Activity-dependent FMRP requirements in development of the neural circuitry of learning and memory

Abstract

The activity-dependent refinement of neural circuit connectivity during critical periods of brain development is essential for optimized behavioral performance. We hypothesize that this mechanism is defective in fragile X syndrome (FXS), the leading heritable cause of intellectual disability and autism spectrum disorders. Here, we use optogenetic tools in the Drosophila FXS disease model to test activity-dependent dendritogenesis in two extrinsic neurons of the mushroom body (MB) learning and memory brain center: (1) the input projection neuron (PN) innervating Kenyon cells (KCs) in the MB calyx microglomeruli and (2) the output MVP2 neuron innervated by KCs in the MB peduncle. Both input and output neuron classes exhibit distinctive activity-dependent critical period dendritic remodeling. MVP2 arbors expand in Drosophila mutants null for fragile X mental retardation 1 (dfmr1), as well as following channelrhodopsin-driven depolarization during critical period development, but are reduced by halorhodopsin-driven hyperpolarization. Optogenetic manipulation of PNs causes the opposite outcome--reduced dendritic arbors following channelrhodopsin depolarization and expanded arbors following halorhodopsin hyperpolarization during development. Importantly, activity-dependent dendritogenesis in both neuron classes absolutely requires dfmr1 during one developmental window. These results show that dfmr1 acts in a neuron type-specific activity-dependent manner for sculpting dendritic arbors during early-use, critical period development of learning and memory circuitry in the Drosophila brain.

Keywords: Channelrhodopsin; Critical period; Dendrite; Drosophila; Fmr1; Fragile X syndrome; Halorhodopsin; Mushroom body; Synapse.

Fig. 1.

Two classes of mushroom body extrinsic input and output neurons. Drosophila brains at 1 day post eclosion show MB pedunculus-medial lobe and vertical lobe arborizing neuron 2 (MVP2) and projection neuron (PN) architecture. (A) Schematic illustrating MVP2 (left) and PN (right) classes. ML, medial lobe. (B) Oli-Gal4 driving UAS-DenMark in MVP2 dendrites (green) within the MB spur (SPU), anterior to the distal-most pedunculus (PED), with anti-fasciclin II (FasII, red) labeling Kenyon cell (KC) axons. Insets show MVP2 dendrites at higher magnification receiving an MB spur input, and an MVP2 axon innervating the MB vertical lobe (VL). D, dorsal; V, ventral. (C) 14-3-3ζ-Gal4 driving UAS-mCD8::GFP in PNs (green) deriving presynaptic input from the antennal lobe (AL) VL1 glomerulus, and innervating KCs in the MB calyx (FasII, red). Insets show FasII-labeled AL and phalloidin (F-actin)-labeled MB calyx, with PN dendritic arbors in the AL and axonal presynaptic microglomeruli in the MB calyx. Scale bars: 50 μm in B; 30 μm in C.

Fig. 2.

MB extrinsic neuron dendritic arbors expand in dfmr1-null mutants. MVP2 dendritic arbors labeled with Oli-Gal4>UAS-DenMark at day 1 post eclosion. Representative images of (A) w¹¹¹⁸ genetic background control (wildtype) and (B) the homozygous dfmr1^50M/dfmr1^50M null mutant (dfmr1), showing dorsoventral dendritic arbor expansion in the absence of FMRP. Bottom panels show a dorsoventral optical section series illustrating every third frame of z-stack images. D, dorsal; V, ventral; A, anterior; P, posterior. Scale bars: 10 μm (top), 20 μm (z-series). (C) Quantification of dendritic arbor volume comparing w¹¹¹⁸ genetic background control (wildtype), heterozygous control (dfmr1^50M/+) and homozygous null mutant (dfmr1^50M/dfmr1^50M). Data are presented as box and whisker plots showing the minimum, 25th percentile, median, 75th percentile and maximum values. N values are as follows: wildtype, 16; dfmr1/+, 13; null, 16. ***P<0.001 (ANOVA).

Fig. 3.

FMRP is required for activity-dependent MVP2 dendritic development. Dorsoventral series of MVP2 dendritic arbors with Oli-Gal4-driven UAS-ChR2(H134R) depolarization during day 1 post eclosion. Representative images show vehicle-fed wild-type control (A), ATR-fed wild-type experimental (B), vehicle-fed dfmr1^50M null (C) and ATR-fed dfmr1^50M null experimental (D) conditions following 24 h of 5 Hz blue light stimuli. On the right, dendritic arbors are shown in a dorsoventral series of 3 µm steps through the z-stack images. Scale bars: 10 μm. (E) Quantification of dendritic arbor volume in all four conditions. N values are as follows: wt^ETOH, 25; wt^ATR, 31; null^ETOH, 32; null^ATR, 25. ***P<0.001; n.s., not significant (ANOVA).

Fig. 4.

PN dendrites exhibit the opposite FMRP-dependent activity dependence. PN dendrites with 14-3-3ζ-Gal4-driven UAS-eNpHR3.0 hyperpolarization during day 1 post eclosion. Representative images show flattened 3D representations of z-stacks from anterior (left) and lateral (right) perspectives for vehicle-fed wild-type control (A), ATR-fed wild-type experimental (B), vehicle-fed dfmr1^50M null (C) and ATR-fed dfmr1^50M null experimental (D) dendritic arbors following 24 h exposure to 5 Hz amber light. Scale bar: 10 μm. (E) Quantification of dendritic arbor volume in the four conditions. N values are as follows: wt^ETOH, 21; wt^ATR, 26; null^ETOH, 16; null^ATR, 15. ***P<0.001; n.s., not significant (ANOVA).

Fig. 5.

Null MVP2 dendritic arbors rescued with conditional dfmr1 expression. Targeted critical period expression of wild-type dfmr1 in MVP2 neurons in an otherwise dfmr1-null background. Compared with null mutant (dfmr1^50M,Oli-Gal4/TM6TbGFP>UAS-DM;dfmr1^50M/TM6) (A), conditional dfmr1 P4 induction (tubP-Gal80ts/CyO;dfmr1^50M,Oli-Gal4/TM6>UAS-DM/CyO;dfmr1^50M,UAS-9557-3/TM6) (B) by temperature shift to 29°C causes a reduction in MVP2 dendritic arbor volume. Gray images (top) show z-stack projections from anterior (left) and lateral (right) perspectives. Color images (bottom) show the absence (A) and presence (B) of FMRP (green) in MVP2 soma labeled with DenMark (red) in an otherwise dfmr1-null brain. Scale bars: 10 μm (upper rows), 5 μm (lower rows). (C) Quantification of dendritic arbor volumes compared to those of constitutive dfmr1 expression controls, including constant restrictive temperature with Gal80ts repressor and expression in the absence of Gal80ts. N values are as follows: null^control, 22; null^P4-rescue, 23; Null^constitutive, 18; Null^{constitutive(-Gal80)}, 15. ***P<0.001 (ANOVA).

Fig. 6.

Critical period dfmr1 induction rescues mutant PN dendritic arbors. Targeted critical period expression of wild-type dfmr1 in PNs in an otherwise dfmr1-null background. Compared to the dfmr1 null (dfmr1^50M,14-3-3ζ-Gal4/TM6TbGFP>UAS-DM;dfmr1^50M/TM6) (A), dfmr1 induction at P4 (tubP-Gal80ts/CyO;dfmr1^50M,14-3-3ζ-Gal4/TM6>UAS-DM/CyO;dfmr1^50M,UAS-9557-3/TM6) (B) by temperature shift to the restrictive 29°C reduces PN dendritic arbor volume. Top rows depict PN dendritic arbors within the phalloidin-labeled (purple) AL. FMRP expression (green) occurs only with conditional expression in DenMark-labeled (red) PNs (B, lower row) in brains otherwise completely deficient for FMRP. The far-right column displays two dendritic arbor examples from temperature-shifted null (A) and conditional rescue (B). Scale bars: 10 μm (upper rows), 5 μm (lower rows). (C) Quantification of dendritic arbor volumes compared to two constitutive dfmr1 expression controls, including constant restrictive temperature with the Gal80ts repressor and expression in the absence of the Gal80ts. N values are as follows: null^control, 15; null^P4-rescue, 14; null^constitutive, 13; null^{constitutive(-Gal80)}, 13. *P<0.05, **P<0.01 (ANOVA).

Fig. 7.

FMRP rescue restores activity-dependent dendritic arbor development. MVP2 dendritic arbors with dfmr1 conditional rescue in an otherwise dfmr1-null background and following ChR2(H134R)-mediated depolarization during the critical period, with and without feeding the essential ATR co-factor. Following dfmr1 P4 conditional induction, tubP-Gal80ts/CyO;dfmr1^50M,Oli-Gal4/TM6>UAS-ChR2(H134R)-mCherry/CyO;dfmr1^50M, UAS-9557-3/TM6 animals were exposed to 24 h of 5 Hz blue light stimulation. (A) Representative images of MVP2 dendritic arbors from anterior (top) and lateral (bottom) perspectives from animals fed with either vehicle control (left column) or ATR (right column). Anti-FMRP (green) in UAS-DenMark (red) labeled soma and dendritic arbors (FMRP granules, arrows) in the otherwise dfmr1-null brains. Scale bar: 10 μm. (B) Quantification of dendritic arbor volumes. N values are as follows: null^{rescue+vehicle}, 18; null^rescue+ATR, 20. The dashed line shows the wild-type MVP2 arbor volume. ***P<0.001.