Author's response

Gary Ward shared

Open questions

1.I enjoyed a lot reading your work. Starting with a broad question – you mention in your introduction the relevance of post-translational modifications for parasite biology, and focus on palmitoylation. Are there other post-translational modifications that are equally relevant or important for Toxoplasma gliding motility, invasion and egress?

The post-translational modification that has been studied most extensively in apicomplexans is phosphorylation, which is widespread and has been implicated in motility, invasion and egress. Focusing on motility, each component of the T. gondii glideosome is known to be phosphorylated on multiple sites. While we don’t know the functional significance of most of these phosphorylations, my lab showed several years ago that phosphorylation of recombinant TgMyoA on three serines near its N-terminus (Ser19, 21 ,29) increases its motor activity. This result was consistent with an elegant study on TgMyoA phosphorylation done in parasites a few years earlier in the Arrizabalaga lab. More recent work demonstrated that phosphorylation of Ser19 also alters the motor function of Plasmodium falciparum MyoA. So a role for MyoA phosphorylation in controlling motor function and parasite motility is at this point well established.  As for post-translational modifications other than phosphorylation, Aoife Heaslip and Ke Hu did some intriguing work suggesting a role for lysine methylation in activating Toxoplasma motility. Palmitoylation and acetylation are also widespread in apicomplexans, including on proteins involved in motility and invasion. Altogether, phosphoproteome, palmitome and acetylome studies have identified more than 25,000 sites of post-translational modification on T. gondii proteins; however, in very few cases do we know anything about their functional significance.

2.Another broad question: how do you expect these post-translational modifications to occur and vary within different tissues within the mammalian host? Namely, how would the parasite location (choice of cell and organ) influence this, and therefore the different phenotypic effects possibly observed in different organs or even hosts?

This is a great question about which we know virtually nothing. In the Plasmodium falciparum MyoA phosphorylation story I mentioned above, Robert-Paganin and colleagues showed that unphosphorylated MyoA could generate more force under load than phosphorylated MyoA, but this ability to generate force came at the expense of the speed with which the motor could displace actin filaments. They suggested that the two phosphorylation states of the motor were “tuned” to function either in invasion (which likely requires the ability to generate more force) or dissemination (where speed may be more important). An interesting model, although purely speculative.

3.More to the specifics – you mention in your manuscript that for a decade, the linear model has dominated the field, but that there are observations that are difficult to reconcile with this model. Are there alternative models already in existence that do reconcile those observations, including the ones you provide here in this manuscript? If not, is there one you would suggest given your findings?

Other models have been proposed, although none (including the linear motor model) can fully explain all observations. Most recently, the Meissner group proposed in intriguing “fountain flow” model, in which an endocytic-secretory cycle drives retrograde membrane flow on the parasite surface, contributing to parasite forward movement. I think the most direct way to discriminate between and directly test the various models would be to visualize the forces the parasite produces as it moves. The Frischknecht and Tardieux labs have done some really nice work in this area, providing insight into parasite force generation on 2D polyacrylamide surfaces. Their data suggest the importance of highly focal zones of attachment to the 2D substrate. We are hoping to eventually move traction force mapping into three dimensional extracellular matrices, where the different models of motility predict distinctly different patterns of force generation. I am most interested in exploring the possibility that more than one motility mechanism exists, and that the parasite uses different mechanisms in different settings (e.g., penetrating a tight junction vs moving along a 2D surface vs moving in 3D through the interstitial extracellular matrix). Many mammalian cells have this capability. For example, cells  may move along planar surfaces via ARP2/3-mediated actin polymerization and extension of their leading edge, but when placed in confined spaces the forward movement of these very same cells is instead driven by polarized water permeation.

4.You investigated possible changes in interactions between glideosome components that could result in compensation to the disrupted interaction between TgGAP45 and TgMLC1 upon blocking palmitoylation, resulting in effective motility. Three questions arising from this work, for me, are: what are your alternative hypothesis on how motility remains intact? Did you observe any differences in other phenomena linked to palmitoylation such as invasion and egress? Do you observe altered infectivity in vivo, which could be missed in in vitro assays?

(1) Motility may remain intact because: (a) the linear motor model or its proposed dependence on TgMLC1-TgGAP45 coupling is wrong; (b) there are redundancies in the system that allow the linear motor to continue generating force even in the absence of TgMLC1-TgGAP45 interaction; or (c) there is more than one way for the parasite to move, and when you block one mechanism another takes over. As we acknowledged in the paper, it could also be that the non-palmitoylatable TgMLC1 still interacts with TgGAP45 with sufficient affinity to couple the motor to the IMC, but insufficient affinity to withstand extraction in non-ionic detergent. It would be very difficult experimentally to completely rule out this possibility. I am going to wait to see what parasite-generated forces look like in 3D (see above) before casting my vote for a particular model or deciding that a completely different one is called for. (2, 3) We did not look at the effect of the mutations on other motility-based phenomena such as invasion or egress, or on disease progression in vivo. These experiments are certainly worth doing, and if a clear phenotype is observed this phenotype would be worth following up. For this particular study we thought we would go straight to the heart of the matter and assay motility itself. We were very surprised to see that the ability of the parasites to move was unaffected by the mutation. Even if it turns out that some other phenotype associated with the mutation is revealed by further work we would still be left with the question of how the mutant parasites can move so well in the 3D motility assay when  TgMLC1 appears to be uncoupled from TgGAP45.

5.From what I understood in the description of the linear model, a question that arose in the context of your work is whether interactions between ELC1 or ELC2 and TgGAP45 would be enough for effective motility, even in the absence of the interaction between MLC1 and TgGAP45?

Based on Figures 5 and S3 in our paper, it seems unlikely that ELC1 binds directly to TgGAP45  when TgMLC1-TgGAP45 interaction is disrupted: in the immunoprecipitates of the mutant parasites, ELC1 continues to associate with TgMyoA/TgMLC1 rather than with TgGAP45. We did not probe the western blots for ELC2, but the functions of ELC1 and 2 are reportedly redundant, and as far as I’m aware there is no direct evidence that either interacts directly with TgGAP45.