An applicable and efficient retrograde monosynaptic circuit mapping tool for larval zebrafish

Reviewer #1 (Public review):

(1) Presentation of Figures in the Response Letter

I would like to note that the figures included in the response letter would benefit from improved organization. For example, Author response image 1 lacks clarity for experimental conditions. From the response letter, my understanding is that a "Labeling rate index", Rg−Rn, was calculated to represent the difference in the rate of increase in labeling between neurons and glial across two time intervals based on experiments shown in Figure 2-figure supplement 1C and G. It seems that a mean convergence index was calculated for each experimental condition at each time point for glial and neurons, and then the differences in mean convergence index increase between time intervals were calculated for glial and neurons. The legend needs more detail to enhance clarity.

Furthermore, the manuscript should clearly distinguish between figures generated from re-analysis of existing data and those based on newly conducted experiments. This distinction should be explicitly stated in the figure legends and/or main text.
I recommend that all response figures containing data integral to the authors' rebuttal be properly integrated into the manuscript's existing supplementary figure set, rather than remaining isolated in the response document. This would enhance clarity and ensure that key supporting data are fully accessible to readers. For instance, Author response image 1 can be integrated with Figure 2-figure supplement.

(2) Glial Cell Labeling and Specificity of Trans-Synaptic Spread

The authors provided a comprehensive and well-reasoned response to the concern regarding the labeling of radial glial cells. The inclusion of a dedicated section in the revised Discussion and response figures (possibly to be integrated with supplementary figures), strengthens the manuscript.

The authors have made an interesting observation in Author response image 2 that glial labeling was frequently observed near the soma and dendrites of starter cells, suggesting that transneuronal labeled glial cells may be synaptically associated with the starter neurons. Also astroglia starter cells lead to infection of nearby TVA-negative astroglia, suggesting astroglia-to- astroglia transmission.

I find the response scientifically satisfactory and appreciate the authors' transparency in addressing the limitations of their approach.

(3) Temperature Effects and Larval Viability

The authors' justification for raising larvae at 36C to improve labeling efficiency is reasonable. The supporting data indicating minimal impact on larval viability within the experimental timeframe are convincing. Referencing prior behavioral studies and including survival data under controlled conditions adds credibility to their claims. I find this issue satisfactorily addressed.

(4) Viral Toxicity and Dosage Considerations, Secondary Starter Cells

The authors present a well-reasoned explanation that viral cytotoxicity is primarily driven by replication and not by viral titer or injection volume. However, the inclusion of experimental data directly testing the effects of higher titer or volume on starter cell viability would have strengthened this point, particularly since such tests are relatively straightforward to perform.

Regarding the potential contribution of secondary starter cells, the authors provide a convincing rationale for why such effects are unlikely under their sparse labeling conditions. However, in cases where TVA and G are broadly expressed-such as under the vglut2a promoter, as shown in Author response image 2-it would be valuable to directly evaluate this possibility experimentally. While the authors' interpretation is reasonable, empirical validation would further strengthen their conclusions.

Reviewer #2 (Public review):

The study by Chen, Deng et al. aims to develop an efficient viral transneuronal tracing method that allows efficient retrograde tracing in the larval zebrafish. The authors utilize pseudotyped-rabies virus that can be targeted to specific cell types using the EnvA-TvA systems. Pseudotyped rabies virus has been used extensively in rodent models and, in recent years, has begun to be developed for use in adult zebrafish. However, compared to rodents, the efficiency of spread in adult zebrafish is very low (~one upstream neuron labeled per starter cell). Additionally, there is limited evidence of retrograde tracing with pseudotyped rabies in the larval stage, which is the stage when most functional neural imaging studies are done in the field. In this study, the authors systematically optimized several parameters of rabies tracing, including different rabies virus strains, glycoprotein types, temperatures, expression construct designs, and elimination of glial labeling. The optimal configurations developed by the authors are up to 5-10 fold higher than more typically used configurations.

The results are convincing and support the conclusions. There are some additional changes that are recommended:

(1) The new data included in the response to reviewer's letter are important to support the main conclusions and should be included in the manuscript.

(2) Line 357-362: This section should include all of the Author response image and associated details. Additionally, the Author response image 3 is at odds with Fig 2-supplement 1G. In Author response image 3, ~75% of glial cells labeled at 4 dpi loses their fluorescence by 10 dpi. However, Figure 2-supplement 1G shows that glial overall labeling increases ~2 fold from 4 dpi to 10 dpi. This would suggest that the de novo labeling rate for glia is much higher than the net labeling rate calculated from the convergence index. The authors should clarify these findings.

Author response:

The following is the authors’ response to the original reviews.

Reviewer #1 (Public Review):

EnvA-pseudotyped glycoprotein-deleted rabies virus has emerged as an essential tool for tracing monosynaptic inputs to genetically defined neuron populations in the mammalian brain. Recently, in addition to the SAD B19 rabies virus strain first described by Callaway and colleagues in 2007, the CVS N2c rabies virus strain has become popular due to its low toxicity and high trans-synaptic transfer efficiency. However, despite its widespread use in the mammalian brain, particularly in mice, the application of this cell-type-specific monosynaptic rabies tracing system in zebrafish has been limited by low labeling efficiency and high toxicity. In this manuscript, the authors aimed to develop an efficient retrograde monosynaptic rabies-mediated circuit mapping tool for larval zebrafish. Given the translucent nature of larval zebrafish, whole-brain neuronal activities can be monitored, perturbed, and recorded over time. Introducing a robust circuit mapping tool for larval zebrafish would enable researchers to simultaneously investigate the structure and function of neural circuits, which would be of significant interest to the neural circuit research community. Furthermore, the ability to track rabies-labeled cells over time in the transparent brain could enhance our understanding of the trans-synaptic retrograde tracing mechanism of the rabies virus.

To establish an efficient rabies virus tracing system in the larval zebrafish brain, the authors conducted meticulous side-by-side experiments to determine the optimal combination of trans-expressed rabies G proteins, TVA receptors, and recombinant rabies virus strains. Consistent with observations in the mouse brain, the CVS N2c strain trans-complemented with N2cG was found to be superior to the SAD B19 combination, offering lower toxicity and higher efficiency in labeling presynaptic neurons. Additionally, the authors tested various temperatures for the larvae post-virus injection and identified 36℃ as the optimal temperature for improved virus labeling. They then validated the system in the cerebellar circuits, noting evolutionary conservation in the cerebellar structure between zebrafish and mammals. The monosynaptic inputs to Purkinje cells from granule cells were neatly confirmed through ablation experiments.

However, there are a couple of issues that this study should address. Additionally, conducting some extra experiments could provide valuable information to the broader research field utilizing recombinant rabies viruses as retrograde tracers.

(1) It was observed that many radial glia were labeled, which casts doubt on the specificity of trans-synaptic spread between neurons. The issues of transneuronal labeling of glial cells should be addressed and discussed in more detail. In this manuscript, the authors used a transgenic zebrafish line carrying a neuron-specific Cre-dependent reporter and EnvA-CVS N2c(dG)-Cre virus to avoid the visualization of virally infected glial cells. However, this does not solve the real issue of glial cell labeling and the possibility of a nonsynaptic spread mechanism.

In agreement with the reviewer’s suggestion, we have incorporated a standalone section in the revised Discussion (page 9) to address the issue of transneuronal glial labeling, including its spatial distribution, temporal dynamics, potential mechanisms, and possible strategies for real resolution.

Regarding the specificity of trans-synaptic spread between neurons, we have demonstrated that our transsynaptic tracing system reliably and specifically labels input neurons. Structurally, we only observed labeling of inferior olivary cells (IOCs) outside the cerebellum, which are the only known extracerebellar inputs to Purkinje cells (PCs), while all other traced neurons remained confined within the cerebellum throughout the observation period (see Figure 2G–I). Functionally, we verified that the traced neurons formed synaptic connections with the starter PCs (see Figure 2J–M). Together, these findings support the conclusion that our system enables robust and specific retrograde monosynaptic tracing of neurons in larval zebrafish.

Regarding the transneuronal labeling of radial glia cells, we observed that their distribution closely correlates with the location of neuronal somata and dendrites (see Author response image 2). In zebrafish, radial glial cells are considered functional analogs of astrocytes and are often referred to as radial astroglia. The adjacent labeled astroglia may participate in tripartite synapses with the starter neurons and express viral receptors that enable RV particle entry at postsynaptic sites. This suggests that rabies-based tracing in zebrafish may serve as a valuable tool for identifying synaptically associated and functionally connected glia. Leveraging this approach to investigate glia–neuron interactions represents a promising direction for future research.

In our system, the glial labeling diminishes at later larval stages, likely due to abortive infection (see Author response image 3 and relevant response). However, the eventual clearance of infection does not preclude the initial infection of glial cells, which may compete with neuronal labeling and reduce overall tracing efficiency. Notably, transneuronal infection of glial cells by RV has also been observed in mammals (Marshel et al., 2010). To minimize such off-target labeling, future work should focus on elucidating the mechanisms underlying glial susceptibility—such as receptor-mediated viral entry— and developing strategies to suppress receptor expression specifically in glia, thereby improving the specificity and efficiency of neuronal circuit tracing.

In addition, wrong citations in Line 307 were made when referring to previous studies discovering the same issue of RVdG-based transneuronal labeling radial glial cells. "The RVdG-based transneuronal labeling of radial glial cells was commonly observed in larval zebrafish29,30".

The cited work was conducted using vesicular stomatitis virus (VSV). A more thorough analysis and/or discussion on this topic should be included.

We thank the reviewer for pointing out the citation inaccuracy. The referenced study employed vesicular stomatitis virus (VSV), which, like RV, is a member of the Rhabdoviridae family. We have revised the text accordingly—from "RVdG-based transneuronal labeling of radial glial cells…" to " Transneuronal labeling of radial glial cells mediated by VSV, a member of the Rhabdoviridae family like RV, has been commonly observed in larval zebrafish" (page 9, line 347).

Several key questions should be addressed:

Does the number of labeled glial cells increase over time?

Yes, as shown in Figure 2—figure supplement 1C and G, the number of labeled radial glial cells significantly increased from 2 to 6 days post-injection (dpi). This phenomenon has been addressed in the revised Discussion section (page 9, line 357).

Do they increase at the same rate over time as labeled neurons?

Although glial cell labeling continued to increase over time, we observed a slowdown in labeling rate between 6 and 10 dpi, as shown in Figure 2—figure supplement 1C and G. Therefore, we divided the timeline into two intervals (2–6 and 6–10 dpi) to compare the rate of increase in labeling between neurons and glia. The rate (R) was defined as the daily change in convergence index. To quantify the difference between neuronal and glial labeling rates, we calculated a labeling rate index: R_g−R_n, where R_g and R_n denote the rates for glia and neurons, respectively) (Author response image1). Our analysis revealed that, between 2 and 6 dpi, glial cells exhibited a higher labeling rate than neurons. However, this trend reversed between 6 and 10 dpi, with neurons surpassing glial cells in labeling rate. These findings have been included in the revised Discussion section (page 9).

Author response image 1.

Labeling rate index of glia and neurons across two time intervals. Data points represent the mean labeling rate index for each tracing strategy within each time interval. *P < 0.05 (nonparametric two-tailed Mann-Whitney test).

Are the labeled glial cells only present around the injection site?

We believe the reviewer is inquiring whether labeled glial cells are spatially restricted to the vicinity of starter neurons. The initial infection is determined by the expression of TVA rather than the injection site. For example, injecting a high volume of virus into the anterior hindbrain resulted in the infection of TVA-expressing cells in distant regions, including the 109 tectum and posterior hindbrain (Author response image 2).

Regarding glial labeling, PC starter experiments showed that labeled glial cells (i.e. Bergmann glia) were predominantly localized within the cerebellum, likely due to the confinement of PC dendrites to this region. When using vglut2a to define starter neurons, glial labeling was frequently observed near the soma and dendrites of starter cells (14 out 114 of 17 cases; Author response image 2). These observations suggest that transneuronal labeled glial cells may be synaptically associated with the starter neurons. We have included this point in the revised Discussion section (page 9).

Author response image 2.

Location of transneuronal labeled glial cells. (a and b) Confocal images showing the right tectum (a) and posterior hindbrain (b) of different WT larvae expressing EGFP and TVA using UGNT in randomly sparse neurons (vglut2a⁺) and infected with CVSdGtdTomato[EnvA] (magenta) injected into the anterior hindbrain. Dashed yellow circles, starter neurons (EGFP⁺/tdTomato⁺); gray arrows, transneuronally labeled radial glia (tdTomato⁺/EGFP⁻); dashed white lines, tectum or hindbrain boundaries. C, caudal; R, rostral. Scale bars, 20 μm.

Can the phenomenon of transneuronal labeling of radial glial cells be mitigated if the tracing is done in slightly older larvae?

Yes, we agree. As elaborated in the following response, we hypothesize that the loss of fluorescence in radial glial cells at later developmental stages is due to abortive infection (see Author response image 3 and associated response). This supports the notion that abortive infection becomes increasingly pronounced as larvae mature, potentially explaining the negligible glial labeling observed in adult zebrafish (Dohaku et al., 2019; Satou et al., 2022). However, as noted in our response to the first comment, the disappearance of fluorescence does not indicate the absence of viral entry. Viral receptors may express on glial cells, allowing initial infection despite a failure in subsequent replication. Consequently, glial infection—though abortive—may still compete with neuronal infection and reduce tracing efficiency.

What is the survival rate of the infected glial cells over time?

We observed the disappearance of glial fluorescence after transneuronal labeling, while we did not observe punctate fluorescent debris typically indicative of apoptotic cell death. Therefore, we favor the hypothesis that the loss of glial fluorescence results from abortive infection rather than cell death. Abortive infection refers to a scenario in which viral replication is actively suppressed by host antiviral responses, preventing the production of infectious viral particles. For example, recent studies have shown that lab-attenuated rabies virus (RV) induces the accumulation of aberrant double-stranded DNA in astrocytes, which activates mitochondrial antiviral-signaling protein (MAVS) and subsequent interferon expression (Tian et al., 2018). This antiviral response inhibits RV replication, ultimately resulting in abortive infection.

In addition, we quantified the proportion of glial cells labeled at 2 dpi and 4dpi that retained fluorescence over time. By 6 dpi (approximately 11 dpf), glial labeling had largely diminished in both groups (Author response image 3). These results suggest that the decline in glial fluorescence is more closely linked to larval age than to the duration of glial infection, supporting the notion of abortive infection. This also addresses the reviewer’s earlier concern and indicates that glial labeling is mitigated in older larvae.

Author response image 3.

Fraction of glial cells with fluorescence retention. (a and b) Proportion of glial cells labeled at 2 dpi (a) and 4 dpi (b) that retained fluorescence over time. Data are from the CVS|N2cG|36°C group. In boxplots: center, median; bounds of box, first and third quartiles; whiskers, minimum and maximum values. n.s., not-significant; *P < 0.05, **P < 0.01 (nonparametric two-tailed Mann-Whitney test).

If an infected glial cell dies due to infection or gets ablated, does the rabies virus spread from the dead glial cells?

In our system, glial cells do not express the rabies glycoprotein (G). Therefore, even if glial cells are transneuronally infected, they cannot support viral budding or assembly of infectious particles due to the absence of G (Mebatsion et al., 1996), preventing further viral propagation to neighboring cells.

If TVA and rabies G are delivered to glial cells, followed by rabies virus injection, will it lead to the infection of other glial cells or neurons?

We have conducted experiments in which TVA and rabies G were specifically expressed in astroglia using the gfap promoter, followed by RVdG-mCherry[EnvA] injection. This resulted in initial infection of TVA-positive astroglia and occasional subsequent labeling of nearby TVA-negative astroglia (Author response image 4), suggesting astroglia-toastroglia transmission. Notably, no neuronal labeling was observed. This glial-to-glial spread is consistent with previous rabies tracing studies reporting similar phenomena involving the interaction of astrocytes with astrocytes and microglia (Clark et al., 2021). However, the underlying mechanism remains unclear, and we have discussed this in response to the first comment.

Author response image 4.

Viral tracing initiated from astroglia. (a) Confocal images of the tectum of a larva expressing EGFP and TVA using UGBT in randomly sparse astroglia (gfap⁺) and infected by SADdG-mCherry[EnvA] (magenta) injected into the anterior hindbrain. (b) Confocal images of the posterior hindbrain of a larva expressing EGFP and TVA using UGNT in randomly sparse astroglia (gfap⁺) and infected by CVSdG-tdTomato[EnvA] (magenta) injected into the anterior hindbrain. Dashed yellow circles, starter astroglia (EGFP+/mCherry+ or EGFP⁺/tdTomato⁺); gray arrows, transneuronally labeled astroglia (tdTomato⁺/EGFP⁻); dashed white lines, tectum or hindbrain boundaries. C, caudal; R, rostral. Scale bars, 20 μm.

Answers to any of these questions could greatly benefit the broader research community.

(2) The optimal virus tracing effect has to be achieved by raising the injected larvae at 36C. Since the routine temperature of zebrafish culture is around 28C, a more thorough characterization of the effect on the health of zebrafish should be conducted.

Yes, 36°C is required to achieve optimal labeling efficiency. Although this is above the standard zebrafish culture temperature (28°C), previous work (Satou et al., 2022) and our observations indicate that this transient elevation does not adversely affect larval health within the experimental time window.

In the previous study, Satou et al. reported no temperature-dependent effects on swimming behavior, social interaction, or odor discrimination in adult fish maintained at 28°C and 36°C. In larvae, both non-injected and virus-injected fish showed a decrease in survival at later time points (7 dpi), with slightly increased mortality observed at elevated temperatures.

In our study, we raised the same batch of non-virus-injected larvae at 28°C and 36°C, and found no mortality over a 10-day period. For CVS-N2c-injected larvae, electrode insertion caused injury, but survival rates remained around 80% at both temperatures (see Figure 3A). Moreover, we successfully maintained CVS-N2c-injected larvae at 36°C for over a month, indicating that elevated temperature does not adversely affect fish health. Notably, higher temperatures were associated with an accelerated developmental rate.

This point was briefly addressed in the previous version and has now been further elaborated in the revised Discussion section (page 8).

(3) Given the ability of time-lapse imaging of the infected larval zebrafish brain, the system can be taken advantage of to tackle important issues of rabies virus tracing tools.

a) Toxicity.

The toxicity of rabies viruses is an important issue that limits their application and affects the interpretation of traced circuits. For example, if a significant proportion of starter cells die before analysis, the traced presynaptic networks cannot be reliably assigned to a "defined" population of starter cells. In this manuscript, the authors did an excellent job of characterizing the effects of different rabies strains, G proteins derived from various strains, and levels of G protein expression on starter cell survival. However, an additional parameter that should be tested is the dose of rabies virus injection. The current method section states that all rabies virus preparations were diluted to 2x10^8 infection units per ml, and 2-5 nl of virus suspension was injected near the target cells. It would be interesting to know the impact of the dose/volume of virus injection on retrograde tracing efficiency and toxicity. Would higher titers of the virus lead to more efficient labeling but stronger toxicities? What would be the optimal dose/volume to balance efficiency and toxicity? Addressing these questions would provide valuable insights and help optimize the use of rabies viruses for circuit tracing.

This is an important concern. Viral cytotoxicity is primarily driven by the level of viral transcription and replication, which inhibits host protein synthesis (Komarova et al., 2007). The RVdG-EnvA typically infects cells at a rate of one viral particle per cell (Zhang et al., 2024), suggesting that increasing viral concentration does not proportionally increase percell infection. Accordingly, viral titer and injection volume are unlikely to influence cytotoxicity at the single-cell level. In our experiments, injection volumes up to 20 nl (i.e., 4 to 10 times the standard injection volume) did not affect starter cell survival. However, higher titers or volumes may increase the number of initially infected starter cells, potentially leading to greater overall mortality in larval zebrafish.

Similarly, given that rabies virus typically infects cells at one particle per cell, increasing viral titer alone is unlikely to enhance tracing efficiency once the virus type is fixed. In contrast, the level of G protein expression significantly influences tracing efficiency (see Figure 2D). However, excessive G protein expression reduces the survival of starter cells (see Figure 3D). Therefore, careful control of G protein levels is essential to balance tracing efficiency and cytotoxicity.

Notably, regardless of whether infected cells undergo apoptosis or necrosis due to cytotoxicity, the resulting disruption of the plasma membrane severely impairs viral budding. As a result, the formation of intact, G protein-enveloped viral particles is prevented, limiting further infection of neighboring neurons.

The latest second-generation ΔGL RV vectors (Jin et al., 2024), which lack both the G and L (viral polymerase) genes, have been shown to markedly reduce cytotoxicity. These improved tracing strategies may be explored in future zebrafish studies to further optimize labeling efficiency and cell viability.

The issue of viral titer and volume has been addressed in the revised Discussion section (page 10).

b) Primary starters and secondary starters:

Given that the trans-expression of TVA and G is widespread, there is the possibility of coexistence of starter cells from the initial infection (primary starters) and starter cells generated by rabies virus spreading from the primary starters to presynaptic neurons expressing G. This means that the labeled input cells could be a mixed population connected with either the primary or secondary starter cells.

It would be immensely interesting if time-lapse imaging could be utilized to observe the appearance of such primary and secondary starter cells. Assuming there is a time difference between the initial appearance of these two populations, it may be possible to differentiate the input cells wired to these populations based on a similar temporal difference in their initial appearance. This approach could provide valuable insights into the dynamics of rabies virus spread and the connectivity of neural circuits.

The reviewers suggestion is valuable. Regarding the use of Purkinje cells (PCs) as starter cells, we consider the occurrence of secondary PCs to be extremely rare. Although previous evidence suggests that PCs can form synaptic connections with one another (Chang et al., 2020), our sparse labeling strategy—typically involving fewer than 10 labeled cells— significantly reduces the likelihood of viral transmission between PC starter cells. In addition, if secondary starter PCs were frequently generated, we would expect increased tracing efficiency at 10 dpi compared to 6 dpi. However, our results show no significant difference (see Figure 2—figure supplement 1C and G).

Given the restricted expression of TVA and G in PCs, even if a limited number of secondary starters were generated, the labeled inputs would predominantly be granule cells (GCs), thereby preserving the cell-type identity of upstream inputs. While this raises a potential concern regarding an overestimation of the convergence index (CI). Notably, within the GC-PC circuit, individual GCs often project to multiple PCs. Consequently, a GC labeled via a secondary PC may also a bona fide presynaptic partner of the primary starter population. This overlap could mitigate the overestimation of CI. Taken together, we believe that the CI values reported in this study provide a reasonable approximation of monosynaptic connectivity.

In scenarios where TVA and G are broadly expressed—for example, under the control of vglut2a promoter—secondary starter cells may arise frequently. In such cases, long-term time-lapse imaging in the zebrafish whole brain presents a promising strategy to distinguish primary and secondary starter cells, along with their respective input populations, based on the timing of their appearance. This approach potentially enables multi-step circuit tracing within individual animals. An alternative strategy is to use an EnvA-pseudotyped, G-competent rabies virus, which allows targeted initial infection while supporting multisynaptic propagation. When combined with temporally resolved imaging, this strategy could facilitate direct labeling of higher-order circuits and allow clear differentiation between multi-order inputs and the original starter population over time.

In conclusion, we find this suggestion compelling and will explore these strategies in future studies to optimize and broaden the application of rabies virus-based circuit tracing.

Reviewer #2 (Public Review):

The study by Chen, Deng et al. aims to develop an efficient viral transneuronal tracing method that allows efficient retrograde tracing in the larval zebrafish. The authors utilize pseudotyped-rabies virus that can be targeted to specific cell types using the EnvA-TvA systems. Pseudotyped rabies virus has been used extensively in rodent models and, in recent years, has begun to be developed for use in adult zebrafish. However, compared to rodents, the efficiency of the spread in adult zebrafish is very low (~one upstream neuron labeled per starter cell). Additionally, there is limited evidence of retrograde tracing with pseudotyped rabies in the larval stage, which is the stage when most functional neural imaging studies are done in the field. In this study, the authors systematically optimized several parameters of rabies tracing, including different rabies virus strains, glycoprotein types, temperatures, expression construct designs, and elimination of glial labeling. The optimal configurations developed by the authors are up to 5-10 fold higher than more typically used configurations.

The results are solid and support the conclusions. However, the methods should be described in more detail to allow other zebrafish researchers to apply this method in their own work.

Additionally, some findings are presented anecdotally, i.e., without quantification or sufficient detail to allow close examinations. Lastly, there is concern that the reagents created by the authors will not be easily accessible to the zebrafish community.

(1) The titer used in each experiment was not stated. In the methods section, it is stated that aliquots are stored at 2x10e8. Is it diluted for injection? Are all of the experiments in the manuscripts with the same titer?

We injected all three viral vectors as undiluted stock aliquots. The titer for SADdGmCherry[EnvA], CVSdG-tdTomato[EnvA], and CVSdG-mCherry-2A-Cre[EnvA]) was 2 × 10⁸, 2 × 10⁸, and 3 × 10⁸ infectious units/mL, respectively. This has been clarified in the updated Methods section (page 12).

(2) The age for injection is quite broad (3-5 dpf in Fig 1 and 4-6 dpf in Fig 2). Given that viral spread efficiency is usually more robust in younger animals, describing the exact injection age for each experiment is critical.

We appreciate the reviewer’s suggestions. For the initial experiments tracing randomly from neurons in Figure 1, the injection age was primarily 3–4 dpf, with a one-day difference. Due to the slower development of PCs, the injection age for experiments related to Figure 2,3, and 4, is mainly 5 dpf. To clarify the developmental stages at the time of injection for each experiment, we have newly added tables (see Figure 1,2—table supplement 2) listing the number of fish used at each injection age for all experimental groups shown in Figure 1 and 2.

(3) More details should be provided for the paired electrical stimulation-calcium imaging study. How many GC cells were tested? How many had corresponding PC cell responses? What is the response latency? For example, images of stimulated and recorded GCs and PCs should be shown.

Yes, these are important details for the paired electrical stimulation-calcium imaging study. We stimulated 33 GCs from 32 animals and detected calcium responses in putative postsynaptic PCs in 15 cases. Among these, we successfully ablated the single GC in 11 pairs and observed a weakened calcium response in PCs following ablation (see Figure 2M). The response latency was determined as the first calcium imaging frame where ΔF/F exceeded the baseline (pre-stimulus average) by 3 times the standard deviation. Imaging was performed at 5 Hz, and as shown in Figure 2L, the calculated average response latency was 152 ± 35 ms (mean ± SEM), indicating an immediate response with calcium intensity from the first post-stimulus imaging frame consistently exceeding the threshold.

We have added additional details to the Results (page 5), Discussion (page 9), and Methods (page 15) sections. A representative image showing both the stimulated GC and the recorded PC has been added to Figure 2 in the revised manuscript (see Figure 2K).

(4) It is unclear how connectivity between specific PC and GC is determined for single neuron connectivity. In other images (Figure 4C), there are usually multiple starter cells and many GCs. It was not shown that the image resolution can establish clear axon dendritic contacts between cell pairs.

In our experiments, sparse labeling typically results in 1–10 starter cells per fish. Regarding the case shown in Figure 4C (right column), only two PC starters were labeled, which simplifies the assignment of presynaptic inputs to individual PCs. Connectivity is determined based on clear axon-dendritic or axon-cell body apposition between GCs and PCs. We have accordingly added more details to the Methods (page 16) section regarding how we determined connectivity between specific PCs and GCs.

Reviewer #2 (Recommendations For The Authors):

To enable broader use of this technique, I would encourage the authors to submit their zebrafish lines, plasmids, and plasmid sequences to public repositories such as ZIRC and Addgene. Additionally, there is no mention of how viral vectors will be shared.

We have deposited the related zebrafish lines at CZRC (China Zebrafish Resource Center) and uploaded plasmid maps and sequences to Addgene. The viral vectors are available through BrainCase (Shenzhen, China). We have included the information in the revised manuscript.

Reviewer #3 (Public Review):

Summary:

The authors establish reagents and define experimental parameters useful for defining neurons retrograde to a neuron of interest.

Strengths:

A clever approach, careful optimization, novel reagents, and convincing data together lead to convincing conclusions.

Weaknesses:

In the current version of the manuscript, the tracing results could be better centered with respect to past work, certain methods could be presented more clearly, and other approaches worth considering.

Appraisal/Discussion:

Trans-neuronal tracing in the larval zebrafish preparation has lagged behind rodent models,limiting "circuit-cracking" experiments. Previous work has demonstrated that pseudotyped rabies virus-mediated tracing could work, but published data suggested that there was considerable room for optimization. The authors take a major step forward here, identifying a number of key parameters to achieve success and establishing new transgenic reagents that incorporate modern intersectional approaches. As a proof of concept, the manuscript concludes with a rough characterization of inputs to cerebellar Purkinje cells. The work will be of considerable interest to neuroscientists who use the zebrafish model.

Reviewer #3 (Recommendations For The Authors):

The main limitations of the work are as follows:

(1) The optimizations might differ for different neurons. Purkinje cells are noteworthy because they develop considerably during the time window detailed here, almost doubling in number between 7-14dpf. Presumably, connectivity follows. This sort of neurogenesis is much less common elsewhere. It would be useful to show similar results in, say, tectal neurons, which would have spatially-restricted retinal ganglion cells labelled.

We acknowledge that Purkinje cells (PCs) undergo significant development between 7–14 dpf, which may influence synaptic connectivity and result in differences in tracing efficiency. However, all experimental conditions were standardized across groups, and the selection of starter PCs was unbiased, typically focusing on PCs in the lateral region of the CCe (corpus cerebelli) subregion, ensuring that the relative comparisons remain valid.

We agree that testing other neuronal populations would be valuable, as tracing efficiency is influenced by multiple factors, such as the number of endogenous inputs, synaptic maturation, and developmentally regulated synaptic strength. Tectal neurons, which receive spatially restricted retinal ganglion cell inputs, would be a suitable choice for further investigation. However, due to the various tectal cell types and the opacity of the eyeball, such studies present additional technical challenges and are beyond the scope of this paper.

(2) The virus is delivered by means of microinjection near the cell. This is invasive and challenging for labs that dont routinely perform electrophysiology. It would be useful to know if coarser methods of viral delivery (e.g. intraventricular injection) would be successful.

Our protocol does not require the level of precision needed for electrophysiology. The procedure can be performed using a standard high-magnification upright (135× magnification, Nikon SMZ18) or inverted fluorescence microscope (200× magnification, Olympus IX51). The virus suspension was loaded into a glass micropipette with a ~10 µm tip diameter and directly microinjected into the target region using a micromanipulator. The procedure was comparable to embryonic microinjection in terms of precision and operational control. Notably, direct contact with the target cells is not necessary, as the injected virus solution can diffuse and effectively infect nearby cells.

We had attempted intraventricular injection as an alternative, but it failed to produce robust labeling, reinforcing the necessity for direct tissue injection.

We have now included additional methodological details in the Methods section (page 13).

(3) Because of the combination of transgenic lines, plasmid injection, and viral type, it is often confusing to follow exactly what is being done for a particular experiment. It would be useful to specify the transgenic background used for each experiment using standard nomenclature e.g. "Plasmids were injected into Tg(elavl3:GAL4) fish." This is particularly important for the experiments in Figure 4: it isnt clear what the background used for the sparse labels was.

Thank the reviewer for bringing this issue to our attention. In order to improve clarity, we have revised the figure legends to explicitly state the transgenic background, injected plasmids, and viral type used in each experiment, particularly for Figure 4.

(4) Plasmids should be deposited with Addgene along with maps specifying the particular "codon-optimized Tetoff" per 388.

We confirm that all plasmids, including those containing codon-optimized Tetoff constructs, have been uploaded to Addgene along with detailed maps.

(5) It would be useful to know if there were more apoptotic cells after transfection -- an acridine orange or comparable assay is recommended, rather than loss of fluorescence.

We appreciate the reviewer’s suggestion to assess apoptosis using acridine orange staining or comparable assays. We agree that such methods can provide more direct detection of apoptotic events. However, we believe that the difference in cytotoxicity is already evident in our current data: SAD-infected cells exhibit greater loss than CVSinfected cells (see Figure 3D). This is consistent with previous observations in mice, where greater toxicity of SAD compared to CVS was demonstrated using propidium iodide (PI) staining in cultured cells (Reardon et al., 2016).

(6) Line 219-228 Hibis lab has described the subtypes of granule cells in detail already; the work should discuss the tracings with respect to previous characterizations instead of limiting that work to a citation.

Thanks for the reminding of this point. We have expanded the Results section (page 6) to discuss the subtypes of GCs and PCs in relation to previously reported characterizations.

(7) "Activities" is often used when "activity" is correct. The use of English in the manuscript is, by and large, excellent, but its worth running the text through software like Grammarly to catch the occasional error.

We have carefully edited the manuscript using professional language editing tools to correct any grammatical issues.

(8) The experiments in 2J-2L would be more convincing if they were performed on inferior olive inputs as well -- especially given the small size of the granule cells.

We acknowledge the reviewers observation that granule cells (GCs) are relatively small, which may underline the finding that, out of 33 stimulated GCs, only 15 were capable of eliciting calcium responses in putative postsynaptic PCs. However, in all 11 pairs where a single GC was successfully ablated, we observed a weakened calcium response in PCs after the ablation (see Figure 2M), suggesting our tracing approach specifically identifies synaptically coupled neurons. We have clarified this point in the revised manuscript (page 5).

We agree that verifying the IO inputs to PCs would strengthen the validity of our findings. However, in our experiments, the probability of tracing upstream IO cells was relatively low. This may be due to the developmental immaturity of the synapse and the fact that each PC typically receives input from a single IO cell. Additionally, the deep and distant anatomical location of the IO presents technical challenges for paired electrical stimulationcalcium imaging study. To address these limitations, we are currently exploring the integration of viral tracing and optogenetics to further investigate IO-PC connectivity in future studies.

(9) It would be useful if the manuscript discussed the efficacy of trans-synaptic labelling. What fraction of granule cell / olivary inputs to a particular Purkinje cell do the authors think their method captures?

This is an important point for assessing the efficacy of our trans-synaptic labeling. Ideally, electron microscopy (EM) data would provide the most precise evaluation. In the absence of EM data, we estimated the number of GCs, IOs and PCs using light microscopy-based cell counting.

At approximately 7 dpf, we manually counted 327 ± 14 PCs and 2318 ± 70 GCs in the Tg(2×en.cpce-E1B:tdTomato-CAAX) and Tg(cbln12:GAL4FF);Tg(5×UAS:EGFP) zebrafish cerebellum, across all subregions (Va, CCe, EG, and LCa). Given the developmental increase in the number of GCs and the fact that some GCs that have exclusively ipsilateral projections, and that a single PC would not receive input from all parallel fibers, we estimate that by 10–14 dpf, a single PC receives approximately 1000– 2000 GC inputs. Under optimal tracing conditions, we observed an average of 20 labeled GC inputs per PC, yielding a capture fraction of ~1–2%. Although this represents only a subset of total inputs, it is consistent with mammalian studies (Wall et al., 2010; Callaway et al., 2015), suggesting inherent limitations of this viral labeling approach.

For IO inputs, we counted 325 ± 26 inferior olivary neurons in Tg(elavl3:H2B-GCaMP6s) fish. A single PC likely receives input from one IO neuron, though an IO neuron may innervate multiple PCs. Accordingly, the observed capture rate for IO inputs was lower (7 out of 248 starters).

Further optimization is required to enhance the tracing efficiency. We have now incorporated a Discussion on this point in the revised manuscript (page 8).