As previously reported, the versions of TcBDF3 were purified from inclusion bodies (Alonso et al., 2016). Briefly, the plasmids pDEST17-TcBDF3, pDEST17-TcBDF3m, and pDEST17-TcBD3 were transformed into Escherichia coli BL21, and the recombinant proteins (fused to a His tag) were obtained by induction with 0.1-mM isopropyl-β-d-thiogalacto-pyranoside (IPTG) overnight at 22°C. Protein was purified from inclusion bodies and solubilized proteins were dialyzed against 0.1-mM phosphate buffer pH 8. TcBD2 was purified under soluble conditions, as reported previously from E. coli BL21 transformed with pDEST17-TcBD2. The recombinant protein was obtained by induction with 0.5-mM IPTG for 3 h at 37°C (Tavernelli et al., 2024). The recombinant protein was purified using a nickel-charged resin (Ni-NTA Agarose, Qiagen) following the manufacturer’s instructions. Correct folding was verified by circular dichroism spectroscopy using a spectropolarimeter (Jasco J-810, Easton, MD, USA).

96-well black microplates with 10-μM recombinant proteins (TcBDF3, TcBDF3m, TcBD3, and TcBD2) and three concentrations of each compound were used (5, 10, and 20 μM). In this assay, A1B4 and iBET-151 (positive control) were used. All determinations were made from the top with an excitation wavelength of 275 and a 340/30 nm emission filter, 50 reads per well, and a PMT-sensitivity setting of 180 in a Synergy HT multidetection microplate reader equipped with time-resolved capable optics. Each condition was tested in triplicates, and normalized maximum fluorescence intensityvs. Concentration was plotted using GraphPad Prism 9.0. The relative fluorescence intensity obtained to the condition without the inhibitor was normalized. Compounds that showed concentration-dependent fluorescence quenching, selecting a cut-off value of 0.9 normalized units, were considered positive binders.

The proteins were buffered in 10-mM HEPES, pH 7.5, and 500 mM NaCl and assayed in a 48-well plate at a final concentration of 2 μM in a volume of 20 μL. Compounds were added at a final concentration between 1 and 200 μM to calculate dissociation constants. In this assay, A1B4, A1B4c, A5B4 and A5B4c were used. SYPRO Orange (Invitrogen) was added as a fluorescence probe at a dilution of 1:1000. The excitation and emission filters for the SYPRO Orange dye were set at 465 and 590 nm, respectively. The temperature was increased with steps of 2°C/min from 25°C to 96°C, and fluorescence readings were taken at each interval in a CFX Opus 96 Real-Time System (Biorad) and analyzed using Maestro Software (Biorad). The melting curves and average melting temperature for each condition were obtained and plotted in temperature vs. concentration graphs in GraphPad Prism 9.0. Plots were fitted using the differential fluorescence scanning (DSF) single binding site equation to obtain the dissociation constants (K_d) in GraphPad Prism 9.0.

FP measurement was performed on a Cary Eclipse spectrophotometer equipped with a polarizer at an excitation wavelength of 488 nm and an emission wavelength of 675 nm in a 3-ml fluorescence cuvette. The essay was set at a final volume of 2 mL, 0.5 μM of BSP-AF488, 100-μM TcBDF3, phosphate pH = 8, glycerol 1%, and DMSO 0.5%. The FP values were plotted against the log of the compound concentrations in a non-linear regression model in GraphPad Prism 9.0 as previously described (Zhang et al., 2002). In this assay, A1B4, A1B4c, and A5B4c were used. Briefly, free and bound probes (BSP-AF488 in the presence of TcBDF3) were included in each replica. Since the intensity of the probe remains similar for all samples, the fraction of the probe bound to TcBDF3 is correlated to the mP value. Thus, the percentage of inhibition can be derived from the equation % inhibition =100[1 − (mP − mPf)/(mPb − mPf)], in which mPf is the free probe control and mPb is the bound probe control. Based on the percentage of inhibition, the inhibitory concentration (IC₅₀), or the inhibitor concentration at which 50% of the bound probe is displaced, is obtained by fitting the inhibition data.

T. cruzi Dm28c that expresses the E. coli LacZ gene epimastigotes, trypomastigotes, and amastigotes were cultured and obtained as previously described (Buckner et al., 1996; Alonso et al., 2021). The compounds were dissolved in DMSO with a final concentration of the solvent below 1% to ensure that DMSO does not interfere with cytotoxicity. The controls included for the amastigotes (intracellular stage) were (a) uninfected cells and (b) infected with 1 μg/mL benznidazole (BZ) (positive control) and 1% DMSO to evaluate the toxicity of the solvent. For epimastigotes and trypomastigotes, the controls used were: 1-μg/ml benznidazole (BZ) (positive control) and 1% DMSO to evaluate solvent toxicity. Results obtained with this assay were verified by manually counting the parasites. The results were expressed as the percentage of inhibition of T. cruzi growth in compound-treated infected cells compared to untreated infected cells for amastigotes. For epimastigotes and trypomastigotes, treated and untreated cultured parasites were compared. Quadruplicates were run in the same microplate and repeated in three experiments. Half-maximal IC₅₀ values were calculated using non-linear regression on GraphPad Prism 9.0, as previously described (Alonso et al., 2021).

The 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) reduction assay determined cell viability after treatment as previously described (García et al., 2018). Optical density (OD) was spectrophotometrically quantified (λ = 540 nm) using a Synergy HT multidetection microplate reader. DMSO was used as blank, and each treatment was performed in triplicates. The 50% cytotoxic concentration (CC₅₀) values were calculated by non-linear regression on GraphPad Prism 9.0, and the selectivity indexes (SI) were determined as the ratio CC₅₀/IC₅₀ in all three life cycle stages.

The initial model of TcBDF3 was generated with AlphaFold2 using ColabFold with the following options: --num-models 5 --num-recycle 10 --rank plddt --num-relax 1 --use-gpu-relax, and the sequence of TcBDF3 from Dm28c (Jumper et al., 2021). A custom multiple sequence alignment with the combination of the alignments produced with the Discoba database and the default alignment of the ColabFold server was used, as described by Wheeler (2021). All models obtained showed high average pLDDT values (93.6, 93.1, 92.7, 92.6, and 92.1) and pTM (0.840, 0.829, 0.828, 0.827, and 0.827). Domain cavity sizes were computed with CAVIAR (Marchand et al., 2021). The structural alignment of the ET domain of TcBDF3 and other ET domains was performed using the ChimeraX matchmaker algorithm guided by Smith–Waterman alignments (Meng et al., 2023).

Docking was done using Glide in extra precision (XP) mode. TcBDF3 structure was obtained from a representative molecular dynamics (MD) simulation snapshot. Ligand docking used Schrödinger software (version 2023-3). Protein preparation employed the Protein Preparation Wizard within Schrödinger software. LigPrep was used to prepare ligands. The grid box of 20 Å was centered in the binding site. The best docking pose of each ligand was selected for further analysis.

Molecular dynamics (MD) simulations were performed using Amber22 software (Case et al., 2023). The starting structure of the protein was obtained using AlphaFold2 (see above), and the structures of the protein with the ligands were taken from the docking studies. Protein and ligand-protein complexes were subjected to the 100 ns MD simulation protocol using the particle mesh Ewald molecular dynamics (PMEMD) CUDA software implemented in Amber22 with the FF14SB force field. Ligand parameters (van der Waals radius, force constants, etc.) were from the gaff database. Partial charges were RESP charges computed using the Hartree–Fock method and 6-31G* basis set (Bayly et al., 1993). The proteins were first immersed in a periodical octahedral box of TIP3P water molecules; structures were then optimized, taken from 0 to 300 K at constant volume for 100 ps, equilibrated at a constant pressure of 1 bar for 200 ps, and finally subjected to 100 ns of production simulation. Temperature control was performed using the Langevin thermostat, and pressure control was performed using the Monte Carlo barostat (Gordon et al., 2009). The binding-free energy of each complex and its decomposition per residue were performed with the Poisson-Boltzmann Surface Area of Molecular Mechanics (MM-PBSA) in AmberTools (Faller and de Pablo, 2001; Wang et al., 2016; Case et al., 2023).

The structural homologous sequences of TcBDF3 were retrieved by combining structural and profile HMM searches with strict cutoffs. First, a PDB file was generated with the extraterminal (ET) domain from TcBDF3 AF2 model, and its atom coordinates were used as a query on the Foldseek search server (van Kempen et al., 2023) against AlphaFold/UniProt50 v4 database (50% sequence identity clusters). This generated a list of all non-redundant proteins that contained an ET fold in its AF2 structure. Only proteins with a probability greater than 0.9 were considered. The HMMER v3.4 search algorithm (Eddy, 2011) was used to detect the Pfam bromodomain (BD) profile HMM PF00439 in the sequences, extracting their starting and end positions and the number of matches (i.e., the number of bromodomains), considering only those with a score bigger than 30; resulting in a list of 307 proteins with at least one BD before ET. No BD was detected at the C-terminal of the ET in any sequence. Then, the last BD of each sequence was concatenated with its corresponding ET, followed by a sequence redundancy filter of 90%, leaving 121 proteins. Each domain was then aligned separately with ClustalO (Sievers and Higgins, 2018) using default settings, followed by removing proteins without complete coverage of the BD sequence and a per-domain realignment. Non-informative positions were removed with ClipKIT (Steenwyk et al., 2020) using the kpic-gappy method. Finally, a maximum likelihood tree was generated using IQ-TREE2 (Minh et al., 2020), with a bootstrap of 1,000, letting ModelFinder find the best substitution model. Tree visualizations were performed with a Dendroscope (Huson and Scornavacca, 2012).

As mentioned in the Introduction, our group previously described A1B4, a compound that binds to TcBDF3 in vitro and displays trypanocidal activity (García et al., 2018). Later, Laurin et al. (2021) reported contradictory results regarding the binding of A1B4 to a recombinant bromodomain portion of TcBDF3 (TcBD3) using waterLOGSY. A possible explanation for the reported difference between Laurin et al.’s results and ours is that they used a different recombinant bromodomain. In contrast, the full-length protein (TcBDF3) was used in our publication. To further evaluate the significance of this difference, a truncated version of TcBDF3 containing only the bromodomain (TcBD3: amino acids 1–151) was used, together with the full-length version to evaluate A1B4 binding using our in-house fluorescence quenching assay (Figure 1). As an additional control experiment, the binding of A1B4 to a mutated version of TcBDF3 (TcBDF3m) that includes two-point mutations (Y123A and L130A) was measured. TcBDF3m was designed to retain its secondary structure while losing its ability to bind to its acetylated ligand in vitro (Alonso et al., 2016). Furthermore, A1B4 binding to the recombinant bromodomain of TcBDF2 (TcBD2) was measured to explore selectivity. TcBDF2 is a nuclear bromodomain-containing protein that recognizes acetylated histone H4, H2B, and H2B.V in T. cruzi (Villanova et al., 2009; Pezza et al., 2022). A1B4 did not bind to TcBDF3m or TcBD2, indicating high selectivity for the bromodomain portion of TcBDF3 (TcBD3) (Figure 1). As a positive binding control, iBET-151 was used, a BET bromodomain inhibitor previously reported to bind to TcBDF3 (Alonso et al., 2016). This assay validated our previous results and confirmed that A1B4 binds to the bromodomain of TcBDF3 with specificity.

In our previous study, the dissociation constant (K_d) reported for A1B4 and TcBDF3 was 1.7 μM calculated using DSF, also known as thermal shift (Table 1). To further validate the parameters obtained, a Fluorescence Polarization (FP) assay recently described by our group for TcBDF2 was used (Tavernelli et al., 2024). FP is a sensitive technique that allows quantitative analysis of the interaction between two molecules. This assay employs a fluorescent probe (AlexaFluor-488) linked to bromosporine (BSP-AF488). Bromosporine is a mammalian bromodomain pan inhibitor, which has been reported to bind to TcBDF3 and TcBDF2 (Alonso et al., 2016). First, the dissociation constant of BSP-AF488 and TcBDF3 was determined and a working window of approximately 140 mP was obtained, which was necessary to proceed with the binding assay (Figure 2A). The FP was then measured using a fixed concentration of TcBDF3 and increasing amounts of A1B4 to obtain a binding activity of 2.4 μM for A1B4, in agreement with the value previously calculated by DSF (Table 1) (Figure 2B). These results confirmed the binding of A1B4 to TcBDF3 and validated this small molecule as a starting point for the search for new 1,3,4-oxadiazoles as parasitic bromodomain inhibitors. Second, a structural analysis of TcBDF3 and TcBD3 was performed in silico to explain the differences reported in the literature.

To investigate the discrepancies in the binding parameters reported for the full-length and truncated versions of TcBDF3, we generated a reliable structural model of TcBDF3 using AlphaFold2, enhanced by the Discoba database. This database was specifically developed to improve the quality of predictions on protein structure within the phylogenetic group that includes trypanosomatids (Figure 3A; Wheeler, 2021). To evaluate the quality of the predicted structure, the pLDDT values per residue were plotted, and high values across the entire protein length for all the models were observed, both over the BD region and the ET region (Figure 3B, left panel). In addition, the high pTM and low domain error values in the PAE matrix graphed in Figure 3B (right panel) indicated a compact globular structure. When the 3D structure for the full-length and the truncated protein was simulated using AlphaFold2 it was observed that the bromodomain portion of TcBDF3 has a slightly less hydrophobic pocket with a more extensive cavity (142 ų) compared to the pocket of the isolated domain (132 ų) (Figure 3C). This could be explained by a conserved α-helix formed by the first 32 amino acids of the N-terminal portion that interact tightly with the BD or ZA loops, which can impact its mobility. The fact that the inner structure is not precisely conserved when the bromodomain is expressed in isolation could explain, in part, the discrepancies in binding observed in the literature (Laurin et al., 2021). It is worth mentioning that, although in the same order of magnitude, the observed binding parameters for A1B4 for the truncated version are slightly higher than that of the complete protein (Supplementary Figure S1).

With these results in hand, the validated methodology was applied to three other compounds related to A1B4: hydrazone A5B4 and 1,3,4-oxadiazoles A1B4c and A5B4c. Hydrazone A5B4 was included as a control because it contains the thiophene portion observed in A1B4 and a furane ring observed in another TcBDF3 binder previously discovered by our group (Ramallo et al., 2018). Although A5B4 was present in the dynamic library that led to the discovery of A1B4, it was not amplified in the presence of TcBDF3 (García et al., 2018). The other 1,3,4-oxadiazoles are more rigid versions of their hydrazone precursors, A1B4 and A5B4, respectively. Furthermore, 1,3,4-oxadiazoles represent an engaging heterocyclic platform found in products that show a wide range of biological properties (Gao and Wei, 2013; Chauhan et al., 2023).

Hydrazone A5B4 was prepared from the corresponding acyl hydrazide and aldehyde in the presence of TFA. 1,3,4-oxadiazoles A1B4c and A5B4c were synthesized from the corresponding hydrazones, by oxidative cyclization with chloramine T or (diacetoxyiodo)benzene, respectively (see Methods section). A comparison of the binding parameters for the hydrazones A1B4 and A5B4 (Table 1) enables us to explain why the latter compound was not amplified in our previous study while A1B4 was. The affinity of A5B4 affinity for TcBDF3 (K_d = 136 μM) is significantly lower than that of A1B4 (K_d = 1.7 μM). The cyclic compounds gave slightly higher dissociation constants but in the same order of magnitude as A1B4 using both DSF and FP (Table 1; Supplementary Figure S2). It should be noted that cyclization of A5B4 led to a significantly improved affinity for TcBDF3 (Table 1). The results were of the same order of magnitude for the full-length and truncated version of TcBDF3 (Supplementary Figure S1).

To evaluate the in vivo effect of the new compounds, the cytotoxic effect against the three life cycle stages of T. cruzi (epimastigotes, trypomastigotes, and amastigotes) in vitro, and a mammalian cell line was determined for all of the small molecules. For these in vivo cell assays, a T. cruzi Dm28c strain expressing the E. coli β-galactosidase as a reporter gene was used (Alonso et al., 2021). Epimastigotes, trypomastigotes, or amastigotes obtained 48 h postinfection of Vero cells monolayers, were incubated with the compounds. As shown in Table 2, A1B4c has a cytotoxic effect of <10 μM in all life cycle stages, improving the parasitic activity of A1B4. A5B4c was more active than A1B4 but to a lesser extent than A1B4c. Depending on the stage, the selectivity indexes (SI) obtained ranged between 11 and 54 for A1B4c, and between 21 and 37 for A5B4c, thereby improving upon those obtained for A1B4 (Table 2).

To complement the activity studies, molecular docking was employed to understand the binding of A1B4, A1B4c, and A5B4c to the bromodomain at the atomistic level. Figure 4 shows the docking results obtained, which determine that the three ligands are located inside the hydrophobic pocket of the bromodomain (responsible for the Ac-Lysine binding) interacting with Val70, Leu75, Tyr80, Ile116, Asn119, Cys120, Tyr123, and Asn124 of TcBDF3. The ligands are detected with the thiophene group toward the outside of the protein. The position of the thiophene group is vital in achieving sulfur-π interactions with Tyr80, and this type of interaction is crucial for the binding processes. The alcohol group of the ligands is placed toward the inside of TcBDF3 and allows interaction with Asn119, Tyr123, and Asn124 (Figure 4).

Furthermore, the TcBDF3-ligand structures obtained from the docking studies were subjected to molecular dynamics (MD) simulations to evaluate the conformation of the complexes as a function of time. Results show that the three ligands remained in the cavity and maintained the interactions described above (Supplementary Figure S3). In addition to the temporal evolution of the complexes, the MD simulations can provide thermodynamic data, so the binding energies for the three complexes were calculated. A more favorable binding for A1B4 compared to A1B4c and A5B4c (ΔG [A1B4/A1B4c] = 4.52 kcal/mol, ΔG [A1B4/A5B4c] = of 5.97 kcal/mol) was estimated. These results correlate with the dissociation constants obtained by DSF and FP. The binding energy was further decomposed into the contribution of each residue, and the results are presented in Supplementary Figure S4, where only the residues of the binding site were considered. In all three cases, the contribution to the binding energy follows the same pattern and is consistent with the binding mode analyzed by the docking studies.

According to Laurin et al., one possible explanation for the lack of binding observed for A1B4 to a truncated version of TcBD3 that does not contain the N-terminal α-helix is that A1B4 could be binding to a second site of TcBDF3, distinct from the hydrophobic pocket of the BD. To test this hypothesis, we explored the TcBDF3 modeled structure in search of additional binding sites using CAVIAR, finding one additional smaller pocket in the C-terminal half of the protein (Supplementary Figure S5). The size of this cavity (101 ų) and the fact that A1B4 and the two cyclic derivatives displaced the bromodomain-specific bromosporine probe in the FP assay using both full-length and our C-terminal truncated version make us dismiss the possibility that they are interacting with this alternative domain. Moreover, docking simulations do not predict the insertion of any of the compounds into the cavity of this alternative domain. Given that our recombinant TcBD3 version contains the conserved N-terminal α-helix, the discrepancy in binding affinity observed by Laurin et al. could be attributed to the absence of this structural element in their truncated protein. This N-terminal α-helix likely contributes to stabilizing the bromodomain structure, mainly through stabilization of the ZA loop, influencing the binding pocket’s conformation and, consequently, the affinity of a small molecule ligand such as A1B4. Since most BD research centers on the development of inhibitors using isolated BDs, it is possible that removing specific regions surrounding the domain could affect the binding affinity. Therefore, it might be advantageous to keep specific non-BD areas, especially those that interact with the ZA or BC loops, as these elements may play a crucial role in stabilizing the BD structure and influencing the ligand binding.

On closer inspection, it was observed that the C-terminal region has an extraterminal (ET) domain fold, found in the BET (bromodomain and extraterminal Domain) family of bromodomains (Taniguchi, 2016). The structural alignment between the ET domain of TcBDF3 and the ET domain of human BRD2 previously characterized is shown in Supplementary Figure S5 with their HHpred values. In addition, the TcBDF1 structure was predicted to have an ET fold (data not shown). These observations suggest that TcBDF1 and TcBDF3 could be part of this family.

To explore the phylogenetic relationship between TcBDF3 and other BET proteins, homologous sequences of TcBDF3 were recovered by combining structural and profile HMM searches with strict cut-off values (Supplementary Figure S6A). The final sequence dataset consisted of 121 non-redundant proteins from the four eukaryotic kingdoms with either 1 or 2 BDs N-terminal to the ET domain (Supplementary Table S1). This set of sequences contains previously undescribed BET proteins from protists and well-characterized mammalian BET proteins. Phylogenetic analysis of BET proteins revealed a distinct pattern in the number of bromodomains present in different major taxonomic groups (Figure 5). Specifically, plant and protist BET proteins consistently exhibit a single bromodomain, while their fungal and metazoan counterparts display variation, some possessing one or two bromodomains. Also, the distance from the end of the last BD to the start of the ET, typically separated by a primarily disordered region, was measured for each protein (Supplementary Figure S6B). The distance was shorter in protists, plants, and fungi compared to metazoans, with TcBDF3 having the shortest distance. This suggests that mammalian 2BD-ET-type BDFs evolved from a single BD-ET ancestor through the duplication of the BD. This duplication was associated with an increased protein size due to the elongation of the disordered region between the BD and the ET motif.

In conclusion, the affinity of compound A1B4 for TcBDF3 was confirmed by two different methods. In contrast, it did not bind to TcBDF3m, designed to retain its secondary structure while losing its ability to bind to its acetylated ligand, nor the recombinant bromodomain of TcBDF2. In addition, two new compounds related to A1B4 that bind to TcBDF3 were described. Simulations indicate that the compounds can be inserted in the hydrophobic pocket, and they do not support its binding to an alternative domain within TcBDF3. The two new active compounds, identified as 1,3,4-oxadiazoles, demonstrated enhanced trypanocidal activity, particularly against trypomastigotes (the infective stage) and amastigotes (the intracellular stage). Interestingly, prior research by different authors identified 1,3,4-oxadiazoles as inhibitors of human CREBBP bromodomains using a fragment-based high-throughput docking method (Xu et al., 2016). However, our study marks the first instance of this functional group acting as a parasitic bromodomain inhibitor.

An ET domain, which is not reported yet in TcBDF3, has been identified. This domain is also present in TcBDF1, its trypanosomatids orthologs, and various BDFs from protists and plants. These proteins represent highly divergent members of the BET family, where they possess only one bromodomain instead of the typical two found in highly studied BET proteins. Phylogenetic analysis suggests that the 1BD-ET domain architecture was present in ancient eukaryotic cells. On the contrary, the 2BD-ET architecture is associated with more complex organisms such as metazoans. This taxonomic divergence in BD composition suggests evolutionary divergence and potentially distinct functional roles of BET proteins across kingdoms. The fact that 1,3,4-oxadiazoles have previously been described as inhibitors of CREBBP but not BET bromodomains reinforces this idea.

On the other hand, the function of 1BD-ET may also differ among protists. The components of multiprotein nuclear complexes containing BDs in trypanosomatids are only beginning to be identified (Staneva et al., 2021; Jones et al., 2022; Vellmer et al., 2022). The Conserved Regulators of Kinetoplastid Transcription (CRKT) Complex, described in Leishmania but also present in T. brucei, include the double BD protein BDF5 along with BDF3 and BDF8. In this report, BDF1 and BDF4 were also detected in the proximity proteome of BDF5 (Jones et al., 2022). The unique features of these species result in a “polybromo complex.” Although the CRKT complex is essential, there may be redundancy in the function of bromodomains within the complex, making it more difficult to identify compounds that bind to these BDs with effective antiparasitic activity. In T. cruzi, the situation is different. Our group demonstrated that TcBDF1 and TcBDF3 are not nuclear. Furthermore, none were found in proximity labeling experiments using TcBDF5 and TcBDF8 TurboID fusion proteins (Rodríguez Araya, unpublished results). Furthermore, TcBDF1 and TcBDF3 have different cellular localizations than their homologs in T. brucei indicating different functions in different trypanosomatids (glycosomal and microtubule-associated, respectively).

The essentiality of these proteins is supported by the phenotypes observed after their overexpression as wild-type and dominant negative mutants, and by the fact that complete knockouts for both Tcbdf1 and Tcbdf3 genes could not be achieved using CRISPR/Cas9 genome editing despite several attempts (unpublished results). In this context, TcBDF3 is notable due to its association with the cytoskeleton and a function that is not yet fully understood and does not seem to be shared by any other BD-containing protein. Furthermore, the classification of TcBDF3 as a member of the BET family is significant, given the success of developing BET BD inhibitors to treat cancers and inflammatory conditions (To et al., 2023). Unlike its mammalian counterparts, TcBDF3’s single BD, unique structure, and cytoplasmic localization make it a distinctive target that could be exploited without affecting human BET proteins, thus reducing potential side effects. The success of 1,3,4-oxadiazoles as inhibitors of TcBDF3 further underscores the potential to target this protein in T. cruzi. By focusing on atypical BET bromodomains, we can develop highly selective inhibitors that disrupt the parasite’s lifecycle, offering a novel and effective strategy against this neglected tropical disease with limited treatment options.