Mushroom body dopamine neurons drive approach and avoidance

To investigate the role of dopamine on acute approach/avoidance behaviors, we generated transgenic flies expressing an optogenetic activator (Chrimson, henceforth “Chr”) [65] in the PAM dopaminergic neurons (PAM DANs) that project to the MB [9,10,66]. The flies were then analyzed in a light-dark choice assay (Fig 1A). We selected R58E02—a transgenic line of a flanking region of the Dopamine transporter (DAT) gene fused with Gal4 [67,68]. This driver is expressed in a large subset of the PAM DANs and has fibers in a broad set of MB neuropil zones [9,10,23,24,59,69] (Figs 1B and S1A and S1 Movie). Valence was calculated as the mean difference between optogenetic test flies and the corresponding controls and displayed as effect-size curves [70]. Flies expressing the Chr opto-activator with the R58E02 transgene were strongly attracted to light (Fig 1E and S2 Movie). By contrast, flies expressing another driver, R15A04, which is expressed in PAM DANs that send fibers to a more restricted set of MB zones [24,59,68] (Figs 1C and S1B and S3 Movie) tended to avoid the light at the highest illumination intensity (70 μm/mm²; Fig 1F). Thus, while activation of the R58E02 DANs drives strong positive valence, flies will avoid activation in the subset defined by R15A04.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Activities in different dopaminergic cells drive valence.

(a) Schematic of the optogenetic assay showing that after an initial dark phase, half of the chamber is illuminated with 2 bands of red light. See Methods for further details. (b, c) Schematic of the 2 DAN drivers, R58E02 and R15A04, with projections to MB synaptic zones. R58E02 is expressed in nearly all PAM types, projecting to α1, β1, β2, β’1, β’2, γ4, and γ5, with weaker expression in γ1, γ2, and the peduncle. R15A04 is expressed in PAMs that project to the α1, β2, β’1, and γ5 zones. (d) Schematic of the hypothetical locomotor modes for valence. Top Flies move slowly in the favored area. Bottom Flies maneuver to remain in the favored area. Either mode increases the time spent in the preferred area. (e) R58E02>Chr flies spent more time in the light zones. The upper panel shows the preference indices (PIs) for test flies (red dots) and driver and responder controls (R58E02/+ and Chr/+, gray dots). The lower panel shows the valence effect sizes (mean differences, ΔPI) between control and test flies, with confidence intervals (black line) and the distribution of ΔPI error (blue curve). The positive ΔPI values indicate a positive valence. See S1 Table for detailed genotypes and statistics. (f) R15A04>Chr flies avoided opto-activation. The negative ΔPI values indicate avoidance. (g) Walking behavior of the subset of flies that entered the choice zone from the dark side and approached the dark–light interface. Only data from flies that approached the choice zone were included. Traces of R58E02>Chr paths (black) are aligned to choice-zone entry, i.e., locked to the time of entering the boundary area. The colored lines show the overall mean trajectory. The horizontal axis is aligned to the middle of the choice zone. Test flies slowed or stopped at the boundary, with their heads on either side of the middle of the light interface. (h) Traces of R15A04>Chr flies as they entered the choice zone from the dark side. Trajectory data were taken from epochs with 70 μW/mm illumination. Data for all panels can be found in the corresponding folder on the Zenodo data repository (https://doi.org/10.5281/zenodo.7747425). MB, mushroom body; PAM, paired-anterior-medial.

https://doi.org/10.1371/journal.pbio.3002843.g001

DAN activation influences locomotion

When traversing a boundary between 2 stimulus areas, walking flies encountering aversive stimuli use reversals or turns to maneuver away [14,51]. A fly that displays a spatial preference for one of 2 areas, however, could hypothetically employ another locomotor mode: slowing down in the favored area (Fig 1D). We thus explored how valence, choice, and speed were associated in R58E02 and R15A04 flies. Regressions between preference, speed ratio, and a choice index showed that in these lines, preference was more strongly determined by differential speed (S1C–S1F Fig). We also inspected locomotion at the boundary by aligning the dark-to-light trajectories in a single experiment. In the subset of flies that entered the boundary choice-zone, trajectory lines drawn over time indicated that as the optogenetic DAN lines traverse into the choice zone from the dark side, the R58E02>Chr flies tend to walk slower and frequently stop in the boundary area (Fig 1G and 1H). These observations indicate a relationship between DAN valence and changes in walking speed.

Olfactory circuits are dispensable to broad DAN attraction

We next investigated whether circuit and molecular components that are essential for memory formation are similarly necessary for valence. We started by investigating the KCs, as DANs instruct odor memory by modulating KC function [52,54]. The expression of a conditioned response in olfactory associative learning (that is, avoidance of or approach towards a conditioned odor) is also reliant on KC activity [71,72]. We asked whether odor memories could be impaired by expressing the light-actuated anion channelrhodopsin, GtACR1 (hereon “ACR1”) in the KCs using the MB247 driver and actuating with green light. As hypothesized, in aversive shock–odor conditioning with ACR1 actuation (Fig 2A), flies expressing the opto-inhibitor in the KCs (MB247>ACR1) failed to learn (Fig 2B). This finding demonstrates that ACR1 sufficiently inhibits KCs to abolish memory formation. In R58E02>Chr flies, green light successfully induced synthetic, optogenetic appetitive memory; however, in R58E02>Chr, MB247>ACR1 flies, light actuation did not induce synthetic memory (Fig 2C), further verifying that ACR1 elicits KC inhibition and can block the memory-inducing effect of R58E02 dopamine cells. We thus demonstrated the successful inhibition of KCs using the opto-inhibitor ACR1, with functional consequences for memory acquisition both using a real stimulus, electric shock, and optogenetic stimulation of the PAMs.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Inhibition of KC activity with GtACR1 prevents the formation of shock-conditioned and DAN-conditioned olfactory memories.

(a) Optogenetic olfactory conditioning protocol: 2 odors were presented in sequence, one paired with green light; the 2 odors were then presented in different arms of the chamber to test the conditioned preference (see Methods for details). (b) Silencing the KCs with ACR1 actuation decreased aversive shock memory by 75.0% [95 CI −44%, −104%]. The genetic controls were MB247-LexA/+ and LexAop-ACR1/+ (gray dots); the test animals were MB247-LexA/LexAop-ACR1 (green dots). Unilluminated flies (both control and test animals) showed robust learning with shock-paired odor (left side, 0 μW/mm²). Illumination with green light reduced the PI of test flies by ΔPI = −0.34 relative to the genetic controls (right side, 28 μW/mm²). (c) Activation of DANs with green light paired with odor, instructed an attractive olfactory memory in R58E02>Chr test flies (orange dots) relative to controls (R58E02-Gal4/+ and UAS-Chr/+, left side gray dots). The contrast between test and control animals is ΔPI = +0.32 (orange curve). Green light induced almost no memory formation in R58E02>Chr; MB247>ACR1 flies (right side, gray dots). The difference attributable to ACR1 inhibition corresponds to a performance reduction of −120% [95 CI −200%, −24%], i.e., a shift from attraction to mild aversion. Sample sizes: N _experiments = 6, 6, 6; N _flies = 144 per genotype. Data for all panels can be found in the corresponding folder on the Zenodo data repository (https://doi.org/10.5281/zenodo.7747425). KC, Kenyon cell.

https://doi.org/10.1371/journal.pbio.3002843.g002

We then questioned whether KC activity was likewise required for acute valence. We implemented a strategy to allow flies to simultaneously activate DANs with Chr and inhibit KCs using ACR1 [73,74]. As both channelrhodopsins are responsive to green light, we tested the ability of green light to both activate DANs and silence KCs. R58E02>Chr flies were attracted to green light (Fig 3A), confirming effective Chr actuation. One possible confound would be if flies responded to KC inhibition with a strong attraction that masked DAN attraction; in a valence test, however, MB247>ACR1 flies exhibited only mild aversion (Fig 3B). We then tested flies carrying the R58E02>Chr, MB247>ACR1 genotype for their optogenetic preference. Even with inhibited KCs, valence remained intact (Figs 3C, S2A, and S2B).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. KCs are dispensable for R58E02 DAN valence.

(a) R58E02>Chr flies are attracted to green light. The left schematic illustrates the expression pattern of R58E02. Flies carrying all 3 transgenes displayed attraction to green light (green dots), resulting in positive valence (black dots and blue curves in the lower panel). Parental type control flies (R58E02-Gal4/+ or MB247-LexA/+; UAS-Chr/+, gray dots) showed a neutral preference for green light. (b) Relative to genetic controls, MB247-LexA>lexAop-ACR1 flies displayed a modest avoidance of green light at high intensities (22 and 72 μW/mm²). The schematic indicates that MB247-LexA drives ACR expression in most MB-intrinsic KCs. (c) In R58E02>Chr/MB247>ACR1 flies, preference for DAN activation mediated by R58E02-Gal4>UAS-Chr was unaffected by simultaneous opto-inhibition of the MB intrinsic cells with MB247-LexA>lexAop-ACR1. Effect sizes (blue curves) show the net effect of comparing test flies carrying all 4 transgenes (green dots), with controls (gray dots). (d) Simultaneous inhibition of ORNs using Orco>ACR1 had little impact on attraction to PAM DAN activation. At 92 μW/mm², the ΔPI was +0.40 [95 CI +0.17, +0.60], P = 0.001. Transgene abbreviations: LexAop = LOP, Gal4 = G4. Data for all panels can be found in the corresponding folder on the Zenodo data repository (https://doi.org/10.5281/zenodo.7747425). KC, Kenyon cell; MB, mushroom body; ORN, olfactory receptor neuron; PAM, paired-anterior-medial.

https://doi.org/10.1371/journal.pbio.3002843.g003

We also performed experiments that showed that the olfactory receptor neurons (ORNs) are not required for PAM valence. Activating PAMs and simultaneously inhibiting ORNs still resulted in positive valence (Figs 3D and S2C). These results indicate that for PAM-mediated acute attraction, components of the olfactory circuit required for memory are dispensable for valence at both the primary sensory level (ORNs) and at higher levels of representation (KCs).

Dopamine receptors in the KCs are inessential for broad PAM attraction

Dopamine-receptor function in the KCs is required for olfactory learning [10,13,54]. Targeted Dop1R1 knockdown in the KCs almost entirely ablates both short-term and long-term learning, while Dop1R2 knockdown has an impact only on long-term learning, and Dop2R knockdown has no effect on either [75]. To determine the dopamine-receptor KC requirements for R58E02-mediated attraction, we knocked down Dop1R1, Dop1R2, and Dop2R in KCs using RNAi transgenes [76]. In a R58E02-LexA > lexAop-Chr background, individual knockdowns of Dop1R1 and Dop1R2 each caused only minor reductions in R58E02-mediated valence, while the knockdown of Dop2R resulted in a moderate reduction in valence (–41%) (Fig 4A–4C). Along with the evidence that KC activity itself is not required, these data showing different receptor dependencies verifies that acute PAM valence and olfactory learning are mediated by distinct mechanisms. Note that the partial dependence of valence on Dop2R contrasts with our finding that acute optogenetic inhibition of KC activity has no effect on R58E02-mediated valence, which could be explained by long-term changes as a result of chronic Dop2R knockdown.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. For driving DAN-mediated attraction, dopamine is a partial contributor.

(a) Schematic of the use of MB247-Gal4 to knockdown receptor expression by RNAi in the R58E02-LexA>lexAop-Chr optogenetic background. (b, c) Knocking down Dop1R1, Dop1R2, and Dop2R in KCs had minor effects on R58E02-LexA>lexAop-Chr light attraction. The gray ribbon indicates the control-valence 95% confidence interval. Experiments used 72 μW/mm² red light. (d) Schematic for the use of R58E02-Gal4 to simultaneously express Chr and knockdown TH expression. (e) Immunohistochemistry of the PAM DAN cluster stained with α-TH (red) and α-YFP (green) in flies expressing Chr-YFP in R58E02 cells. Yellow rings indicate the co-localization of α-TH and α-YFP signals in cells in the PAM cell-body cluster at 3 optical slices. (f) Immunohistochemistry images of the DANs with TH-RNAi co-expression, showing that cells with an α-YFP signal (R58E02 cells) have a greatly lower α-TH signal. (g, h) Knocking down TH expression with TH-RNAi has a moderate effect on R58E02 valence across 4 intensities. For example, at 70 μW/mm² the valence is +0.79 ΔPI in R58E02>Chr flies (g) and is reduced to +0.59 ΔPI in flies carrying the UAS-TH-RNAi knockdown transgene (h). (i) Averaging summary of the effects of reducing dopamine on R58E02-mediated valence with either gene knockdown (UAS-TH-RNAi, with or without UAS-Dicer) or a chemical inhibitor of TH activity (3-Iodo-L-tyrosine, 3IY). Each dot represents the percentage effect size of light intensity in an experiment (i.e., the R58E02>Chr; TH-RNAi experiment was replicated 3 times). Across all 3 intensities in 5 experiments, dopamine depletion resulted in an average ~46% reduction in valence. The vertical line indicates the 95% confidence interval. Data for all panels can be found in the corresponding folder on the Zenodo data repository (https://doi.org/10.5281/zenodo.7747425). KC, Kenyon cell; PAM, paired-anterior-medial.

https://doi.org/10.1371/journal.pbio.3002843.g004

Dopamine has a partial role in R58E02 valence

As PAMs are dopaminergic, we expected that depletion of dopamine from the PAMs would result in a loss of most if not all valence. We tested this hypothesis by depleting dopamine in DANs by several methods. First, we used RNAi against tyrosine hydroxylase (TH), an essential enzyme for dopamine synthesis [77] and encoded in flies by the TH gene (also referred to as pale). Compared to flies with intact TH expression (Fig 4E), flies with reduced TH in the R58E02 DANs exhibited only a modest valence reduction (Fig 4F). This partial valence reduction was observed in the 3 higher light intensities and across 3 replications of this experiment (Figs 4E and 4F and S3A–S3D). To confirm that TH knockdown was effective, we performed immunohistochemical staining of the DANs in R58E02 > TH-RNAi flies, which showed that TH expression was markedly reduced (Fig 4G and 4H and S4 and S5 Movies).

Second, we aimed to amplify the RNAi transgene’s efficacy with the simultaneous overexpression of Dicer2 endonuclease [76]; this approach resulted in overall valence (e.g., at 70 μW/mm²) that was not substantially different from the RNAi alone, i.e., a partial reduction (Figs 4I and S3E). Third and finally, we depleted dopamine systemically by feeding flies 3-iodotyrosine (3-IY), a competitive inhibitor of TH [78,79]. This pharmacological intervention also resulted in R58E02>Chr valence exhibiting only a partial reduction (S3F and S3G Fig). To gain a summary estimate of the overall effect of removing dopamine from the PAMs on valence, we averaged the results across all 3 DA-depleting interventions (TH-RNAi alone, TH-RNAi with Dicer2, and 3-IY). Averaging the 3 interventions in each of the 3 light intensities indicated that dopamine depletion reduces valence to 44%, 58%, and 60% of control R58E02-mediated valence in 5, 22, and 70 mW/mm² illumination, respectively (Fig 4I). Averaging across the intensities gives 54% of control levels, i.e., an overall reduction of –46% (Fig 4I). Thus, over a range of light conditions and with several different lesions, dopamine appears to mediate roughly half of R58E02 valence.

Broad DAN attraction relies on glutamate and octopamine

Single-cell RNA sequencing data have shown that PAMs express neurotransmitter-related genes aside from those pertaining to dopamine [27]. As dopamine did not fully account for R58E02 valence, we hypothesized that the dopamine-independent component of R58E02 valence might depend on other neurotransmitters. We knocked down genes involved in the synthesis or vesicular transport of 4 other transmitters: vGlut for glutamate, Gad1 for GABA, vAChT for acetylcholine, and Tβh for octopamine. Of these, knockdown of vGlut and Tβh in the R58E02 cells produced effects on valence that were comparable in magnitude to the dopamine reduction: –48% reduction for vGlut and –38% for Tbh (Fig 5A).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Large reductions in PAM valence require knockdown of multiple neurotransmitters.

(a) A knockdown screen for neurotransmitters that contribute to R58E02 valence. R58E02>Chr flies were crossed with RNAi transgenes targeting factors required for 5 transmitters: TH (dopamine, replicating the prior experiment), vGlut (glutamate), GAD1 (GABA), vAchT (acetylcholine), and TβH (octopamine). The vGlut and TβH knockdowns showed a reduction in valence comparable to the TH knockdown. Simultaneous knockdown of TH with either TβH or vGlut resulted in a further reduction in the effect size. (b) A combinatorial approach using 3-IY to systematically deplete dopamine along with RNAi-mediated knockdown of vGlut and TβH reveals a progressive reduction of valence with depletion of each neurotransmitter, with valence being virtually depleted when 3-IY, vGlut-RNAi and TβH-RNAi were simultaneously present (ΔPI = +0.10 [95% CI −0.16, +0.37]). Data for all panels can be found in the corresponding folder on the Zenodo data repository (https://doi.org/10.5281/zenodo.7747425). PAM, paired-anterior-medial; TH, tyrosine hydroxylase.

https://doi.org/10.1371/journal.pbio.3002843.g005

Because the knockdowns of TH, vGlut, and Tβh each resulted in a partial reduction of valence, we hypothesized that knocking down these genes in two-way combinations would elicit additive deficits. Indeed, the combined knockdown of TH and vGlut via the co-expression of both inhibitory RNAs produced a further reduction in valence, as did the simultaneous knockdown of TH and Tβh (Fig 5A); however, neither two-way combination succeeded in ablating valence completely. We thus hypothesized that all 3 neurotransmitters needed to be depleted to ablate valence. As the co-expression of all 3 RNAis with Chr proved to be technically prohibitive, we instead used the competitive inhibitor 3-IY to systematically reduce dopamine in conjunction with RNAi knockdowns. [Note that pseudo-replicate knockdowns of vGlut or Tbh, using different food, yielded somewhat different valence scores (Fig 5A and 5b).] Simultaneous depletion of 2 transmitters—by combination of 3-IY with vGlut-RNAi, 3-IY and Tβh-RNAi, or both vGlut and Tβh-RNAis—showed further reductions in valence (Fig 5B). Finally, depletion of all 3 transmitters using 3-IY, vGlut-RNAi and Tβh-RNAi virtually ablated valence (Fig 5B). These results are consistent with the idea that R58E02 valence is reliant on a combination of dopamine, glutamate, and octopamine.

PAM-β valence is partially dependent on dopamine

The opposing valence of R58E02 and R15A04 suggests valence heterogeneity in different subsets of PAM DANs. Indeed, numerous studies have found that valence-related behaviors like food-seeking, courtship, sleep, and appetitive memory are dependent on different MB sub-compartments and specific DAN subsets [27,33,42,51,53,59,80–83]. To identify the PAM types that drive valence, we screened 20 split-Gal4 lines [49,51] and identified several with valence, both negative and positive (Fig 6A and 6B). Of these lines, we focused on MB213B, as it had the strongest positive valence (Fig 6B). This line expresses in the PAM-β1 and PAM-β2 types (PAM4 and PAM10, respectively) with minor expression in the PAM11 (PAM-α1) cells [49] (Fig 6C and S2 Table and S8 Movie). As valence in R58E02>Chr flies is only partially dependent on dopamine, we interrogated MB213B>Chr dopamine dependency. Knocking down TH in PAM-β cells produced variable results (S5A–S5C Fig); to resolve these replicate differences, we used meta-analytic averaging to calculate weighted mean-difference estimates at the 22 and 70 μW/mm² light intensities [84]. This averaging showed that when compared to non-RNAi flies (Fig 6D), knockdown of TH elicited robust (though incomplete) valence reduction: –68% and –65%, respectively (Fig 6E and 6F). These effects are larger than the TH-RNAi knockdown effects for R58E02>Chr, which were –42% and –40% for the same light intensities (Fig 5F).

PAM-β valence is partially dependent on glutamate

PAM-β cells have been shown via single-cell RNA-seq to express several other neurotransmitter-related genes, including vGlut, vAChT, and Gad1 [27]. As glutamate was required for R58E02 valence (Fig 5) and is known as a co-transmitter in dopaminergic cells [85], we examined the effect of vGlut knockdown in the specific PAM-β driver. The MB213B > vGlut-RNAi; Chr flies displayed valence that was reduced but—like the TH knockdown—not completely abolished (Fig 6G). Thus, it appears that for valence mediated by the MB213B driver, glutamate and dopamine transmission each make partial contributions. These results show that as for R58E02 valence, at least 1 specific PAM subset similarly drives acute valence using multiple transmitters.

Silencing activity in β-lobe DANs drives a negative valence response

We next asked whether flies would respond to inhibition of the MB213B cells [56]. To drive inhibition, we expressed ACR1 in the MB213B cells [73,74]. The MB213B>ACR1 flies avoided green light, i.e., showed negative valence. That ACR1 actuation of the PAM-β cells has a behavioral effect indicates that at least a subset of these cells was active in control flies during the experiment (Fig 6H).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Valence mediated by dopaminergic PAM-β neurons is dependent on both dopamine and glutamate.

(a, b) An optogenetic activation screen of 22 PAM-DAN lines identified MB213B as the specifically expressing line with the strongest positive valence. Light preference was tested with 72 μW/mm² red light. The table shows the PAM cell types in which each driver expresses; “M” denotes multiple cell types; see S2 Table for further details. (c) The expression pattern of Chr-YFP with driver MB213B, showing projections to both zones of the β lobe (zones β1 and β2). (d) Replication of the MB213B>Chr screen experiment confirmed that these flies are attracted to optogenetic light at the 2 highest intensities. (e, f) Meta-analysis of 3 replicates of MB213B > TH-RNAi; Chr yielded weighted ΔPI values of +0.17 at 22 μW/mm² and +0.25 at 70 μW/mm² (orange curves). (g) Expressing vGlut-RNAi with the MB213B driver similarly resulted in reduced (but not ablated) valence. (h) MB213>ACR1 flies avoided the green-illuminated area. Data for all panels can be found in the corresponding folder on the Zenodo data repository (https://doi.org/10.5281/zenodo.7747425). PAM, paired-anterior-medial.

https://doi.org/10.1371/journal.pbio.3002843.g006

DANs and MBONs drive valence with different locomotion patterns

In olfactory learning, appetitive DANs are thought to be paired with aversive MBONs [86,87]. The concurrent release of dopamine with odor experience then causes weakening of the connections between the KCs and the MBONs, thus reducing the aversive response to the presented odor and driving approach [52,53,82,88]. Many MBONs themselves drive acute valence and receive synaptic input from DANs [50,51]. Considering that MBON types were shown to drive acute aversion, we hypothesized that optogenetic PAM activity might be inhibiting these aversive MBONs, and thereby rendering these output neurons the direct mediators of acute PAM valence.

To address this hypothesis, we screened a panel of MBON drivers for valence (Figs 7A and S6A–S6D). We found that one line, VT999036, had the strongest valence response. This line drives expression in MBONs that project to the γ lobe, termed MBON-γ1γ2 and MBON-γ4γ5 cells, also known as types MBON20 and MBON21, respectively [23,49] (Figs 7B, S6E, and S6F and S6 Movie and S2 Table). Activation of VT999036 caused strong aversive valence, with flies turning away from the light at light–dark boundaries (Fig 7C and 7D and S7 Movie). The “dark preference” of VT999036>Chr flies had little correlation with light–dark speed differences, but was correlated with a choice index, i.e., fly trajectories at the boundary (Fig 7F and 7G). These locomotion patterns stand in contrast with R58E02>Chr and R15A04>Chr flies, in which valence correlated with optogenetic speed differences (Figs 1G–1H and S1C–S1F).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. MBONs drive valence via choice effects, not speed effects.

(a) A screen of optogenetic Chr valence in 15 MBON-related lines (split-Gal4 and Gal4 drivers). Orange markers show the valence scores (black dots) and distributions (curves) of each cross, comparing test flies with controls. See S1 Table for effect sizes. The matrix key shows driver identifiers in the top row, and MBON cell types in which each driver expresses; M denotes multiple cell types. See S2 Table for further details. (b) Schematic of VT999036 projections to MBON synaptic-zone subsets. VT999036 drives expression in 2 MBON types, MBON20 and MBON21 [49]. (c) VT999036>Chr flies avoided opto-activation at the 2 highest illumination intensities (22 and 70 μW/mm²). The valence curve was produced using the same data from the screen summary. (d) When VT999036>Chr flies move through the choice point, they tend to turn away from the light. (e) In VT999036>Chr flies, a relationship between preference and speed ratios was absent. (f) In VT999036>Chr flies, choice index and preference were related. (g) Summary of regressions of DAN and MBON driver valence. Coefficients of determination for DAN lines (blue dots) and MBON lines (orange dots) are shown for 4 locomotor metrics as compared to valence (ΔPreference). The 4 metrics are Δchoice, Δspeed ratio, and the effect sizes of the dark-light and light-dark choice-point exit probabilities (ΔPEDL and ΔPELD, respectively). (h) Combined schematic of combined R58E02 and VT999036 projections to the MB. The 2 expression patterns overlap in the γ4 and γ5 regions, corresponding to MBON21. (i) Knocking down Dop1R1, Dop1R2, and Dop2R in the MBONs of VT999036-Gal4 had modest effects on R58E02>Chr light preference. The gray ribbon indicates the control-valence 95% confidence interval. Data from 70 μW/mm² red light. Data for all panels can be found in the corresponding folder on the Zenodo data repository (https://doi.org/10.5281/zenodo.7747425). MBON, mushroom body output neuron.

https://doi.org/10.1371/journal.pbio.3002843.g007

These observations led us to ask whether the speed/choice dissociation observed in the R58E02 and VT999036 lines was part of a general trend for DAN and MBON lines. We analyzed metrics for all lines in the MBON and DAN screens (Figs 6A, 6B, S4A, S4B, and S6A–S6D) and used them in regressions of the 2 screens’ valence and speed-ratio effect sizes. This analysis indicated that PAM-mediated valences were weakly determined by choice and strongly determined by speed differences (Figs 7G and S7A–S7D). By contrast, MBON-mediated valence was weakly associated with speed and more strongly determined by choice and choice-point exit probabilities (Figs 7G and S7E–S7H). Because lines for both neuronal categories can drive attraction and avoidance, this difference is not easily explained by valence polarity. Nevertheless, these data suggest that MBON and DAN activities have differential effects on 2 navigational properties: turning and speed.

Dopamine-receptor knockdowns in strongly valent MBONs produce modest reductions in broad-PAM valence

As VT999036 produced the strongest aversive valence in our MBON screen, we hypothesized that valence caused by broad PAM activation might be mediated via MBONs marked by VT999036, given the hypothesized pairing of appetitive DANs with aversive MBONs. The projections of R58E02 and VT999036 overlap in the γ4 and γ5 synaptic zones, corresponding to the MBON21 cells (Fig 7H). Compared to the control effect size of ΔPI = +0.53 observed in R58E02-LexA>LexAOp-Chrimson flies, individual knockdowns of Dop1R1 and Dop1R2 in VT999036 each produced a modest reduction in valence: approximately –30% and –21%, respectively (Fig 7I). Knocking down Dop2R resulted in only a –7% change in valence (Fig 7I). Notably, the summed effect of the Dop1R1 and Dop1R2 knockdowns (–51%) is comparable to the approximately –46% average reduction of valence observed earlier with dopamine-depleting interventions (Fig 4I). These results suggest that the 2 DopR1-like receptors in the MBON21 cells contribute to R58E02 valence.

Co-zonal PAM and MBON valences are not coherently related

Broad-PAM R58E02 attraction is linked to VT999036 aversion (as shown by targeted receptor knockdowns; Fig 7I) and both project to the γ4 and γ5 synaptic zones; these observations lend credence to the idea that acute positive DAN valence is mediated by inhibiting their postsynaptic aversive MBONs. To explore if this pattern extended beyond this instance, we explored whether PAM:MBON driver pairs with shared MB synaptic zones have an inverse valence relationship. Interestingly, the valence scores (ΔPIs) of co-zonal PAM and MBON did not show a consistent trend, contradicting the idea that co-zonal PAMs and MBONs have systematically opposing valence effects (S8 Fig). For example, MB213B activity is consistently attractive, and these cells project to both of the 2 β zones (Fig 6A, 6B, and 6D) [49]. Yet, 2 β-zone-projecting MBON drivers have variously positive and negative valence: MB399C (MBON02, zones β2 and β’2, ΔPI = +0.25) drives attraction; and MB434B (MBON05 and MBON06, zones β1 and γ4, ΔPI = –0.27) drives avoidance (S8 Fig) [49]. It therefore seems that acute valence resulting from PAM stimulation does not necessarily arise from the inhibition of co-zonal aversive MBONs, or at least there is no clear evidence supporting a simple, general PAM:MBON valence relationship.

Differences between appetitive learning and acute DAN valence

In this study, we provide evidence that a dopaminergic system known to instruct learning also drives acute valence. The PAM DANs instruct appetitive odor learning [9,10], such that subsequent encounters with the same odor will elicit an increased approach behavior. Three features distinguish DAN-mediated olfactory learning from DAN-evoked acute valence. First, unlike classical Pavlovian learning—which requires an association between dopamine and a sensory stimulus—acute DAN valence occurs in an experimental environment that is otherwise largely featureless, such that the optogenetic illumination is the only salient sensory stimulus. Thus, the strong elicited valence observed is consistent with non-associative DAN functions that operate independently of external stimuli [44,53]. Second, while KCs are critical to DAN-mediated sensory associative learning [71,72], activated DAN acute valence has little-to-no reliance on KCs. Third, while learning has a strong dependency on Dop1R1 receptors in the KCs [13,22,75], R58E02 PAM valence does not have critical dependencies on either dopamine-receptor function in KCs or PAM dopamine synthesis. Thus, DAN-dependent learning and DAN valence seem to be distinct processes that act through different circuits and signaling systems.

Differences between DAN and MBON valence

Paired-posterior-lateral 1 (PPL1) DANs can drive optogenetic valence behavior [50]. Findings from our study exclude KCs as the downstream neurons through which PAM DANs affect valence. This finding suggests that MBONs are the possible mediators, as this class of neurons are major output cells of the MB [49,86,89,90] and receive synaptic input from DANs [50].

Under the assumption that DAN valence is mediated by MBONs, we explored some features of valent locomotion as driven by the 2 cell types. The DAN lines primarily influenced optogenetic light preference by affecting walking speed, while the MBONs, as previously shown [51], had their primary effect by changing trajectories at the light–dark interface, a trend that was independent of valence polarity. There are at least 4 possible explanations for this difference. First, some of the driver lines (e.g., VT999036) capture cells without MB projections, and these could (in some cases) contribute to valence. Second, DANs may have a symmetrical influence on walking, while MBONs have an asymmetrical effect on walking, perhaps via an algorithm similar to those in Braitenberg vehicles [91]. Third, assuming that DAN valence is mediated by MBON activity, it might be the case that MBONs have speed effects when quiescent, but drive turning when highly active, e.g., through specific downstream circuits with a distinct responsiveness to MBON activity. Fourth, the MBON screen might not have included the downstream cells responsible for DAN valence, i.e., other downstream circuits mediate DAN effects. On this fourth point: the drivers that express in co-zonal DAN and MBON types do not necessarily specify cellular subtypes that share synapses [60,92].

The specific function of DAN→MBON synapses in acute opto-valence remains unclear. Some MBON subtypes can modulate locomotor dynamics like walking speed and turning [51]; however, the extent to which DANs drive valent locomotion through MBONs is not known. Moreover, the idea that DAN valences are exclusively mediated by DAN→MBON synapses in the MB is likely to be incorrect. A connectomic analysis has shown that PAMs send axons to other neuropils [92], suggesting that at least some of the valence effects could be mediated through PAM signaling to non-MB areas.

Appetitive DANs have non-associative functions

Overall, the speed/turning dissociation is hard to explain with current information. One possibility is that the visual associative pathway could send a sensory input to MBONs, thus providing a signal that involves the PAM DANs, but does not require the KCs. We do not consider this likely. While green light (like the light used to actuate ACR1) can indeed function as a visual conditioned stimulus, inhibition of KCs by MB247-Gal4 abolishes visual learning [31]. In our hands, inhibition using the identical MB247 driver left valence mostly intact, so the possibility of valence being equivalent to a conditioned visual response seems unlikely. Moreover, while olfactory-learning scores and R58E02 optogenetic valence are of comparable magnitudes [93], visual-learning scores are typically <50% of this study’s valence effects [31,94]. These observations indicate that valence and visual learning are distinct.

The experimental results cast DAN-driven acute valence and olfactory sensory learning as separable processes, wherein DANs perform 2 distinct functions. In response to rewarding (or punishing) stimuli, PAM DANs seem to (1) write olfactory memories to Kandelian-type KC synapses for future reference; and (2) instruct immediate changes in locomotor behavior. Imaging of the MB has shown that DAN activity is closely connected with motor states and locomotion [44,53,95]. Inhibition experiments have revealed that DAN activities guide innate odor avoidance [96] and odor navigation [44]. Similarly, reward-related behaviors in mammals can be separated into consummatory, motivational, and learning components, of which the latter 2 are attributable to dopamine function [97]. Organizing parallel signals of a single reward circuit into distinct motivational and associative dopaminergic synapses could ensure coherence between valence (the present) and subsequent learned sensory responses (the future). From an evolutionary perspective, we speculate that the motor-function arm predates the associative arm [98], which was later inserted when the first sensory systems (taste and olfaction) developed.

DAN valence relies on glutamate and octopamine

Knocking down TH function reduced R58E02 valence by roughly half, suggesting the involvement of other neurotransmitters. Screening 4 other transmitter pathways implicated glutamate transport and octopamine synthesis; combined lesions showed that reduction of any 2 of TH, Vglut, and Tbh reduced R58E02 valence by roughly two-thirds, while reduction of all 3 resulted in essentially no optogenetic valence. With the narrow PAM driver MB213B, knockdowns revealed that dopamine and glutamate were required by the PAM-β cells for normal valence. Previous single-cell RNA sequencing studies revealed the co-expression of vGlut, Gad1, Tbh, and vAChT in subpopulations of dopaminergic neurons [27,99]. Together, these findings implicate dopamine, glutamate, and sometimes octopamine as co-transmitters in eliciting PAM-DAN optogenetic valence.

These findings are in agreement with a growing body of literature demonstrating the presence and utility of multiple transmitters in dopamine neurons and MB KCs. Recent work in PAMs has demonstrated roles for co-transmitters: memory updating requires the co-release of nitric oxide [27], and appetitive memory formation is dampened by the co-release of GABA [100]. In PPL DANs, the dampening of aversive memory formation requires both GABA and glutamate [100]. Glutamate signaling from glia has also been implicated in aversive-memory formation, and octopamine-receptor knockdowns in KCs dampen aversive and appetitive-memory formation [101–104]. All of these prior studies focused on olfactory memory, whereas the experiments reported here demonstrate the non-associative functions of PAMs. More broadly, dopamine and glutamate co-transmission has been well-documented in mice, including in dopaminergic midbrain neurons, where dopamine and glutamate are released from distinct terminals [105–107]. Other neuromodulatory neurons also exhibit dual-transmission: some octopaminergic neurons in flies also transmit glutamate [108]. The use of multiple neurotransmitters thus seems to be a common mechanism by which neuromodulatory neurons exert their effects.

The molecular mechanisms by which dopamine, Vglut and Tbh act together to mediate valence are currently unknown. In Drosophila, synaptic vesicles at DAN terminals undergo hyper-acidification in response to neuronal activity; this process is driven by glutamate transport into these vesicles by Vglut, which in turn increases dopamine loading [85]. As a result, the effect of Vglut knockdown might be due to lower co-release of glutamate and/or reduced loading of dopamine into synaptic vesicles. As shown for some octopaminergic neurons that release octopamine and glutamate from separate termini [108], it is possible that dopamine and glutamate are also released from distinct PAM termini.

Normal R58E02 valence also requires Tbh. While octopamine is known to be important in both appetitive and aversive memory [103], octopamine co-transmission from PAMs is novel. How octopamine-dependent, PAM-driven behaviors differ from DA- and glutamate-dependent behaviors are interesting topics for future studies.

Technical note on interpreting knockdown data

Beyond the TH-RNAi immunostaining, we did not assess the molecular effects of the other RNAi manipulations, and it is likely that these RNAi transgenes produced partial knockdowns [76]. Previous work has shown that the Vglut-RNAi line used in this paper produces a ~70% mRNA knockdown, whereas the Gad1-RNAi line produces an ~82% protein knockdown [42,109]. The remaining lines (Tbh-RNAi and vAChT-RNAi) were sourced from the Transgenic RNAi Project: most of these lines produce knockdowns greater than 50% [110]. Nonetheless, that each of vGlut-RNAi and Tbh-RNAi produced a substantial reduction in R58E02 valence demonstrates that the corresponding neurotransmitters have a role in R58E02 valence. For vAChT-RNAi and Gad1-RNAi, neither produced appreciable changes to R58E02 valence, but we cannot discern whether this was due to poor RNAi efficiency or that those transmitters have no role. Even though the individual knockdown effect sizes could be underestimated, nevertheless, the sum of the TH, Vglut, and Tbh knockdown effect sizes is approximately 100%, and the combination lesions indicate that reducing all 3 ablates valence almost completely (Fig 5B).

Ongoing DAN activities shape behavior

That flies avoid silencing their PAM-β dopaminergic cells indicates that stimulus-independent DAN activities influence behavior. In the context of classical conditioning, dopamine is thought of as a transient, stimulus-evoked signal; this result indicates that some PAM neurons are active even in a chamber lacking odor, food, shock, or other salient stimuli. Physiological recordings show that, along with sucrose responses, PAM-γ activities correlate with motion and guide odor-tracking behavior, supporting the idea that PAM–DAN activities both respond to and steer locomotor behavior [44,53]. In our optogenetic preference screen, activity in various PAM DAN populations were rewarding, aversive or showed little preference effect, indicating that there is a diversity of functions between different DAN types, consistent with findings for learning and memory [9–11,14,19–27]. The bidirectionality of the attractive and aversive effects of increasing and decreasing activity in this dopaminergic system is similar to valence responses to activation and inhibition of dopaminergic cells in the mammalian ventral tegmental area [8,111,112] and is reminiscent of the increases and decreases in activity in that area that occur during positive and negative reward prediction errors, respectively [113]. Whole-brain imaging in the nematode has shown that global brain dynamics track closely with locomotion [114], suggesting that overarching brain function is to coordinate motor function. That the MB DANs drive preference-related locomotion suggest that their valence roles have 2 timescales: informing responses to future experiences and steering current behavior.

Discordance between PAM and MBON valence implies complex interactions

The knockdown of Dop1R1 and Dop1R2 in VT999036 (MBON20, MBON21) produced reductions of valence like that of presynaptic dopamine depletion from the PAMs. Given that VT999036 is highly aversive, it is tempting to hypothesize that PAM-mediated attractive valence acts by the inhibition of aversive MBONs, much in the way that memory is thought to be expressed [52,53,88]. Given the effects of the dopamine-receptor knockdowns, this paradigm might be the case for VT999036. When we examined the relationship between PAM and MBON valences across a range of drivers, however, it became clear that this was not a general pattern; that is to say, there was no consistent relationship between a PAM’s valence and the corresponding valence from its co-zonal MBONs.

The valence resulting from stimulation of a PAM subset seems to not arise from co-zonal MBONs, at least not exclusively. For memory, MBONs provide feedback to PAMs, not only via recurrent connections in the same zone but between zones as well [50,92,115–117]. Furthermore, MBONs do not have a one-to-one relationship with PAMs; there are more MBON cell types than PAM cell types and many MBONs sample information from multiple zones of the MB [49,50]. As such, when a population of PAMs is activated, it likely induces a complex network state that results in valence-related behaviors that may be mediated by an MBON outside of the PAM-targeted zone, or even by the overall effect of broader brain activity. Moreover, we cannot exclude the possibility that valence induced by stimulating different PAM subsets may work via diverse synaptic mechanisms, for example, inhibiting aversive MBONs versus activating attractive MBONs. Further work is required to elucidate the circuit mechanisms required for PAM stimulation-driven valence, including imaging.

Possible explanations for valence variability

Of this study’s many limitations, of note is the sometimes pronounced variability in results between experimental iterations. For example, R58E02 > Chr flies showed variable valence when tested by different experimenters at different times (Figs 1E and 4E). Similarly, MB213B > Chr; TH-RNAi flies showed variable valence (S5A–S5C). One likely variance contributor is sampling error, a routine issue in behavioral data [118]. For several dopamine-depletion experiments, we adopted replication and meta-analysis to mitigate sampling error and estimate effect sizes with greater precision [93]. A second possibility is that neurotransmitter dependencies might vary between iterations due to uncontrolled changes during development. For example, the loss of dopamine may sometimes lead to developmental compensation, such as neurotransmitter switching or circuit adjustments. In mammals, transmitter switching in dopaminergic cells can occur as a result of stimuli such as odor or light stress [119–121]. In Drosophila, the expression of vGlut in DANs increases as dopamine is depleted either pharmacologically or due to aging [122]. It is possible that, as RNAi depletes TH, the PAM-β cells switch to glutamate as a substitute transmitter (but see the notes on effect-size surfeit above). A third possible explanation of the variability is that due to one or more uncontrolled variables, some neurons’ valence is susceptible to internal state. Imaging has shown that PAM-γ cells display activities that vary depending on state, such as starvation and walking [53,56]. If, for example, PAM-β baseline activity is high, it might be expected that the valence due to further activation will be small. Conversely, in a low-baseline activity state, stimulation of these cells would be expected to result in a large effect. Recent studies have described high variability between different optogenetic valence assays [61,64], suggesting that optogenetic valence is not generalizable between different behavioral tasks. As different sensory modalities and/or associative circuits may be employed for different tasks, this finding is not unexpected. Like all cognitive constructs, the broader property of so-called “valence” will benefit from systematic investigations by multiple groups across behavioral paradigms.

Valence modulation beyond PAM-DAN cells

The driver R58E04-Gal4 mainly expresses in PAM-DAN cells, but it also shows expression in the optic lobes [10,24], raising the possibility that cells outside the PAM-DANs are responsible for some part of the valence. While this possibility could explain the residual valence after inducing dopamine lesions, we consider it to be unlikely. First, the cells in the optic lobe have been reported to be glia: “[R58E02] strongly labels the PAM cluster neurons and glial cells in the optic lobes with little expression elsewhere” [10]. Second, our screen of the PAM–DAN split-Gal4s shows that several of them give positive valence, generalizing the phenomenon beyond the R58E02 driver itself. For example, the MB213B split-Gal4 driver (which does not have any detectable expression outside of the PAM-DANs) has a positive valence that—like R58E02—is only partially reduced by dopamine knockdown. Third, the combination-knockdown experiments indicate that valence can be attributed to multiple transmitters, so an additional cause seems redundant. These observations support the idea that the valence observed is attributable to PAM-DAN activity.

Questions beyond the study’s scope

Besides the questions we tackled in this study, there are other important areas worth considering. These include studying how the optogenetic activation of DANs and MBONs can give rise to distinct locomotor patterns. The role of co-transmission of different neurotransmitters in modulating these behavioral outcomes is another key aspect to explore. Furthermore, why certain DANs drive attraction while a subset of the larger DAN population induces aversion remains an intriguing question. Additionally, in the absence of olfactory stimuli, investigating the neurophysiological signal that dopamine acts upon to modulate walking speed is of interest. These questions extend beyond the current study and offer valuable avenues for future research.

Overall, our findings reveal that the PAM system, in parallel with its associative functions, can instruct acute valence behavior using distinct mechanisms. Associative learning shows a strong dependence on odor stimuli, KCs, and Dop1R1; by contrast, acute PAM valence does not have strong requirements for a salient olfactory stimulus, Orco-cell function, KC activity, or Dop1R1 reception in the KCs. The results reveal that the PAM neurons utilize multiple neurotransmitters such that dopamine is not acting alone, with glutamate and octopamine also having substantial roles in acute valence. The locomotor features of PAM valence seem distinct from MBON valence: while PAM valence is primarily mediated by a reduction in speed, MBON valence is mediated by a change in turning behavior. These findings provide insight into the diverse roles and mechanisms of PAM neurons in mediating both associative and non-associative valence responses, highlighting the complexity of this neuromodulatory circuit in regulating behavior.

Fly strains

Flies were cultured on a standard fly medium [123] at 25°C and 60% humidity in a 12 h light: 12 h dark cycle. Wild-type flies were a cantonized w¹¹¹⁸ line. The DAN and MBON split-Gal4 lines described in [49] were a gift from Gerry Rubin (Howard Hughes Medical Institute), except for VT041043-Gal4 [109] and VT49126-Gal4 [23], which were obtained from the Vienna Drosophila Resource Center (VDRC). VT999036 was a gift from Barry Dickson (Howard Hughes Medical Institute). The Gal4 transgenic lines were obtained from the Bloomington Drosophila Stock Center (BDSC) and included: R58E02-Gal4 [10], R15A04-Gal4 [68], 20x-UAS-CsChrimson [65], 13X-LexAOp2-GtACR1 [124], MB247-Gal4 [125], MB247-LexA [126]. R53C03-Gal4 [49], R76B09-Gal4 [23], R52G04-Gal4 [51], NP5272-Gal4 [16] were obtained from the Kyoto Stock Center (DGRC). The RNAi lines used were: Dop1R1 (KK 107058), Dop1R2 (KK 105324), Dop2R (GD 11471), vGlut (KK104324), and TH (KK 108879), obtained from VDRC; as well as Gad1 (BDSC_51794), Tβh (BDSC_76062), and vAChT (BDSC_80435) from the Transgenic RNAi Project [110]. Supporting table in S1 Table provides detailed descriptions of genotypes shown in each figure.

Transgenic animal preparation

Gal4, UAS-CsChrimson, and UAS-ACR1 crosses were maintained at 25°C and 60% humidity, in darkness. Groups of 25 newly eclosed flies were separated into vials for 2 to 3 days (in the dark at 25°C) before behavioral phenotyping. Control flies were generated by crossing the driver or responder line with a wild-type w¹¹¹⁸ strain (originally bought from VDRC), and raising the progeny under identical regimes to those used for the test flies. A stock solution of all-trans-retinal was prepared in 95% ethanol (w/v) and mixed with warm, liquefied fly food. Each vial was covered with aluminum foil and incubated at 25°C in the dark. Before optogenetic experiments, 3- to 5-day-old male flies were fed 0.5 mM all-trans-retinal (Sigma) for 2 to 3 days at 25°C in the dark.

Drug treatment

Male flies (3 to 5 days old) were placed on 1% agar containing 5% sucrose, 10 mg/mL 3-iodo-L-tyrosine (3-IY, Sigma), and 0.5 mM all-trans-retinal for 2 to 3 days at 25°C in the dark prior to behavioral testing. Control flies were fed on the same food but with 3-IY omitted.

Immunohistochemistry

Immunohistochemistry was performed as previously described [74]. Briefly, brains were dissected in phosphate-buffered saline (PBS) and fixed in PBS with 4% paraformaldehyde (Electron Microscopy Sciences) for 20 min. Samples were washed 3 times with PBT (PBS + 1% Triton X-100) and blocked with 5% normal goat serum for 1 h. Samples were then incubated with primary antibodies overnight at 4°C. After 3 additional washes with PBT, samples were incubated with secondary antibodies overnight at 4°C. The following primary and secondary antibodies were used: mouse α-DLG1 (4F3 α-DISCS LARGE 1, Developmental Studies Hybridoma Bank, 1:200 dilution), rabbit α-TH (AB152, Millipore, 1:200 dilution), chicken α-GFP (Abcam, ab13970, 1:1,000), Alexa Fluor 488 rabbit α-GFP-IgG (A-21311, Invitrogen, 1:200), Alexa Fluor 568 goat anti-mouse (A-11004, Invitrogen, 1:200), Alexa Fluor goat anti-chicken 488 (A-11039, Invitrogen, 1:200), and Alexa Fluor goat anti-rabbit 568 (A-11036, Invitrogen, 1:200).

Confocal laser microscopy and neuroanatomy

Confocal images were acquired under a Zeiss LSM 710 microscope at a z-step of 0.5 μm using 20×, 40×, or 63× objectives. Images were analyzed using ImageJ software. Black and white images are a maximum projection intensity (MIP) of the green channel. The stacks were visualized and analyzed with the FIJI distribution (www.fiji.sc) of ImageJ (NIH). Outlines of α-Dlg1 expression in the mushroom body were traced with Adobe Illustrator. Projection patterns and zonal identity were assigned as previously described [49]. When not verified by microscopy, cell types and projection patterns were classified by review of published reports (S2 Table) [10,23,24,49,59,109].

Optogenetic response assay

Behavior experiments were performed as previously described [74]. Each behavioral arena was cut with 55 × 4 mm stadium/discorectangle geometry; 15 such arenas were cut from 1.5 mm-thick transparent acrylic. During the behavioral assay, arenas were covered with a transparent acrylic lid. As previously described [3], flies were anesthetized on ice before loading into each chamber in the dark. The arena multiplex was kept under infrared (IR) light at 25°C for 2 to 3 min before starting the assay. Flies were aroused by shaking the arenas just before starting the experiment. All behaviors were recorded under IR light. The multiplex was illuminated with red or green light from a mini-projector positioned above the arena (Optoma ML750). For CsChrimson experiments, flies were illuminated with 4 red-light intensities: 1.3, 5, 22, and 70 μW/mm². For ACR1 experiments, the flies were illuminated with 4 green-light intensities: 1.6, 7, 28, and 92 μW/mm². The colored light intensity was varied by changing the level of the respective RGB component of the projected color. For each experiment, the arenas were illuminated for 60 s with equal-sized quadrants to produce a banded light-dark-light-dark pattern.

Video tracking

The behavior arena was imaged with a monochrome camera (Guppy-046 B, Allied Vision) with 2 IR long-pass filters in series (IR Filter IR850, Green.L). Videos were processed in real time with CRITTA software written in LabView [74]. The x-y coordinates of each fly’s head were individually tracked (at 25 frames per second) using CRITTA’s tracking feature. CRITTA was also used to control the timing, hue and intensity of the illumination, and to count the number of flies in each quadrant for each video frame. The light borders were identified and calibrated using a function of the CRITTA plugin, which illuminates quadrants at low intensity and captures an image of the arenas (with camera IR filters removed). The plugin software calculates the horizontal intensity profile of each arena and finds the center of each light–dark boundary using an edge-detection algorithm. The light-border drift between presented experiments was 330 μm (95CI 230 μm; 430 μm). Between the light and dark regions was a light gradient that was a mean 670 μm wide with a range of 420 to 1,040 μm. This gradient was measured from 45 images of boundaries from 15 chambers and scored as all the pixels falling between high (light-on) and low-intensity light regions.

Olfactory conditioning

Conditioning was performed as previously described [74,127]. Briefly, each behavior chamber was 50 mm long, 5 mm wide, and 1.3 mm high; the floor and ceiling of each chamber were composed of transparent shock boards made from indium tin oxide electrodes printed on glass (Visiontek UK). Odorized air was pumped into the ends of each arena at 500 ml/min. The odors were 4-methylcyclohexanol (MCH) at 9 parts per million (ppm) and 3-octanol (OCT) at 6 ppm, as measured with a photoionization detector (RAE systems, ppbRAE3000). The air exited the chamber via 2 vents located in the middle, creating 2 odor partitions in the conditioning area. Each experiment was performed with 4 to 6 flies. For opto-conditioning, flies were presented with either OCT or MCH odor paired with green light (515 nm, 28 μW/mm²), followed by another odor without visible light (IR light only). During shock conditioning, the presentation of either OCT or MCH was coupled with 12 electric shocks of 1 s duration at 60V [14]. Conditioned-odor preference (memory) was tested by the presentation of both odors, one from each side. For each of the 2 odors, a half performance index (PI) was calculated according to the fly position coordinates during the last 30 s of each testing phase; for each iteration, data from odor pairs were averaged to obtain a full PI [128].

Preference and speed analysis

Custom Python scripts were used for data processing, analysis, and visualization. The scripts integrated several routines from NumPy, pandas, matplotlib, and seaborn. For every fly, the x-y coordinates of the head location (recorded at 25 frames per second) underwent rolling-window smoothing, using a centered 1 s-wide triangular window. The following metrics were obtained (for every fly) for the last 30 s of each test session:

While all flies could be assigned a PI, the log₂ Speed Ratio (LSR) could be computed only for flies that moved in both light and dark regions during the illumination epoch. Flies that remained stationary for the entire illumination epoch, or remained in only the light or dark zones, were excluded from the speed ratio calculation. Flies that started and remained in either the dark or light zone throughout an epoch, but still moved within the zone, were assigned an extreme PI (–1.0 or +1.0, respectively).

Choice-zone trajectory analysis

A choice trajectory was defined as any transit in and out of a choice zone defined to extend 3 mm in either direction from all 3 light borders. Trajectories were identified for every fly that approached the choice zone, and the following metrics were computed:

The above 3 metrics were computed for all flies that entered a choice zone at least once during the illumination epoch. Flies that did not make such a crossing during the epoch (i.e., remained on one side for the epoch duration) were necessarily excluded from boundary trajectory analysis. Note that as flies could enter a choice zone without ever subsequently crossing from dark to light (or vice versa), not all flies with a choice index could also be assigned a speed ratio; this would include flies that consistently made choice-zone reversals without crossing a light–dark boundary. Thus, the choice-zone analysis necessarily excluded flies that never crossed a light–dark boundary.

Effect-size regression

To compare the valence effect size (Δ Preference or ΔPI) with locomotion effect sizes, Δ values were calculated for 4 locomotion metrics.

Each Δ value was calculated for the 2 highest illumination intensity epochs (22 and 70 μW/mm2) for all PAM or MBON lines. As the locomotion metrics require that a fly crosses the light–dark boundary at least once, some flies were necessarily censored from this analysis.

Each set of effect sizes was subjected to regression against the corresponding Δ Preference values. Regression was performed with the linear least-squares method of the SciPy library. For both mean differences and coefficients of determination (R²), distributions and 95% confidence intervals were obtained from 3,000 resamples, using bias correction and acceleration [129] with the scikits.bootstrap package.

Meta-analysis of dopamine-depleting interventions

The dopamine loss-of-function experiments with R58E02>Chr were established using 3-iodo-tyrosine, TH-RNAi, or TH-RNAi with Dicer2 at several light intensities, resulting in a total of 15 valence experiments. The effect of reducing dopamine function in the R58E02 cells was estimated as a percentage of wild-type behavior [93,130]. Using data from 3 replicates of the R58E02>Chr experiment, a mean ΔPI (controls) valence was first calculated via simple averaging for the 5, 22, and 70 μW/mm² light conditions. The ΔPIs for each dopamine-depleting intervention were then expressed as a percentage of the control value [93] for each of 3 light intensities.

The data used for calculating the mean ΔPI (controls) value are presented in Figs 4E, S3A, and S3B. The data used for calculating % of control for each intervention are presented in Figs S3G (3-iodo-tyrosine); 4F, S3C, and S3D (TH-RNAi); S3E (TH-RNAi with Dicer2).

Meta-analysis of experimental replicates

For replicates of the MB213> TH-RNAi; Chrimson experiment, an inverse-variance meta-analysis was performed with a fixed-effects model. A weighted effect size was calculated as follows:

Where:

Correlation analysis for PAM:MBON preference index effect sizes

PAM:MBON driver pairs for analysis in S8 Fig were identified on the basis of both drivers in each pair having major anatomical staining in the same MB zone. For lines from the FlyLight split-Gal4 collection, major staining was defined as a staining intensity of 3 or greater, as defined in a previous report [49]. For other, non-split Gal4 driver lines, the presence of major zonal projections was derived from other previously published reports (S2 Table). Regression was performed with the linear least-squares method of the SciPy library.

Statistics

Effect sizes were used for data analysis and interpretation. Summary measures for each group were plotted as a vertical gapped line: the ends of the line correspond to the ±standard deviations of the group, and the mean itself plotted as a gap. Effect sizes were reported for each driver line as mean differences between controls and test animals for all the behavioral metrics [131]. Two controls (driver and responder) were grouped together and the averaged mean of the 2 controls was used to calculate the mean difference between control and test flies. In text form, the mean differences and their 95% confidence intervals are presented as “mean [95 CI lower bound, upper bound].” The mean differences are depicted as black dots, with the 95% confidence interval indicated by error bars. Where possible, each error bar is accompanied with a filled curve displaying the distribution of mean differences, as calculated by bootstrap resampling. Bootstrapped distributions are robust for non-normal data [132].

P-values were calculated by the Mann–Whitney rank method in SciPy and presented pro forma only: following best practice, no significance tests were conducted [131]. The behavioral sample sizes (typically N = 45, 45, 45) had a power of >0.8, assuming α = 0.05 and an effect size of Hedges’ g = 0.6 SD. All error bars for mean differences represent the 95% confidence intervals.

Genotypes and statistics for control and test flies for each panel are provided in the supporting table (S1 Table), as are details of the mega-analysis of mutant effect sizes (Fig 4G) incorporating data from across the study.