Background

During their life cycle, Trypanosoma brucei parasites undergo various developmental transitions. These transitions involve changes in nutrient-specific metabolism, morphology, organelle organization and structure, and stage-specific surface protein expression, which facilitates survival and transmission. In the mammalian host, these forms include long slender bloodstream forms, which can differentiate into stumpy bloodstream forms through a quorum sensing process. Although these 2 extremes are well identified, there are possibly multiple intermediate stages between both forms which have not been well defined. Stumpy forms remain arrested in the cell cycle until ingested by a tsetse fly. In the fly midgut the stumpy forms undergo a further differentiation event and re-enter the cell cycle as tsetse-midgut procyclic forms. Procyclic, slender, and stumpy forms, differ at the transcript and protein level. However, understanding the detailed progression between slender and stumpy cells has been hampered due to the asynchrony of this differentiation step. To address this, single-cell RNA sequencing offers the opportunity to study individual cells in a heterogeneous population, to decipher in detail this developmental process. In their current work, Briggs et al (1) applied single cell transcriptomics (scRNA-seq) to dissect the asynchronous differentiation of slender to stumpy forms, deriving a temporal map of the transition between these forms, based on individual cells.

Key findings and developments

scRNA-seq identifies transcriptionally distinct long slender and short stumpy form T. brucei.  To model stumpy differentiation in vitro the authors used a pleomorphic line, and began by treating parasites with oligopeptide-rich BHI broth, which can induce T. brucei bloodstream form differentiation in a titratable manner. To capture the transcriptomes of slender, intermediate and stumpy forms, they combined parasites after 0, 24, 48, or 72h after 10% BHI treatment in equal numbers. scRNA-Seq was then performed using the Chromium Single Cell 3’ workflow (10x genomics) and Illumina Sequencing, of two independent biological replicates with a total of 9344 cells examined. Medians of 1051 and 1439 genes were detected per cell. Cells from both experiments were integrated and visualized using UMAP. The authors identified four distinct groups containing transcriptionally similar cells, with two of those groups being clear slender and stumpy-like cells. The four groups were termed slender A, slender B, stumpy A and stumpy B. Differential expression analysis of the transcripts between the 4 groups showed significant overlap between the genes of group slender A and B, and between stumpy A and B. However, the study showed 183 markers unique to the slender A group, 95 to slender B, 55 to stumpy A, and 9 to stumpy B. Gene ontology enrichment analysis revealed the association of each cluster’s marker genes with distinct biological processes. Altogether, the authors emphasize that a distinct cluster representative of ‘the intermediate’ stage transcriptome between slender and stumpy forms was not evident.

Trajectory analysis of long slender to short stumpy differentiation. Given the overlap detected in the clustering analysis, the authors conducted trajectory interference and pseudotime analysis to study gene expression changes during stumpy development in detail. Here, individual cells were re-plotted as a PHATE (potential of heat-diffusion for affinity-based transition embedding) map, which allows preservation of the continual progression of developmental processes. They found that slender A and B clusters remained separate, while stumpy A and stumpy B showed significant overlap. This allowed identification of 2001 genes differentially expressed as a function of pseudotime, which were grouped into 9 modules of co-expressed genes showing similar patterns of expression throughout differentiation. 66% of those genes had been previously shown to be significantly differentially expressed between slender and stumpy populations isolated from low and peak parasitemias in vivo. Altogether, the authors highlight the advantage of scRNA-Seq to reveal transient events in an asynchronous developmental trajectory. Moreover, GO term enrichment for biological processes associated with each gene module revealed a potential order of biological events during slender-to-stumpy development. Besides of the annotated genes, 635 hypothetical genes were identified as differentially expressed during slender to stumpy differentiation. Altogether, pseudotime analysis allowed identification of novel genes differentially expressed during bloodstream form differentiation, as well as each gene’s detailed expression pattern.

Transcript abundance during the bloodstream slender cell cycle. Given that replicating slender bloodstream form cells were captured in the experiments described above, the authors went on to explore if the scRNA-Seq data could reveal greater detail than what is known, on gene expression changes during the cell cycle. Each cell was assigned to a cell cycle phase using marker genes previously identified in bulk RNA-Seq analyses. Slender A and B cells were grouped closer to cells of the same phase, with parasites most distal to Stumpy A and B labelled as late G1, followed by S and G2/M phase cells. Slender B cells most proximal to stumpy A contained all 4 cycle phases, although early G1 cells were enriched here. Interestingly, stumpy A and B cells were marked in a variety of cell cycle phases. An important finding of this section of the work was a) the identification of genes driving the cell cycle, and b) the identification of 3 genes previously shown to be involved in stumpy development with differential expression patterns in slender cells- namely, RBP7B (increased in late G1 cells through to G2/M), PPC2 (decreased in late G1/S phase parasites), and ZC3H20 (dropped in expression in late G1/S phase).

ZC3H20 null parasites fail to differentiate in response to BHI. ZC3H20 peaks in expression at the slender B to stumpy transition in pseudotime, and it has been previously shown to be required for differentiation in vivo and in vitro. Based on this, the authors used a ZC3H20 null T. brucei line to investigate where parasites fail in their development to stumpy forms with respect to transcriptome changes, and aimed to identify mRNA targets of ZC3H20 itself. Incubation of ZC3H20 KO in 10% BHI broth showed that these parasites continued to replicate beyond time points where WT cells had arrested, and after 72h of culture, failed to express PAD1. Moreover, consistent with their inability to produce stumpy forms, ZC3H20 KO failed to differentiate into procyclic cells. scRNA-Seq was then performed on ZC3H20 KO cells at 0, 24, 48, or 72h after 10% BHI treatment- as done for WT cells. Clustering the ZC3H20 KO and WT integrated cells resulted in 6 distinct clusters: stumpy A and B, and 4 slender clusters, called slender A.1, A.2, B.1, and B.2. While 77.3% of WT cells were found in clusters stumpy A or B, only 0.3% of ZC3H20 KO cells were in either, consistent with the near complete ablation of stumpy formation in the mutant parasites. The B.2 cohort was comprised almost entirely of ZC3H20 KO cells. Marker gene analysis between clusters identified 94 marker genes upregulated in slender B.2 cells, 18 of which were unique to this cluster.

Trajectory comparison between WT and ZC3H20 KO cells reveals functional separation of downregulation and upregulation of transcripts during differentiation. The authors then compared transcriptomic changes in ZC3H20 KO and WT parasites after BHI treatment, by inferring a trajectory from the WT and ZC3H20 KO integrated parasites. This identified a branched trajectory – while early in pseudotime WT and ZC3H20 KO parasites are transcriptionally similar and arrange on the same lineage, later there was a clear branching in their development, whereby WT cells ended in stumpy forms, and ZC3H20 KO in slender B.2. 587 genes of the 2001 identified as differentially expressed during stumpy development in WT cells, significantly changed in expression in ZC3H20 KO cells across the truncated trajectory. ZC3H20 KO cells failed to upregulate transcripts later in development that are required for stumpy formation, and this point of dysregulation coincided with the peak of ZC3H20 expression during normal WT differentiation.

The authors then aimed to identify regulators of early stumpy development, and so looked for genes which changed significantly in abundance at the start of the trajectory to a point downstream of the ZC3H20 branch. 234 genes changed in transcript abundance between these points, and were associated with trajectory progression. RDK2 and PAD2 were associated with both trajectories, but had different patterns of expression. 83 genes were differentially expressed only in the trajectory of WT parasites. Conversely, 35 genes with early altered expression were associated with the truncated ZC3H20 KO development only. Altogether, comparison of the differentiation of WT and ZC3H20 KO cells through scRNA-Seq allowed the identification of direct and indirect targets of ZC3H20 altered specifically during differentiation; the failure point of ZC3H20 KO cells during differentiation; and putative immediate early regulators of differentiation.

What I like about this preprint

I think the question addressed in this work is one that remained outstanding in the field of T. brucei, namely, little is known about the intermediate stages of development of the parasite from slender to stumpy forms. I think the use of scRNA-Seq in T. brucei research, and in this work allowed opening various avenues of research, relevant to the whole field.

 

References

1. Briggs et al, Single cell transcriptomic analysis of bloodstream form T. brucei reconstructs cell cycle progression and differentiation via quorum sensing. bioRxiv, 2020.

 

Posted on: 23 December 2020 , updated on: 8 January 2021

doi: https://doi.org/10.1242/prelights.26638

Read preprint (No Ratings Yet)