The trypanosome transcriptome is remodelled during differentiation but displays limited responsiveness within life stages

Abstract

Background: Trypanosomatids utilise polycistronic transcription for production of the vast majority of protein-coding mRNAs, which operates in the absence of gene-specific promoters. Resolution of nascent transcripts by polyadenylation and trans-splicing, together with specific rates of mRNA turnover, serve to generate steady state transcript levels that can differ in abundance across several orders of magnitude and can be developmentally regulated. We used a targeted oligonucleotide microarray, representing the strongly developmentally-regulated T. brucei membrane trafficking system and approximately 10% of the Trypanosoma brucei genome, to investigate both between-stage, or differentiation-dependent, transcriptome changes and within-stage flexibility in response to various challenges.

Results: 6% of the gene cohort are developmentally regulated, including several small GTPases, SNAREs, vesicle coat factors and protein kinases both consistent with and extending previous data. Therefore substantial differentiation-dependent remodeling of the trypanosome transcriptome is associated with membrane transport. Both the microarray and qRT-PCR were then used to analyse transcriptome changes resulting from specific gene over-expression, knockdown, altered culture conditions and chemical stress. Firstly, manipulation of Rab5 expression results in co-ordinate changes to clathrin protein expression levels and endocytotic activity, but no detectable changes to steady-state mRNA levels, which indicates that the effect is mediated post-transcriptionally. Secondly, knockdown of clathrin or the variant surface glycoprotein failed to perturb transcription. Thirdly, exposure to dithiothreitol or tunicamycin revealed no evidence for a classical unfolded protein response, mediated in higher eukaryotes by transcriptional changes. Finally, altered serum levels invoked little transcriptome alteration beyond changes to expression of ESAG6/7, the transferrin receptor.

Conclusion: While trypanosomes regulate mRNA abundance to effect the major changes accompanying differentiation, a given differentiated state appears transcriptionally inflexible. The implications of the absence of a transcriptome response in trypanosomes for both virulence and models of life cycle progression are discussed.

Figure 1

Overview of developmental transcriptome changes in the membrane trafficking system of trypanosomes. Panel A: Scatter plot of raw data for all 3600 spots on a representative microarray used for developmental expression experiments. Cy5 fluorescence is plotted on the Y-axis (bloodstream, BSF) and Cy3 fluorescence on the X-axis (procyclic, PCF). Spots with a BSF/PCF ratio above two are highlighted in red, while spots with a PCF/BSF ratio above two are highlighted in green. Panel B: Clustering of the data for eight microarray experiments comparing BSF to PCF, representing four biological replicates plus relevant dye-swaps. The scale indicates the colour scheme for the z-score of the data, i.e. how far and in what direction, the ratio for each spot deviates from the mean for each array, expressed in units of standard deviation; bright red indicates significant upregulation in BSF, bright green indicates significant upregulation in PCF, dark colours or black indicate no differential expression between the two developmental stages. For each target gene, the four replicate spots on the array were averaged. White indicates data points rejected because of too much variability in the replicate spots. A small but significant number of ORFs exhibit strong developmental regulation.

Figure 2

Significantly developmentally expressed trypanosome genes grouped by functional class. Genes represented by array oligonucleotides were grouped by function, based on sequence similarity to annotated sequences, domain annotation, GO terms and additional criteria. The proportion in each functional class significantly upregulated in BSF are shown in red and in PCF in green. The remaining fraction of genes in each class are shown in grey. The first bar illustrates the proportion of developmentally regulated transcripts in the entire studied gene cohort. For raw data [see Additional file 6].

Figure 3

Relative steady state mRNA levels of selected trypanosome genes. Fluorescence intensity data for eight microarray experiments comparing BSF to PCF, representing four biological replicates plus dye-swaps, was normalised to the total signal intensity of each slide, averaged and plotted as bar-graphs against ORF designation or accession number. Error bars indicate the standard deviation, dark bars are BSF and light bars PCF. Panel A: Small GTPases. ARL and Rab annotations are based on [39], Ras-like annotations based on [29] and additional data at GeneDB. Panel B: Cell surface proteins. Panel C: SNARE family proteins. Annotations based on domain architecture and similarity to sequences described for L. major [59]. Panel D: Putative secretory pathway proteases. ORFs were included for predicted proteases that bear an ER targeting sequence and/or a predicted trans-membrane domain. In each cohort, a limited number of transcripts are highly expressed.

Figure 4

Scatter plots of transcriptome data from trypanosomes subjected to altered culturing conditions and other challenges. In each case, Cy5 fluorescence is plotted on the Y-axis and Cy3 fluorescence on the X-axis. Spots with a Cy5/Cy3 or Cy3/Cy5 ratio above two, i.e. significantly differentially expressed are highlighted in black, spots with a ratio below two are shown in grey. Panel A: BSF (Cy5) vs PCF (Cy3); data are identical to Figure 1 and reproduced for comparison. Panel B: BSF in 30% FBS (Cy5) vs BSF (Cy3). Panel C: BSF (Cy5) vs BSF in 0% FBS, 5 mg/ml BSA (Cy3). Panel D: BSF (Cy5) vs BSF in 0% FBS, 5 mg/ml, BSA 0.3 mg/ml Tfn (Cy3). Panel E: BSF in 1 mM DTT for 1 hr (Cy5) vs BSF (Cy3). Panel F: BSF in 1 mM DTT for 4 hr (Cy5) vs BSF (Cy3). Panel G: BSF in 5 μg/ml tunicamycin for 4 hr (Cy5) vs BSF (Cy3). Panel H: BSF in 5 μg/ml tunicamycin for 24 hr (Cy5) vs BSF (Cy3). Panel I: VSG RNAi in BSF, induced for 24 hr (Cy5) vs uninduced (Cy3). Panel J: VSG RNAi in BSF, induced for 72 hr (Cy5) vs uninduced (Cy3). Panel K: CLH RNAi in BSF, induced for 24 hr (Cy5) vs uninduced (Cy3). Panel L: BSF (Cy5) vs MITat1.1 grown in vivo in rats (Cy3). The scatter plots for most experiments (panels B-K) indicate few transcripts that fall outside of the region considered as constitutive, in marked contrast to the scattergrams obtained for developmental changes (panel A) and in the comparison between in vitro versus in vivo BSF cultures (panel L).

Figure 5

Limited differential expression in Rab5-overexpressing PCF cells. Panel A: Western blot analysis for the clathrin heavy chain (CLH) and Rab5A, for wildtype BSFs and PCFs, PCFs overexpressing Rab5A (5AWT), Rab5A^QL (GTP-locked mutant) and Rab5B^QL (GTP-locked mutant) cell lines. BiP was used as a loading control. Panel B: Scatter plots of raw data for representative microarrays for BSF (Cy5) vs PCF (Cy3) experiment (plot as in Figure 1 for comparison), and Rab5-overexpessing lines (Cy5) vs PCF (Cy3) as indicated. Cy5 fluorescence is plotted on the Y-axis and Cy3 fluorescence on the X-axis. Spots with a Cy5/Cy3 or Cy3/Cy5 ratio above two are highlighted in black, spots with a ratio below two are shown in grey. Panel C: Relative expression levels of Rab5A, Rab5B, CLH, RabX3, COPIε, and adenylate kinase 3 (AK3) in wildtype SMBs and PCFs, and in the Rab5-overexpressor PCF lines, as assessed by qRT-PCR.

Figure 6

Transcriptional flexibility and inflexibility in differentiation and responsiveness. Upper panel: flexible system. Gene cohorts 1 and 3 are developmentally regulated, and either highly expressed or not expressed; examples of these types of gene products are the trypanosome surface antigens, VSG in the bloodstream form (red) and procyclin in the insect stage (green). The vast majority of genes fall into cohort 2, where, for example, either small or large changes to transcription could result from alterations to the environment (light and dark blue), or a more continually altering transcriptional profile is present that may seek to track changing conditions (purple). This behavior may propagate from one life stage to the next (light and dark blue) or be lost (purple) resulting in altered transcriptional flexibility for genes between life stages. Such a profile is found in higher eukaryotes, including humans and yeast, and probably also many protists, including E. gracilis. Lower panel: inflexible system. In this model gene cohorts 1 and 3 behave as before, but transcription of the genes in cohort 2 remains unchanged. The relative levels of mRNAs from the genes in this cohort may remain constant following differentiation (light blue) or be significantly altered (dark blue and purple). Such a profile is observed here for T. brucei and has been reported previously for P. falciparum, and is potentially a result of a parasitic life style where the host is responsible for provision of a homeostatic environment.