On 2013 Jun 27, Markus Meissner commented:

Reply to Gary Ward:

Hi Gary. Thank you for your comments. We are glad you find the review helpful. For now, I prefer to focus on the first two comments and just mention a few information on the collar. 1) Regarding Non-essential and non-important. I agree with you. In fact all the discussed genes are probably essential in vivo. Although we only have preliminary data for the ama1KO that shows that this parasite is highly attenuated, similar data were obtained for a knockdown of MyoA and MIC2 (Meissner et al., Science 2002; Huynh et al., 2006). I have no doubt that other mutants behave similar. Furthermore, we know that MLC1, GAP45 and act1 are essential and depletion leads to parasite death. However, depletion does NOT lead to a block in host cell invasion. Therefore, none of the discussed genes can be defined as "essential for the process of host cell invasion". I agree that they are important for invasion. Invasion is however a multistep process (attachment, reorientation, junction formation, penetration) and we need to know which step is influenced by removal of the respective gene in order to obtain a clear picture of the (real) invasion mechanism. 2) Regarding interpretation of KO data you propose a third possibility, which relies on the accumulation of compensatory mutations or changes in gene expression. This explanation is certainly appealing and valid for some of the effects seen. For example it could be argued that deletion of an attachment factor can be compensated by upregulation of another, as seen in Plasmodium. Especially in case of non-essential genes (in vitro), long term adaptations are a strong possibility and like you we do observe this effect for some of the KOs. However, in case of conditional KOs for essential genes (i.e. actin) we observe that the whole culture dies within 10 days due to the loss of the apicoplast (Andenmatten et al., 2013). During this whole period the parasites remain able to invade. How would you imagine a scenario, where a gene remains essential for two processes (apicoplast division and egress) but not for invasion? I would strongly argue against a scenario, where an actin-like protein can form a filament (which has been never demonstrated in apicomplexans) that can be used by MyoA only during invasion but not during egress.

Regarding the Collar, we are not sure at this point what exactly it reflects. In the provided images wild type parasites were analysed and it will be certainly interesting to investigate if KO parasite show an increase in collar formation. Clearly, we need to analyse the collar in detail in the future.

On 2013 Jun 24, Gary Ward commented:

Thanks to the authors for this timely reanalysis of the current model of apicomplexan invasion. Recent work from the Meissner group has shown that several of the parasite proteins previously thought to be essential for invasion can be knocked out, casting some doubt on the model. I agree with much of what they write, but have three comments:

1) Non-essential does not mean non-important. In several of these knockouts, invasion and motility are dramatically impaired (e.g., ~80-85% inhibition of invasion) and defects of this magnitude can translate into greatly reduced virulence in vivo (e.g., Meissner et al, Science [2002] 298, 837). While the new data show clearly that MyoA and AMA1 are not essential for invasion, they at the same time show them to be important for invasion. Whether they are important for the reasons predicted by the current model is a separate question.

2) In the authors' view, the data either show that the "parasites use a single entry mechanism and hence the current invasion model is wrong and needs to be replaced by a new model, [or] the current model is overall valid but an additional, motor independent invasion mechanism is at work that facilitates host cell invasion in KO mutants of the glideosome". Redundant mechanisms are a common feature of parasite biology (e.g., erythrocyte-binding proteins of malaria parasites, Trends in Parasitology [2012] 28, 23), so I agree with these two possibilities. However, I would add a third: by knocking out a parasite protein that plays an important role in a process that is so critical to the parasite lytic cycle, enormous selective pressure is put on the parasites to come up with a way around the problem. Compensatory mutations or changes in the expression of other proteins that can "fill in" for the missing protein may occur. In fact, we have seen an improvement over time in the ability of some of these mutants to invade, which likely represents some advantageous mutation or change in gene expression that is being selected for in culture. Such changes will be hard to track down and even harder to rule out, but the phenotype observed in these knockouts will be due to some combination of the knockout itself and whatever the parasite does to overcome it.

3) The collar seen around the invading tachyzoite in Fig. 2 is intriguing, and unlike what we typically see by EM during invasion of human foreskin fibroblasts. How rare are these events, and are they specific to this cell type? This profile could reflect an alternative invasion mechanism, as proposed. On the other hand, these cells might have an unusually impenetrable cortical cytoskeleton and the parasites push but can't get in. In this case the moving junction would in fact really be a moving junction and would be pulled along the body of the stationary parasite, much like occurs during Cryptosporidium invasion and consistent with the standard model of invasion.

On 2013 Jun 24, Gary Ward commented:

None

On 2013 Jun 24, Gary Ward commented:

None

On 2013 Jun 24, Gary Ward commented:

Thanks to the authors for this timely reanalysis of the current model of apicomplexan invasion. Recent work from the Meissner group has shown that several of the parasite proteins previously thought to be essential for invasion can be knocked out, casting some doubt on the model. I agree with much of what they write, but have three comments:

1) Non-essential does not mean non-important. In several of these knockouts, invasion and motility are dramatically impaired (e.g., ~80-85% inhibition of invasion) and defects of this magnitude can translate into greatly reduced virulence in vivo (e.g., Meissner et al, Science [2002] 298, 837). While the new data show clearly that MyoA and AMA1 are not essential for invasion, they at the same time show them to be important for invasion. Whether they are important for the reasons predicted by the current model is a separate question.

2) In the authors' view, the data either show that the "parasites use a single entry mechanism and hence the current invasion model is wrong and needs to be replaced by a new model, [or] the current model is overall valid but an additional, motor independent invasion mechanism is at work that facilitates host cell invasion in KO mutants of the glideosome". Redundant mechanisms are a common feature of parasite biology (e.g., erythrocyte-binding proteins of malaria parasites, Trends in Parasitology [2012] 28, 23), so I agree with these two possibilities. However, I would add a third: by knocking out a parasite protein that plays an important role in a process that is so critical to the parasite lytic cycle, enormous selective pressure is put on the parasites to come up with a way around the problem. Compensatory mutations or changes in the expression of other proteins that can "fill in" for the missing protein may occur. In fact, we have seen an improvement over time in the ability of some of these mutants to invade, which likely represents some advantageous mutation or change in gene expression that is being selected for in culture. Such changes will be hard to track down and even harder to rule out, but the phenotype observed in these knockouts will be due to some combination of the knockout itself and whatever the parasite does to overcome it.

3) The collar seen around the invading tachyzoite in Fig. 2 is intriguing, and unlike what we typically see by EM during invasion of human foreskin fibroblasts. How rare are these events, and are they specific to this cell type? This profile could reflect an alternative invasion mechanism, as proposed. On the other hand, these cells might have an unusually impenetrable cortical cytoskeleton and the parasites push but can't get in. In this case the moving junction would in fact really be a moving junction and would be pulled along the body of the stationary parasite, much like occurs during Cryptosporidium invasion and consistent with the standard model of invasion.

On 2013 Jun 27, Markus Meissner commented:

Reply to Gary Ward:

Hi Gary. Thank you for your comments. We are glad you find the review helpful. For now, I prefer to focus on the first two comments and just mention a few information on the collar. 1) Regarding Non-essential and non-important. I agree with you. In fact all the discussed genes are probably essential in vivo. Although we only have preliminary data for the ama1KO that shows that this parasite is highly attenuated, similar data were obtained for a knockdown of MyoA and MIC2 (Meissner et al., Science 2002; Huynh et al., 2006). I have no doubt that other mutants behave similar. Furthermore, we know that MLC1, GAP45 and act1 are essential and depletion leads to parasite death. However, depletion does NOT lead to a block in host cell invasion. Therefore, none of the discussed genes can be defined as "essential for the process of host cell invasion". I agree that they are important for invasion. Invasion is however a multistep process (attachment, reorientation, junction formation, penetration) and we need to know which step is influenced by removal of the respective gene in order to obtain a clear picture of the (real) invasion mechanism. 2) Regarding interpretation of KO data you propose a third possibility, which relies on the accumulation of compensatory mutations or changes in gene expression. This explanation is certainly appealing and valid for some of the effects seen. For example it could be argued that deletion of an attachment factor can be compensated by upregulation of another, as seen in Plasmodium. Especially in case of non-essential genes (in vitro), long term adaptations are a strong possibility and like you we do observe this effect for some of the KOs. However, in case of conditional KOs for essential genes (i.e. actin) we observe that the whole culture dies within 10 days due to the loss of the apicoplast (Andenmatten et al., 2013). During this whole period the parasites remain able to invade. How would you imagine a scenario, where a gene remains essential for two processes (apicoplast division and egress) but not for invasion? I would strongly argue against a scenario, where an actin-like protein can form a filament (which has been never demonstrated in apicomplexans) that can be used by MyoA only during invasion but not during egress.

Regarding the Collar, we are not sure at this point what exactly it reflects. In the provided images wild type parasites were analysed and it will be certainly interesting to investigate if KO parasite show an increase in collar formation. Clearly, we need to analyse the collar in detail in the future.