Chikungunya virus requires an intact microtubule network for efficient viral genome delivery

Chikungunya virus (CHIKV) is a re-emerging mosquito-borne alphavirus, which has rapidly spread around the globe thereby causing millions of infections. CHIKV is an enveloped virus belonging to the Togaviridae family and enters its host cell primarily via clathrin-mediated endocytosis. Upon internalization, the endocytic vesicle containing the virus particle moves through the cell and delivers the virus to early endosomes where membrane fusion is observed. Thereafter, the nucleocapsid dissociates and the viral RNA is translated into proteins. In this study, we examined the importance of the microtubule network during the early steps of infection and dissected the intracellular trafficking behavior of CHIKV particles during cell entry. We observed two distinct CHIKV intracellular trafficking patterns prior to membrane hemifusion. Whereas half of the CHIKV virions remained static during cell entry and fused in the cell periphery, the other half showed fast-directed microtubule-dependent movement prior to delivery to Rab5-positive early endosomes and predominantly fused in the perinuclear region of the cell. Disruption of the microtubule network reduced the number of infected cells. At these conditions, membrane hemifusion activity was not affected yet fusion was restricted to the cell periphery. Furthermore, follow-up experiments revealed that disruption of the microtubule network impairs the delivery of the viral genome to the cell cytosol. We therefore hypothesize that microtubules may direct the particle to a cellular location that is beneficial for establishing infection or aids in nucleocapsid uncoating.

> The dotted lines in figure S1B and S3B represents 75% cell viability. We now added this information to the corresponding figure legends (lines 638 & 655).

Why have authors chosen to assess the effects of nocodazole at 10 hrs post infection?
> Growth curve analysis showed that initial CHIKV particle production is seen at 8 hpi and continues to increase thereafter. We chose 10 hpi as this increases the sensitivity yet in the used assays still reflects 1 round of replication. Line 268. The LS3 infection kinetics are described in detail by Scholte et al. (2013Scholte et al. ( + 2015 and confirmed experimentally in our lab (data not shown). Comparable date was obtained for CHIKV-LR OPY1 strain (Bouma et al., 2020).
Using DiD as a fusion reporter does detect hemifusion, but it also reports full fusion. In these assays the authors cannot distinguish between hemifusion and full fusion. So I think it misleading to use hemifusion throughout the paper, and I suggest they use 'hemifusion/full fusion' instead. > We agree with the reviewer that this can be misleading. Yet we do not know if the hemifusion intermediate always proceeds to full fusion. We now tried to explain the terminology more clearly in line 180-184 Can the authors be certain that the fast directed movement described in Fig 2A is intracellular and not on the cell surface? I.e. could it be viral surfing? > In order to assess this question, we investigated fast-directed movement in cells transfected with Clathrin(-YFP) or the early endosome marker Rab5(-GFP). The results (shown in Fig. 3D) show that this type of movement is seen after the process of clathrin-mediated endocytosis and prior to arrival to Rab5-positive early endosomes. Mostly likely transport occurs directly after the clathrin coat dissociates from the intracellular vesicle. Therefore, we conclude this type of movement occurs intracellular and is not viral surfing.
Line 291 -there does seem to be a 30% change in MFI. > The reviewer is correct and we adjusted the text accordingly. Line 293. > Correct, both particles fused in the periphery. The location of fusion is determined by eye. When a virion fuses closer to the nucleus than the plasma membrane fusion was determined as perinuclear. Vice versa, when the particle was closer to the plasma membrane than the nucleus fusion was determined as periphery. The image shown in Figure 2a was primarily selected for visual purposes, as it contains multiple trajectories in which fast directed movement was observed. By chance, the fusion events of both trajectories occurred in the periphery. Line 171/172. distinct patterns prior to membrane fusion in the early endosomes. Using nocodazole to disrupt the microtubule network, they observed reduced number of infected cells, restriction of fusion to the cell periphery, and impaired delivery of the viral genome into the cytoplasm.
The microtubule network has been previously implicated in the entry of multiple other viruses, and thus the innovation here is limited to the demonstration of this requirement for CHIKV and the discovery of the two distinct trafficking patterns. While the data demonstrating the two distinct viral populations is interesting, it remains unclear what is the functional relevance of this difference and what is the mechanistic role mediated by microtubules in CHIKV entry.
Major comments: 1. The magnitude of the effect shown in most figures is small (less than a 2 fold difference in many cases -e.g. Fig. 1A, Fig. 4B), which makes it unclear what the biological relevance of the findings is. > Fig. 1A shows 70-80% reduction in infectivity. In 4B we used shorter incubation times which might slightly reduce the observed effect. In our view given the sensitivity of the applied assays a 2-fold reduction in infectivity is biologically relevant.

The number of infected cells is very low (2.9% in an MOI of 1, 9.2% in an MOI of 20 as per line 288). Why did the authors choose to infect the cells only for 30 min (when a more standard infection time is 1 hour for alphaviruses)? Perhaps a longer infection time will increase infection rate and improve the dynamic range?
> In standard infectivity experiments the incubation time is 1.5 hrs (see line 211). However in case of indicated experiments we have shortened it because our previous paper (Hoornweg et al., 2015) demonstrated that fusion events occur very rapidly after addition of the virus to cells.

The authors should demonstrate a dose-response effect for the various phenotypes shown with nocodazole treatment.
> We do not fully understand the remark of the reviewer. We used a non-toxic nocodazole concentration that completely disrupted the microtubule network. At lower concentrations, the microtubule network is at least partially intact and it is hard to draw conclusions from this. For similar reasons, single virus tracking was only done at the concentration were the microtubule network is disrupted.

Is it possible the cell cycle arrest at G2/M induced by nocodazole accounts for the observed reduction in infection and trafficking of viral particles?
> Nocodazole is indeed known to be used in cell biology to synchronize the cell division cycle as cells arrest at G2/M-phase when treating for a prolonged time (>12-16h). This arrest is a direct consequence of microtubule disruption, as the cell cannot assemble the mitotic spindle.
In most experiments we treat the cells for 2h prior to infection. Tracking experiments are recorded within 30 min. Therefore, there is no reason to believe that the altered trafficking behavior is an indirect consequence of cell cycle arrest but rather a direct effect of the disruption of the microtubule network.

The distance and velocity traveled by individual viral particles needs to be shown in figures 2 and 3.
> Unfortunately, due to cropping of the movies for proper trajectory analysis, the original dimensions of the images analyzed were not recognized by the analysis software. Therefore, the exact velocity of the particles is unknown and the distance travelled by these particles cannot be calculated. Consequently, the relatively velocity over time as reported by the analysis software is shown in figures 2 and 3. Although the exact velocity cannot be given, these graphs, together with the provided images and scale bars, show that during the burst of directed movement the viral particles shown in Fig. 2a travel approximately 2,5-3,5 µm in 4-5 seconds. From this, one can calculate a velocity of 0,5-0,8 µm/s during directed movement, which is in line with the velocity reported for microtubule-dependent movement in literature (on average 0,3-1,5 µm/s (Kulic et al., 2008)).
6. In Fig. 4A, the Western blot should include all 3 conditions (control, and two treatments). > The Western blot is performed to confirm cell fractionation in each experiment. To detect this, we add an additional well to each experiment (purely for WB). As the fractionation protocol is independent of the treatment condition we did not perform WB for all three treatment conditions. In our view this is not necessary and due to the lock-down of our lab due to COVID19 we are unable to perform these experiments at this moment in time.
Additional comments: 1. Graphs in all figures -would be better to show controls in each figure rather than showing just the % (or alike) difference relative to control. > We believe this is a matter of taste. We prefer to show % inhibition but will change it if the above is the policy of the journal.
2. Would be good to avoid single column graphs (as in 1D). These can be just mention in the text. > We do think these graphs are informative as they show the experimental replicates and standard deviations, therefore we for now did not remove this data from the figures.
3. Figure 1E -a schematic showing time of drug addition would be helpful. Thank you, added 4. Figure 3D -the legend should explain what before, during and after means. Thank you, added 5. Embedding figure legends in the text makes it harder to review. We followed the guide lines given by PLOS Neglected Tropical Diseases, as they stated "Insert figure captions in manuscript text, immediately following the paragraph where the figure is first cited (read order). Don't include captions as part of the figure files themselves or submit them in a separate document".
Reviewer #2: Overall, this work provides a modest step forward in understanding CHIKV entry into cells. Unfortunately, there is really no indication of how microtubule based transport of CHIKV favors penetration and infection. This report describes for the first time that CHIKV particles traffic along microtubules during cell entry and that gRNA release is impaired in nocodazole-treated cells. The exact role of nocodazole is indeed not uncovered and this is primarily due to the fact there are no methods available to capture the transient and dynamic events of CHIKV nucleocapsid delivery/uncoating. --------------------