puromycin, phleomycin, hygromycin, blasticidin, and tetracycline were maintained at 1 μg ml⁻¹.

To construct PT1 RBP6 overexpression cell line, RBP6 was amplified from wild type genomic DNA using primers RBP6F and RBP6R (See Table 1 for sequence) and cloned into pRPa vector (Alsford et al., 2005) using HindIII—BamH1. The resulting construct was linearized with Asc1. To construct the constitutive VEX1 overexpression construct pRPΔopVEX1, we digested pRPaVEX1 ^i6m and pRPΔop with HindIII—BamH1 and ligated. pRPaVEX1^i6m, pRPaVEX1^isl and pNATVEX1^12myc are described in (Glover et al., 2016).

Immunofluorescence analysis was carried out using standard protocols as described previously (Glover et al., 2013). Rabbit α-myc (Cell Signalling, # 71D10) (1: 200) and rabbit α- CRD (1: 500) [Davids-Biotechnologies, (Zamze et al., 1988)]. Secondary anti-sera used were goat α-rabbit AlexaFlour^® 555, goat α-rabbit AlexaFlour^® 488 (1:1,000). Samples were mounted in Vectashield (Vector Laboratories) containing 4, 6-diamidino-2-phenylindole (DAPI). In T. brucei, DAPI-stained nuclear and mitochondrial DNA can be used as cytological markers (Woodward and Gull 1990); Images were captured using a ZEISS Imager 72 epifluorescence microscope with an Axiocam 506 mono camera and images were processed in ImageJ.

RNA samples were taken at 0-, 4- and 8-days post RBP6 and VEX1 overexpression or knockdown induction. RNA was extracted from 50 ml of culture at 2 × 10⁶ cells/ml. RNA-seq was carried out on a BGISeq platform at The Beijing Genome Institute (BGI). Reads were mapped to a hybrid genome assembly consisting of the T. brucei 427 reference genome plus the bloodstream VSG-ESs (Hertz-Fowler et al., 2008; Cross et al., 2014; Muller et al., 2018). Bowtie 2-mapping was used with the parameters --very-sensitive --no-discordant --phred33. Alignment files were manipulated with SAMtools (Li et al., 2009). Per-gene read counts were derived using the Artemis genome browser (Carver et al., 2012); MapQ, 0. Read counts were normalised using edgeR and differential expression was determined with classic edgeR. RPKM values were derived from normalised read counts in edgeR (Robinson et al., 2010).

The expression levels of RBP6, VEX1, and mVSG genes was analysed by RT-qPCR using Luna Universal qPCR MasterMix (NEB) with 500 nM of primers. All primer pairs are listed in Table 1. RNA was extracted using a Qiagen RNeasy Kit and the samples were treated with DNase 1 for 1 h according to manufactures instructions and eluted in 30 µL of RNase free water. The samples were quantified using a Nanodrop (ThermoFisher). cDNA was prepared using SuperScript IV (ThermoFisher) following the supplier instructions from 1–2 µg RNA with a polyT primer. For each pair of primers (used at 500 nM), triplicates of each sample were run per plate (Hard-shell PCR Plates 96 well, thin wall; Bio-Rad), which were sealed with Microseal “B” Seals (BioRad). All experiments were run on a CFX96 Touch Real-time Detection system with a C1000 Touch Thermal cycler (Bio-Rad), using the following PCR cycling conditions: 95°C for 1 min, followed by 40 cycles of 95°C for 15 s and 60°C for 30 s (fluorescence intensity data collected at the end of the last step). Data was then analysed by relative quantification using the ΔΔCt method (CFX Maestro software—Bio-Rad) and Cq determination regression was used. In all cases, product abundance was determined relative to an actin control locus.

Tsetse flies (Glossina morsitans) were maintained at 27°C and 70% hygrometry in Roubaud cages, in groups of 50 male flies per cage. Pleomorphic trypanosome cell lines were maintained at 1 × 10⁵ cells/ml density in HMI-9 medium plus 10% FBS at 37°C with 5% CO₂. In vitro stumpy differentiation was induced in HMI-9, supplemented with 10% FBS without antibiotics, by adding 8-pCPT-2′-O-Me-5′-AMP (5 μM) (BioLog-Life Science Institut) to the culture 48 h before fly infection (Laxman et al., 2006). On the day of infection, trypanosomes were resuspended at 10⁶ cells per ml in SDM79 with no antibiotics supplemented with 10 mM glutathione prior infection (MacLeod et al., 2007). Flies were fed on infected media through a silicone membrane and maintained until dissection by feeding three times per week on sheep’s blood in heparin. Flies were starved for 2 days before dissection at day 28. Imaging was carried out using a ZEISS Imager 72 epifluorescence microscope with an Axiocam 506 mono camera. Single images (for DIC) or multichannel stacks (for fluorescence) of images every 0.24 µm were acquired. When maximum intensity Z-projections are presented, they were generated using Fiji (Schlindelin et al., 2012).

The transition between developmental stages is accompanied by changes in gene expression, including the silencing or activation of both the BESs and MESs. The bloodstream form to procyclic stage differentiation is marked by the replacement of the VSG surface coat with EP and GPEET procyclins (Acosta-Serrano et al., 2001). Given the VEX complex association with the VSG transcription compartment in the bloodstream form cells and that it redistributes upon differentiation (Glover et al., 2016; Faria et al., 2019; Faria et al., 2021), we wanted to ask whether VEX1 was required for silencing BES and MESs in the procyclic stage cells. We first assessed expression of the 8 mVSG genes (Muller et al., 2018) and VEX1 by RT-qPCR in VEX1 overexpression or knock down backgrounds (Figures 1A,B). VEX1 RNAi led to a greater derepression of the 8 mVSG genes (Figure 1A; Supplementary Figure S1A) compared to VEX1 overexpression (Figure 1B; Supplementary Figure S1B)—average fold change of 15,04 compared to 5,82 respectively. Given this striking difference between expression of the mVSG genes according to the VEX1 expression level, we wanted to determine whether this was similar for the BES VSGs. Transcriptomic analysis of VEX1 silenced by RNAi or overexpressed for 96 h showed similar patterns to the RT-qPCR, where RNAi led to a greater derepression of the 8 mVSG genes than overexpression, suggesting a primarily silencing function for VEX1 in procyclic stage cells. When we analysed BES linked genes, we found that overall, there was a subtler derepression as compared to the mVSG genes (Figures 1C,D), but this was more pronounced following VEX1 RNAi, again suggesting that VEX1 is primarily required for silencing in this life cycle stage. Our data point to VEX1 having a role as a negative regulator of VSG expression in procyclic stage cells, where knocking down expression of VEX1 from the cell has a stronger effect compared to overexpression. This suggests that VEX1 is required for efficient silencing of mVSG ES promoters in procyclic stage cells.

As we have shown VEX1 is required for efficient silencing of expression-site linked VSG genes in insect stage cells, we then asked whether the distribution of VEX1 in the nucleus changed depending on the life cycle stage in the tsetse fly, and specifically in metacyclic cells where a VSG is expressed. To assess VEX1 localisation in metacyclic cells, we used the inducible RBP6 expression system (Kolev et al., 2012). In this system, metacyclic cells are produced spontaneously rather than in a temporal order (Kolev et al., 2012; Ramey-Butler et al., 2015), allowing us to capture both early stage and mature metacyclic cells. We found, as previously shown, that induction of RBP6 stimulates metacyclogenesis in vitro and leads to the expression of mVSGs and the production of metacyclic cells in culture (Figure 2A). We assessed the expression of 5 mVSG genes and RBP6 by RT-qPCR and found a significant increase in mVSG653, mVSG1954, mVSG531 and mVSG639 expression over day 2 to day 4 (Figure 2A; Table 2). We then scored the number of metacyclic cells in culture using a pan-VSG antibody (anti-CRD). Between day 5 and 7 up to 25% of the culture were metacyclic cells (Figure 2B). Following in vitro differentiation to insect stage cells, the active BES relocalises to the nuclear periphery (Landeira and Navarro 2007), while the VEX complex is redistributed from one—two nuclear foci to a multifocal distribution that appears to be concomitant with all telomeres (Glover et al., 2016). This differentiation step is also associated with a substantial increase in protein abundance VEX2 (Faria et al., 2019). A similar pattern was observed in the uninduced procyclic stage cells (Figure 3A), where VEX1 forms a multi-focal pattern in the nucleus. Following 5 days of RBP6 overexpression, metacyclic cells were identified in culture based on cell morphology and the position of the nucleus and kinetoplast. We found that in these metacyclic cells there were significantly more VEX1 forming one to two foci (Figures 3A,B; Procyclic cells 95% formed multiple foci versus 68% of metacyclic cells with 1-2 foci), suggesting that the pattern of VEX1 accumulation in the nucleus is life cycle stage specific. We then wanted to confirm these findings in vivo. Tsetse flies (Glossina morsitans morsitans) were infected with Antat 1.1E bloodstream form cells with natively tagged VEX1. As was seen in the in vitro RBP6 system, VEX1 formed a multi focal pattern in midgut trypomastigote cells (procyclic and mesocyclic forms, Figures 3C,E; Supplementary Figure S2). Approximately 60% of metacyclic cells had one to two foci (Figures 3D,E; Supplementary Figure S3), and in approximately 40% we could detect 3 or more VEX1 foci, these may be early metacyclic where monoallelic mVSG expression is not yet established (Hutchinson et al., 2021). We also examined additional life cycle stages found throughout in the tsetse fly and found that VEX1 was distributed across the nucleus in a multi-focal pattern. Therefore, in metacyclic cells one to two VEX1 foci per nucleus was the dominated pattern (Figure 4; Supplementary Figure S3). We, therefore, see a similar pattern of VEX1 distribution both in vitro and in vivo. The VEX1 localisation we see in metacyclic is reminiscent of that seen in bloodstream form cells (Glover et al., 2016; Faria et al., 2019; Faria et al., 2021) where the VEX complex is required for singular VSG expression.

The transition from epimastigote to metacyclic cells in the tsetse fly salivary gland results in the activation of MESs promoters and expression of mVSG genes (Graham and Barry 1995; Sharma et al., 2009). To determine the role of VEX1 in metacyclogenesis we established a doubly inducible TET ON system to simultaneously modulate VEX1 expression levels and induce metacyclogenesis via RBP6 overexpression (Supplementary Figure S1C). Initially we assessed the expression of two mVSG genes by RT-qPCR (Figure 5A; Table 3) at 4- and 8-days post induction along with RBP6 and VEX1. In the parental cell line, with RBP6 overexpression only, we see a 5-fold increase in RBP6 expression and a 7–13 -fold increase in mVSG397 and mVSG653 expression (Figure 5A; Table 3). Strikingly, a reduction in VEX1 resulted in poor mVSG expression as compared to the parental cell line (RBP6 overexpression only), and surprisingly we also see a 6-fold reduction in RBP6 expression (Figure 5A; Table 3). Conversely, when VEX1 and RBP6 are overexpressed, we see an 8-fold increase in RBP6 expression and between 18–26-fold increase in mVSG397 and mVSG653 expression (Figure 5A; Table 3). We noted that the increase in VEX1 overexpression cell line was subtle, but even in this context has a dramatic effect on both mVSG and RBP6 expression levels (Figure 5A; Table 3).

To further investigate the role of VEX1 in metacyclogenesis we wanted to see what global changes were associated with VEX1 RNAi or overexpression during metacyclogenesis. For this we performed transcriptomics analyses at day 4 and 8. Our analysis revealed that during metacyclogenesis, VEX1 RNAi resulted in reduced mVSG expression and a concomitant reduction in the level of RBP6 (Figure 5B) as we saw with the RT-qPCR (Figure 5A). This suggests that initiation of mVSG expression is crucial to this life cycle differentiating process. Using a pan-VSG antibody, we found that in cultures at day 4 and 8 cultures, no VSG expressing cells were seen (data not shown). We noted that there was a cohort of genes whose transcripts increased in abundance at day 4 and 8 (Figure 5B) following RBP6 overexpression and VEX1 RNAi. Analysis revealed a specific increase in abundance of silent BES linked genes, normally only expressed from the single active BES in bloodstream form cells (Supplementary Figure S4A). Differentiation from bloodstream form to procyclic form trypanosomes results in cessation of transcription of the ESAG and VSG genes from the active BES, however multiple BES promoters remain active a relatively equivalent level (Pays et al., 1989; Zomerdijk et al., 1990; Rudenko et al., 1994). Although BES protomers are active in procyclic cells, transcription terminates before the ESAG7 gene (Zomerdijk et al., 1990; Rudenko et al., 1994). We then looked with more detail at which BES-linked genes were significantly upregulated, we found that ESAG 10, 7, and 6 were 4 to 6-fold upregulated at all BESs, but this level of upregulation was not sustained (Supplementary Figure S4C, S5) suggesting transcription does not proceed across the whole BES. In fact, the BES-linked VSG genes were only on average 2-fold upregulated. We then looked at other Pol I transcribed genes and found that the procyclin associated genes PAG1, PAG4, and PAG5 (Tb927.10.10240, Tb927.10.10210, Tb927.10.10230 respectively) showed an increase in transcript abundance (by 4, 6, 5-fold respectively on day 4) but not EP1, EP2 Procyclin or GPEET (Supplementary Figure S4A). This suggests that during metacyclogenesis, VEX1 is necessary for expression of mVSG genes and maintain silencing of BESs, but additional factors are required for full BES silencing in this life cycle stage. In contrast, following VEX1 overexpression we see an increase in expression of mVSG genes and only moderate increase in BES linked VSG genes (Figure 5C), suggesting that VEX1 promotes mVSG expression. Unlike with VEX1 RNAi, we see a more restrained increase in transcript abundance of BES linked genes and the procyclin and procyclin associated genes (Supplementary Figure S4B). Across the BES, ESAG 10, 7, and 6 again showed the highest fold change, but by only 2-fold change on average, which was significantly lower than in the VEX1 RNAi (Supplementary Figure S4D, S5). Strikingly though, mVSG genes show an average of 6-fold increase in transcript abundance (Supplementary Figure S4D), revealing a positive role for VEX1 in mVSG transcription in metacyclic cells.