Single dopaminergic neuron DAN-c1 in Drosophila larval brain mediates aversive olfactory learning through D2-like receptors

Reviewer #1 (Public Review):

Summary:
Both flies and mammals have D1-like and D2-like dopamine receptors, yet the role of D2-like receptors in Drosophila learning and memory remains underexplored. The paper by Qi et al. investigates the role of the D2-like dopamine receptor D2R in single pairs of dopaminergic neurons (DANs) during single-odor aversive learning in the Drosophila larva. First, they use confocal imaging to screen driver strains with expression in only single pairs of dopaminergic neurons. Next, they use thermogenetic manipulations of one pair of DANs (DAN-c1) to implicate DAN-c1 activity during larval aversive learning. They then use confocal imaging to demonstrate expression of D2R in the DANs and mushroom body of the larval brain. Finally, they show that optogenetic activation during training phenocopies D2R knockdown in these neurons: aversive learning is impaired when DAN-c1 is targeted, while appetitive and aversive learning are impaired when the mushroom body is manipulated. Qi et al. thus propose a model in which D2R limits excessive dopamine release to facilitate successful olfactory learning.

Strengths:
The paper reproduces prior findings by Qi and Lee (2014), which demonstrated that D2R knockdown in DL1 DANs or the mushroom body impairs aversive olfactory learning in Drosophila larvae. The authors extended this previous work by screening 57 GAL4 drivers to identify tools that drive expression in individual DANs and used one of the tools, the R76F02-AD; R55C10-DBD driver, to manipulate DAN-c1 neurons with greater specificity. They also show that GFP-tagged D2R is expressed in most DANs and the mushroom body. Although the authors only train larvae with a single odor, they demonstrate that driving D2R knockdown in DAN-c1 neurons impairs aversive learning, as do other loss-of-function manipulations of DAN-c1 neurons.

Weaknesses:
The authors claim to have identified drivers that label single DANs in Figure 1, but their confocal images in Figure S1 suggest that many of those drivers label additional neurons in the larval brain. It is also not clear why only some of the 57 drivers are displayed in Figure S1.
Critically, R76F02-AD; R55C10-DBD labels more than one neuron per hemisphere in Figure S1c, and the authors cite Xie et al. (2018) to note that this driver labels two DANs in adult brains. Therefore, the authors cannot argue that the experiments throughout their paper using this driver exclusively target DAN-c1.
Missing from the screen of 57 drivers is the driver MB320C, which typically labels only PPL1-γ1pedc in the adult and should label DAN-c1 in the larva. If MB320C labels DAN-c1 exclusively in the larva, then the authors should repeat their key experiments with MB320C to provide more evidence for DAN-c1 involvement specifically.
The authors claim that the SS02160 driver used by Eschbach et al. (2020) labels other neurons in addition to DAN-c1. Could the authors use confocal imaging to show how many other neurons SS02160 labels? Given that both Eschbach et al. and Weber et al. (2023) found no evidence that DAN-c1 plays a role in larval aversive learning, it would be informative to see how SS02160 expression compares with the driver the authors use to label DAN-c1.
The claim that DAN-c1 is both necessary and sufficient in larval aversive learning should be reworded. Such a claim would logically exclude any other neuron or even the training stimuli from being involved in aversive learning (see Yoshihara and Yoshihara (2018) for a detailed discussion of the logic), which is presumably not what the authors intended because they describe the possible roles of other DANs during aversive learning in the discussion.
Moreover, if DAN-c1 artificial activation conveyed an aversive teaching signal irrespective of the gustatory stimulus, then it should not impair aversive learning after quinine training (Figure 2k). While the authors interpret Figure 2k (and Figure 5) to indicate that artificial activation causes excessive DAN-c1 dopamine release, an alternative explanation is that artificial activation compromises aversive learning by overriding DAN-c1 activity that could be evoked by quinine.
The authors should not necessarily expect that D2R enhancer driver strains would reflect D2R endogenous expression, since it is known that TH-GAL4 does not label p(PAM) dopaminergic neurons. Their observations of GFP-tagged D2R expression could be strengthened with an anti-D2R antibody such as that used by Lam et al., (1999) or Love et al., (2023).
Finally, the authors could consider the possibility other DANs may also mediate aversive learning via D2R. Knockdown of D2R in DAN-g1 appears to cause a defect in aversive quinine learning compared with its genetic control (Figure S4e). It is unclear why the same genetic control has unexpectedly poor aversive quinine learning after training with propionic acid (Figure S5a). The authors could comment on why RNAi knockdown of D2R in DAN-g1 does not similarly impair aversive quinine learning (Figure S5b).

Reviewer #2 (Public Review):

Summary:
The study wanted to functionally identify individual DANs that mediate larval olfactory
learning. Then search for DAN-specific driver strains that mark single dopaminergic neurons, which subsequently can be used to target genetic manipulations of those neurons. 56 GAL4 drivers identifying dopaminergic neurons were found (Table 1) and three of them drive the expression of GFP to a single dopaminergic neuron in the third-instar larval brain hemisphere. The DAN driver R76F02-AD;R55C10-DBD appears to drive the expression to a dopaminergic neuron innervating the lower peduncle (LP), which would be DAN-c1.
Split-GFP reconstitution across synaptic partners (GRASP) technique was used to investigate the "direct" synaptic connections from DANs to the mushroom body. Potential synaptic contact between DAN-c1 and MB neurons (at the lower peduncle) were detected.
Then single odor associative learning was performed and thermogenetic tools were used (Shi-ts1 and TrpA1). When trained at 34{degree sign}C, the complete inactivation of dopamine release from DAN-c1 with Shibirets1 impaired aversive learning (Figure 2h), while Shibirets1 did not affect learning when trained at room temperature (22{degree sign}C). When paired with a gustatory stimulus (QUI or SUC), activation of DAN-c1 during training impairs both aversive and appetitive learning (Figure 2k).
They examined the expression pattern of D2R in fly brains and were found in dopaminergic neurons and the mushroom body (Figure 3). To inspect whether the pattern of GFP signals indeed reflected the expression of D2R, three D2R enhancer driver strains (R72C04, R72C08, and R72D03-GAL4) were crossed with the GFP-tagged D2R strain.
D2R knockdown (UAS-RNAi) in dopaminergic neurons driven by TH-GAL4 impaired larval aversive learning. Using a microRNA strain (UAS-D2R-miR), a similar deficit was observed. Crossing the GFP-tagged D2R strain with a DAN-c1-mCherry strain demonstrated the expression of D2R in DAN-c1 (Figure 4a). Knockdown of D2R in DAN-c1 impaired aversive learning with the odorant pentyl acetate, while appetitive learning was unaffected (Figure 4e). Sensory and motor functions appear not affected by D2R suppression.
To exclude possible chronic effects of D2R knockdown during development, optogenetics was applied at distinct stages of the learning protocol. ChR2 was expressed in DAN-c1, and blue light was applied at distinct stages of the learning protocol. Optogenetic activation of DAN-c1 during training impaired aversive learning, not appetitive learning (Figure 5b-d).
Knockdown of D2Rs in MB neurons by D2R-miR impaired both appetitive and aversive learning (Figure 6a). Activation of MBNs during training impairs both larval aversive and appetitive learning.
Finally, based on the data the authors propose a model where the effective learning requires a balanced level of activity between D1R and D2R (Figure 7).

Strengths:
The work is well written, clear, and concise. They use well documented strategies to examine GAL4 drivers with expression in a single DAN, behavioral performance in larvae with distinct genetic tools including those to do thermo and optogenetics in behaving flies. Altogether, the study was able to expand our understanding of the role of D2R in DAN-c1 and MB neurons in the larva brain.

Weaknesses:
Is not completely clear how the system DAN-c1, MB neurons and Behavioral performance work. We can be quite sure that DAN-c1;Shits1 were reducing dopamine release and impairing aversive memory (Figure 2h). Similarly, DAN-c1;ChR2 were increasing dopamine release and also impaired aversive memory (Figure 5b). However, is not clear what is happening with DAN-c1;TrpA1 (Figure 2K). In this case the thermos-induction appears to impair the behavioral performance of all three conditions (QUI, DW and SUC) and the behavior is quite distinct from the increase and decrease of dopamine tone (Figure 2h and 5b).

The study successfully examined the role of D2R in DAN-c1 and MB neurons in olfactory conditioning. The conclusions are well supported by the data, with the exception of the claim that dopamine release from DAN-c1 is sufficient for aversive learning in the absence of unconditional stimulus (Figure 2K). Alternatively, the authors need to provide a better explanation of this point.
The study provides insight into the role of D2R in associative learning expanding our understanding and might be a reference similar to previous key findings (Qi and Lee, 2014, https://doi.org/10.3390/biology3040831).

Reviewer #3 (Public Review):

It is a strength of the paper that it analyses the function of dopamine neurons (DANs) at the level of single, identified neurons, and uses tools to address specific dopamine receptors (DopRs), exploiting the unique experimental possibilities available in larval Drosophila as a model system. Indeed, the result of their screening for transgenic drivers covering single or small groups of DANs and their histological characterization provides the community with a very valuable resource. In particular the transgenic driver to cover the DANc1 neuron might turn out useful. However, I wonder in which fraction of the preparations an expression pattern as in Figure 1f/ S1c is observed, and how many preparations the authors have analyzed. Also, given the function of DANs throughout the body, in addition to the expression pattern in the mushroom body region (Figure 1f) and in the central nervous system (Figure S1c) maybe attempts can be made to assess expression from this driver throughout the larval body (same for Dop2R distribution).

A first major weakness is that the main conclusion of the paper, which pertains to associative memory (last sentence of the abstract, and throughout the manuscript), is not justified by their evidence. Why so? Consider the paradigm in Figure 2g, and the data in Figure 2h (22 degrees, the control condition), where the assay and the experimental rationale used throughout the manuscript are introduced. Different groups of larvae are exposed, for 30min, to an odour paired with either i) quinine solution (red bar), ii) distilled water (yellow bar), or iii) sucrose solution (blue bar); in all cases this is followed by a choice test for the odour on one side and a distilled-water blank on the other side of a testing Petri dish. The authors observe that odour preference is low after odour-quinine pairing, intermediate after odour-water pairing and high after odour-sucrose pairing. The differences in odour preference relative to the odour-water case are interpreted as reflecting odour-quinine aversive associations and odour-sucrose appetitive associations, respectively. However, these differences could just as well reflect non-associative effects of the 30-min quinine or sucrose exposure per se (for a classical discussion of such types of issues see Rescorla 1988, Annu Rev Neurosci, or regarding Drosophila Tully 1988, Behav Genetics, or with some reference to the original paper by Honjo & Furukubo-Tokunaga 2005, J Neurosci that the authors reference, also Gerber & Stocker 2007, Chem Sens).
As it stands, therefore, the current 3-group type of comparison does not allow conclusions about associative learning.

A second major weakness is apparent when considering the sketch in Figure 2g and the equation defining the response index (R.I.) (line 480). The point is that the larvae that are located in the middle zone are not included in the denominator. This can inflate scores and is not appropriate. That is, suppose from a group of 30 animals (line 471) only 1 chooses the odour side and 29, bedazzled after 30-min quinine or sucrose exposure or otherwise confused by a given opto- or thermogenetic treatment, stay in the middle zone... a P.I. of 1.0 would result.

Unless experimentally demonstrated, claims that the thermogenetic effector shibire/ts reduces dopamine release from DANs are questionable. This is because firstly, there might be shibire/ts-insensitive ways of dopamine release, and secondly because shibire/ts may affect co-transmitter release from DANs.
To implicate a role of dopamine in DANs, previous work used e.g. RNAi against the dopamine-synthesizing TH enzyme (Rohwedder et al, cited).

It is not clear whether the genetic controls when using the Gal4/ UAS system are the homozygous, parental strains (XY-Gal4/ XY-Gal4 and UAS-effector/ UAS-effector), or as is standard in the field the heterozygous driver (XY-Gal4/ wildtype) and effector controls (UAS-effector/ wildtype) (in some cases effector controls appear to be missing, e.g. Figure 4d, Figure S4e, Figure S5c).

As recently suggested by Yamada et al 2024, bioRxiv, high cAMP can lead to synaptic depression (sic). That would call into question the interpretation of low-Dop2R leading to high-cAMP, leading to high-dopamine release, and thus the authors interpretation of the matching effects of low-Dop2R and driving DANs.

Author response:

Reviewer #1 (Public Review):

Weakness #1: The authors claim to have identified drivers that label single DANs in Figure 1, but their confocal images in Figure S1 suggest that many of those drivers label additional neurons in the larval brain. It is also not clear why only some of the 57 drivers are displayed in Figure S1.

As introduced in the results section, we screened 57 driver strains based on previous studies, either they were reported identifying a single (a pair of) dopaminergic neuron (DAN) in larvae or identifying only several DANs in the adult brain indicating the potential of identifying single dopaminergic neuron in larvae. In Figure 1, TH-GAL4 was used to cover all neurons in the DL1 cluster, while R58E02 and R30G08 were well known drivers for pPAM. Fly strains in Figure 1h, k, l, and m were reported as single DAN strains in larvae4, while strains in Figure 1e, f, g were reported identifying only several DANs in adult brains5,6. We examined these strains and only some of them labeled single DANs in 3rd instar larval brains (Figure 1f, g, h, l and m). Among them, only strains in Figure 1f and h labeled single DAN in the brain hemisphere, without labeling other non-DANs. Other strains labeled non-DANs in addition to single DANs (Figure 1g, l and m). Taking ventral nerve cord (VNC) into consideration, strain in Figure 1h also labeled neurons in VNC (Figure S1e), while strain in Figure 1f did not (Figure S1c).

In summary, the strain in Figure 1f (R76F02AD;R55C10DBD, labeling DAN-c1) is a strain we screened labeling only a single DAN in the 3rd instar larval brains. Others (Figure 1g, h, l, and m) we still describe them as strains labeling single DANs, but they also label one to several non-DANs. In Figure 1, we mainly showed the strains labeling single DANs. The labeling patterns of other screened driver strains were summarized in Table1. Since all brain images of the rest 47 strains are available, we will state in Fig S1 that additional brain images can be provided upon request.

Weakness #2: Critically, R76F02-AD; R55C10-DBD labels more than one neuron per hemisphere in Figure S1c, and the authors cite Xie et al. (2018) to note that this driver labels two DANs in adult brains. Therefore, the authors cannot argue that the experiments throughout their paper using this driver exclusively target DAN-c1.

Figure S1c shows single DA neuron in each brain hemisphere. Additional GFP (+) signals were often observed, but not from cell bodies of DANs because they were not stained by a TH antibody. These additional GFP (+) signals were mainly neurites, including axonal terminals, but could be false positive signals or weakly stained non-neuronal cell bodies. This conclusion was based on analysis of a total of 22 larval brains. We will add this in the text or Fig S1 caption. Enlarged insert of GFP (+) signals will be added also to Figure S1c.

Weakness #3: Missing from the screen of 57 drivers is the driver MB320C, which typically labels only PPL1-γ1pedc in the adult and should label DAN-c1 in the larva. If MB320C labels DAN-c1 exclusively in the larva, then the authors should repeat their key experiments with MB320C to provide more evidence for DAN-c1 involvement specifically.

We thank the reviewer for the suggestion. MB320C mainly labels PPL1-y1pedc in the adult brain, with one or two other weakly labeled cells. It will be interesting to investigate the pattern of this driver in 3rd instar larval brains. If it only covers DAN-c1, we can try to knock-down D2R in this strain to check whether it can repeat our results. This will be an interesting fly strain to test, but we believe that it will not be necessary for our current manuscript as DAN-c1 driver is very specific (for details, refer to our response to Reviewer#3). However, this line will be very useful for future experiments.

Weakness #4: The authors claim that the SS02160 driver used by Eschbach et al. (2020) labels other neurons in addition to DAN-c1. Could the authors use confocal imaging to show how many other neurons SS02160 labels? Given that both Eschbach et al. and Weber et al. (2023) found no evidence that DAN-c1 plays a role in larval aversive learning, it would be informative to see how SS02160 expression compares with the driver the authors use to label DAN-c1.

We did not have our own images showing DANs in brains of SS02160 driver cross line. However, Extended Data Figure 1 in the paper of Eschbach et al. (2020) shows strongly labeled four neurons on each brain hemisphere9, indicating that this driver is not a strain only labeling one neuron, DAN-c1.

Weakness #5: The claim that DAN-c1 is both necessary and sufficient in larval aversive learning should be reworded. Such a claim would logically exclude any other neuron or even the training stimuli from being involved in aversive learning (see Yoshihara and Yoshihara (2018) for a detailed discussion of the logic), which is presumably not what the authors intended because they describe the possible roles of other DANs during aversive learning in the discussion.

We agree that the words ‘necessary’ and ‘sufficient’ are too exclusive for other neurons. As mentioned in the Discussion part, we do think other dopaminergic neurons may also be involved in larval aversive learning. We are going to re-phrase these words by replacing them with more logically appropriate words, such as ‘important’, ‘essential’, or ‘mediating’.

Weakness #6: Moreover, if DAN-c1 artificial activation conveyed an aversive teaching signal irrespective of the gustatory stimulus, then it should not impair aversive learning after quinine training (Figure 2k). While the authors interpret Figure 2k (and Figure 5) to indicate that artificial activation causes excessive DAN-c1 dopamine release, an alternative explanation is that artificial activation compromises aversive learning by overriding DAN-c1 activity that could be evoked by quinine.

This is a great point! Yes, we cannot rule out the possibility that artificial activation compromises aversive learning by overriding DAN-c1 activity that could be evoked by quinine. The experimental results with TRPA1 could be caused by depletion of dopamine, or DA inactivation due to prolonged depolarization or adaptation. However, we still think that our hypothesis on the over-excitation of DAN-c1 is more consistent with our experimental results and other published data. Our justification is as follows:

(1) Associative learning occurs only when the CS and US are paired. In wild type larvae, a specific odor (conditioned stimulus, CS, such as pentyl acetate) depolarizes a subset of Kenyon cells in the mushroom body, while gustatory unconditioned stimulus (US, quinine) induces dopamine release from DAN-c1 to the lower peduncle (LP) compartment in the mushroom body (Figure 7a). Only when the CS and US are paired, calcium influx caused by CS and Gas activated by D1R binding to dopamine will turn on a mushroom body specific version of adenylyl cyclase, rutabaga, which is the co-incidence detector in associative learning (Figure 7d).

(2) Rutabaga transforms ATP into cAMP, activating PKA signaling pathway and modifying the synaptic strength from mushroom body neurons (MBN, also called Kenyan cells) to the mushroom body output neurons (MBON, Figure 7d). This change in synaptic strength will lead to learned responses when the same odor appears again.

(3) In our work, we found D2R is expressed in DAN-c1, and knockdown D2R in DAN-c1 impairs larval aversive learning. As D2R reduces cAMP level and neuronal excitability3, we hypothesized that knockdown of D2R in DAN-c1 would remove the inhibition of D2R auto-receptor, and lead to more dopamine (DA) release when US (quinine) was delivered compared to the wild type larvae. The elevated DA release along with calcium influx caused by CS increases the cAMP level in MBN, which leads to the learning deficit (over-excitation, Figure 7b). Mutant larvae with excessive cAMP, dunce, showed aversive learning deficiency, supporting our hypothesis2.

(4) Our results of TRPA1 can be explained by this over-excitation hypothesis. When DAN-c1 is activated (34C) in distilled water group, the artificial activation mimicked the gustatory activation of quinine. The larvae showed the aversive learning responses towards the odor (Figure 2k DW group). When DAN-c1 is activated (34C) in sucrose group, the artificial activation mimicked the gustatory activation of quinine, so the larvae showed a learning response combining both appetitive and aversive learning (Figure 2k SUC group).

(5) When DAN-c1 is activated (34C) in quinine group, the artificial activation and the gustatory activation of quinine lead to elevated DA release from DAN-c1. During training, this elevated DA caused over-excitation of MBN, leading to failure of aversive learning (Figure 2k QUI group), which had a similar phenotype compared to larvae with D2R knockdown in DAN-c1.

(6) Similarly, optogenetic activation of DAN-c1 during aversive training, leads to elevated DA release from DAN-c1 (both gustatory activation of quinine and artificial activation). This would also cause over-excitation of MBN, and lead to failure of aversive learning. Artificial activation in other stages (resting or testing) won’t cause elevated DA release during training, so the aversive learning was not affected (Figure 5b).

(7) However, when optogenetic activation was applied during training, we did not observe aversive learning responses in the distilled water group, or a reduction in the sucrose group (Figure 5c, Figure 5d). Our explanation is that the optogenetic stimulus we applied is too strong, DAN-c1 has already released elevated DA in both groups. So, the aversive learning in these groups has already been impaired, they just showed the corresponding learning responses to distilled water or sucrose.

(8) We also applied this over-excitation to activate MBNs. As MBN takes over both appetitive and aversive learnings, over-excitation of MBNs led to deficit in both types of learning, which follows our hypothesis (Figure 6).

In summary, we hypothesized that DAN-c1 restricts DA release via activation of D2R, which is important for larval aversive learning. D2R knockdown or artificial activation of DAN-c1 during training would induce elevated DA release, leading to over-excitation of MBNs and failure of aversive learning.

Weakness #7: The authors should not necessarily expect that D2R enhancer driver strains would reflect D2R endogenous expression, since it is known that TH-GAL4 does not label p(PAM) dopaminergic neurons.

Just like the example of TH-GAL4, it is possible that the D2R driver strains may partially reflect the expression pattern of endogenous D2R in larval brains. When we crossed the D2R driver strains with the GFP-tagged D2R strain, however, we observed co-localization in DM1 and DL2b dopaminergic neurons, as well as in mushroom body neurons (Figure S3 c to h). In addition, D2R knockdown with D2R-miR directly supported that the GFP-tagged D2R strain reflected the expression pattern of endogenous D2R (Figure 4b to d, signals were reduced in DM1). In summary, we think the D2R driver strains supported the expression pattern we observed from the GFP-tagged D2R strain, especially in DM1 DANs.

Weakness #8: Their observations of GFP-tagged D2R expression could be strengthened with an anti-D2R antibody such as that used by Lam et al., (1999) or Love et al., (2023).

Love et al., (2023) used the antibody from Draper et al.10. We have tried the same antibody, but we were not able to observe clear signals after staining. Maybe it is not specific for the neurons in the fly larval brain, or our staining protocol did not fit with this antibody.

Unfortunately, we were not able to find Lam (1999) paper.

Weakness #9: Finally, the authors could consider the possibility other DANs may also mediate aversive learning via D2R. Knockdown of D2R in DAN-g1 appears to cause a defect in aversive quinine learning compared with its genetic control (Figure S4e). It is unclear why the same genetic control has unexpectedly poor aversive quinine learning after training with propionic acid (Figure S5a). The authors could comment on why RNAi knockdown of D2R in DAN-g1 does not similarly impair aversive quinine learning (Figure S5b).

We also think that other DANs may be involved in aversive learning. We re-analyzed the learning assay data, seemingly D2R knockdown in DAN-g1 with miR partially affected aversive learning when trained with pentyl acetate (Figure S4e). We are going to build single statistic panels for DAN-g1 and DAN-d1. However, neither larvae with D2R knockdown in DAN-g1 using miR trained with propionic acid (Figure S5a), nor larvae with D2R knockdown in DAN-g1 using RNAi trained with pentyl acetate (Figure S5b) showing aversive learning deficit. We will add paragraphs about this in both Results and Discussion sections.

Reviewer #2 (Public Review):

Weakness#1: Is not completely clear how the system DAN-c1, MB neurons and Behavioral performance work. We can be quite sure that DAN-c1;Shits1 were reducing dopamine release and impairing aversive memory (Figure 2h). Similarly, DAN-c1;ChR2 were increasing dopamine release and also impaired aversive memory (Figure 5b). However, is not clear what is happening with DAN-c1;TrpA1 (Figure 2K). In this case the thermos-induction appears to impair the behavioral performance of all three conditions (QUI, DW and SUC) and the behavior is quite distinct from the increase and decrease of dopamine tone (Figure 2h and 5b).

The study successfully examined the role of D2R in DAN-c1 and MB neurons in olfactory conditioning. The conclusions are well supported by the data, with the exception of the claim that dopamine release from DAN-c1 is sufficient for aversive learning in the absence of unconditional stimulus (Figure 2K). Alternatively, the authors need to provide a better explanation of this point.

Please refer to our response to Weakness #6 of Public Reviewer #1.

Reviewer #3 (Public Review):

Weakness #1: It is a strength of the paper that it analyses the function of dopamine neurons (DANs) at the level of single, identified neurons, and uses tools to address specific dopamine receptors (DopRs), exploiting the unique experimental possibilities available in larval Drosophila as a model system. Indeed, the result of their screening for transgenic drivers covering single or small groups of DANs and their histological characterization provides the community with a very valuable resource. In particular the transgenic driver to cover the DANc1 neuron might turn out useful. However, I wonder in which fraction of the preparations an expression pattern as in Figure 1f/ S1c is observed, and how many preparations the authors have analyzed. Also, given the function of DANs throughout the body, in addition to the expression pattern in the mushroom body region (Figure 1f) and in the central nervous system (Figure S1c) maybe attempts can be made to assess expression from this driver throughout the larval body (same for Dop2R distribution).

We thank the reviewer for the positive comments and the suggestions. For the strain R76F02AD; R55C10DBD, we examined 22 third instar larval brains expressing GFP or Syt-GFP and Den-mCherry, all of them clearly labeled DAN-c1. Half of them only labeled DAN-c1, the rest have 1 to 5 weak labeled soma without neurites. Barely 1 or 2 strong labeled cells appear. These non-DAN-c1 neurons are seldom dopaminergic neurons. In VNC, 8 out of 12 do not label cells, 3 have 2-4 strong labeled cells. These data supported that R76F02AD;R55C10DBD exclusively labeled DAN-c1 in 3rd instar larval brains.

For the question about the pattern of R76F02AD; R55C10DBD and the expression pattern of D2R in larval body, it is an interesting question. However, our main focus was on the central nervous system and the learning behaviors in fruit fly larvae, we may investigate this question in the future.

Weakness #2: A first major weakness is that the main conclusion of the paper, which pertains to associative memory (last sentence of the abstract, and throughout the manuscript), is not justified by their evidence. Why so? Consider the paradigm in Figure 2g, and the data in Figure 2h (22 degrees, the control condition), where the assay and the experimental rationale used throughout the manuscript are introduced. Different groups of larvae are exposed, for 30min, to an odour paired with either i) quinine solution (red bar), ii) distilled water (yellow bar), or iii) sucrose solution (blue bar); in all cases this is followed by a choice test for the odour on one side and a distilled-water blank on the other side of a testing Petri dish. The authors observe that odour preference is low after odour-quinine pairing, intermediate after odour-water pairing and high after odour-sucrose pairing. The differences in odour preference relative to the odour-water case are interpreted as reflecting odour-quinine aversive associations and odour-sucrose appetitive associations, respectively. However, these differences could just as well reflect non-associative effects of the 30-min quinine or sucrose exposure per se (for a classical discussion of such types of issues see Rescorla 1988, Annu Rev Neurosci, or regarding Drosophila Tully 1988, Behav Genetics, or with some reference to the original paper by Honjo & Furukubo-Tokunaga 2005, J Neurosci that the authors reference, also Gerber & Stocker 2007, Chem Sens).
As it stands, therefore, the current 3-group type of comparison does not allow conclusions about associative learning.

We adopted this single odor larval learning paradigm from Honjo’s papers1,2. In these works, Honjo et al. first designed and performed this single odor paradigm for larval olfactory associative learning. To address the reviewer’s question about the potential non-associative effects of the 30-min quinine or sucrose exposure, we would like to defend it primarily based on results from Honjo et al. (2005 and 2009). They applied the odorant to the larvae after training, only the ones had paired training with both odor and unconditioned stimulus (quinine or sucrose) showed learning responses. Larvae exposed 30 min in only odorant or unconditioned stimulus did not show different response to the odor compared to the naïve group1,2. To validate this paradigm induces associative learning responses, they also tested the paradigm from three aspects:

(1) The odor responses are associative. Honjo et al. showed only when the odorant paired with unconditioned stimulus would induce corresponding attraction or repulsion of larvae to the odor. Neither odorant alone, unconditioned stimulus alone, nor temporal dissociation of odorant and unconditioned stimulus would induce learning responses.

(2) The odor responses are odor specific. When applied a second odorant that was not used for training, larvae only showed learning responses to the unconditioned stimulus paired odor. This result ruled out the explanation of a general olfactory suppression and indicates larvae can discriminate and specifically alter the responses to the odor paired with unconditioned stimulus. Although the two-odor reciprocal training is not used, these results can show the association of unconditioned stimulus and the corresponding paired odor.

(3) Well known learning deficit mutants did not show learned responses in this learning paradigm. Honjo et al. tested mutants (e.g., rut and dnc) showing learning deficits in the adult stage with two odor reciprocal learning paradigm. These mutant larvae also failed to show learning responses tested with the single odor larval learning paradigm.

(4) In our study, we used two distinct odorants (pentyl acetate and propionic acid), as well as two D2R knockdown strains (UAS-miR and UAS-RNAi for D2R). We obtained similar results for larvae with D2R knockdown in DAN-c1. In addition, our naïve olfactory, naïve gustatory, and locomotion data ruled out the possibilities that the responses were caused by impaired sensory or motor functions. Comparison with the control group (odor paired with distilled water) ruled out the potential effects if habituation existed. All these results supported this single odor learning paradigm is reliable to assess the learning abilities of Drosophila larvae. And the failure of reduction in R.I when larvae with D2R knockdown in DAN-c1 were trained in quinine paired with the odorant is caused by deficit in aversive learning ability. We will add a paragraph to address this in the Discussion part.

Weakness #3: A second major weakness is apparent when considering the sketch in Figure 2g and the equation defining the response index (R.I.) (line 480). The point is that the larvae that are located in the middle zone are not included in the denominator. This can inflate scores and is not appropriate. That is, suppose from a group of 30 animals (line 471) only 1 chooses the odor side and 29, bedazzled after 30-min quinine or sucrose exposure or otherwise confused by a given opto- or thermogenetic treatment, stay in the middle zone... a P.I. of 1.0 would result.

It is a good question. We gave 5 min during the testing stage to allow the larvae to wander in the testing plate. Under most conditions, more than half of larvae (>50%) will explore around, and the rest may stay in the middle zone (will not be calculated). We used 25-50 larvae in each learning assay, so finally around 10-30 larvae will locate in two semicircular areas. Indeed, based on our raw data, a R.I. of 1 seldom appears. Most of the R.I.s fall into a region from -0.2 to 0.8. We should admit that the calculation equation of R. I. is not linear, so it would be sharper (change steeply) when it approaching to -1 and 1. However, as most of the values fall into the region from -0.2 to 0.8, we think ‘border effects’ can be neglected if we have enough numbers of larvae in the calculation (10-30).

Weakness #4: Unless experimentally demonstrated, claims that the thermogenetic effector shibire/ts reduces dopamine release from DANs are questionable. This is because firstly, there might be shibire/ts-insensitive ways of dopamine release, and secondly because shibire/ts may affect co-transmitter release from DANs.

Shibirets1 gene encodes a thermosensitive mutant of dynamin, expressing this mutant version in target neurons will block neurotransmitter release at the ambient temperature higher than 30C, as it represses vesicle recycling1. It is a widely used tool to examine whether the target neuron is involved in a specific physiological function. We cannot rule out that there might be Shibirets1 insensitive ways of dopamine release exist. However, blocking dopamine release from DAN-c1 with Shibirets1 has already led to learning responses changing (Figure 2h). This result indicated that the dopamine release from DAN-c1 during training is important for larval aversive learning, which has already supported our hypothesis.

For the second question about the potential co-transmitter release, we think it is a great question. Recently Yamazaki et al. reported co-neurotransmitters in dopaminergic system modulate adult olfactory memories in _Drosophila_11, and we cannot rule out the roles of co-released neurotransmitters/neuropeptides in larval learning. Ideally, if we could observe the real time changes of dopamine release from DAN-c1 in wild type and TH knockdown larvae would answer this question. However, live imaging of dopamine release from one dopaminergic neuron is not practical for us at this time. On the other hand, the roles of dopamine receptors in olfactory associative learning support that dopamine is important for Drosophila learning. D1 receptor, dDA1, has been proven to be involved in both adult and larval appetitive and aversive learning12,13. In our work, D2R in the mushroom body showed important roles in both larval appetitive and aversive learning (Figure 6a). All this evidence reveals the importance of dopamine in Drosophila olfactory associative learning. In addition, there is too much unknow information about the co-release neurotransmitter/neuropeptides, as well as their potential complex ‘interaction/crosstalk’ relations. We believe that investigation of co-released neurotransmitter/neuropeptides is beyond the scope of this study at this time.

Weakness #5: It is not clear whether the genetic controls when using the Gal4/ UAS system are the homozygous, parental strains (XY-Gal4/ XY-Gal4 and UAS-effector/ UAS-effector), or as is standard in the field the heterozygous driver (XY-Gal4/ wildtype) and effector controls (UAS-effector/ wildtype) (in some cases effector controls appear to be missing, e.g. Figure 4d, Figure S4e, Figure S5c).

Almost all controls we used were homozygous parental strains. They did not show abnormal behaviors in either learnings or naïve sensory or locomotion assays. The only exception is the control for DAN-c1, the larvae from homozygous R76F02AD; R55C10DBD strain showed much reduced locomotion speed (Figure S6). To prevent this reduced locomotion speed affecting the learning ability, we used heterozygous R76F02AD; R55C10DBD/wildtype as control, which showed normal learning, naïve sensory and locomotion abilities (Figure 4e to i).

For Figure 4d, it is a column graph to quantify the efficiency of D2R knockdown with miR. Because we need to induce and quantify the knockdown effect in specific DANs (DM1), only TH-GAL4 can be used as the control group, rather than UAS-D2R-miR.

For the missing control groups in Figure S4e and S5c, we have shown them in other Figures (Figure 4e). We will re-organize the figures to make them easier to understand.

Weakness #6: As recently suggested by Yamada et al 2024, bioRxiv, high cAMP can lead to synaptic depression (sic). That would call into question the interpretation of low-Dop2R leading to high-cAMP, leading to high-dopamine release, and thus the authors interpretation of the matching effects of low-Dop2R and driving DANs.

We will read through this paper and try to add it as possible explanations for the learning mechanisms. As we introduced in the Discussion section, the learning mechanism is quite complex, mixing both non-linear neuronal circuits and multiple signaling pathways, in responding to complex environmental learning contexts. We will try to develop a better hypothesis with the best compatibility to accommodate our results with published data.

Reference

(1) Honjo, K. & Furukubo-Tokunaga, K. Induction of cAMP response element-binding protein-dependent medium-term memory by appetitive gustatory reinforcement in Drosophila larvae. J Neurosci 25, 7905-7913 (2005). https://doi.org/10.1523/JNEUROSCI.2135-05.2005

(2) Honjo, K. & Furukubo-Tokunaga, K. Distinctive neuronal networks and biochemical pathways for appetitive and aversive memory in Drosophila larvae. J Neurosci 29, 852-862 (2009). https://doi.org/10.1523/JNEUROSCI.1315-08.2009

(3) Neve, K. A., Seamans, J. K. & Trantham-Davidson, H. Dopamine receptor signaling. J Recept Signal Transduct Res 24, 165-205 (2004). https://doi.org/10.1081/rrs-200029981

(4) Saumweber, T. et al. Functional architecture of reward learning in mushroom body extrinsic neurons of larval Drosophila. Nat Commun 9, 1104 (2018). https://doi.org/10.1038/s41467-018-03130-1

(5) Aso, Y. & Rubin, G. M. Dopaminergic neurons write and update memories with cell-type-specific rules. Elife 5 (2016). https://doi.org/10.7554/eLife.16135

(6) Xie, T. et al. A Genetic Toolkit for Dissecting Dopamine Circuit Function in Drosophila. Cell Rep 23, 652-665 (2018). https://doi.org/10.1016/j.celrep.2018.03.068

(7) Hartenstein, V., Cruz, L., Lovick, J. K. & Guo, M. Developmental analysis of the dopamine-containing neurons of the Drosophila brain. J Comp Neurol 525, 363-379 (2017). https://doi.org/10.1002/cne.24069

(8) Aso, Y. et al. The neuronal architecture of the mushroom body provides a logic for associative learning. Elife 3, e04577 (2014). https://doi.org/10.7554/eLife.04577

(9) Eschbach, C. et al. Recurrent architecture for adaptive regulation of learning in the insect brain. Nat Neurosci 23, 544-555 (2020). https://doi.org/10.1038/s41593-020-0607-9

(10) Draper, I., Kurshan, P. T., McBride, E., Jackson, F. R. & Kopin, A. S. Locomotor activity is regulated by D2-like receptors in Drosophila: an anatomic and functional analysis. Dev Neurobiol 67, 378-393 (2007). https://doi.org/10.1002/dneu.20355

(11) Yamazaki, D., Maeyama, Y. & Tabata, T. Combinatory Actions of Co-transmitters in Dopaminergic Systems Modulate Drosophila Olfactory Memories. J Neurosci 43, 8294-8305 (2023). https://doi.org/10.1523/jneurosci.2152-22.2023

(12) Selcho, M., Pauls, D., Han, K. A., Stocker, R. F. & Thum, A. S. The role of dopamine in Drosophila larval classical olfactory conditioning. PLoS One 4, e5897 (2009). https://doi.org/10.1371/journal.pone.0005897

(13) Kim, Y. C., Lee, H. G. & Han, K. A. D1 dopamine receptor dDA1 is required in the mushroom body neurons for aversive and appetitive learning in Drosophila. J Neurosci 27, 7640-7647 (2007). https://doi.org/10.1523/JNEUROSCI.1167-07.2007