Identification of a novel SNF2 ATPase complex

To identify candidate SNF2-like proteins in T. brucei, we carried out a homology-based search in the TriTrypDB and identified 15 candidates that putatively belong to the SNF2 superfamily. In particular, we were interested in candidates with an insertion in the DEXH/Q motif between the DEAD-Box and helicase C motif, which has been shown to be characteristic for SNF2 proteins of the SWR-1-like subfamily [12] (S1A Fig). One putative candidate, Tb927.11.10730, matched this criterion. To test whether this candidate is a component of a SWR1-like remodeller complex, we wanted to carry out immunoprecipitations using a tagged version of the protein. Neither N- nor C-terminal tagging of Tb927.11.10730 proved possible, however. As an alternative strategy, we carried out co-immunoprecipitations (co-IPs) using the DNA helicase RuvB, which is an important and conserved part of both the INO80 and SWR1 complexes [51, 55]. Similar to other model organisms, two versions of the RuvB DNA helicase can be found in T. brucei, Tb927.4.1270 and Tb927.4.2000. In the following we will refer to these putative helicases as RuvB1 and RuvB2, respectively. Mass spectrometry (MS) analysis of the RuvB2 co-IP successfully identified the SNF2 candidate Tb927.11.10730 together with 14 additional proteins (Table 1 and S1B Fig) that were significantly enriched (p<0.01) and had high nuclear enrichment scores (NES) [56]. These proteins included known conserved components of the SWR1 remodeller such as YEATS domain-containing proteins and actin-like proteins.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Summary of the RuvB2 co-IP.

15 proteins with a positive or unknown nuclear enrichment score were identified by MS in the RuvB2-HA co-IP. The nuclear enrichment score (NES; (52)) indicates the probability of a nuclear localisation based on cell fractionation combined with quantitative MS analysis. The “Annotation” column indicates the curated annotation that was found for the corresponding accession number on the TriTyp database. In the “p-value” column a probabilistic confidence measure (P-value) is assigned to each identified protein. The fold enrichment compared to the WT control is stated for every identified protein. The column "TrypTag localisation” indicates the cellular localisation of N- and C-terminally tagged proteins.

https://doi.org/10.1371/journal.ppat.1010514.t001

To confirm these proteins as components of the complex and to identify factors that might have been missed in the RuvB2 co-IP, another round of co-IPs was performed. The proteins tagged for this reciprocal approach were Tb927.10.11690, a protein containing a YEATS domain (referred to as Trypanosoma brucei SWR1 complex protein 1 (hereafter TbSWRC1), Tb927.11.5830, which contained an YL1 domain (hereafter TbSWRC2) and the Trypanosoma-specific hypothetical protein Tb927.7.4040 (hereafter TbSWRC4). TbSWRC1 and TbSWRC2 were promising candidates, as parts of the three-dimensional structure of TbSWRC1 and TbSWRC2 could be modelled with high confidence based on Yaf-9 (yeast AF9) and Vps72/SWC2, respectively, employing the Phyre2 homology modelling web tool [57]. In yeast, the YEATS domain of Yaf-9 is essential for histone binding and enables the accurate assembly of the SWR1 complex on the chromatin fibre [58, 59]. The conserved YL1 domain of Vps72 (vacuolar protein sorting 72 homologue)/ SWC2 (SWR1 complex protein 2) acts as a molecular lock to prevent H2A.Z eviction in the SWR1 complex and mediates, together with Arp5 proper nucleosome repositioning by the INO80 complex [55, 60]. The two proteins play a key role in both SWR1 and INO80 function and are conserved in SWR1 and INO80 complexes in several species [51]. TbSWRC4 was chosen as a third candidate as it appears to be exclusively found in Trypanosoma brucei, Trypanosoma cruzi and Blechomonas ayalai, thus representing a phylogenetically restricted complex member. Neither a BLAST search with the NCBI database, the EMBL InterPro database nor a homology modelling led to reliable information about a possible homologue of TbSWRC4 in other organisms. This second set of co-IPs enriched 13 proteins (Table 2) as well as histones H2A, H2B, H3, H4 and the histone variant H2A.Z, which strongly suggested a role in nucleosome assembly.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Characterisation of the novel SWR1 complex components.

13 proteins with a positive or unknown NES were identified by MS analysis in at least three of the four co-IP experiments. The initial co-IP was performed with RuvB2 (Tb927.4.2000, the reciprocal co-IPs with the proteins Tb927.10.11690, Tb927.11.5830 and Tb927.7.4040 were performed to confirm the RuvB2 co-IP data). The “Annotation” column indicates the curated annotation that was found for the corresponding accession number in the TriTryp database. The”identified domains” column displays the domains that were found by BLAST search using the NCBI database. The Phyre2 modelling column indicates proteins that were identified by homology modelling. Coverage (Cov.) indicates the coverage in percent between query and template. The confidence (Conf.) represents the relative probability in percent (from 0 to 100) that the match between query and template is a true homology. The nuclear enrichment score (NES; (52)) indicates the probability of a nuclear localisation based on cell fractionation combined with quantitative MS analysis. The column "TrypTag localisation” indicates the cellular localisation of N- and C-terminally tagged proteins. The last column shows in which of the four co-IPs the protein could be identified.

https://doi.org/10.1371/journal.ppat.1010514.t002

In addition to the proteins that were identified in the initial RuvB2 co-IP, the reciprocal co-IPs could identify the conserved hypothetical protein Tb927.9.8510 (referred to as TbSWRC3 in the following). Five proteins originally found in the RuvB2 co-IP were not enriched in the other three co-IPs, indicating they are more closely associated with RuvB2. Furthermore, Phyre2 modelling suggested that Tb927.6.2570, a potential SUMO-interacting motif-containing protein (referred to as TbARP2 in the following) is probably a third actin-like protein [57]. TbSWRC3 was modelled as a potential SET (Su(var)3-9, enhancer-of-zeste and trithorax)-methyltransferase by Phyre2 [57]. A putative catalytic site specific to SET-methyltransferases could be annotated including three potential binding sites for the co-factor S-adenosylmethionine. The 13 proteins that were identified in at least three of the four co-IP experiments assemble the trypanosome SWR1-like complex (Table 2). The complex composition differs slightly from the Saccharomyces cerevisiae SWR1 complex (Fig 1). Nevertheless, sharing of certain components such as a BCNT- and a Znf-HIT1-domain containing protein clearly hint towards the complex being a SWR1-like remodeller. Domains that are typical for an INO80 complex like the PAPA-1-domain or an HMG-Box are absent (Fig 2) [51]. However, further characterisation of the molecular functions of this novel complex and especially of the SNF2 ATPase were necessary to support a SWR1-like complex.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. The Trypanosoma brucei SNF2 ATPase complex exhibits characteristics of a SWR1 complex.

A comparison of the modular composition of the newly-identified SNF2 complex with the SWR1 and INO80 complexes of S. cerevisiae. INO80 complex S. cerevisiae: The Nhp-10 module forms a platform for nucleosome interaction. The Arp8 module is the nucleosome binding module, while the Arp5 module, which contains the YL1-domain protein Ies6 is responsible for the nucleosome remodelling step [14]. Ies2 with the PAPA-1-domain plays a structural role within the complex [14, 112]. INO80 specific domains are highlighted in bolt. SWR1 complex S. cerevisiae: Proteins listed under key features are essential for H2AZ incorporation [68]. The C-module of the SWR1 complex, which contains the Bromo-, YEATS- and SANT/DAMP1-domain mediates nucleosome affinity, while the N-module with the YL1- and Zinc finger ZNHIT1-domain is involved in the histone variant exchange reaction [49, 68]. SWR1 specific domains are highlighted in bolt letters [51, 68]. SWI2/SNF2 complex T. brucei: The identification of a BCNT- as well as a ZNHIT1 domain (highlighted in bold letters) hint toward the complex being a SWR1-like complex [51, 68]. Structure of the SWI2/SNF2 complex and its modules is only putative. The potential interaction interface of the species-specific proteins TbSWRC3 and TbARPx is unknown.

https://doi.org/10.1371/journal.ppat.1010514.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Loss of TbSWR1 leads to cell death and a reduction of chromatin-associated H2A.Z.

(A) Growth of parasites was monitored for 96 hours after RNAi-mediated depletion of TbSWR1 (Tb927.11.10730) using tetraycline (tet). The parental 2T1 cell line was used as a control (n = 3). (B) Quantification of live/dead staining with propidium iodide of TbSWR1-depleted cells at the indicated timepoints post-induction. Analysis was done by flow cytometry (n = 3). (C) Western blot analysis of the insoluble nuclear fraction with antibodies specific for histone H3 and the histone variant H2A.Z. Lysates from an equal number of cells (2x10⁶ per lane) were analysed for each timepoint. (D) Quantification of chromatin-associated H3 (dark blue), Ty1-H2A.Z (turquoise) and H2AZ (grey) (N = 3 for all depicted experiments; *** = p-value <0.001; ** = p-value 0.001–0.01; * = p-value 0.01–0.05).

https://doi.org/10.1371/journal.ppat.1010514.g002

Depletion of TbSWR1 leads to a reduction of chromatin-associated H2A.Z

Ultimately, as the INO80 complex represents the functional counterpart of the SWR1 complex and vice versa, RNAi-mediated depletion of the SNF2 subunit can reveal whether chromatin association of H2A.Z will increase, indicating an INO80-like function, or if H2A.Z incorporation will decrease, implying an SWR1 functionality. We thus depleted the SNF2 ATPase (Tb927.11.10730) by tetracycline-inducible RNAi (Fig 2). The depletion resulted in a severe growth phenotype with an almost complete growth arrest 24h post-induction (Fig 2A). FACS analysis of propidium iodide-stained cells showed almost unchanged viability of the population 24h after induction, and approximately 70% living cells after 48h, however (Fig 2B). Cell fractionation experiments were performed to determine whether the amount of nucleosome-associated H2A.Z decreased in comparison to core histones following depletion of the SNF2 ATPase (Fig 2C and 2D). Proteins in the insoluble chromatin fraction were analysed by Western blot with antibodies specific for histone H3 and the variant H2A.Z. While levels of H2A.Z and a Ty1-tagged version of H2A.Z were significantly reduced over time, H3 levels appeared not to be affected, indicating that this SNF2 ATPase is indeed associated with efficient incorporation of H2A.Z into chromatin (Fig 2C and 2D). A second clone (TbSWR1 c4) produced a similar albeit less severe phenotype (S1C Fig). Cell cycle analysis of TbSWR1-depleted cells showed a reduction of G1 and to a lesser extent G2 cells during the time course and a continued accumulation of anucleate parasites (S1D Fig). As both the YEATS domain and the YL1 domain in the SWR1 complex play a key role in H2A.Z incorporation [49], we also investigated how the loss of TbSWRC1 and TbSWRC2 affected chromatin-associated H2A.Z levels. Though SNF2 ATPase depletion resulted in stronger effects, the reduction of chromatin-associated H2A.Z was also observable upon depletion of the YEATS domain-containing protein TbSWRC1 and YL1 domain protein TbSWRC2 (S1D Fig and S1E Fig), emphasizing the contribution of the additional complex members in H2A.Z incorporation. As fractionation experiments showed a significant reduction of chromatin-associated H2A.Z, we were confident that this novel SNF2 ATPase is indeed a SWR1 homologue in trypanosomes. To analyse the effect of TbSWR1 depletion more precisely, a ChIP-seq assay using a cell line with a Ty1-tagged H2A.Z allele was performed. Previous studies showed that Ty1-tagged H2A.Z is efficiently incorporated into nucleosomes at TSSs and our fractionation experiments showed that Ty1-H2A.Z is chromatin-associated (Figs 2 and S1C Fig). To avoid strong secondary effects, we chose TbSWR1 c4, which had a milder phenotype, and an early time point after RNAi induction (S1C Fig). In line with our fractionation experiment, we could demonstrate a reduction of Ty1-H2A.Z at TSSs after 48 h post induction of TbSWR1 depletion (Fig 3) clearly establishing the SWR1-like functionality.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Loss of TbSWR1 (Tb927.11.10730) leads to a reduced H2A.Z deposition at TSS.

(A) ChIP-Seq analysis of distribution of Ty1-tagged H2A.Z before (black) and after (green) RNAi-mediated depletion of TbSWR1 (48h post induction) revealed a reduction of chromatin associated Ty1-H2A.Z. Depicted is a representative region of chromosome 5. Data (n = 2) were normalised to the total number of reads and plotted as counts per million reads (CPM). (B) Average Ty1-H2A.Z signal across non-divergent TSSs. Peaks derived from two non-tetracycline induced reference samples depicted in blue and black. Peaks derived from two samples after 48 h of RNAi-mediated depletion of TbSWR1 depicted in green and dark-green.

https://doi.org/10.1371/journal.ppat.1010514.g003

Depletion of HAT2-associated Bdf3 reduces chromatin-associated H2A.Z

SWR1 regulation and H2A.Z acetylation by the NuA4 complex has been extensively described in yeast [10, 29, 30, 61]. Recent studies in T. brucei linked acetylation of H4 and H2A.Z. with the two histone acetyltransferases HAT2 and HAT1 (S1G Fig). Since H2A.Z incorporation into nucleosomes was linked to HAT2 activity [54] we decided to investigate the complex composition of HAT2. Co-IP experiments with HA-tagged HAT2 identified a complex that consists of Bdf3, HAT2 and several other proteins including factors with an ENT (EMSY N-terminal)-domain and an FHA (fork head-associated)-domain (S1A Table and S1J Fig). Bdf3 was particularly interesting because these factors play a key role in linking H4 acetylation to SWR1 activity and it has been associated with H2A.Z in trypanosomes already. ChIP-seq experiments with Bdf3 revealed a co-localization with histone H4K10 acetylation and H2A.Z in TSSs of T. brucei [52]. We performed a reciprocal Bdf3 co-IP and could confirm the HAT2 complex composition in trypanosomes with 10 out of 11 proteins reproducibly enriched (S1A Table and S1J Fig). Furthermore, RNAi-mediated depletion of Bdf3 caused a severe growth defect similar to the phenotype observed after TbSWR1 depletion. Importantly, the amount of chromatin-associated H2A.Z significantly decreased after Bdf3 depletion (Fig 4), suggesting that it is also involved in H2A.Z deposition.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Loss of Bdf3 leads to cell death and a reduction of chromatin-associated H2A.Z.

(A) Growth of parasites was monitored for 96 hours after RNAi-mediated depletion of Bdf3 using tetracycline (tet) induction. The parental 2T1 cell line was used as a control (n = 3). (B) Quantification of live/dead staining with propidium iodide of Bdf3-depleted cells. Analysis was done by flow cytometry (n = 3). (C) Western blot analysis of the insoluble nuclear fraction with antibodies specific for histone H3 and the histone variant H2A.Z. Lysates from an equal number of cells (2x10⁶ per lane) were analysed for each timepoint. (D) Quantification of chromatin-associated H3 (dark blue), Ty1-H2A.Z (turquoise) and H2AZ (grey); (N = 3 for all depicted experiments; *** = p-value <0.001; ** = p-value 0.001–0.01; * = p-value 0.01–0.05).

https://doi.org/10.1371/journal.ppat.1010514.g004

Loss of TbSWR1 leads to reduced RNA levels

The contribution of H2A.Z to transcription regulation has been extensively described in various organisms and a recent study from the Siegel laboratory showed that H2A.Z acetylation by HAT1 is required for RNAP II transcription regulation in T. brucei [54, 62] and that loss of H2A.Z acetylation affects the transcription of PTUs [54]. To determine the effect of SWR1 complex components on transcription of PTUs, we inserted a luciferase reporter construct into the tubulin array of several RNAi cell lines. Luciferase reporter assays have been used before in T. brucei to study transcriptional activity [63]. A significant decrease of luciferase activity following TbSWR1 depletion could already be observed after 24 h. Luciferase activity decreased to approximately 60% and 30% after 36 h and 48 h induction of RNAi, respectively (Fig 5A). Cell death did not contribute substantially to decrease of luciferase signals because approximately 90% and more than 60% of the parasites were still alive after 36 and 48 hours (Fig 5B), respectively, suggesting that that the loss of luciferase activity is directly caused by the depletion of TbSWR1. Depletion of H2A.Z itself showed comparable results but with faster kinetics (Fig 6). Although 96% of the cells were alive 24 h post-induction, a reduction of the luciferase to approximately 50% compared to un-induced parasites could be observed. For comparison and as a positive control, the loss of luciferase activity was also investigated in RNAP II-depleted cells. A loss of activity of 85% after 24 h of RNAi induction could be observed while approximately 85% of the cells were still alive (Fig 5D and 5E). In line with the two acetyltransferases HAT1 and HAT2 being linked to transcription and H2A.Z regulation, a similar reduction of mRNA levels was observed. HAT1 depletion caused a reduction to 56% and 71% in two independent clones 48 h post induction (S1H Fig). We could also detect a loss of luciferase activity to approximately 70% and 80% in HAT2-depleted cells compared to uninduced cells (S1I Fig). In both experiments, we observed only a mild growth defect and a proportion of dead cells below 10% indicating that the loss of activity is not caused by dead cells (S1H Fig and 1I Fig).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Depletion of TbSWR1 caused a decrease of reporter luciferase activity within a PTU.

A luciferase reporter construct was integrated into the tubulin array in both RNAi cell lines. The TbSWR1 RNAi clone 6 was used for transfection (Figs 3 and S1C Fig). Samples for the luciferase assay were normalised to cell numbers. (A) Luciferase activity was monitored for 48 h after induction of RNAi in two independent clones. Values of non-induced cells were set to 1. (B+E) Live/dead staining of each RNAi cell line was performed in triplicates at the same time points. (C+F) Growth of parasites was monitored for 96 hours after RNAi-mediated depletion of TbSWR1 and RBP2 using tetracycline (tet) induction. The parental 2T1 cell line was used as a control (n = 3). (D) As a positive control, luciferase activity of the same reporter construct was measured in an RNAP II RNAi cell line. (N = 3 for all depicted experiments; *** = p-value <0.001; ** = p-value 0.001–0.01; * = p-value 0.01–0.05).

https://doi.org/10.1371/journal.ppat.1010514.g005

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Depletion of the histone variant H2A.Z leads to a decreased luciferase activity within a PTU.

A luciferase reporter construct was integrated into the tubulin array of a H2A.Z (Tb927.7.6360) RNAi cell line. Samples for the luciferase assay were normalised to cell numbers. (A) Luciferase activity was monitored for 48 h after induction of RNAi in two independent clones. Values of non-induced cells were set to 1; (N = 3); (B) Live/dead staining of each RNAi cell line was performed in triplicates at the same time points. (C) Growth of parasites was monitored for 96 hours (N = 3) after RNAi-mediated depletion of H2A.Z using tetracycline (tet) induction. Growth of the parental 2T1 cell line was measured for 96h as a control. (N = 3 for all depicted experiments; *** = p-value <0.001; ** = p-value 0.001–0.01; * = p-value 0.01–0.05).

https://doi.org/10.1371/journal.ppat.1010514.g006

Since these luciferase experiments provide only indirect information about polymerase activity, we next assessed steady-state mRNA levels directly in two independent TbSWR1 and one RNAP II RNAi cell lines (Fig 7). Total RNA was extracted from TbSWR1- and RNAPII-depleted cells for Northern blot analysis. A clear reduction of total mRNA and ribosomal RNA could be detected 48 hours after TbSWR1 depletion. Analysis of RNA derived from RNAP II-depleted cells showed similar results but as observed before with faster kinetics. These Northern blot analyses confirmed the initial observations demonstrating that mRNA and rRNA levels decreased approximal 2-fold 48 h after RNAi induction, with a stronger effect for TbSWR1 c6 compared to TbSWR1 c4 (Fig 7), in agreement with the reduction of H2A.Z in chromatin (Fig 2). Our data indicate that TbSWR1 has a conserved role in transcription regulation which is typical for all SWR1-like proteins described so far.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. Loss of TbSWR1 leads to a reduction of mRNA and rRNA.

(A) Representative Northern blot of two TbSWR1-depleted cell lines. RNAP II-depleted cells were used as a control. The samples were normalised to cell numbers. Upper and middle panel were hybridised with a probe specific for the spliced leader RNA (mini exon, red) the two lower panels with probes specific for ribosomal RNA (rRNA, green) (B) Quantification of results from three replicates showing total mRNA levels in red and 5.8S rRNA in green. Values of un-induced RNAP II RNAi cell lines were set to 100%.

https://doi.org/10.1371/journal.ppat.1010514.g007

Depletion of TbSWR1 affects chromatin structure

In other organisms, cessation of transcriptional activity causes massive changes in chromatin structure and nuclear architecture [64, 65]. Electron microscopy images of TbSWR1-depleted cells showed that after 24 h of RNAi-mediated depletion of TbSWR1 an increased condensation of chromatin in the nucleus in comparison to non-induced cells could be observed (Figs 8 and S1L Fig). Such a condensation could be seen in several albeit not all cells. A knockdown of the RNAP II and H2A.Z led to a comparable phenotype but with different kinetics. While RNAP II-depleted cells showed a condensation 16 h post induction, H2A.Z depleted cells exhibited chromatin condensation after 24 h of RNAi induction (Figs 8 and S1L Fig). Overall, the temporal behaviour of chromatin condensation seems similar in TbSWR1- and H2A.Z-depleted cells, supporting the function of TbSWR1 as a histone remodelling complex important for H2A.Z deposition.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 8. Loss of TbSWR1 and RNAP II leads to chromatin condensation.

Representative electron microscopy images of the nucleus of TbSWR1-, H2A.Z- and RNAP II depleted and non-depleted parasites as indicated. Depletion of the proteins resulted in large black patches of condensed chromatin. Scale bar, 200 nm for TbSWR1 RNAi and H2A.Z RNAi images, 500 nm for RNAP II (RBP2) RNAi images. Images of TbSWR1 RNAi cells were obtained using a TEM (transmission electron microscope), images of RNAP II RNAi cells with a STEM (scanning transmission electron microscope).

https://doi.org/10.1371/journal.ppat.1010514.g008

In summary, purification, molecular characterization and RNAi-mediated depletion of critical components of a novel SWR1 chromatin remodeller complex could shed some light on the so far insufficiently investigated mechanism of transcriptional regulation in African trypanosomes.

Composition of the TbSWR1 and HAT2 complexes

Co-IP experiments with 4 different subunits combined with quantitative proteomics provided a clear picture and was supplemented with a thorough in silico analysis of the identified proteins, with particular reference to homology with the SWR1 complex from S. cerevisiae. We identified homologues for most of the essential SWR1 complex components (reviewed in [68]), such as an SNF2 family helicase (TbSWR1), actin-related proteins (TbARP1-3), a YEATS domain containing protein (TbSWRC1), and a YL1-domain containing protein (TbSWRC2). The BCNT-domain and the Znf-HIT1-domain that are characteristic for a SWR1 complex [51] could be identified in TbSWRC5 and TbSWRC6 (Fig 2) as well. We also found that the composition of the SWR1 complex in trypanosomes has some distinct differences. The presence of a DNA methyltransferase 1-associated protein 1 (DAMP1)-domain [51] could not be verified in the TbSWR1 complex so far. A fragment of a DAMP-1-domain was annotated in TbSWRC4 by Phyre2 homology modelling but only with an insufficient confidence. To our surprise, a bromodomain protein—typically part of SWR complexes in other species—was missing from our co-IP data. In S. cerevisiae, Bdf1 is able to bind acetylated histone H4 [69] and is part of the SWR1 complex. While it is possible that the absence of a bromodomain factor is due to different interaction affinities and it was lost during purification, our data suggest that the core TbSWR1 complex may not contain such a bromodomain protein. TbSWRC1, which has a YEATS domain and was shown to effectively bind to acetylated histone H3 [59], may bind acetylated histone H4 instead. Besides the absence of a Bdf homologue in the complex, the identification of a third actin-related protein (S1B Fig) and a potential SET domain-containing protein (named TbSWRC3; S1B Fig) was unexpected. SET methyltransferases play important roles in transcription regulation (reviewed in [70]), suggesting that TbSWRC3 represents an important part of the transcriptional regulatory machinery in T. brucei. Depending on the substrate of TbSWRC3, the additional Arp protein might be necessary to create a larger nucleosome interaction interface or add a specific functionality. The SWR1 complex and H2A.Z have been shown to be involved in many cellular processes including DNA repair [71–73]. As part of the NuA4 complex in S. cerevisiae, Arp4 plays an important role in DNA damage repair by recognizing γH2A [74]. Besides a putative role in correct TbSWRC3 positioning, the additional Arp protein in the TbSWR1 complex might be a γH2A recognition module. Given that H2A.Z was shown to be actively incorporated at DNA double strand-breaks [71], the TbSWR1 complex might be directly recruited to sites of DNA damage. As T. brucei lacks non-homologous end joining (NHEJ), H2A.Z might be required to mark areas around DSBs as nucleosome free regions to expose the DNA for the repair process by homologous recombination or micro-homology end joining [75].

In course of investigations that focused on Tb927.9.2910, a protein annotated as a NuA4 subunit, we were able to identify a potential protein complex assembled around HAT1 and HAT3. Co-IP experiments with Tb927.9.2910 identified three proteins that could be linked to the NuA4 complex after Phyre2 modelling: Tb927.1.650 (an Eaf3/MORF4 homologue), Tb927.7.5310 (a Yaf9/GAS41 homologue) and Tb927.10.14190 (a homologue of Epl1). A reciprocal co-IP with the Eaf3/MORF4 homologue could also precipitate HAT1 (but not HAT3) and the three proteins homologous to NuA4 subunits (S1B Table). These data indicate the presence of at least one HAT complex with similarities to a NuA4 complex in T. brucei (S1K Fig).

In contrast, the T. brucei HAT2 complex composition appears to be significantly different compared to the NuA4/TIP60 complex in S. cerevisiae [76]. We could show that Bdf3, a factor that has been shown to locate at the TSSs [77] is involved in H2A.Z incorporation as a part of the HAT2 complex (Figs 4 and S1J Fig), but the function of the other novel complex components remain elusive. We identified two hypothetical proteins without conserved domains and three proteins, which either possess an ENT domain, an FHA domain, or a BTB domain (S1A Table). BTB domains and the ENT domain are associated with chromatin structure regulation [78–80]. None of trypanosome factors appear to have homologues in the NuA4/TIP60 complex [76]. FHA domains are known for recognition of phosphorylate threonine. Phospho-threonine can be found on a broad range of proteins, including kinases, transcription factors and RNA-binding proteins [81, 82]. The presence of an FHA-domain could hint towards phosphorylation events that are important for H2A.Z incorporation. Based on our data, we cannot assess the nature of a potential molecular interaction between HAT2 and the TbSWR1 complex. Interestingly, proximity labelling assays with TbSWRC2 showed that both complexes are in close proximity to each other. Given that the FHA and EMSY domain could be linked with DNA damage repair and chromatin remodelling, it will be interesting to analyse the HAT2 complex structure and the function of its different subunits to learn more about its role in H2A.Z incorporation and other nuclear processes.

Transcription regulation by the TbSWR1 complex

H2A.Z is involved in transcription regulation in many organisms including T. brucei [22, 32, 54]. Northern blot analysis confirmed a loss of RNAPI- and RNAPII-dependent transcripts after the downregulation of TbSWR1 (Fig 7), presumably due to the decrease in H2A.Z incorporation. Consequently, HAT1 loses its substrate and therefore cannot initiate transcription [54]. The observed downregulation of total mRNA and rRNA is not seen for the spliced leader mRNA, which appears to be mainly unaffected by TbSWR1 depletion (Fig 7). This is in agreement with a promotor-dependent and mechanistically different transcription of the spliced leader compared to that of PTUs [83]. The loss of rRNA was surprising, but since a similar loss was observed after RNAP II downregulation (Fig 7), it is likely due to secondary effects. Many genes of RNAP I subunits are encoded in PTUs, which are transcribed by RNAP II. A lack of RNAP II activity might therefore result in a reduced amount of RNAP I transcripts, which finally leads to a reduced amount of rRNA.

Changes in chromatin structure after loss of H2A.Z and RNAP II

The chromatin condensation we observed (Fig 8) might be a direct consequence of the loss of RNAP II activity. It has been described previously that condensation of chromatin can be induced in mouse and human cells by treatment with alpha-amanitin, an RNAP II and RNAP III inhibitor [64, 65, 84]. Alpha-amanitin treatment also blocks RNAP II and III transcription in trypanosomes [85]. Surprisingly, it also affected genes that were under the regulation of an exogenous T3 promoter indicating that open chromatin structure is dependent on transcription activity [85]. Given that a knockdown of TbSWR1, RNAP II and H2A.Z led to a reduction of luciferase activity and mRNA abundance (Figs 5, 6 and 7), it is likely that the chromatin condensates we observed is a consequence of the RNAP II transcription shutdown.

The identification and subsequent characterization of a SWR1-like complex in T. brucei closes the gap in transcription regulation between HAT2 and HAT1, because we identified the remodeller responsible for H2A.Z incorporation.

Interestingly, a protein complex such as INO80, which mediates the removal of histone variants still remains elusive in T brucei. Despite the fact that RuvB1 and RuvB2 are shared subunits of the INO80 and the SWR1 complex in other organisms [68], there is no hint of an INO80 complex in our co-IP data. Given the high degree of conservation of the two SNF2 ATPase complexes between different species, a RuvB co-IP should have identified at least a second SNF2-like protein. In the absence of an INO80 complex and with trypanosome-specific subunits the TbSWR1 complex could have a dual function, also facilitating H2A.Z removal. Alternatively, locus-specific removal of H2A.Z might not be necessary considering the constitutive transcription initiation in T. brucei.

In addition to TbSWR1 only one additional chromatin remodelling complex, the ISWI complex, has so far been identified in T. brucei [86]. As we have shown here, epigenetic complexes in trypanosomes may have evolved differently compared to yeast and humans. It is thus crucial to identify other chromatin-remodelling complexes to understand fundamental biological processes in trypanosomes such as DNA repair and recombination, chromatin dynamics during developmental differentiation or monoallelic transcription of surface proteins.

Trypanosoma brucei culture

Monomorphic Trypanosoma brucei Lister 427 bloodstream form (BSF) MiTat 1.2 (clone 221a) strain, and a derivative ‘2T1’ strain [87] that contains a puromycin-tagged ribosomal spacer for directed integration of the RNAi construct and expresses a Tet repressor protein, were cultivated in HMI-9 medium [88] with 10% heat-inactivated fetal calf serum (FCS; 56°C for 1 h) at 37°C and 5% CO2. Strains were cultured with their attendant drug selection with the following concentrations used: 2.5 μg/ml G418 (neomycin), 5 μg/ml hygromycin, 0.1 μg/ml puromycin, 5 μg/ml blasticidin, 2.5 μg/ml phleomycin. RNAi was induced with 1 μg/ml tetracycline. Reduction of mRNA was monitored by RNA/FISH (S1M Fig). Growth rates were monitored for 96 h and cell densities were determined every 24 h using a Coulter Counter Z2 particle counter (Beckman Coulter). Procyclic forms (PCF; strain 427) were cultured in modified SDM-79 with 10% heat-inactivated FCS (Sigma) at 27°C. BSF and PCF parasites were transfected as previously described [89], with independent clones obtained by limiting dilution.

MNase-ChIP sequencing

The MNase-ChIP experiments were performed with 2T1 cells and the 2T1 TbSWR1 RNAi cell line with a single Ty1-tagged H2A.Z allele. Functionality of the Ty1-tagged H2A.Z has been described in [54]. In brief, 2 × 10⁸ cells were harvested, crosslinked in 1% formaldehyde, and subsequent lysed using 200 μM digitonin (final concentration). Chromatin was fragmented by incubating the cells with 4 U μl⁻¹ MNase (Sigma-Aldrich) for 10 min at 25°C. For a detailed ChIP protocol see ref. [90]. Immunoprecipitation was performed using Dynabeads M-280 sheep anti-mouse coupled to 10 μg monoclonal, purified anti-Ty1 (BB2) mouse antibody [91], overnight (~16 h) at 4°C in the presence of 0.05% (final concentration) sodium dodecyl sulfate (SDS). Immunoprecipitated material was washed with RIPA-Buffer (50 mM HEPES-KOH (pH 7.5), 500 mM LiCl, 1 mM EDTA, 1% (vol/vol) IGEPAL CA-630, 0.7% (wt/vol) Na-Deoxycholate. LiCl and Na-Deoxycholate dissolved separately in, mix and add remaining components. Store at 4°C.) and eluted with 200μl elution buffer (50 mM Tris-HCl (pH 8.0), 10 mM EDTA, 1% (wt/vol) SDS) at 65°C for 30 min. Cross-links were reversed at 65°C for ~16 h in the presence of 300 mM NaCl (final concentration). 4 μl of 10 mg/ml RNaseA were added to the ChIP sample and incubated at 37°C for 2 h. 4 μl of 10 mg/ml proteinase K were added to the ChIP sample and incubated at 55 C for 2 h in a heat block. The tubes were centrifuged at 10,000 x g for 10 min at RT. DNA was purified with the Macherey & Nagel NucleoSpin Gel and PCR Clean-up kit (the NTB buffer was used instead of the NTI buffer due to the high SDS concentration within the samples). Sample purification was performed according to the manufacturer’s instructions. The ChIP sample was eluted with 16 μl and the input sample with 26 μl of NE buffer. The DNA library preparation was performed using NuGEN´s Ovation Ultralow System V2 (M01379 v5). Libraries were prepared with a starting amount 7ng of DNA and were amplified in 8 PCR cycles. Libraries were profiled with a 2100 Bioanalyzer (Agilent technologies) and quantified using the Qubit dsDNA HS Assay Kit, in a Qubit 2.0 Fluorometer (Life technologies) set to high sensitivity. All 14 samples were pooled in equimolar ratio and sequenced on a NextSeq500 High output Kit, PE for 2x 42 cycles plus 8 cycles for the index read. The ChIPseq metadata is available at https://www.ncbi.nlm.nih.gov/bioproject/PRJNA744383

Bioinformatics analysis

Differential binding analysis

Further filtering and an exploratory analysis was performed in an R framework including ggplot2 [100]. Overall experimental quality was assessed via IP strength [101]. Differential binding comparisons were performed with DiffBind package [102] and differential peaks were selected with a 5% FDR. Finally, differential peaks were functionally annotated with ChIPseeker package [103] were a -3000/ + 3000 region around the TSS was considered as promoter.

Extraction of chromatin-associated proteins

The amount of incorporated H2A.Z was determined by Western blot analysis after fractionation of the cells. H3 was used as a control. We followed the extraction protocol described by Kraus et al. [54]. Analysis of chromatin-associated proteins was performed in 2T1 cells with a tagged Ty1-H2A.Z allele. Cells were harvested by centrifugation (4°C, 1500 × g for 10 min) and washed in 1 ml of trypanosome dilution buffer (TDB; 5 mM KCl, 80 mM NaCl, 1 mM MgSO₄, 20 mM Na₂HPO₄, 2 mM NaH₂PO₄, 20 mM glucose, pH 7.4) followed by an additional centrifugation (4°C, 1500 × g for 10 min). The cell pellet was solubilized in CSK-buffer (100 mM NaCl, 0.1% Triton X-100, 300 mM Sucrose, 1 mM MgCl₂, 1 mM EGTA, 10 mM PIPES (pH 6.8; with NaOH) supplemented to contain 1x concentration of Roche cOmplete Protease Inhibitor Cocktail EDTA-free) and incubated for 10 min on 4°C. To separate the soluble from the insoluble fraction, the suspension was centrifuged (4°C, 2550 × g for 5 min). The supernatant was removed and the pellet was resuspended with CSK-buffer and centrifuged (4°C, 2550 x g for 5 min). The pellet with the chromatin fraction was resuspended in Laemmli buffer supplemented with 2.5% β-mercaptoethanol and protease inhibitor cocktail. Proteins were denatured at 90°C for 10 min. For detection of H2A.Z and H3 a polyclonal affinity-purified H2A.Z rabbit antibody (kindly provided by Nicolai Siegel [54]) and a polyclonal H3 rabbit antiserum were used [91], respectively.

Co-immunoprecipitation

Prior to the immunoprecipitation (IP), 30 μl of Protein G Sepharose Fast Flow beads (GE Healthcare) were washed with 1 ml phosphate-buffered saline (PBS) (4°C, 1000 x g for 1 min) and twice in PBS/1% bovine serum albumin (BSA). Unspecific binding sites were blocked by incubation with PBS/1% BSA for 1 h at 4°C on an orbital mixer. The beads were pelleted by centrifugation (4°C, 500 x g for 1 min), supernatant removed, and the antibody diluted in PBS added. Beads and antibody were incubated overnight at 4°C. Unbound antibody was removed by washing 3x in 1 ml PBS/0.1% BSA. Before adding the lysate for immunoprecipitation, the beads were washed with 1 ml IP-Buffer buffer (150 mM NaCl, 0.5% IGEPAL CA-630, 20 mM Tris–HCl, pH 8.0, 10 mM MgCl₂, 1 mM dithiothreitol (DTT), protease inhibitor cocktail (Roche cOmplete)) and subsequent centrifugation (at 4°C, 1000 x g for 1 min). Per IP, 1x10⁸ cells were harvested and washed with ice-cold TDB (BSF), or with PBS (PCF), and incubated with 1 ml IP buffer for 20 min on ice. Cells were lysed by sonication (5 cycles, each 30 s on and 30 s off) using a Biorupter (Diagenode). A centrifugation step (at 4°C; 20.000 x g for 15 min) followed to separate the soluble from the insoluble fraction. The soluble fraction was then added to the Protein G sepharose beads coupled to either HA 12CA5 mouse monoclonal IgG (Sigma Aldrich) or anti-Ty1 (BB2) mouse monoclonal IgG [91] and incubated at 4°C for 3 h. Beads were washed two times with 1 ml IP buffer for 10 min at 4°C. Proteins were eluted by incubating the beads in 50 μl sample buffer (1xNuPAGE LDS Sample buffer with 100 mM DTT) at 70°C for 10 min. Eluates were then analysed by mass spectrometry (MS).

Label-free quantitative mass spectrometry analysis

Samples were run on a Novex Bis-Tris 4–12% gradient gel (Thermo) with ‘3-(N-morpholino) propanesulfonic acid’ (MOPS) buffer (Thermo) for 10 min at 180 V. The gel was stained with Coomassie blue G250 dye (Biozym) prior to cutting each gel lane into pieces, the gel lanes were minced and destained in 50% EtOH/water. The gel pieces were dehydrated with pure acetone, reduced with 10 mM DTT (Sigma Aldrich) and alkylated with 55 mM iodoacetamide (Sigma Aldrich) in the dark. The dried gel pieces were rehydrated with 1 μg trypsin for an in-gel digestion overnight at 37°C. On the following day the digested peptides were desalted and stored on StageTips [104] for further analysis. Using a C18 reverse phase column that was previously packed in-house with Reprosil C18 (Dr. Maisch GmbH) the peptides were separated along a 105 min gradient using an EasyLC 1000 UHPLC system. The column was enclosed into a column oven (Sonation) and peptides were sprayed into a Q Exactive Plus mass spectrometer (Thermo), which was operating in a data-dependent top 10 acquisition mode. Spray voltage was set to approximately 2.4 kilovolt (kV). The acquired raw files were processed with MaxQuant (version 1.5.8.2) [105] using the Trypanosoma brucei protein database downloaded from TriTrypDB and activated LFQ quantitation. Contaminants, reverse hits and protein groups that were only identified by site and protein groups with less than two peptides (one of them unique) were removed prior to bioinformatics analysis. For enrichment, the median of the log2 LFQ intensity values of the replicates was calculated and the p-value was determined by a Welch t-test between the IP and the control sample. The volcano plot was plotted using the R environment. The MS proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the data identifier PXD026796. The MS data are summarized in S2 Table.

Western blot analysis and antibodies

Protein extracts of 3x10⁶ to 8x10⁶ cells were separated on 15% sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) gels and transferred onto polyvinylidene difluoride (PVDF) membranes. Membranes were blocked in 5% milk powder in PBS at 4°C overnight. Primary antibodies were applied in PBS/1% milk/0.1% Tween-20 solution for 1 h at RT. After three washes with PBS/0.1% Tween 20, IRDye 800CW- and 680LT-coupled secondary antibodies (LI-COR Biosciences) were used to detect the corresponding primary antibodies. Secondary antibodies were incubated in PBS/1% milk/0.1% Tween20/0.02% SDS solution for 1 h at RT. After the incubation, blots were washed with PBS/0.1% Tween 20. Blots were analysed using a LI-COR Odyssey Imager (LI-COR Biosciences). Images were quantified with the Image Studio Lite Software. Background subtraction was performed by defining areas for normalization above and below the measured area. The polyclonal anti-TbH3 rabbit antibody was described in [106]. The monoclonal mouse BB2 anti-Ty1 antibody was derived from a hybridoma cell line and described in [91]. The monoclonal mouse anti-HA 12CA5 antibody was obtained from Sigma Aldrich.

Cell cycle analysis

5x10⁶ BSF cells were centrifuged (at 4°C, 1.500 x g for 10 min) and washed once with 5 ml ice-cold TDB. After subsequent centrifugation (at 4°C, 1.500 x g for 10 min) the cells were resuspended in 1 ml ice cold PBS/2 mM EDTA and fixed by dropwise adding 2.5 ml ice cold 100% ethanol or methanol. Cells were fixed at 4°C for 1 h. After centrifugation (at RT, 1.500 x g for 10 min) cells were washed with 1 ml PBS/EDTA, centrifuged (at RT, 1.500 x g for 10 min) and resuspended in 1 ml PBS/EDTA. To the suspension 1 μl RNase (10 μg/μl) and 10 μl propidiumiodide (1 μg/μl) were added and incubated for 30 min at 37°C. Samples were stored at 4°C in the dark until analysis with a BD FACSCalibur using the FL-2 detector channel.

Live/dead staining

1x10⁶ BSF cells were centrifuged (at 4°C, 1.500 x g for 10 min) and washed twice with 1 ml ice-cold TDB (stored on ice). After each washing step, the cells were centrifuged (at 4°C, 1.500 x g for 10 min). The cells were resuspended in 400 μl TDB and incubated with 1 μl propidium iodide (1 mg/ml with a final concentration of 2.5 μg/ml) for 10 min on ice, in the dark. After the staining, the cells were analysed with a BD FACSCalibur.

Luciferase assay

Luciferase assays were performed using the dual luciferase assay system (Promega). 1x10⁶ cells were centrifuged (1500 x g for 10 min at 4°C) and washed with ice-cold PBS. After centrifugation (at 4°C, 1.500 x g for 10 min) the supernatant was discarded and the cells were resuspended in 155 μl passive lysis buffer. The 155 μl were immediately transferred into a transparent flat bottom 96-well plate. 55 μl of stop&glow solution were added in the dark and the lid was covered to prevent light incidence. The plate was incubated for 2 minutes at RT. The samples were analysed with an Infinite 200M plate reader (TECAN). Each well was measured for a duration of 1s.

Northern blot analysis

Northern blots were done as previously described [107]. mRNA was prepared with the RNeasy kit (Qiagen). 18S rRNA and 5.8S rRNA were detected with antisense oligos coupled to IRDye 800, namely 5’-CCTTCGCTGTAGTTCGTCTTGGTGCGGTCTAAGAATTTC-3’ and 5’-ACTTTGCTGCGTTCTTCAACGAAATAGGAAGCCAAGTC-3’, respectively. Total mRNA and SL RNA were detected by an oligo antisense to the mini-exon sequence (5’-CAATATAGTACAGAAACTGTTCTAATAATAGCGTT-3’), coupled to IRDye 700. Blot images were obtained with the Odyssey Infrared Imaging System (LI-COR Biosciences) and quantified with the Image Studio Lite Software. Background subtraction was performed by defining areas for normalization above and below the measured area (for 5.8S rRNA) or manually by defining a square in between the lanes (total mRNA).

EM sample preparation and imaging

The EM sample preparation protocol can be found in [108] and was adapted from [109–111]. 3x10⁷ BSF (bloodstream form) cells were centrifuged (750 g, 3 min, RT). All but 2 ml medium were removed and 2 ml heat-inactivated fetal calf serum was added as a cryoprotectant. Cells were centrifuged (750 g, 3 min, RT) and the pellet was transferred to a polymerase chain reaction (PCR) tube and further compacted (10 s, minifuge). A drop of the final pellet (around 1.5 μl) was transferred to the freezing container (specimen carriers type A, 100 μm, covered with specimen carriers type B, 0 μm, Leica Microsystems). High pressure freezing was done in an EM HPM100 (Leica Microsystems) at a freezing speed >20 000 Ks⁻¹ and a pressure >2100 bar. The samples were stored in liquid nitrogen until freeze substitution in an EM AFS2 freeze substitution system (Leica Microsystems).

For embedding in Epon, samples were incubated in 0.1% (w/v) tannic acid and 0.5% (v/v) glutaraldehyde in anhydrous acetone at −90°C for 96 h (with one change in solution after 24 h), washed four times for 1 h with anhydrous acetone at −90°C and fixed in 2% OsO₄ (w/v) in anhydrous acetone at −90°C for 28 h. Then the temperature was gradually raised to −20°C within 14 h, kept at −20°C for 16 h and gradually raised to 4°C within 4 h. Afterwards samples were immediately washed with anhydrous acetone at 4°C four times at 0.5 h intervals, followed by gradually increasing the temperature to 20°C within 1 h. Subsequently, samples were transferred for embedding into increasing concentrations of Epon (50% Epon in acetone for 3 h at room temperature, 90% Epon in acetone overnight at 4°C, followed by two times 100% Epon at room temperature for 2 h, all solutions were freshly prepared). Epon infiltrated samples were polymerized for 72 h at 60°C.

For staining and contrasting, Epon-embedded sections were incubated in 2% aqueous uranyl acetate for 10 min followed by incubation in Reynolds lead citrate for 5 min. LR White-embedded sections were incubated in 2% aqueous uranyl acetate for 5 min followed by incubation in Reynolds lead citrate for 1.5 min. A 200 kV JEM-2100 (JEOL) transmission electron microscope or a 30 kV JEOL JSM-7500F Scanning Electron Microscope equipped with a TemCam F416 4k x 4k camera (Tietz Video and Imaging Processing Systems) was used for imaging.