PREreview of Geometry effects on protein mobility in a synapse

This review resulted from the graduate-level course "How to Read and Evaluate Scientific Papers and Preprints" from the University of São Paulo, which aimed to provide students with the opportunity to review scientific articles, develop critical and constructive discussions on the endless frontiers of knowledge, and understand the peer review process.

Review by José Renato Silva dos Santos, Rodolfo Pereira, Sophia Fernandes Dias de Lima.

1. introduction and context

This preprint discusses how the geometry of the presynaptic terminal and the dynamics of protein binding/unbinding to synaptic vesicles can influence the interpretation of Fluorescence Recovery After Photobleaching (FRAP) experiments. The relevance of the topic is due to the fact that FRAP is widely used to measure diffusion coefficients and immobilized fractions of proteins in synapses, being a crucial tool to elucidate the dynamics of essential components for neuronal functioning.

In general, the study is structured in such a way as to compare different photodegradation (bleaching) protocols and to demonstrate that recovery times can reflect not only the effective diffusion of proteins, but also intra-synaptic redistribution phenomena and protein influx through the axon. These findings may be relevant to researchers in cellular neuroscience, synaptic physiology and related areas.

1. Overall Remarks

The problem raised by the authors, namely the difficulty in properly interpreting FRAP recovery times in synapses, is an interesting issue. The ability to distinguish whether a shorter recovery time is the result of a higher diffusion coefficient or local protein redistribution (and not necessarily high mobility) contributes to the advancement of fluorescence microscopy analysis techniques.

The text clearly presents the motivation for studying the influence of synaptic geometry on protein mobility. However, some points could be explored in greater depth, especially in relation to the initial hypotheses and the theoretical basis (stochastic reaction-diffusion). It would be useful to include additional explanations, in accessible language, for readers who are not experts in computer modeling.

Although the work is theoretical in nature, internal "controls" could have been included. For example, simulating simplified synapses (without organelles) or fixed proteins, in order to evaluate the behavior of the model under pure diffusion conditions, would help to demonstrate the robustness of the results.

As far as the results are concerned, the distinction between the mechanism of axonal influx and internal redistribution is a noteworthy point. The figures that show the fit between simulated and experimental data are very enlightening. However, it would be useful to include visual diagrams or tables that relate the variation in parameters (D, k, T) to the behavior of the recovery curves.

1. Detailed suggestions

Initially, the authors set out the problem of FRAP experiments on synapses and the hypothesis that recovery times can alternatively reflect intrinsic mobility or redistribution due to geometry. However, the text does not demonstrate a precise delimitation of the theoretical parameters adopted (the Reaction-Diffusion Master Equation - RDME model, for example) nor does it present a more in-depth contextualization of the limitations of current experimental methods. It would be interesting to include a subsection right after the presentation of the problem, describing in detail the mathematical foundations of the RDME model and, if possible, relating it to studies already consolidated in the area.

In section II, when the three photobleaching protocols are described (snapshot across the synapse, local snapshot and 80 ms continuous bleaching), the authors present the recovery curves and synthetic images (figs. 1 and 2). At this point, the manuscript does not make it clear what the specific parameters of each protocol are. For example, in the description of continuous bleaching, there is no precise indication of the diameter of the beam, the intensity of the laser or the time interval between measurements. In addition, the selection of the region of interest (ROI) is mentioned too briefly, with no discussion of the possible variations in this criterion and its impact on the analysis of the results. It would be interesting to have a comparative table summarizing the parameters used in each protocol, showing: bleaching time, ROI diameter, spatial resolution and ROI selection criteria. Include comments in the text that indicate, for example: "In Figure 1A, it can be seen that the continuous bleaching protocol (80 ms) results in an increase in the effective beam size due to diffusion during the pulse, a fact that can be quantified by comparing the D parameters adopted for high versus low mobility proteins." In addition, it would be useful to detail the implications of using different protocols so that the reader understands how methodological choices affect recovery times (for example, pointing out that a bleaching that covers the entire synapse favors axial influx, while a local bleaching emphasizes internal redistribution).

Other parameters that they could have taken into account in the simulations are the irradiation time (in the paper they only do simulations for instantaneous irradiation and 80ms) since varying this time could give a better understanding of how the proteins diffuse through the area, and so the region of interest would be variable depending on the degree of "spreading" of the protein in question. It would be nice to know how they chose this 80ms time as well, was it a limitation of the technique or was it according to the ROI they wanted to study at the time?

In addition, it would have been good to see longer simulation times, especially in the case of total irradiation, where we couldn't see much recovery in the time shown, and it would have added to the work to see the complete recovery time in images. Along the same lines of images, it would have been interesting to see simulations with smaller deltas, in which we could follow the path of the protein through the synapse. Especially when the axon influx is predominant, since in the methods they state that the synapse format used was an ellipse with two connections to the axon, one at each of the "tips", and this architecture certainly influences the distribution of proteins.

The architecture of the synapse, once again, influences how the distribution and arrival of proteins in the axon takes place, which is why where the irradiation is carried out is extremely important. This is not mentioned in the texts, but in the microscopy images it always seems to be in the middle. Why not also test irradiation in different places? Closer to one of the connections with the axon (which should make the contribution of the axon to the recolonization of the synapse by fluorescent proteins greater) or even "higher up" or "lower down", further away from the center (which probably hinders the redistribution of proteins, since the synapse loses symmetry and the irradiated area loses part of its circumference because it is at the edge of the synapse and has no non-irradiated proteins in one of the directions).

They also don't specify how much of a contribution each recovery method is responsible for. In table 1, where this information is placed, the proteins are only separated by "redistribution", "axon" or "both", but the authors never specify how they separate these contributions. How predominant does one of the methods have to be over the other to have been placed as the main one, and how equally present do they have to be to be indicated as having both present? Perhaps the use of percentage contribution could improve the understanding of this separation.

In addition, a very important part of comparing the contribution of the axon and redistribution is that it depends on whether the synapse has been fully irradiated or only partially irradiated, a fact that he proves at the beginning of the text and provides images in figure 1, when he talks about these two irradiation methods. However, throughout the text he says that the redistribution method (with the recovery times (τ_syn and τ_axon)) that is predominant depends on the effective length L̃, which compares the size of the synapse with the size of the irradiated area. It is even pointed out that when the protein is small its entire area is irradiated and this makes the axon the main contributor, but it doesn't make sense to compare protein sizes only in this way, with the same irradiation size, when this size is large enough to irradiate some of the proteins studied completely. It would also be good to see a comparison with the relative area of the protein that is irradiated, as this could give us a result that is a little less variable with size, i.e. more predictable. The authors could clarify this point by inserting graphs with appropriate scales where it is possible to see, for example, how varying D (from 0.1 to 20 µm²/s) quantitatively changes the τ_syn values, with specific comments in the text such as: "In the case of protein X, when D was changed from 0.5 to 5 µm²/s, a reduction in recovery time from 16.2 s to 10.2 s was observed, showing a change in the predominant recovery mechanism." They could also detail in legends and in the body of the text how the data is grouped around the value of L̃ ≈ 1, emphasizing the transition between regimes and suggesting practical implications for the experimental design.

Also in this section, the analysis of simulations that incorporate vesicle bonds (parameters k and T) is presented in a consolidated form, but without a proper discussion of the limitations of the model for approximating the effective diffusion coefficient (Deff). However, the use of the mean field model for Deff (Deff = D / (1 + ρT-k)) is not critically discussed. To remedy this problem, it would be useful to discuss the conditions under which this approximation is valid, as well as the possible errors introduced when the binding/unbinding times are not fast enough in relation to the experiment time. And also include a figure or table comparing the results obtained with and without the approximation, allowing the degree of discrepancy in extreme cases to be visualized.

In the Conclusions section (p. 21-24), the authors state that faster recovery times do not always indicate greater mobility and that the geometry of the synaptic terminal is a determining factor. The discussion lacks a more direct approach that relates computational findings to experimental practice. For example, there is no "guide" to advise researchers on how to adjust experimental parameters as a function of synapse size. We suggest drafting a final section entitled "Implications for Experimental Practice", in which specific recommendations are listed, such as:

• "For synapses with L̃ < 1, it is recommended to use local bleaching protocols in order to prevent axonal influx from masking local diffusion."

• "When the immobilized fraction exceeds 20%, check that the ROI selection criterion is not incorporating regions with volumetric exclusion (e.g. presence of mitochondria)."

This section should contain illustrative examples (even if summarized) and suggest directions for future work that could test these hypotheses experimentally.

It should also be noted that although the manuscript describes the computational methods in good detail, there is no indication that the data and source code for the simulations will be made available. It would be interesting to include a paragraph in the Acknowledgements section or in an Appendix indicating that the code and data will be made available in a public repository (e.g. GitHub or Zenodo), ensuring transparency and allowing the reproducibility of the results.

1. Specific points

Image 2: It would be interesting if they did something similar to what was presented in Image 1, but considering more than one format, in order to compare the graphs of the three methods in different geometries. In the text, the images in Figure 2 are mentioned without specifying the corresponding letter, which makes it difficult to understand. In addition, when referring to the letter C, the author states that the average is represented by the orange dot, but in the image it appears indicated by a black dot, which is difficult to distinguish amid the dark blue. Perhaps using a different format could make this average more obvious.

 Image 4:  In this case, "C" in the caption is presented in a visually different way to items "A" and "B", which causes a certain inconsistency. In addition, the X axis of image "B" is incorrectly positioned - located between "B" and "C" - which can confuse interpretation. Perhaps a slight reduction in size would help to make it clearer that this axis belongs to image "B". In the caption, it would be interesting to better detail the part referring to image "C" and explain that "immobile fraction" corresponds to the "IF" shown in the image.

References:  In some passages he uses references both in numerical format and mentioning the author's name followed by "et al.", which can lead to inconsistencies in formatting.

• Final considerations

Overall, the manuscript makes a relevant contribution to the interpretation of FRAP experiments on synapses, demonstrating that recovery times can be significantly influenced by geometry and protein binding mechanisms. However, for the work to reach its full potential and become a methodological reference, it is essential that the authors:

• Delineate more precisely the experimental and computational parameters (by means of comparative tables and quantitative graphs) that highlight the effects of each variable;

• Test other variables, exemplified above, which could provide more varied data;

• Develop a critical discussion about the limits of the model, especially with regard to the approximation of the effective diffusion coefficient;

• Relate the findings to practical implications for FRAP experiments, offering direct recommendations to guide the design of future studies;

• Make available the data and code used in the simulations to ensure transparency and reproducibility of the results.

The article is well-written and well-explained, and with these suggested improvements, it will certainly gain in clarity and impact. We believe that by adopting these improvements, the manuscript will not only become more robust from a methodological point of view, but will also offer a more direct and practical contribution to the scientific community investigating protein dynamics in synaptic terminals.

Competing interests

The authors declare that they have no competing interests.