If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
# Present address: EMBL Grenoble, 71 avenue des Martyrs, CS 90181, 38042 Grenoble Cedex 9, France.
KarinaB. Sabalette
Footnotes
# Present address: EMBL Grenoble, 71 avenue des Martyrs, CS 90181, 38042 Grenoble Cedex 9, France.
Affiliations
Instituto de Investigaciones Biotecnológicas, Universidad Nacional de San Martín - Consejo Nacional de Investigaciones Científicas y Técnicas, General San Martín, 1650, Prov. de Buenos Aires, ArgentinaEscuela de Bio y Nanotecnologías (EByN), Universidad Nacional de San Martín
Department of Genomics, Instituto de Investigaciones Biológicas Clemente Estable, Av. Italia 3318, Montevideo, CP 11600, UruguayInstituto de Biología, School of Sciences, Universidad de la República, Montevideo, Uruguay
Department of Genomics, Instituto de Investigaciones Biológicas Clemente Estable, Av. Italia 3318, Montevideo, CP 11600, UruguayInstituto de Biología, School of Sciences, Universidad de la República, Montevideo, Uruguay
Instituto de Investigaciones Biotecnológicas, Universidad Nacional de San Martín - Consejo Nacional de Investigaciones Científicas y Técnicas, General San Martín, 1650, Prov. de Buenos Aires, ArgentinaEscuela de Bio y Nanotecnologías (EByN), Universidad Nacional de San Martín
Trypanosomes regulate gene expression mainly by using post-transcriptional mechanisms. Key factors responsible for carrying out this regulation are RNA-binding proteins (RBPs), affecting subcellular localization, translation, and/or transcript stability. Trypanosoma cruzi U-rich RBP 1 (TcUBP1) is a small protein that modulates the expression of several surface glycoproteins of the trypomastigote infective stage of the parasite. Its mRNA targets are known but the impact of its overexpression at the transcriptome level in the insect-dwelling epimastigote cells has not yet been investigated. Thus, in the present study, by using a tetracycline-inducible system, we generated a population of TcUBP1-overexpressing parasites and analyzed its effect by RNA-seq methodology. This allowed us to identify 793 up- and 371 down-regulated genes with respect to the wild-type control sample. Among the up-regulated genes, it was possible to identify members coding for the TcS superfamily, MASP, MUCI/II, and protein kinases, whereas among the down-regulated transcripts, we found mainly genes coding for ribosomal, mitochondrial, and synthetic pathway proteins. RNA-seq comparison with two previously published datasets revealed that the expression profile of this TcUBP1-overexpressing replicative epimastigote form resembles the transition to the infective metacyclic trypomastigote stage. We identified novel cis-regulatory elements in the 3'-untranslated region of the affected transcripts and confirmed that UBP1m -a signature TcUBP1 binding element previously characterized in our lab- is enriched in the list of stabilized genes. We can conclude that the overall effect of TcUBP1 overexpression on the epimastigote transcriptome is mainly the stabilization of mRNAs coding for proteins that are important for parasite infection.
Trypanosomes are interesting models to study unusual mechanisms of gene expression regulation. Unlike most eukaryotes, trypanosomatids lack control at the level of transcription initiation for each individual gene. In contrast, transcription by RNA polymerase II is polycistronic and transcript synthesis initiates from a few sites on each chromosome (
). Owing to these biological constraints, these microorganisms control protein levels mainly by post-transcriptional events. The fate of mRNAs in the cell depends on the set of RBPs associated with them, and these molecular interactions can also be organized into larger mRNP complexes forming stress granules or P-bodies (
Integrative analysis of the Trypanosoma brucei gene expression cascade predicts differential regulation of mRNA processing and unusual control of ribosomal protein expression.
Transcriptome-wide analysis of trypanosome mRNA decay reveals complex degradation kinetics and suggests a role for co-transcriptional degradation in determining mRNA levels.
). Over the years we have contributed, in part, to a better understanding of these mechanisms in Trypanosoma cruzi, an early branching eukaryotic unicellular parasite causing Chagas disease (
Transcriptome-wide analysis of the Trypanosoma cruzi proliferative cycle identifies the periodically expressed mRNAs and their multiple levels of control.
A 43-nucleotide U-rich element in 3’-untranslated region of large number of Trypanosoma cruzi transcripts is important for mRNA abundance in intracellular amastigotes.
). Particularly, the first RNA-seq transcriptome and translatome for this parasite showed that translation regulation plays a critical role in governing gene expression profiles during T. cruzi differentiation (
). At all these regulatory points, RBPs can intervene as crucial trans-acting factors and mediate parasite differentiation in both T. cruzi and T. brucei (21, 22).
The present study focused on T. cruzi U-rich RBP 1 (TcUBP1), one of the first trypanosome RNA-recognition motif (RRM)-containing proteins described. TcUBP1 is an exclusive-trypanosomal RBP having a single RRM (
) with the characteristic β1α1β2β3α2β4 fold. It is expressed in all stages of the parasite life cycle and regulates the abundance of a large number of genes containing U-rich elements (
A 43-nucleotide U-rich element in 3’-untranslated region of large number of Trypanosoma cruzi transcripts is important for mRNA abundance in intracellular amastigotes.
). Some of the ribonucleoprotein complexes containing TcUBP1 are developmentally regulated, as determined by profile expression of target transcripts and RT-PCR analysis of co-immunoprecipitated RNAs (
The ability of T. cruzi to survive in the mammalian host is in part due to the expression of a plethora of surface proteins and signaling genes, which include the trans-sialidase and trans-sialidase like (TcS) superfamily, mucins, and mucin-associated surface proteins, among others (
Genomic analyses, gene expression and antigenic profile of the trans-sialidase superfamily of Trypanosoma cruzi reveal an undetected level of complexity.
). Interestingly, TcUBP1, in synchrony with nutritional deficiency, is known to mediate differentiation of T. cruzi epimastigotes into infective metacyclic trypomastigotes (
). TcS members are surface glycoprotein-coding genes expressed only in trypomastigote forms, but the in vivo interaction of TcUBP1-TcS RNAs occurs in both replicative and infective cells. In this regard, ectopic overexpression of TcUBP1 in replicative forms resulted in >10-fold up-regulated expression of numerous TcS mRNAs and changes in their subcellular localization from the posterior zone to the perinuclear region of the cytoplasm, as is typically observed in the infective stages. This fact has led to the hypothesis that TcUBP1 can promote a switch toward profile expression of infective trypomastigotes in T. cruzi by increasing the mRNA levels and translation rates of an RNA regulon for trypomastigote surface glycoproteins during parasite development. The post-transcriptional paradigm of RNA regulons was first posited by Keene and Lager almost two decades ago (
), and suggests that by recognizing structural and/or sequence RNA elements, cells can co-regulate subsets of transcripts with a shared physiological function.
For an RBP of interest, identifying the in vivo binding sites is a critical step towards understanding its function. However, the complete influence of TcUBP1 overexpression in the determination of the parasite transcriptome is not known and its precise binding sites have not been described. Thus, the aim of the present study was to perform an RNA-seq analysis on epimastigote samples overexpressing UBP1.
RESULTS
Identification of differentially expressed genes after TcUBP1 overexpression
To gain comprehensive insights into the regulatory role of TcUBP1, we analyzed the impact of TcUBP1 overexpression on the T. cruzi CL-Brener transcriptome. For this, TcUBP1-GFP-induced epimastigotes (UBP1-OE) or control wild-type samples (WT) were subjected to RNA-seq analysis (see Experimental Procedures). After assembly and annotation, we identified a total of 9,039 genes (Fig. 1A). The expression levels of each gene of the UBP1-OE and WT populations were calculated by mapping clean read sets onto the reference transcriptome of the CL Brener Esmeraldo-like strain (TriTrypDB-59_TcruziCLBrenerEsmeraldo-like_Genome.fasta). The data from different libraries were normalized using the normalization method in the software package DESeq2 (
Figure 1Hierarchical clustering of differentially expressed genes in UBP1-OE samples compared with wild-type epimastigotes defined by DESeq2 with FDR less than 0.05. Groups on vertical represent the clustered genes based on gene expression, the horizontal line represents the single gene and color of the line indicates the gene expression in UBP1-OE. A) Heatmap and complete linkage clustering using all replicates per group, 5741 genes were clustered. B) Heatmap of 1,164 significant genes with |log2 fold change|>1 with 793 up-regulated and 371 down-regulated genes. C) Most correlated samples (n=20 genes). The Z-score scale bar represents relative expression +/- SD from the mean.
The distribution pattern of transcript expression in UBP1-induced versus WT parasite populations was analyzed in detail. Results showed that 64% of the total genes (5,741) were significantly expressed, with FDR-adjusted p-values lower than 0.05, and that 13% of the genes (1,164) were differentially expressed (|log2 fold change|>1, FDR-adjusted p-value < 0.05; Table S1). The percentages of up- and down-regulated genes in UBP1 tetracycline-induced parasites were 8.8% and 4.1% respectively. In addition, the expression patterns for all genes (A), for the significantly expressed genes (B), and for the most correlated genes (i.e. genes that were found overexpressed in one sample and underexpressed in the other and vice versa) (C), in control and OE parasites are shown in Fig. 1. The light blue, white, and orange colors indicate less expressed, medium-level expressed, and highly expressed genes, respectively (Fig. 1A). By analyzing the complete expression profile of the up- and down-regulated genes, we concluded that the global effect of UBP1-OE is mostly to stabilize the transcriptome, since nearly ∼800 genes were two-fold up-regulated and less than half of the genes were two-fold down-regulated (|log2 fold change| > 1, FDR-adjusted p-value < 0.05) (Fig. 1B).
A Venn diagram was generated to show a representation of the differentially expressed genes mentioned above (Fig. 2A). This included 793 up-regulated and 371 down-regulated genes in TcUBP1-induced samples (Table S1). The number of up-regulated genes in UBP1-OE parasites was two times higher than that of down-regulated genes. In addition, results showed that 33 genes, most of which coded for hypothetical proteins, were expressed exclusively in UBP1-OE samples, and that 11 genes, mostly related to the chromosome organization process, were expressed exclusively in WT parasites (Table S2). As expected, TcUBP1 (TcCLB.507093.220) was 71 times higher in the OE samples (FDR-adjusted p-value = 1.1E-58), showing the largest difference between the OE and WT samples. This value reflected the expected overexpression of TcUBP1 as a consequence of the pTcINDEX induction with tetracycline. A volcano plot of gene expression in UBP1-induced and WT parasites is shown in Fig. 2B, where significantly expressed genes are separated from the non-significantly expressed genes by different color codes. The 20 most statistically significant up- and down-regulated genes are toward the top, labeled with gene symbols together with TcUBP1. Also, the Top 10 list with the most differentially over- or under-expressed genes (based on fold change values) is shown in Table I and also depicted in Fig. 2C.
Figure 2A) Venn diagram representing the total number of differentially expressed genes (DEGs) between UBP1-OE and WT. UBP1 up-regulated (log2 >1, FDR <0.05) or down-regulated (log2 <1, FDR <0.05) genes are shown. B) Volcano plot showing the differential expression analysis of genes in UBP1-OE and wild-type parasites. Black and red dots show non-significant and significant DEGs, respectively. TcUBP1 and most significant up and down-regulated genes are labeled. UBP1, the most expressed gene, is plotted out of scale with and x-value of 6.16 log2 fold change. Up-regulated: TS_1 (TcCLB.504099.50); MASP_1 (TcCLB.507237.190); HYPO1 (TcCLB.507859.31); HYPO2 (TcCLB.507859.15); TS_2 (TcCLB.510005.20); MASP_2 (TcCLB.506965.130); HYPO3 (TcCLB.506227.205); PRENYLS (TcCLB.507879.10); TS_3 (TcCLB.506211.150); TS_4 (TcCLB.511311.20); GLYTRA (TcCLB.510071.30); down-regulated: PORIN1 (TcCLB.511687.10); PEPT (TcCLB.511181.50); ATPase9 (TcCLB.503579.70); UBIQ (TcCLB.511575.120); PORIN2 (TcCLB.504225.20); PORIN3 (TcCLB.511687.19); SMUG (TcCLB.511685.10); HYPO4 (TcCLB.503897.120); HSP10 (TcCLB.508209.100). C) Chart of top-10 most up/down-regulated genes based on log2 fold change.
TcUBP1 overexpression leads to up-regulation of cell-surface trypomastigote glycoproteins and down-regulation of ribosomal and mitochondrial proteins
Gene ontology (GO) analyses using TriTrypDB performed on genes over- and under-expressed in UBP1-OE parasites showed a distribution of 18 and 107 GO overrepresented terms, respectively. The enrichment chart was plotted showing each significant GO term and the percentage of genes present in our differentially expressed genes compared to the background for each category (Fig. 3). The complete distribution is provided in Table S3. A plot for all the three GO domains: biological process, molecular function, and cellular process is presented in Fig. 3A (up-regulation) and Fig. 3B (down-regulation). The GO analysis of differentially expressed genes with significant differences revealed that they are involved in critical biological processes and cellular components, such as pathogenesis, cell adhesion, and protein phosphorylation (in the case of up-regulated genes), and in ribosomes, GTPase activity and mitochondria (in the case of down-regulated genes).
Figure 3Differentially expressed genes enrichment analysis. GO classification of differentially expressed genes, the graph shows up to ten GO terms with most genes annotated. First, second, and third charts indicated GO terms clustered in the biological process (BP), cellular component (CC) and molecular function (MF) terms, respectively. The size of the dots diameter indicates the number of differential genes; color depth indicates significance; abscissa indicates enrichment abundance; and the ordinate indicates different pathways.
DAVID (Database for Annotation, Visualization and Integrated Discovery) enrichment analysis classified all the enriched protein domains into three categories: InterPro, Pfam and Smart. DAVID annotation products were recovered using the online GeneID Conversion tool. Out of 793 genes in the up-regulated group, 791 were accepted by DAVID for the analysis and assigned to nine clusters, whereas all the 371 genes in the down-regulated group were accepted and assigned to eight clusters (Table S4). Based on FDR-adjusted p-values, among the top enriched domains for the up-regulated group, the trypanosome sialidase, protein kinase and RNA-binding domains had the largest number of genes (Table II). For the down-regulated genes, the most abundant classes were found to be the mitochondrial substrate/solute carrier, 40S ribosomal protein and small GTP-binding protein domains. The results obtained using the graphical tool of the ShinyGO web application are shown in Fig. S1.
Table IIGene ontology clusters defined by DAVID server. List of top four clusters enriched in the up- or down-regulated genes after UBP1 overexpression
We then investigated transcript expression by carrying out a comparative analysis of several functional gene groups. Based on the data presented above, we manually classified the majority of sequences obtained from UBP1-OE parasites into 16 general categories: cell-surface glycoproteins (A), ribosomal proteins (B), RNA transcription (C), cell division and DNA synthesis (D), protein kinases (E), protein phosphatases (F), flagellar proteins (G), chaperones (H), lipid, fatty acid and ATP biosynthesis (I), biogenesis, cell organization and cellular motors (J), cellular signaling and processing (K), ATPases (L), glycolysis and carbohydrate metabolism (M), disperse gene family proteins (N), mitochondrial transcripts (O), and RNA-binding proteins (P). The overall gene distribution of transcripts among these groups was analyzed using violin plots showing expression values (log2 fold change OE/WT) (Fig. 4).
Figure 4Violin plots displaying the expression distribution of the genes within 16 different functional categories in the UBP1-OE cell transcriptome. Two transversal gray lines separate three groups of expression intensity: low expressed genes in red [<-1.5-fold UBP1-OE/WT (<-0.58 log2 fold change)], highly expressed genes in blue [>1.5X UBP1-OE/WT (>0.58 log2 fold change)] and middle expressed genes in green [>-1.5-fold and <1.5-fold UBP1-OE/WT (>-0.58 and <0.58 log2 fold change)]. Categories in the figure are: Membrane glycoproteins, Ribosome proteins, RNA transcription, Cell division and DNA synthesis, Protein kinases, Protein phosphatases, Flagellar proteins, Chaperones; Lipid, fatty acid and ATP biosynthesis; Biogenesis, cell organization and cellular motors; Cellular signalling and processing; ATPase; Glycolysis and carbohydrates metabolism; N-DGF, Mitochondrial transcripts and RNA-binding proteins.
Results confirmed that among the most abundant transcripts in the UBP1-OE transcriptome are those coding for cell-surface glycoproteins, protein kinases/phosphatases and RNA-binding proteins (Fig. 4A, E, F and P) and that among the least abundant transcripts are those coding for ribosomal proteins, mitochondrial transcripts and some Dispersed Gene Family hits (Fig. 4B, N and O), with the cluster of ribosomal proteins having the highest number of down-regulated hits. The dispersed gene family is large, with many of its members predicted to have transmembrane domains and reported to be more abundant in the amastigote stage than in trypomastigotes and epimastigotes (
Clearly, the most abundant cluster among the up-regulated genes in UBP1-OE samples was that of surface membrane-associated proteins. Within this group, we identified 171 trans-sialidase/trans-sialidase-like genes, 108 mucin-associated surface proteins and 88 mucins (Fig. 5). Particularly, six of these transcripts: TcCLB.504099.50 (TS_1), TcCLB.507237.190 (MASP_1), TcCLB.510005.20 (TS_2), TcCLB.506965.130 (MASP_2), TcCLB.507879.10 (TS_3) and TcCLB.511311.20 (TS_4) were among the 10 most significantly up-regulated RNAs (labeled in Fig. 2B; p-value < 1E-80 and log2 fold change > 1.9). Notably, we also observed up-regulation of three trans-sialidase-like mRNAs that had been previously reported to be up-regulated in UBP1-OE parasites (
): TcCLB.506455.30 (GP85 [trans-sialidase, Group II, putative], log2 fold change = 2.36, FDR-adjusted p-value = 5.43e-16), TcCLB.510163.60 (C71 [trans-sialidase, Group V, putative], log2 fold change = 3.16, FDR-adjusted p-value = 3.71e-57), and TcCLB.508285.60 (SA85 [trans-sialidase, Group II, putative], log2 fold change 2.61, FDR-adjusted p-value = 9.88e-33) (Fig. S2). These three transcripts harbor a known structural TcUBP1 RNA-binding element in their 3′-UTRs, previously described in our laboratory and termed UBP1m (
). In addition, in the up-regulated list, we observed 33 protein kinases and 15 protein phosphatases (see Discussion).
Figure 5Bars chart displaying the number of up- and down-regulated genes within seven categories in the UBP1-OE transcriptome. Categories in the figure are: MASP, mucin-associated surface proteins; TcS, trans-sialidase/trans-sialidase like; MUCI/II, mucin genes; Protein kinases, Protein phosphatases, Mitochondrial transcripts and Ribosomal proteins.
By contrast, many mRNAs coding for ribosomal and mitochondrial proteins were down-regulated in UBP1-OE parasites. We observed transcriptional down-regulation of 27 ribosomal protein-coding genes and 24 mitochondrial transcripts (Fig. 5). Particularly, five of these genes were among the 10 significantly down-regulated transcripts (p-value < 1E-75 and log2 fold change < -1.9): PORIN1 (TcCLB.511687.10); PEPT (TcCLB.511181.50); ATPase9 (TcCLB.503579.70); PORIN2 (TcCLB.504225.20); and PORIN3 (TcCLB.511687.19). In this Top 10 list, we also found the known TcUBP1-mRNA target TcSMUGS (TcCLB.511685.10) (Fig. 2B), as has already been reported (
) to compare the expression profiles of TcUBP1-overexpressing parasites with those of the four T. cruzi stages. In order to compare between sets of RNA-seq data from different experiments, we used an ad hoc pipeline to map the reads from these labs with the reference Esmeraldo-like CL Brener genome and then compared the fold change of the expression values. We calculated the percentage of regulated transcripts in UBP1-OE parasites among the most up- and down-regulated genes in a pairwise comparison between the metacyclic trypomastigote (MT), cell-derived trypomastigote (Trypo), epimastigotes (Epi) and amastigotes (Ama) stages. We clustered the different fold change values for each pairwise comparison into groups of up- and down-regulated genes with >1.5-, >2-, >2.83-, >4-, or >8-fold change differences between two stages (UBP1-OE versus Epi, MT versus Epi, Trypo versus Epi, Trypo versus Ama, and Epi versus Ama).
When analyzing the whole range (<4- to >8-fold change), we found that the UBP1-OE transcriptome showed highest similarity with the Trypo/Epi and MT/Epi datasets (genes over-represented in Trypo or MT with respect to Epi). The expression profile of UBP1-OE coincided 43.0% with the MT/Epi and 43.9% with the Trypo/Epi ratios. Particularly, for the up-regulated genes (>1.5- to >4-fold change), the Trypo/Epi comparison showed >60% similarity to UBP1-OE. The third dataset that was more similar to UBP1-OE was Trypo/Ama, which also displayed average percentage values of 58% in the up-regulated genes. No significant coverage was found for any of the up- or down-regulated transcripts in the Epi/Ama comparison. The similarity between datasets indicates that the transcriptome of UBP1-induced parasites has an expression profile that resembles that of the trypomastigote and metacyclic trypomastigote infective forms (ANOVA with post-hoc Tukey test, p-value = 0.00509). This can be visualized by different statistically significant colored clusters in the heatmap depicted in Fig. 6A (Tukey multiple comparisons: MT/Epi - Epi/Ama, p-value = 0.0121; Trypo/Ama - Epi/Ama, p-value = 0.0096; and Trypo/Epi - Epi/Ama, p-value = 0.0417).
Figure 6Comparison of UBP1-OE transcriptome with RNA-seq datasets of infective forms. A) heat map representation of the percentages of shared genes between UBP1-OE and different pairwise comparisons. Data of genes with >1.5, 2, 2.8, 4, or 8-fold change differences were extracted from Smircich et al. (
) for columns T/E (trypomastigote versus epimastigote), T/A (trypomastigote versus amastigote) and E/A (epimastigote versus amastigote). The brown/orange color indicates a high correlation, whereas the yellow color indicates a low correlation. B, PC analysis plot displaying the same samples as in A) along PC1 and PC2, which describe 64% and 25.9% of the variability, respectively. PC analysis was applied to 1737 genes with log2 fold change data for all the pairwise comparisons. C) Hierarchical clustering was performed using R based on Poisson distance. A/E (amastigote versus epimastigote), A/T (amastigote versus trypomastigote), E/MT (epimastigote versus metacyclic trypomastigote), E/T (epimastigote versus trypomastigote). The dark blue color indicates a high correlation, whereas the light blue/white color indicates a low correlation.
Next, we obtained fold change values for 1,737 genes from the RNA-seq experiments (Table S5). This RNA-seq expression table was used to perform a principal component analysis (PCA) to compare the dispersion of the different datasets. The horizontal axis (PC1) describes 64.2% of the variability, and, considering this component, the sample UBP1-OE/Epi is distinctly located closer to the MT and Trypo experiments than to Epi/Ama. Thus, similar to that shown in Fig. 6A, this analysis showed that the expression profile of the UBP1-OE population is more similar to that of the infective stages (MT/Epi, Trypo/Epi and Trypo/Ama) than to that of the replicative stage (Epi/Ama) (Fig. 6B).
These expression values were then used to calculate the Pearson correlation of all the samples, to which we also added the expression values of Ama/Epi, Ama/Trypo, Epi/Trypo and Epi/MT. The column corresponding to UBP1-OE/Epi is boxed. Again, the highest correlation was observed with the Trypo/Epi (0.5576), Trypo/Ama (0.4972) and MT/Epi (0.4610) datasets (Fig. 6C). No significant correlation was found between UBP1 and any of the remaining RNA-seq datasets. The analysis of shared genes, PCA, and correlation between the different experiments analyzed showed that UBP1-overexpressing parasites have an expression profile that resembles that of infective forms of T. cruzi.
Identification of cis-elements in the 3′-UTR of genes regulated by TcUBP1 overexpression
We next searched this transcriptome for the occurrence of a known structural UBP1 RNA-binding element, UBP1m -previously described in our laboratory (
)-, and also de novo sequence motifs. The most abundant mRNA targets previously identified for TcUBP1 encode for energy metabolism and cell-surface membrane glycoproteins. As mentioned above, the transcriptome analysis showed that, in UBP1-OE cells, these groups are either over- or under-represented: cell-surface trypomastigote glycoproteins are up-regulated, and mitochondrial transcripts coding for proteins related to energy metabolism are down-regulated.
With this result in mind, we decided to analyze how many of the mRNAs impacted by TcUBP1 overexpression could be direct interacting targets. For this purpose, we used the presence of the characteristic binding element UBP1m (
) as a target criterion. We then evaluated the motif coverage of UBP1m in the up- and down-regulated genes in the UBP1-OE sample. We looked at all transcripts that were expressed with fold changes ranging from < -8X to > 8X.
The down-regulated genes showed no significant differences in the presence of the UBP1m motif. Similarly, we observed that not regulated genes, found to be between the -1.5-fold and 1.5-fold categories (log2 fold change = -0.58 to log2 fold change = 0.58), presented a 5% UBP1m coverage (147 out of 2,829). However, as the fold change increased in the up-regulated genes, the abundance of the UBP1m element also increased (see Fig. 7A). Genes with > 5-fold up-regulation (log2 fold change > 2.32) showed an 8% (4 out of 46) presence of UBP1m, genes up-regulated > 6-fold showed 15% motif coverage (4 out of 27) and genes with > 8-fold showed the highest motif coverage (18%; 2 out of 11). This is consistent with the idea that the UBP1m motif might have a stabilizing effect on the mRNAs containing it. This all makes sense given that the UBP1m was originally detected in UBP1-immunoprecipitated mRNAs. Therefore, mRNAs stabilized by UBP1 (and containing the motif) were easily purified, whereas those destabilized and possibly containing other elements did not precipitate. We concluded that the UBP1 binding motif was enriched in the group of up-regulated genes (with log2 fold change values >2.32) compared with all the remaining groups (ANOVA analysis, post-hoc Tukey, p-value = 1.47E-05).
Figure 7RNA motifs identified in transcripts up-regulated by TcUBP1 overexpression. A) Bubble (left) and bar (right) charts of the percentage of sequences harboring the experimental motif UBP1m within the 3′-UTRs in each group (from <-3 to >3-fold change OE versus WT). B) Predicted RNA elements, logo consensus graphic and Z-score in the group of up-regulated genes (>4-fold change) defined by TRAWLER. Hits predicted to bind the motifs were obtained by a home-made bioinformatic pipeline 1 (see text).
After that, by using a cutoff value of 4-fold change, we detected 89 up- and 14 down-regulated genes in UBP1-OE parasites. For each group, a length of 350 nt downstream from the coding sequences was downloaded using TcruziDB to obtain sequences resembling the 3′-UTR, in agreement with data previously reported for trypanosomes (
). The up-regulated list was partitioned into two subgroups: set 1 (composed of 45 genes) and set 2 (composed of 44 genes). Thus, we first searched motifs in set 1 and then in set 2. To this end, we ran the motif prediction tool TRAWLER (http://trawler.monash.edu.ar) with default parameters, using set 1 as input sequences and the down-regulated group as the background list. Results showed three candidate motifs: family_1 (5'-TVTMTATATATATATATABR-3', Z-score: 64.78), family_2 (5'-NNTTTRCTTTB-3', Z-score: 79.50) and family_3 (5'-CTCYTSCY-3', Z-score 61.92). The WebLogo representation indicated that the family_1 motif is rich in AT content, whereas the family_3 motif has a CT-rich sequence composition (Fig. 7B). When the relative frequencies of these motifs within the 3′-UTR of both the input (set 1) and the control (set 2) datasets were examined using the FIMO software, an enrichment of all elements was observed: family_1 showed 7.47 total hits/kbps in set 1 and 5.24 total hits/kbps in set 2; family_2 showed 3.86 total hits/kbps in set 1 and 3.11 total hits/kbps in set 2; and family_3 showed 1.64 total hits/kbps in set 1 and 1.03 total hits/kbps in set 2.
We next generated a homemade bioinformatic pipeline to predict the RNA-binding of motifs to sequence proteins based on integrated published data sources.
For this purpose, we used the sequence elements as a query to search for proteins with the ability to bind them according to Tomtom motif analysis (https://meme-suite.org/meme/tools/tomtom, MEME Suite). Thus, heterologous interacting proteins were predicted as binders by searching against the RNAcompete database composed of 244 eukaryotic RBPs, with a p-value < 0.004 (
). Then, highly similar T. cruzi proteins were identified as binders of these motifs by using the previous RNAcompete hits as queries in BLASTP searches against a T. cruzi RBP database composed of 285 sequences with RRM (
) (blastp E_value < 1E-08 and subject coverage >=50%, Table S6). After this step, an RNA-protein interaction prediction software that uses only primary sequence information was systematically run on all previously obtained candidates to select those reliable interactions that have binding probabilities > 0.5, by using an SVM classifier (
). The results are shown in Fig. 7B, with five putative RBP targets for family_1 (UBP1, UBP2, RBP5A/B, PABP1 and DRBD11B), three for family_2 (UBP1, UBP2 and RBP3) and only one for familiy_3 (DRBD3A). Of note, TcUBP1 was predicted to bind both family_1 and family_2 RNA motifs. Not surprisingly, the polypyrimidine-tract binding protein DRBD3A/PTB1 (TcCLB.506649.80) was predicted to bind the C/T-rich family_3 sequence. Moreover, we validated our predictions by using molecular docking experiments. To this, we used the TcUBP1 protein structure predicted by the AlphaFold database (Q4E1N5.pdb) and obtained the 3D structures of the RNA motifs with the 3DRNA software (https://bio.tools/3dRNA). We then ran the HDOCK docking server (
) and checked the RNA-protein interactions using the transcript sequences UAUAUAUAUAUAUAUAUAUA (as family_1 RNA ligand) and UUUGCUUUU (as family_2 RNA ligand). As positive controls, we used two experimental sequences reported to be binders of TcUBP1: UBP1m (5'-UGGCGCAUCCAUGCCUGGAUGCGCCG-3') (
J. G. De Gaudenzi and A. C. C. Frasch, unpublished results.
In all the cases, we obtained HDOCK confidence scores >0.75, suggesting that these molecular interactions occur (Table III). Taken together, these results suggest that these two family_1 and family_2 motifs, identified in the 3′-UTRs of the up-regulated transcripts, might be involved in the interaction with TcUBP1.
Table IIIProtein-RNA docking of TcUBP1 and cis-regulatory elements
RNA motif
Nucleotide sequence (from 5' to 3')
Confidence score
family_1
UAUAUAUAUAUAUAUAUAUA
Model_1 =0.9366, very high probability
family_2
UUUGCUUUU
Model_1 =0.8745, very high probability
UBP1m
UGGCGCAUCCAUGCCUGGAUGCGCCG
Model_1 =0.9421, very high probability
UBP1m28
UUUUGGAGGAAGUUUUUUUUGGGG
Model_1 =0.90, very high probability
RNA motifs, nucleotide sequence and HDOCK confidence score obtained for model 1 using TcUBP1 AlphaFold predicted structure Q41NE5 as receptor, and RNA element (family_1, family_2, UBP1m or UBP1m28) as ligand.
Open chromatin analysis in Trypanosoma cruzi life forms highlights critical differences in genomic compartments and developmental regulation at tDNA loci.
). Nonetheless, it is broadly accepted that transcription by RNA polymerase II in these pathogens deviates from the standard eukaryotic paradigm. In T. cruzi, there is no dedicated promoter for each gene, resulting in polycistronic transcription, and thus gene expression regulation depends heavily on large post-transcriptional networks (
). In the present work, we used an in vitro system based on the inducible expression of a GFP-tagged UBP1 to monitor transcriptome changes during the differentiation of T. cruzi from non-infectious epimastigotes to infectious metacyclic trypomastigotes. In addition, we performed the bioinformatic analysis of two RNA-seq samples, with three biological replicates each (Fig. S3), highlighting the differential transcript abundance and providing a data source to understand how this parasite becomes infectious.
Several lines of evidence support the role of certain RBPs as key regulators of trypanosome differentiation (
The hnRNP F/H homologue of Trypanosoma brucei is differentially expressed in the two life cycle stages of the parasite and regulates splicing and mRNA stability.
). In a previous work, we showed that TcUBP1 binds to structural binding elements highly enriched in transcripts coding for surface-cell virulence factors associated with the metacyclic trypomastigote developmental stage (
). In epimastigotes, translation of these transcripts is diminished and thus localized in the posterior zone of the cell until a stimulus such as the ectopic expression of TcUBP1-GFP triggers the metacyclogenesis program, up-regulating and mobilizing these trypomastigote stage-specific mRNAs to polysomes (
In the present study, when analyzing exclusive transcripts in a given experimental condition, we found that the mRNAs expressed only in WT parasites are mostly related to chromosome organization, while those exclusively expressed in TcUBP1-GFP parasites code for mitochondrial RNA processing (Table S2). Our results also evidenced that 3,035 out of 9,039 genes (34%) showed significant differences in the mRNA steady-state levels in TcUBP1-OE parasites compared to WT parasites (|log2 fold change| > 0.58, FDR 0.05). It can be easily noticed that a high number of genes coding for trypomastigote cell-surface glycoproteins are stabilized in the transcriptome of the UBP1-transgenic epimastigotes (Fig. 5). The transcriptome difference between UBP1-OE versus Epi-WT is similar to the one observed between the quiescent infective metacyclic trypomastigote MT versus Epi (Fig. 6).
In agreement with the data described for the closely related parasite T. brucei (
), in the present study, we obtained a profile expression resembling that of the quiescent infectious trypomastigote parasites by overexpressing a single RRM protein, UBP1, in non-infectious epimastigotes. This conclusion is based upon two results. On the one hand, we focused on the up-regulation of RNA abundances of numerous cell-surface trypomastigote glycoproteins, including members of the TcS superfamily (Figs. 4A and 5). This UBP1-OE transcriptome confirms our data on the glycoprotein RNA regulon of TcUBP1-expressing parasites (
). On the other hand, we noticed that the genes coding for ribosomal proteins were down-regulated in the UBP1-OE parasites (Figs. 4B and 5). This decrease in the number of ribosomal protein-coding mRNAs is consistent with the translational repression previously reported for metacyclic trypomastigotes (
) are detected in the transcriptome of UBP1-overexpressing epimastigotes. These results further support post-transcriptional control as a critical regulatory mechanism required for parasite differentiation.
It has been reported that translation is strongly regulated during the T. cruzi cell cycle, causing variation in specific protein levels (
). In humans, multifunctional RBPs can regulate more than a single aspect of RNA metabolism. Schneider-Lunitz and coworkers have identified dozens of RBPs that influence mRNA abundance and translation efficiency of their targets (
). TcUBP1 could be also acting with this dual functionality. Previous results of our lab have demonstrated that an endogenous TcUBP1 fraction is associated with polysomes (
Insights into the Regulation of mRNA Processing of Polycistronic Transcripts Mediated by DRBD4/PTB2, a Trypanosome Homolog of the Polypyrimidine Tract-Binding Protein.
). Regarding this, and as a consequence of TcUBP1 overexpression in epimastigotes, we also observed a change in the subcellular localization of cell-surface trypomastigote glycoprotein-coding transcripts, resembling the typical distribution of the metacyclic trypomastigote infective stage (
). Moreover, in our previous work, we also detected that trypomastigotes derived from TcUBP1 transgenic epimastigotes have an increased capacity for infection, an effect that has already been seen to be associated with increased protein expression of surface glycoproteins (
). All these observations could also suggest a possible regulatory role of TcUBP1 in the translation rate of trypomastigote-specific mRNAs. To investigate the degree of protein synthesis regulation, future proteomics and ribosome profiling researches should be performed.
According to our present results, after UBP1 overexpression, more protein kinases than protein phosphatases are affected. In T. brucei, the MAP kinase MAPKL1 (Tb927.10.10870) regulates proteins involved in mRNA metabolism (
), whereas, in UBP1-OE parasites, six CMGC family protein kinases with sequence similarities to MAPKL1 are up-regulated. Thus, the transcripts of these protein kinases could be part of the downstream cascade involved in the phosphorylation network of T. cruzi. In contrast, the transcript levels of the TcAMPKs involved in autophagy and parasite nutrient sensing (
Gene regulatory networks provide key strategies to identify RNA regulons and candidate RBPs for functional studies and/or molecular targets for disease control (
Gene expression network analyses during infection with virulent and avirulent Trypanosoma cruzi strains unveil a role for fibroblasts in neutrophil recruitment and activation.
). The short sequence elements identified in this work could be signature marks for clusters of differentially up-regulated genes (Fig. 7). In a preliminary computational work, we described a community of an RNA-protein interaction network composed of 26 T. cruzi RRM proteins (
Notably, among these proteins, we found UBP1, UBP2, RBP5A/B and TcPABP1, which are five out of the seven different trans-factors identified for family_1 and family_2 cis-regulatory sequences. Regarding the mRNA expression levels of RBPs in TcUBP1-GFP-expressing parasites, we observed that 15 genes were three-fold up-regulated and that five genes were two-fold down-regulated (see Table S7). Among the up-regulated RBP genes, two were associated with TcUBP1 by being part of the same RNA-protein community described above: TcRBP9A (TcCLB.511127.10) and TcRBP26A (TcCLB.506795.10). These two proteins may be controlled by TcUBP1 and could provide positive feedback by co-regulating, together with TcUBP1, mRNA targets related to the trypomastigote-specific form. TcUBP1 is expressed in all the life cycle stages of T. cruzi and is involved in the formation of distinct regulatory complexes. TcUBP1 has been previously reported as an interacting partner of the cytoplasmic DRBD2-mRNP complex in epimastigotes, together with UBP2, DRBD3 and PABP2, among others (
TcUBP-1, an mRNA Destabilizing Factor from Trypanosomes, Homodimerizes and Interacts with Novel AU-rich Element- and Poly (A)-binding Proteins Forming a Ribonucleoprotein Complex.
Journal of Biological Chemistry.2002; 277: 50520-50528
By overexpressing TbRBP6 in non-infectious procyclic trypanosomes, Kolev and co-workers recapitulated in vitro the generation of infective metacyclic forms observed in the tse tse fly (
). Since the T. cruzi homologs of these two regulators are among the RBP up-regulated genes listed in Table S7: TcRBP6A (TcCLB.506693.30) and TcRBP10B (TcCLB.510507.50), it is tempting to speculate that TcUBP1 is upstream in the regulatory cascade that triggers parasite differentiation.
In summary, the transcriptome data presented here obtained by overexpressing TcUBP1-GFP in the non-infectious epimastigote T. cruzi stage provide a comprehensive picture of the mRNA steady-state level of the differentiation process toward the infective stage. Our results deepen the knowledge of previous reports of our lab and show that the levels of TcUBP1 trigger a post-transcriptional regulatory program that occurs during parasite differentiation, to transform replicative epimastigotes into infective quiescent metacyclic trypomastigotes.
EXPERIMENTAL PROCEDURES
Plasmid construction, parasite cultures and transfection
The DNA construct pTcINDEX-TcUBP1-GFP previously used in Sabalette et al. (
) was used for parasite transfections. Protein expression values in Tet+ induced epimastigote samples after 96 h were determined relative to non-induced controls (Tet-) by Western blot analysis of GFP levels normalized to total protein loading, as measured by Coomassie Blue staining. T. cruzi epimastigotes, from the CL Brener strain, were cultured in BHT medium containing 10% heat-inactivated fetal calf serum (BHT 10%) at 28°C. All parasite cultures were performed in plastic flasks without shaking, unless otherwise stated. Parasites were transfected by electroporation subsequently with pLew vector and pTcINDEX constructions and selected with 500 μg/ml of G418 and 250 μg/ml Hygromycin. For induction of recombinant proteins from the pTcINDEX vector, parasites were incubated in BHT 10% containing 0.5 μg/ml tetracycline for 96 h at 28°C with shaking.
RNA preparation and RNA-seq
Total RNA was prepared from approximately 107 epimastigote un-induced cells and 4-day Tet+ induced cells that express TcUBP1. RNA from three biological replicates was prepared using the TRIzol reagent from Invitrogen according to the manufacturer’s instructions. The quality of RNA samples were checked on 1% agarose gel and quantified using NanoDrop 2000 spectrophotometer (Thermo Scientific, USA). Additional quality assessment for the integrity of RNA samples, isolation of poly (A)+ mRNA, library preparation and sequencing on DNBSeq platform were performed at the BGI Americas Corporation.
Overall quality parameters of the RNA-seq data
The RNA-seq bio-analyzer library profile of both samples was generated on Agilent 2100 instrument. The samples were next used for paired-end (PE) deep sequencing and the libraries were sequenced using 2 × 100 PE chemistry on DNBSeq platform for generating ∼5.3 GB of data per sample. After trimming of low-quality sequences, in total, ∼24M reads (∼11 GB) were obtained for each UBP1-OE and WT samples. To minimize genetic heterogeneity we choose the reference genome CL Brener Esmeraldo-like strain (TriTrypDB-59_TcruziCLBrenerEsmeraldo-like_Genome.fasta), which has a genome size of 32.53 Mbp. The obtained mapped read numbers for UBP1-OE and WT samples were 42,880,869 and 42,539,799 reads, respectively.
Read processing and data analysis
Read processing and data analysis were performed. The short reads below than 50 bases were dropped to exterminate the sequencing artifacts and the quality of reads was evaluated using FASTQC toolkit (score >35) (
). The high-quality reads were de novo assembled using bowtie2 with parameter --very sensitive-local. Samtools were used to index the output and the quantitative assessment of reads was performed with featureCounts with parameters '-p -t "CDS" -g "ID" -T 40' (
). The obtained count value was used to identify the differentially expressed gene transcripts using the criteria of at least two-fold change (|log2 fold change| >1) in the sequence count between OE and WT-samples and the Benjamini-Hochberg false discovery rate (FDR) adjusted p-value < 0.05. The FPKM values for each transcript were log-transformed and normalized, which was subsequently used to calculate the matrix distance with Euclidean distance and complete-linkage methods. The R statistics package pheatmap was used to construct the heatmap (https://cran.r-project.org/web/packages/pheatmap.html). The differentially expressed genes were used for GO terms/KEGG pathway enrichment analyses using hypergeometric test equivalent to one-tailed Fisher’s exact test with a FDR value of 0.05 using TriTrypDB. Volcano, GO enrichment and violin plots were conducted using R with the package ggplot2 (
). All RNA-seq raw data files for UBP1-OE and WT samples used in this study are available as FASTQ files of 100 bp paired-end reads in the National Center for Biotechnology Information (NCBI) Sequence Read Archive (SRA) database with the following study number: PRJNA907231.
Functional annotation of gene lists
Gene ontology (GO) analysis was carried out for the differentially expressed genes from the TriTrypDB database (http://www. tritrypdb.org). The GO sequence distribution was analyzed for all the three GO domains: biological processes, molecular function, and cellular component. All the genes for T. cruzi were taken as reference set and the differentially expressed genes for both lists were taken as test set (up- or down-regulated after UBP1-OE). The GO annotations were extracted and visualized as bubble charts using ggplot2 in R (
). Also, to categorize gene lists into overrepresented functionally related groups, DAVID (Database for Annotation, Visualization and Integrated Discovery, version 6.8) functional annotation clustering tool was used (
RNA-seq raw data files used in this study are available as FASTQ files of 100 bp paired-end reads in the National Center for Biotechnology Information (NCBI) Sequence Read Archive (SRA) database with the following study number: PRJNA907231.
Funding and additional information
This work was supported by Agencia Nacional de Promoción Científica y Tecnológica Grant PICT 2019-00737 and CONICET Grant PIP-GI 0248 (to J. G. D. G.); and funds from the Research Career of CONICET (to J. G. D. G.).
Conflict of interest
The authors declare that they have no conflicts of interest with the contents of this article.
ACKNOWLEDGMENTS
We are indebted to Liliana Sferco and Agustina Chidichimo for parasite cultures. We thank Vanina Campo for reading the manuscript, valuable comments and advice.
Integrative analysis of the Trypanosoma brucei gene expression cascade predicts differential regulation of mRNA processing and unusual control of ribosomal protein expression.
Transcriptome-wide analysis of trypanosome mRNA decay reveals complex degradation kinetics and suggests a role for co-transcriptional degradation in determining mRNA levels.
Transcriptome-wide analysis of the Trypanosoma cruzi proliferative cycle identifies the periodically expressed mRNAs and their multiple levels of control.
A 43-nucleotide U-rich element in 3’-untranslated region of large number of Trypanosoma cruzi transcripts is important for mRNA abundance in intracellular amastigotes.
Genomic analyses, gene expression and antigenic profile of the trans-sialidase superfamily of Trypanosoma cruzi reveal an undetected level of complexity.
Open chromatin analysis in Trypanosoma cruzi life forms highlights critical differences in genomic compartments and developmental regulation at tDNA loci.
The hnRNP F/H homologue of Trypanosoma brucei is differentially expressed in the two life cycle stages of the parasite and regulates splicing and mRNA stability.
Insights into the Regulation of mRNA Processing of Polycistronic Transcripts Mediated by DRBD4/PTB2, a Trypanosome Homolog of the Polypyrimidine Tract-Binding Protein.
Gene expression network analyses during infection with virulent and avirulent Trypanosoma cruzi strains unveil a role for fibroblasts in neutrophil recruitment and activation.
TcUBP-1, an mRNA Destabilizing Factor from Trypanosomes, Homodimerizes and Interacts with Novel AU-rich Element- and Poly (A)-binding Proteins Forming a Ribonucleoprotein Complex.
Journal of Biological Chemistry.2002; 277: 50520-50528
Unpublished observations and personal communications
Author contributions
K. B. S., P.S. and J. G. D. G. methodology; J. G. D. G. writing original draft; J. G. D. G. investigation; P.S., J.R.S. and J. G. D. G. formal analysis; P.S., J.R.S. and J. G. D. G. conceptualization; J. G. D. G. funding acquisition; J. G. D. G. visualization; J. G. D. G. project administration; J. G. D. G. writing-review and editing; J. G. D. G. supervision.
00265-X&id=gr1.jpg){kind=link}
00265-X&id=gr2.jpg){kind=link}
00265-X&id=gr3.jpg){kind=link}
00265-X&id=gr4.jpg){kind=link}
00265-X&id=gr5.jpg){kind=link}
00265-X&id=gr6.jpg){kind=link}
00265-X&id=gr7.jpg){kind=link}