Flagellar cAMP signaling controls trypanosome progression through host tissues

The unicellular parasite Trypanosoma brucei is transmitted between mammals by tsetse flies. Following the discovery that flagellar phosphodiesterase PDEB1 is required for trypanosomes to move in response to signals in vitro (social motility), we investigated its role in tsetse flies. Here we show that PDEB1 knockout parasites exhibit subtle changes in movement, reminiscent of bacterial chemotaxis mutants. Infecting flies with the knockout, followed by live confocal microscopy of fluorescent parasites within dual-labelled insect tissues, shows that PDEB1 is important for traversal of the peritrophic matrix, which separates the midgut lumen from the ectoperitrophic space. Without PDEB1, parasites are trapped in the lumen and cannot progress through the cycle. This demonstrates that the peritrophic matrix is a barrier that must be actively overcome and that the parasite’s flagellar cAMP signaling pathway facilitates this. Migration may depend on perception of chemotactic cues, which could stem from co-infecting parasites and/or the insect host.

1. The title needs changing -the work here is not about navigation through tissues. 2. Given the work on genetics of motility phenotype defects in other systems such as Chlamydomonas a rigorous assessment of motility is critical for the statement to be made that this is about flagellar signalling and not a motility defect. This is too important a result that the only data presented on the mutant movement is the graph of mean square displacement. They have the videos so Ione would also expect to see the traces of the cell movement analysed in terms of directional issues, processivity, etc. It is perfectly possible for the MSDS graphs to be similar but the mutant cells to exhibit other motility phenotypes. Also, the standard assessment of motility should also include an analysis of high speed movies of the cells to show that the flagella behave in the same way. The results here can still be explained by some subtle motility defect. 3. The lines 176 -182 suggest that the worry is that the KO cells have become late procyclic forms since the Social Motility phenotype is restricted to early forms. The authors suggest this is not an issue since the KO express the GPEET marker. However, does this not raise the more fundamental issue of whether the mutant cells in the tsetse (albeit fewer) that are able to establish the late infections are indeed late procyclics. Ie that the mutants are defective in transition to the late forms in the fly and so there are less of them because of replication of this form not because of "navigation"? The type of parasites seen in the infections of both mutant and wild type surely need assessing because of this issue? 4. There is variability between the infection rates on different experiments (3a vs 5a), which is concerning with a significant infectivity effect seen in 3a but not 5a. Is there a general loss of fitness in the KO or not, the results make it difficult to be sure? Moreover, the use of t-tests seems inappropriate for analysing changes in categories a chi squared test would be better. The proventriculus effect does appear to be consistent between the experiments though. 5. The use of a northern blot to confirm the KO seems odd as this just shows lack of expression not loss of the gene -a Southern or PCR test is needed to show definitively that the gene is gone and not that its expression has stopped. In that phase (but possibly not others or not in vivo) of the life cycle. 6. I was confused by the timings in the discussion -early procyclic forms according to their data are lost by day 7 (fig. 5c -GPEET expression) yet on line 189 day 7 is when the trypanosomes cross the PM and enter the ectoperitrophic space -if this is the time that SOMO is required for movement across the PM why are they switching away from expression of GPEET to late procyclic markers and to a cell type not capable of SOMO? 7. Reading the literature about early and late parasites it appears that glycerol is required to maintain early PCFs yet here this stock of 427 cells are maintained as early PCFs without glycerol, which seems counterintuitive? 8. Given the authors earlier work and that of others on genetics of establishment of infections in the salivary gland how do the authors square the idea of mass movements via SoMo with the idea that there are large bottlenecks and a founder effect. 9. It would've been very valuable to consider the tsetse infection phenotype of true (paralysed) motility defects relative to the SoMo defect they show. One issue is the general baseline for motility of any kind in this life cycle transition process. What is the magnitude of the defect relative to a cell that can't swim? This is essentially a question about the biological significance of the result.
Reviewer #2 (Remarks to the Author): Trypanosoma brucei is transmitted by tsetse flies. Once inside the fly, the parasites have to overcome several challenges in order to survive and develop, including colonization of the midgut ectoperitrophic space. This step is essential for establishing an infection in the tsetse. How trypanosomes reach this compartment is still debatable, but the most accepted hypothesis is that they must cross the fly's peritrophic matrix (PM). In addition, it is well accepted that in vitro grown procyclic trypanosomes (the equivalent forms to the ones found in the tsetse gut) undergo social motility (SoMo). The molecular bases of trypanosomes SoMo are still unknown, but elegant work published by the same authors have shown that this phenomenon occurs only in early procyclic cells and that it may involve multiple pathways/genes, including cAMP signalling systems. In addition, whether SoMo is relevant for parasite colonization of the fly's gut has been a burning question within the tryp community. Thus, this is a relevant and important topic for understanding vector-parasite interactions in general.
In this MS the authors investigated the in vivo role of SoMo by infecting tsetse flies with a T. brucei phosphodiesterase PDEB1 deletion mutant, which shows a SoMo phenotype in vitro. By using live confocal microscopy analyses of labelled tsetse tissues and fluorescent parasites, the authors show evidence that PDEB1 mutants remained trapped in the tsetse gut lumen, suggesting a possible role in the migration to the ectoperitrophic space. However, although there is no doubt that this mutant shows an infection phenotype in the fly, the main problem of this paper is the bias interpretation that this is due to the lack of SoMo and not as direct consequence of having a defect in cAMP metabolism. Of course, defects in cAMP metabolism could lead to a SoMo phenotype, but so other other mutations that in principle are not metabolically related. Thus, given the lack of clarity and experimental evidence on the importance of trypanosome SoMo during colonization of the tsetse gut, the paper becomes merely descriptive and rather reports on the in vivo essentiality of another trypanosome gene.
Below, I list several other concerns and suggestions that need to be considered by the authors in order to show a stronger evidence of the potential role of trypanosome SoMo in the tsetse: 1) The environment of the tsetse midgut (including the type and availability of some nutrients) is very different to the nutritional environment of any procyclics culture media. This means that the authors may have overlooked a growth phenotype of the PDEB1 mutants if these parasites are cultured in chemically defined media. I wonder if at least the authors have tested the growth phenotype of this mutant in media with no or little glucose in it, which would be a closer environment to the tsetse gut.
2) It is not clear which parasite stage was used for tsetse infections, although everything indicates it was procyclics, in which case I don't think is the right one for these experiments. For the purpose of investigating the role SoMo during the early stages of a trypanosome infection in the tsetse, the authors should have used instead BSF. In fact, it'd be cleaner to use a tet-inducible conditional null instead of add back mutants, as sometimes addb back clones differs in infectivity. The time in culture of PCF is critical for tsetse infectivity and tissue tropism, even if these parasites have been in culture only for few days.
3) In general, the quality of the fluorescent microscopy figures is not that great (e.g. Figure 7). The authors need to use other dies and tools to better define the different fly tissues. The same happens with the videos of infection by WT parasites which, in some cases, it is difficult to know what tissues or what aspects of the tissues the authors are pointing out. 4) Importantly and in relation to the previous comment, why the authors did not include videos of PDEB1 mutants in the fly? To me, this is absolutely essential to demonstrate their point. If trypanosome do not cross the PM, where do they go? Do they get stuck in the PM or are they expelled in the frass?
5) It seems that only one biological replicate was made per experiment and using very low tsetse numbers. These experiments need more biological replicates to determine its significance.
6) The number of GPEET+ cells increases in KO cell lines at 7dpi (Fig 5C). Although GPEET expression may have nothing to do with SoMo (but may indicate that the cells are still as "early" procyclics), this delayed expression may also suggest a potential delay in PV colonization by PDEB1 mutants. Therefore, the PVs from flies infected with mutant trypanosomes should have been scored also at later time points to make sure that infection of this organ does not occur at a slower pace due to an overall delayed differentiation process in this KO cells.
Other comments: 1) Figure one is both wrong and low quality. First, the tissues used look broken; i.e. the midgut in panel B has a discontinues nuclei staining whereas the proventriculus shows a strange morphology and the nuclei staining is not defined as shown for the midgut section and looks like autofluorescence. In addition, in panel A, the proventriculus is misplaced and looks connected to the midgut by a fine thread, which I believe represents the anterior midgut.
2) Why using different statistical tests to determine significance in Figure 3A and 3B?
3) Why does the ectopic PDEB1 shows a lower apparent molecular weight with the same probe? Also, in Figure 4A, the second band on the left I believe is PDEB2 and not PDEB1 (as shown in Figure 2A). 4) A recent report (Schuster S. et al 2017, e-Life) showed evidence of "collective motion" occurring at different times of a trypanosome infection and in different tsetse tissues. Despite its potential importance to explain trypanosome migration within the tsetse, this publication was completely ignored by the authors.

T. brucei in vitro in vivo
The authors have added some new data. However, there are issues -some about the data needed to provide a firm conclusion in this (acknowledged) difficulty of showing a clear causal link to one phenomenon as an explanation for a general journal like Nature. The second issue is the balance of the writing that really does not balance out the possibilities of the various possibilities and conclusions.
The title is not a simple issue of replacing navigation with progression ...the issue is that the paper does not address movement through tissues. It deals with movement from one part of the insect gut to another. Also,ot only that it is not the HOST that is being studied --it is the VECTOR. Of course, one can take it that because there is an infective cycle of these parasites in the insect that it is a "host". However, to do so would be to go against all usage of the Host-Vector terminology in the protozoal pathogens. In addition a general reader would surely expect this paper to be about transfer of the pathogen from blood to brain in its host pathology aspect ...which is absolutely not what the paper is about.
The big thing is that as the authors acknowledge in the discussion is that you can't easily connect this inability to establish a proventriculus infection with a loss of SOMO...
There are a few of points that follow from this and from the new data (not in order of importance: 1) Given that they have done the PCR analysis of the KOs and shown it in the responses to the reviewers they should include that data in the supplementary manuscript and not have that as data not shown.
2) The authors say there is not a motility defect in the PDEB1 KO . That is right in one sense the KO cells are able to move in culture; however, the movement of KO is different from the parental cells and so in my opinion there is a change in motility that the authors have not acknowledged as an explanation for the phenotype.
3) As the authors have the addback cell lines and since the motility assays were requested, it would be good to run the in vitro motility analyses on these cells to see if they restore the movement to parental levels. Since not provided I recognise that this is yet more work but the judgment is about how to ensure clearer discussion and conclusion.
4) The focus on SOMO in the discussion is understandable but it does seem constantly odd that the possibility of this being a simple chemotactic response i.e. the KO is unable to respond to a cue from the host is reduced to the phrase 'and other activities that are necessary'. They allude to this earlier in the results when talking about the change in motility but not really in the discussion. Given that this is surely absolutely as plausible as the SOMO option I feel it should be discussed as well -chemotaxis has been postulated to be important in the related parasite Leishmania in the sand fly. One could discuss this trypanosome -vector system in the context of other protozoan parasite -vector interactions where movement, chemotaxis and differentiation triggers have been studied.
Reviewer #3 (Remarks to the Author): The different comments and concerns of reviewer 2 are adequately addressed by the authors. The work presented in this manuscript does not show a direct significance of SoMo (observed behavior in cultured early procyclics) for the progress of the in vivo parasite development in the tsetse midgut. This is now better and more unambiguously phrased by the authors in the discussion part. However, in the abstract a suggestive and 'biased' interpretation of the in vivo relevance of SoMo still remains. Therefore, the last sentence of the abstract should be re-phrased as follows: "These results show that cAMP signaling is crucial for successful transmission and correlates with the parasite's capacity to migrate through host tissues". The inclusion of SoMo here in this last sentence is still too suggestive for its in vivo relevance which cannot be justified from the experimental results presented in this study.