Article Figures & Data
Figures
- Figure 1.
Ca2+ activity is depressed at the MBON-γ5β′2a dendrites following odor-DAN pairing. A, Anterior view of Drosophila brain (via BioRender). Mushroom body colored dark gray. Hashed box marks location of illustration of Chrimson88-expressing subset of PAM DAN axons and GCaMP6f-expressing MBON-γ5β′2a dendrites. B, Anterior view of the Drosophila brain (via Virtual Fly Brain) corresponding to A. Left hemisphere, Blue, boundaries of β′ lobe; white, boundaries of γ lobe (Bogovic et al., 2020). Right hemisphere, Yellow, EM tracing skeleton of MBON-γ5β′2a dendrites (Xu et al., 2020). Both hemispheres: green: expression pattern of VT-006202 PAM DAN axons (Tirian and Dickson, 2017). C, Conventional circuit model for associative olfactory learning; location corresponds to the solid black box in the γ compartment illustrated in A. (1) A KC axon is activated by olfactory input, allowing Ca2+ (yellow dots) influx through voltage-gated Ca2+ channels (yellow rectangles). (2) Ca2+ influx promotes release of acetylcholine (black dots) via vesicular exocytosis, exciting the MBON dendrites. (3) Coincident DAN activation leads to activation of dopamine receptors (orange ellipsis) that activate rutabaga, a Ca2+-dependent adenylyl cyclase (blue pentagon). (4) Rutabaga synthesizes cAMP, initiating downstream signal transduction thought to promote synaptic plasticity. D, Odor-DAN pairing protocol: pre-pairing, post-pairing, and post-reversal odor presentations are 1 s in duration separated by a 30-s interstimulus interval, repeated three times. Pairing and reversal each combine photostimulation (via 4 1-ms pulses delivered at 2 Hz) with Odor A and Odor B, respectively. E, Odor A mean response profile (±SEM, shaded area; N = 12) at pre-pairing (black), post-pairing (red), and post-reversal (blue). Bold horizontal line indicates odor presentation. F, Same as C but for Odor B. G, Minimum-to-maximum box plots of mean ΔF/F of 5-s response window beginning at odor onset. Fine lines indicate data from individual flies; bold lines indicate mean; horizontal bar within box indicates median. Repeated-measures two-way ANOVA to compare effect of pairing across odors (interaction p = 0.0044) and within odors at each stage of the experiment (Dunnett’s multiple comparison test, from left to right: p = 0.0004, 0.0308, 0.4577, 0.0351). Asterisks indicate statistical significance: *p < 0.05, **p < 0.01, ***p < 0.001. “ns” indicates not significant, p > 0.05.
- Figure 2.
Odor-evoked acetylcholine release is depressed at the mushroom body lobes with compartment specificity following odor-DAN pairing. A, Anterior view of Drosophila brain. Mushroom body colored dark gray. Hashed box marks location of illustration of Chrimson88-expressing subset of PAM DAN axons and ACh3.0-expressing KC axons. B, Minimum-to-maximum box plots of mean ΔF/F of response window of axons within γ5 compartment (N = 14). Fine lines indicate data from individual flies; bold lines indicate mean; horizontal bar within box indicates median. Repeated-measures two-way ANOVA to compare effect of pairing across odors (interaction p = 0.0390) and within odors at each stage of the experiment (Dunnett’s multiple comparison test, from left to right: p = 0.0020, 0.4176, 0.3889, 0.0265). Right, Upper, Odor A mean response profile (±SEM, shaded area) from γ5 compartment at pre-pairing (black), post-pairing (red), and post-reversal (blue). Right, Lower, Same as upper but for Odor B. Typical field of view and selection of ROIs are shown in Extended Data Figure 2-1. C, Same as B but for β′2 compartment (N = 14). Effect of pairing across odors (interaction p = 0.4433). D, Same as B but for γ4 compartment (N = 14). Effect of pairing across odors (interaction p = 0.5274). Asterisks indicate statistical significance: *p < 0.05, **p < 0.01. “ns” indicates not significant, p > 0.05.
- Figure 3.
Bulk Ca2+ activity of KC axons does not change following odor-DAN pairing. A, Anterior view of Drosophila brain. Mushroom body colored dark gray. Hashed box marks location of illustration of Chrimson88-expressing subset of PAM DAN axons and GCaMP6f-expressing KC axons. B, Minimum-to-maximum box plots of mean ΔF/F of response window of axons within γ5 compartment (N = 11). Fine lines indicate data from individual flies; bold lines indicate mean; horizontal bar within box indicates median. Repeated-measures two-way ANOVA to compare effect of pairing across odors (interaction p = 0.0227) and within odors at each stage of the experiment (Dunnett’s multiple comparison test, from left to right: p = 0.1668, 0.1129, 0.4829, 0.7307). Right, Upper, Odor A mean response profile (±SEM, shaded area) from γ5 compartment at pre-pairing (black), post-pairing (red), and post-reversal (blue). Right, Lower, Same as upper but for Odor B. C, Same as B but for β′2 compartment (N = 11). Effect of pairing across odors (interaction p = 0.1477). D, Same as B but for γ4 compartment (N = 11). Effect of pairing across odors (interaction p = 0.0409) and within odors at each stage of the experiment (Dunnett’s multiple comparison test, from left to right: p = 0.2121, 0.2212, 0.4735, 0.7854). Asterisks indicate statistical significance: *p < 0.05. “ns” indicates not significant, p > 0.05.
- Figure 4.
Effects of odor-DAN pairing at MBON-γ1pedc and relevant MB compartments. A, Anterior view of Drosophila brain. Mushroom body colored dark gray. Hashed box marks location of illustration of Chrimson88-expressing subset of PPL1 DAN axons and sensor-expressing neurites: corresponding with B, GCaMP6s-expressing dendrites of MBON-γ1pedc; corresponding with C, D, ACh3.0-expressing KC axons; corresponding with E, F, GCaMP6f-expressing KC axons. B, Minimum-to-maximum box plots of mean ΔF/F of response window of MBON-γ1pedc dendrites (N = 10). Fine lines indicate data from individual flies; bold lines indicate mean; horizontal bar within box indicates median. Repeated-measures two-way ANOVA to compare effect of pairing across odors (interaction p < 0.0001) and within odors at each stage of the experiment (Dunnett’s multiple comparison test, from left to right: p = 0.0022, 0.0149, 0.9348, <0.0001). Right, Upper, Odor A mean response profile (±SEM, shaded area) from MBON-γ1pedc dendrites at pre-pairing (black), post-pairing (red), and post-reversal (blue). Right, Lower, Same as upper but for Odor B. C, Same as B but for ACh release in γ1pedc compartments (N = 15). Effect of pairing across odors (interaction p = 0.0290) and within odors at each stage of the experiment (Dunnett’s multiple comparison test, from left to right: p = 0.0167, 0.7553, 0.9986, 0.0006). D, Same as C but for γ2 compartment (N = 14). Effect of pairing across odors (interaction p = 0.2616). E, Same as B but for Ca2+ activity in γ1pedc compartments (N = 13). Effect of pairing across odors (interaction p = 0.1207). F, Same as E but for γ2 compartment (N = 9). Effect of pairing across odors (interaction p = 0.5037). Asterisks indicate statistical significance: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. “ns” indicates not significant, p > 0.05.
- Figure 5.
Diverse effects of odor-DAN pairing on Ca2+ activity of individual presynaptic boutons of KCs. A, Anterior view of Drosophila brain. Mushroom body colored dark gray. Hashed box marks location of illustration of Chrimson88-expressing subset of PAM DANs and KCs sparsely expressing jGCaMP7f. B, Left, Maximum z-projection of a typical field of view from a bouton imaging experiment; γ5 outlined in teal, β′2 in orange; scale bar, 5 μm; five example boutons circled in yellow (a-e). Right, Response profiles of each example bouton; scale bar, 5 s and z score of 5. C, Minimum-to-maximum boxplots of mean z score of the 10-s window following odor presentation for all boutons within the γ5 compartment (paired n = 263 boutons, N = 10 flies; odor-only n = 222, N = 11). Fine lines indicate data from individual flies; bold lines indicate mean; horizontal bar within box indicates median. Repeated-measures two-way ANOVA to compare effect of pairing across odors (interaction p = 0.0125) and within odors at each stage of the experiment (Šídák’s multiple comparisons test, from left to right: p = 0.9100, 0.0048). Right, Upper, Paired odor mean response profile (±SEM, shaded area) at pre-pairing (black) and post-pairing (red). Right, Lower, Same as upper but for odor-only condition. D, Same as C but for all boutons within the γ4 compartment (paired n = 150, N = 11; odor-only n = 178, N = 10). Repeated-measures two-way ANOVA to compare effect of pairing across odors (interaction p = 0.3038). E–H, All γ5 boutons are divisively filtered into groups based on conservative thresholds applied to their individual response profiles. Positive (E, n = 251; mean total z score > 1.6449, corresponding to p = 0.05 threshold), negative (F, n = 30; ≥1 s of mean z score > 1.6449), biphasic (G, n = 25; ≥1 s of mean z score > 1.6449 and < 1.6449), and unresponsive (n = 179; all remaining boutons). I, Example mean pre-pairing responses taken from seven boutons categorized as “Positive” to illustrate within-group diversity. J, Rescaled pre-pairing responses to paired odor in boutons that will undergo changes of at least 20% after pairing, either via depression (n = 52; blue line) or potentiation (n = 51; orange line). K, Same as I but for the odor-only condition (will be depressed, n = 37; will be potentiated, n = 66). Further analyses based on temporal characteristics of γ5 bouton responses are shown in Extended Data Figure 5-1. Asterisks indicate statistical significance: *p < 0.05, **p < 0.01. “ns” indicates not significant, p > 0.05.
- Figure 6.
Plasticity of γ5 boutons is bidirectionally biased by the magnitude of naive odor responses. A, Pre-pairing mean z score plotted against post-pairing mean z score for all γ5 boutons responsive to odor at either pre-pairing or post-pairing. Orange points represent paired odor responsive boutons (n = 116); blue points represent odor-only condition responsive boutons (n = 135); solid lines represent linear regression lines compared by ANCOVA (p < 0.0001); hashed black line marks line of identity. B, Same as A but for odor responsive boutons in γ4 compartment (paired n = 84, odor-only n = 111); linear regression lines compared via ANCOVA (p = 0.2176). C, Minimum-to-maximum box plots of mean z score of “strong responder” boutons within γ5 compartment (paired n = 71, odor-only n = 105). Fine lines indicate data from individual flies; bold lines indicate mean; horizontal bar within box indicates median. Repeated-measures two-way ANOVA to compare effect of pairing across odors (interaction p < 0.0001) and within odors at each stage of the experiment (Šídák’s multiple comparisons test, from left to right: p < 0.0001, p = 0.0775). Lower, Left, Paired odor mean response profile (±SEM, shaded area) at pre-pairing (black), post-pairing (red), and post-reversal (blue). Lower, Right, Same as left but for odor-only condition. D, Same as C but for “strong responder” boutons in γ4 compartment (paired n = 61, odor-only n = 85). Repeated-measures two-way ANOVA to compare effect of pairing across odors (interaction p = 0.4769). E, Same as C but for “weak responder” boutons in γ5 compartment (paired n = 45, odor-only n = 30). Repeated-measures two-way ANOVA to compare effect of pairing across odors (interaction p = 0.0305) and within odors at each stage of the experiment (multiple comparisons from left to right: p < 0.0001, 0.0037). F, Same as B but for “weak responder” boutons in γ4 compartment (paired n = 23, odor-only n = 26). Repeated-measures two-way ANOVA to compare effect of pairing across odors (interaction p = 0.4380). We also performed pairing experiments in genetic control flies, and the data are shown in Extended Data Figure 6-1. Asterisks indicate statistical significance: *p < 0.05, **p < 0.01, ***p < 0.0001. “ns” indicates not significant, p > 0.05.
- Figure 7.
Heterogeneous plasticity exhibited by boutons on likely-spiking axons. A, Pre-pairing and post-pairing mean total z score of paired γ5 boutons (left y-axis) grouped according to their parent axons (1–12 on x-axis). White bar graph (right y-axis) denotes the percentage of boutons on the parent axon that responds to odor. The hashed vertical line separates “likely spiking axons” from “likely nonspiking axons” based on an 80% threshold. B, Same as A but for odor-only γ5 boutons. C, Maximum z-projection of axon #2, which is likely spiking in response to odor that was paired with odor-DAN activation. Bouton ROIs are marked by open circles whose color corresponds to the percent change of their naive odor response after pairing. Red hashed line marks boundary around the γ5 compartment; axon to the left of the bounded area belongs to the γ4 compartment; left to right is proximal to distal; scale bar is 3 μm. D, Same as C but for axon #3. E, Same as C but for axon #4. F, Same as C but for axon #14, which is likely spiking in response to odor that was presented in odor-only condition.
Tables
Figure Genotype Figure 1C–E 20XUAS-IVS-Syn21-OpGCaMP6f-p10(attP8); VT006202-LexA(attP40); 13×LexAop2-IVS-Syn21-Chrimson88:tdT-3.1-p10(VK00005)/SS01308-split-GAL4 Figure 2B–D VT006202-LexA(attP40); 13×LexAop2-IVS-Syn21-Chrimson88:tdT-3.1-p10(VK00005)/R13F02-GAL4(attP2), UAS-ACh3.0(VK00005) Figure 3B–D 20XUAS-IVS-Syn21-OpGCaMP6f-p10(attP8); VT006202-LexA(attP40); 13×LexAop2-IVS-Syn21-Chrimson88:tdT-3.1-p10(VK00005)/R13F02-GAL4(attP2) Figure 4B 20XUAS-IVS-Syn21-OpGCaMP6s-p10(attP8); VT45661-LexA(JK22C); 13×LexAop2-IVS-Syn21-Chrimson88:tdT-3.1-p10(VK00005)/MB112C-split-GAL4 Figure 4C,D VT45661-LexA(JK22C); 13×LexAop2-IVS-Syn21-Chrimson88:tdT-3.1-p10(VK00005)/R13F02-GAL4(attP2), UAS-ACh3.0(VK00005) Figure 4E,F 20XUAS-IVS-Syn21-OpGCaMP6f-p10(attP8); VT45661-LexA(JK22C); 13×LexAop2-IVS-Syn21-Chrimson88:tdT-3.1-p10(VK00005)/R13F02-GAL4(attP2) Figures 5C–K, 6A–F, 7A–F; Extended Data Figures 5-1A–F, 6-1A–F (“Paired”) 20XUAS-IVS-phiC31(attP18); VT006202-LexA(attP40)/20×UAS-SPARC-S-jGCaMP7f(CR-P40); 13×LexAop2-IVS-Syn21-Chrimson88:tdT-3.1-p10(VK00005)/R13F02-GAL4(attP2) Extended Data Figure 6-1A–F (“G.C.”) 20XUAS-IVS-phiC31(attP18); 20×UAS-SPARC-S-jGCaMP7f(CR-P40); 13×LexAop2-IVS-Syn21-Chrimson88:tdT-3.1-p10(VK00005)/R13F02-GAL4(attP2) The genotype of each experimental fly is reported according to the corresponding data, listed by figure panel.
Extended Data Figure 2-1
Typical field of view during imaging experiment to show ROI selection guided by GFP-based sensor and tdTomato-tagged opsin signals. A, ACh3.0 signal pseudocolored in green. Boundaries of ROIs corresponding to γ4, γ5, and β′2 marked in orange. Scale bar, 5 μm. B, Chrimson88:tdTomato signal pseudocolored in magenta. Note that we intentionally drew ROI boundaries conservatively to avoid contamination of signals between compartments. Download Figure 2-1, EPS file.
Extended Data Figure 5-1
Response profiles of responsive γ5 boutons undergoing plasticity lack predictive features in their temporal patterns. A, Rescaled prepairing responses to paired odor in boutons that will undergo changes of at least 20% after pairing, either via depression (n = 52; blue line) or potentiation (n = 51; orange line). Reproduced and modified from Figure 5J. B, Same as A but for the odor-only condition (will be depressed, n = 37; will be potentiated, n = 66). Reproduced and modified from Figure 5K. C, Mean rescaled value during early phase of prepairing odor responses. Repeated-measures two-way ANOVA to compare effect of pairing across odors (interaction p = 0.8913). D, Same as C but for middle phase. Repeated-measures two-way ANOVA to compare effect of pairing across odors (interaction p = 0.7855). E, Same as C but for late phase. Repeated-measures two-way ANOVA to compare effect of pairing across odors (interaction p = 0.1069). F, Principal component analysis of all paired odor-responsive γ5 boutons; mean prepairing odor responses segmented into 1-s bins; plasticity rates <20% (unfilled circles), depression ≥20% (filled blue), potentiation ≥20% (filled orange). Download Figure 5-1, EPS file.
Extended Data Figure 6-1
Comparing paired odor responsive boutons from experimental flies against paired responsive boutons from genetic control (G.C.) flies. A, Prepairing mean z score plotted against postpairing mean z score for all γ5 boutons responsive to odor. Orange points represent paired odor responsive boutons from experimental flies (n = 116 boutons, N = 10 flies); black points represent paired odor responsive boutons from G.C. flies (n = 129, N = 11); solid lines represent linear regression lines compared by ANCOVA (p = 0.3005); hashed black line marks line of identity. B, Same as A but for odor responsive boutons in γ4 compartment (paired n = 84, N = 11; G.C. n = 116, N = 11); linear regression lines compared by ANCOVA (p = 0.1438). C, Minimum-to-maximum box plots of mean total z score of “strong responder” boutons within γ5 compartment (paired n = 71, G.C. n = 93). Fine lines indicate data from individual flies; bold lines indicate mean; horizontal bar within box indicates median. Repeated-measures two-way ANOVA to compare effect of pairing across odors (interaction p = 0.0469) and within odors at each stage of the experiment (multiple comparisons from left to right: p = 0.0016, 0.6231). Lower, Left, Paired odor mean response profile (±SEM, shaded area) at prepairing (black), postpairing (red), and postreversal (blue). Lower, Right, Same as left but for paired odor in G.C., flies. D, Same as C but for “strong responder” boutons in γ4 compartment (paired n = 61, G.C. n = 78). Repeated-measures two-way ANOVA to compare effect of pairing across odors (interaction p = 0.3248). E, Same as C but for “weak responder” boutons in γ5 compartment (paired n = 45, G.C. n = 36). Repeated-measures two-way ANOVA to compare effect of pairing across odors (interaction p = 0.0051) and within odors at each stage of the experiment (multiple comparisons from left to right: p < 0.0001, 0.0397). F, Same as C but for “weak responder” boutons in γ4 compartment (paired n = 23, G.C. n = 38). Repeated-measures two-way ANOVA to compare effect of pairing across odors (interaction p = 0.3309). Download Figure 6-1, EPS file.
In this issue
