The establishment of variant surface glycoprotein monoallelic expression revealed by single-cell RNA-seq of Trypanosoma brucei in the tsetse fly salivary glands

  1. Sebastian Hutchinson
  2. Sophie Foulon
  3. Aline Crouzols
  4. Roberta Menafra
  5. Brice Rotureau
  6. Andrew D. Griffiths
  7. Philippe Bastin

  • Curated by eLife

    eLife logo

    Evaluation Summary:

    All reviewers think the study will be really valuable for the field, especially after re-writing to include a detailed comparison with results that were previously published. We all appreciate the clear identification of a gamete sub-population, and also thought that the discovery of low activation of all VSG expression sites was intriguing and will be of considerable interest to those in the field.

    (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. The reviewers all agreed to share their names with the authors.)

Reviewed by

Read the full article

Listed in

Log in to save this article

Abstract

The long and complex Trypanosoma brucei development in the tsetse fly vector culminates when parasites gain mammalian infectivity in the salivary glands. A key step in this process is the establishment of monoallelic variant surface glycoprotein ( VSG ) expression and the formation of the VSG coat. The establishment of VSG monoallelic expression is complex and poorly understood, due to the multiple parasite stages present in the salivary glands. Therefore, we sought to further our understanding of this phenomenon by performing single-cell RNA-sequencing (scRNA-seq) on these trypanosome populations. We were able to capture the developmental program of trypanosomes in the salivary glands, identifying populations of epimastigote, gamete, pre-metacyclic and metacyclic cells. Our results show that parasite metabolism is dramatically remodeled during development in the salivary glands, with a shift in transcript abundance from tricarboxylic acid metabolism to glycolytic metabolism. Analysis of VSG gene expression in pre-metacyclic and metacyclic cells revealed a dynamic VSG gene activation program. Strikingly, we found that pre-metacyclic cells contain transcripts from multiple VSG genes, which resolves to singular VSG gene expression in mature metacyclic cells. Single molecule RNA fluorescence in situ hybridisation (smRNA-FISH) of VSG gene expression following in vitro metacyclogenesis confirmed this finding. Our data demonstrate that multiple VSG genes are transcribed before a single gene is chosen. We propose a transcriptional race model governs the initiation of monoallelic expression.

Article activity feed

  1. eLife

    Evaluation Summary:

    All reviewers think the study will be really valuable for the field, especially after re-writing to include a detailed comparison with results that were previously published. We all appreciate the clear identification of a gamete sub-population, and also thought that the discovery of low activation of all VSG expression sites was intriguing and will be of considerable interest to those in the field.

    (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. The reviewers all agreed to share their names with the authors.)

  2. eLife

    Reviewer #1 (Public Review):

    Other reviewers will concentrate on the VSG expression aspect, so I will comment only on the remaining data on differential expression.

    Last year, Vigneron et al published single-cell data for trypanosomes from salivary glands, and identified markers for epimastigotes, and clusters that they designated as early and late metacyclics. They used the 10x system, which is probably more sensitive and has fewer barcode errors than the than the "inDrop" system that was used in the current submission (see https://doi.org/10.1016/j.molcel.2018.10.020). There are also separate comparisons available for bulk RNASeq of different Tsetse compartments, as well as for in vitro-induced formation of epimastigotes and metacyclic cells (including one extremely recent study from the Tschudi lab). In comparison with Vigneron et al, the major achievement of this paper, apart from then VSG analysis, is the identification of a separate "gamete" cluster. The current submission, however, seriously lacks detailed acknowledgement of, and comparison with, existing published datasets. I found enormous discrepancies with the Vigneron et al study - perhaps this is technical but I can't tell. This could readily be amended, without any additional data.

  3. eLife

    Reviewer #2 (Public Review):

    In vivo characterisation of the establishment of metacyclic T. brucei monoallelic VSG expression is experimentally challenging due to the heterogeneous nature of the trypanosome population in the salivary glands. In this paper, Hutchinson, Foulon et al profile metacyclic VSG gene activation during metacyclogenesis in vivo by using single-cell RNA sequencing technology.

    scRNAseq of cells isolated from the salivary glands of infected tsetse identified 5 populations of cells corresponding to different developmental stages- midgut, epimastigotes, gametes, pre-metacyclic and metacyclic cells. 86.7% of pre-metacyclic cells were demonstrated to express more than one VSG and this multi-VSG expression was resolved to monoallelic expression in 86.8% of cells in the metacyclic population. This observation was corroborated by reanalysis of VSG expression in a previously generated salivary gland trypanosome dataset (Vignon, O'Neill et al. 2020). Single molecule RNA-FISH was used to confirm the sequencing data findings. The authors propose a model whereby a transcriptional race initiates in pre-metacyclic cells and, through the recruitment of other factors, a transcriptional threshold is reached for a single VSG that is expressed in the metacyclic cell. The results are an interesting new insight into the dynamics of metacyclic VSG gene activation and the data presented supports the conclusions drawn.

    The methodologies used are cutting-edge and well suited to answer the research question. The authors choose inDrop scRNA-seq technology, as opposed to the 10X approach that has been adopted in the only other scRNA-seq study of salivary gland trypanosomes (Vignon, O'Neill et al. 2020). Conceptually both methods are similar, they are droplet microfluidics-based technologies and rely on 3' mRNA barcoding, thus producing datasets with a 3' bias. The authors clearly demonstrate that this method is equally as robust for measuring the transcriptome of single trypanosomes (in fact they detect more UMIs/genes per cell). Where the authors excel is their attention to detail when a) assessing the impact of the inDrop experimental procedure on the viability of their cells (though additional controls could be added, the cells appear to be minimally stressed) and b) rigorously testing that artefacts (e.g barcode swapping, sequencing depth, ambient RNA) are not responsible for multiple VSG expression. This thorough approach adds strength to their results.

    The bioinformatic pipelines developed by the authors to facilitate the analysis of strains for which no reference strain is available could prove useful for the community as the study of field strains of trypanosomes is prioritised.

    The authors' main finding is independently confirmed in three different trypanosome strains/experimental settings, which adds strength to their results (particularly the smRNA-FISH confirmation of the sequencing data). The second of these approaches (reanalysing the published dataset) is slightly less convincing than the others. Only 17 pre-metacyclic cells are identified, and these appear to differ in expression of VSG and EP1 procyclin relative to metacyclics and epimastigotes when compared to the authors own pre-metacyclic data. Though these pre-metacyclic cells clearly do express more than one VSG at a low level, we do not have a background level of VSG transcription to compare to.

    The authors identify a putative population of gamete cells from their data based on marker gene expression, thus presenting a more complete picture of salivary gland trypanosome developmental stages than was previously published in another scRNA-seq dataset (Vignon, O'Neill et al. 2020). However why this population, and an additional midgut population, was identified in this study but not the previous has not been discussed in the text. Furthermore, the authors do not expand on their claim that their data 'raises questions as to whether there are two potential modes of development in the salivary glands'.

  4. eLife

    Reviewer #3 (Public Review):

    Hutchinson et al present the second paper on scRNAseq from tsetse salivary glands to hit the literature over the last year. They use a different technique from Vigneron et al (PNAS 2020), and though the results are generally congruent, they focus on an aspect of biology missed by the previous authors. Specifically, they demonstrate that metacyclic VSGs, which are monoallelically expressed in metacyclics, are in fact co-expressed in pre-metacyclic cells, suggesting that the establishment of monoallelic expression is developmentally regulated (and proposing specific mechanisms as to how, based on that).

    Specific comments:

    Figure 1 sets up the technique and shows nice separation of cultured BSF, PCF and Ramos B cells. Two questions: is it concerning that VSG transcripts also show up (at lower intensities) in PCFs (and vice versa with EP - Figure 1E)? And a related question: what is the limit of detection (ie how many transcripts are reliably captured per cell and at what depth?) From the figure it would appear that the number of Ramos transcripts per cell is higher than per tryp (naturally) but does this mean better coverage for BSF and PCF transcripts? This is important, as a Ramos "spike" is used throughout to normalize amount of "ambient" RNA in the drop.

    Figure 2 shows convincingly the subtypes of trypanosomes present in the salivary glands. And Figure 3 notes transcriptional profiles within those subtypes. Again it would be nice to know the absolute cutoff here (is InDrop capturing 50% of all transcripts? more? it's important, if this is not only meant as a validation pipeline but also a discovery pipeline, including discovery of intra-subtype heterogeneity, as was done in the gamete cluster - Figure 4).

    Figure 5 is in many ways the most exciting. It is particularly interesting that the authors reprocess data from a recently published paper (Vigneron, O'Neill et al. (2020)) that uses 10X rather than InDrop, to arrive at the identical finding (missed by Vigneron et al) that the mVSG transcriptional program initiates early and becomes monoallelically restricted only in the final, metacyclic stage (a finding which they then also validate in vitro, with RBP6 over expression - that does not look as stunning, however, mostly due to lack of antibodies against all 6 mVSGs and the inability to do single cell proteomics).

    Overall these are well done experiments, and the conclusions are justified. It is an important addition to the literature.

  5. Version published to 10.1371/journal.ppat.1009904
  6. Version published to 10.1101/2021.03.01.433049v1 on bioRxiv