Kinesin Khc-73/KIF13B modulates retrograde BMP signaling by influencing endosomal dynamics at the Drosophila neuromuscular junction

Abstract

Retrograde signaling is essential for neuronal growth, function and survival; however, we know little about how signaling endosomes might be directed from synaptic terminals onto retrograde axonal pathways. We have identified Khc-73, a plus-end directed microtubule motor protein, as a regulator of sorting of endosomes in Drosophila larval motor neurons. The number of synaptic boutons and the amount of neurotransmitter release at the Khc-73 mutant larval neuromuscular junction (NMJ) are normal, but we find a significant decrease in the number of presynaptic release sites. This defect in Khc-73 mutant larvae can be genetically enhanced by a partial genetic loss of Bone Morphogenic Protein (BMP) signaling or suppressed by activation of BMP signaling in motoneurons. Consistently, activation of BMP signaling that normally enhances the accumulation of phosphorylated form of BMP transcription factor Mad in the nuclei, can be suppressed by genetic removal of Khc-73. Using a number of assays including live imaging in larval motor neurons, we show that loss of Khc-73 curbs the ability of retrograde-bound endosomes to leave the synaptic area and join the retrograde axonal pathway. Our findings identify Khc-73 as a regulator of endosomal traffic at the synapse and modulator of retrograde BMP signaling in motoneurons.

Fig 1. Khc-73 mutants have normal synaptic structure and function.

(A) Khc-73 synapse structure at muscle 4. Postsynaptic Dlg stain (green) and presynaptic neuron HRP stain (red). Control (Khc-73¹⁰⁰). Scale bar is 10μm. (B) Quantification of bouton number in Control (Khc-73¹⁰⁰) and Khc-73 (Khc-73¹⁴⁹) mutants at muscle 4 n = 17, 18 NMJs. (C) Muscle surface area of muscle 4 in Control (Khc-73¹⁰⁰) and Khc-73 (Khc-73¹⁴⁹). Muscle 4 n = 17, 18. (D) Representative traces of EJCs and mEJCs from third instar larval NMJ in precise excision Khc-73¹⁰⁰ (top) and Khc-73 mutant (Khc-73¹⁴⁹) (bottom). (E) Quantification of mEJC, EJC and QC for control w¹¹¹⁸, Khc-73¹⁰⁰, Khc-73¹⁹³ and Khc-73¹⁴⁹. N = 9, 14, 7 and 11 NMJs. Error Bars are SEM. Student’s t-test. ns-no statistical significance.

Fig 2. Khc-73 mutants have fewer active zones.

(A) Terminal boutons on muscle 4 NMJs stained with anti-nc82 (Brp) (green) and anti-GluRIII (red) in control (Khc-73¹⁰⁰) and Khc-73(Khc-73¹⁴⁹) third instar larva. Scale Bar is 5μm. (B) Quantification of active zones at muscle 4 NMJ with neuron and muscle rescue. Wild-type (Khc-73¹⁰⁰), Khc-73 (Khc-73¹⁴⁹), Khc-73; genomic rescue (Khc-73¹⁴⁹; CH321-36I16/+). N = 36, 19, and 17 NMJs. (C) Khc-73 neuron rescue of Brp puncta. Khc-73 control (OK371-Gal4, Khc-73¹⁴⁹/Khc-73¹⁴⁹) and Neuron rescue (OK371-Gal4, Khc-73¹⁴⁹/Khc-73¹⁴⁹; UAS-HA-Khc-73(K014)/+). N = 20, 18 NMJs. (D) Khc-73 muscle rescue of Brp puncta. Khc-73 control (Khc-73¹⁴⁹; MHC-Gal4/+), and Khc-73 muscle rescue (Khc-73¹⁴⁹; MHC-Gal4/UAS-HA-Khc-73(K014)). N = 6 and 6 NMJs. Error bars are SEM. One-Way ANOVA and Student’s t-test. *P<0.05, **P<0.01, ns-no significance.

Fig 3. Khc-73 genetic interaction with BMP pathway.

(A) Terminal boutons on muscle 4 of third instar larva in Medea/+ (Med^G112/+), Medea (Med^C246/Med^G112) and Khc-73/+; Medea/+ (Khc-73¹⁹³/+; Med^G112/+). Brp (green) and HRP (red). Scale bar is 5μm. (B-E) Quantification of (B) Brp puncta per NMJ, (C) Brp puncta per bouton, (D) HRP synaptic area, (E) Number of boutons at muscle 4 NMJ for Medea/+ (Med^G112/+), Khc-73/+ (Khc-73¹⁴⁹/+), Medea (Med^C246/Med^G112), Khc-73/+; Medea/+ (Khc-73¹⁹³/+; Med^G112/+) and Khc-73/+; Medea/+ (Khc-73¹⁴⁹/+; Med^C246/+), N = 31, 17, 17, 21 and 14 NMJs. (F) Terminal boutons on muscle 4 in wit/+ (wit^A12 /+) and Khc-73/+; wit/+ (Khc-73¹⁴⁹/+; wit^A12/+) mutant larva. (G-J) Quantification of (G) Brp puncta per NMJ, (H) Brp puncta per bouton, (I) HRP Synaptic area, (J) Number of boutons at muscle 4 NMJs for Control (Khc-73¹⁰⁰/+), wit/+ (wit^A12/+) and Khc-73/+; wit/+ (Khc-73¹⁴⁹/+; wit^A12/+). N = 11, 8 and 10. Error bars are SEM. One-Way ANOVA and Student’s t-test. *P<0.05, **P<0.01, ***<0.001.

Fig 4. Enhanced BMP signaling suppresses loss of Brp puncta in Khc-73 mutants.

(A) Brp puncta in terminal boutons of muscle 4 NMJs in Control (BG380-Gal4/+; UAS-luciferase/+), TKV^ACT (BG380-Gal4/+; UAS-TKV^ACT/+), Khc-73 (BG380-Gal4/+; Khc-73¹⁴⁹/Khc-73¹⁴⁹), Khc-73; TKV^ACT (BG380-Gal4/+; Khc-73¹⁴⁹; UAS-TKV^ACT/+) larvae. (B) Quantification of BRP puncta per NMJ in genotypes in (A). N = 18, 17, 13, 12 NMJs. (C) Overexpression of activated form of TKV in motoneurons does not enhance BRP intensity at the NMJ. Control (BG380-Gal4/+; UAS-luciferase/+); TKV^ACT (BG380-Gal4/+; UAS-TKV^ACT). N = 19, 17 NMJs. Error bars are SEM. Student’s t-test. ***<0.001. ns-no significance. Scale bar is 5μm.

Fig 5. Synaptic proteins Brp and synaptotagmin accumulate in axons of Khc-73 and mad mutants.

(A) Brp puncta in axons of Control (Khc-73¹⁰⁰/+), Khc-73 (Khc-73¹⁴⁹) and mad (mad^K00237). (B) Quantification of Brp density (Number of Brp puncta/μm³) expressed as a percentage of control. N = 19, 10, 10. (C) SYT puncta in axons of Control (Khc-73¹⁰⁰/+), Khc-73 (Khc-73¹⁴⁹) and mad (mad^K00237). (D) Quantification of SYT density (Number of Brp puncta/μm³) expressed as a percentage of control. N = 19, 10, 10. (E) Brp puncta in axons of Control (BG380-Gal4/+), TKV^ACT (BG380-Gal4/+; UAS-TKV^ACT/+), Khc-73 (BG380-Gal4/+; Khc-73¹⁴⁹), Khc-73; TKV^ACT (BG380-Gal4/+; Khc-73¹⁴⁹; UAS-TKV^ACT/+). (F) Quantification of Brp puncta axon density for genotypes in (E). N = 10, 5, 10 and 9 larvae. Error Bars are SEM. Student’s t-test. **P<0.01, ***P<0.001. ns-no statistical significance. Scale bar is 5μm.

Fig 6. Khc-73 suppression of BMP pathway activation at the NMJ.

(A) Representative traces of EJC and mEJCs of control (BG380-Gal4/+), Motor Neuron>TKV^ACT (BG380-Gal4/+; UAS-TKV^ACT/+) and Motor Neuron>TKV^ACT, Khc-73 (BG380-Gal4/+; Khc-73¹⁴⁹; UAS-TKV^ACT/+). (B) Quantification of mEJC, EJC and QC of genotypes shown in (A). n = 18, 18 and 20. Error bars are SEM. One-Way ANOVA and Student’s t-test. **P<0.01, ***<0.001.

Fig 7. Khc-73 is required for retrograde BMP signaling at the NMJ.

(A) Representative traces of EJC and mEJCs of Control (MHC-Gal4/+), Muscle>Gbb (UAS-Gbb⁹⁹/+; MHC-Gal4/+) and Muscle>Gbb, Khc-73 (Khc-73¹⁹³, UAS-Gbb⁹⁹/Khc-73¹⁴⁹, +; MHC-Gal4/+). (B) Quantification of mEJC, EJC and QC for genotypes shown in (A). n = 10, 10 and 7. (C) pMad staining in the motoneuron nuclei of ventral nerve cord in MHC Control (MHC-GAL4/+), MHC>Gbb (UAS-Gbb⁹⁹/+; MHC-GAL4/+) and MHC>Gbb, Khc-73 (Khc-73¹⁹³, UAS-Gbb⁹⁹/Khc-73¹⁴⁹, +; MHC-GAL4/+) larvae. (D) Quantification of the mean fluorescence intensity of nuclei for genotypes in (C). n = 302(7), 372(9) and 5(210), Nuclei (larvae) respectively. Error Bars are SEM. One way ANOVA. Student’s t-test. *P<0.05, **P<0.01, ***P<0.001.

Fig 8. Khc-73 induced enhancement of synaptic release is suppressed by heterozygosity in BMP receptor wit.

(A) Representative traces for EJCs and mEJCs in control (OK371-Gal4/+; UAS-luciferase/+), Khc-73 OE (OK371-Gal4/UAS-Khc-73) and Khc-73 OE; wit^HA4/+ (OK371-Gal4/UAS-Khc-73; wit^HA4/+). (B) Quantification of mEJC, EJC and QC for genotypes in control (OK371-Gal4/+; UAS-luciferase/+), wit^HA4/+ (OK371-Gal4/+; wit^HA4/+), Khc-73 OE (OK371-Gal4/UAS-Khc-73) and Khc-73 OE; wit^HA4/+ (OK371-Gal4/UAS-Khc-73; wit^HA4/+). N = 10, 10, 10 and 10. Error Bars are SEM. Student’s t-test. *P<0.05, **P<0.01, ***P<0.001. ns-no statistical significance.

Fig 9. BMP receptors Wit and TKV accumulate at Khc-73 mutant NMJs.

(A) Representative western blot of Wit levels in brain tissue from control (w¹¹¹⁸) and Khc-73 (Khc-73¹⁴⁹) mutants. Actin loading control. (B) Quantification of Wit protein band intensity normalized to actin of (A) and shown as percentage of control. n = 3 blots. (C) Representative western blot of Wit levels in muscle tissue from control (w¹¹¹⁸) and Khc-73 (Khc-73¹⁴⁹) mutants. Actin loading control. (D) Quantification of Wit protein band intensity normalized to actin for (C) and shown as percentage of control. n = 3 blots. (E) Live image of muscle 4 NMJs in unfixed larvae for Control (BG380-Gal4/+; OK371-Gal4/ UAS-Wit-GFP) and Khc-73 (BG380-Gal4/+; Khc-73¹⁴⁹, OK371-Gal4 / Khc-73¹⁴⁹, UAS-Wit-GFP). Scale bar is 10μm. (F) Quantification of mean fluorescence intensity as percentage of control for genotypes in (E). N = 11, 6. (G) Live image of muscle 6/7 NMJs in unfixed larvae for Control (BG380-Gal4/+; OK371-Gal4/UAS-Wit-GFP) and Khc-73 (BG380-Gal4/+; Khc-73¹⁴⁹, OK371-Gal4/ Khc-73¹⁴⁹, UAS-Wit-GFP). Scale bar is 20μm. (H) Quantification of mean fluorescence intensity as percentage of control for genotypes in (G). N = 10, 9 NMJs. Error Bars are SEM. Student’s t-test. *P<0.05, **P<0.01. ns-no statistical significance.

Fig 10. Activation of BMP signaling is suppressed in Khc-73 mutant larvae.

(A) pMad staining at the NMJ for Control (BG380-GAL4/+), Motoneuron>WIT (BG380-GAL4/+; UAS-Wit/+) and Motoneuron>Wit, Khc-73 (BG380-GAL4/+; UAS-Wit, Khc-73¹⁴⁹/+, Khc-73¹⁴⁹). (B) Quantification of mean pMad fluorescence intensity for genotypes in (A). N = 22, 44 and 10 NMJs. (C) pMad staining at the NMJ for Control (MHC-Gal4/+), Muscle>Gbb (UAS-Gbb⁹⁹/+; MHC-Gal4/+) and Muscle>Gbb, Khc-73 (Khc-73¹⁹³, UAS-Gbb⁹⁹/Khc-73¹⁴⁹, +; MHC-Gal4/+) (D) Quantification of mean pMad fluorescence intensity for genotypes in (C). N = 12, 12, and 15 NMJs. Error bars are S.E.M. Student’s t-test. *P<0.05, ***P<0.001. Scale bar is 10μm.

Fig 11. EM analysis of Khc-73 mutant larvae.

(A) EM micrographs of Control (Khc-73¹⁰⁰) and Khc-73 (Khc-73¹⁹³). Magnification 25000x. Scale bar is 0.5 μm. (B) Quantification of synapse length in Control (Khc-73¹⁰⁰) and Khc-73 (Khc-73¹⁹³). N = 79, 153 synapses. (C) Quantification of docked and clustered synaptic vesicles at each synapse for Control (Khc-73¹⁰⁰) and Khc-73 (Khc-73¹⁹³). N = 79, 153 synapses. (D) Quantification of synaptic vesicle diameter for Control (Khc-73¹⁰⁰) and Khc-73 (Khc-73¹⁹³). N = 2082, 3750 synaptic vesicles. (E) Mean number of vesicles >60nm in diameter at the synapse. N = 79, 153 synapses. (F) Quantification of subsynaptic reticulum (SSR) depth at the synapse. N = 79, 153 synapses. (G) EM micrographs of multivesicular bodies in Khc-73 mutant. Magnification of MVB denoted by arrow in left panel (right panel). Scale bar 1 μm. (H) Proportion of boutons with indicated number of multivesicular bodies for control (Khc-73¹⁰⁰) and Khc-73 (Khc-73¹⁹³). N = 33, 28 boutons. Error bars are S.E.M. Student’s t-test. *P<0.05, ns-no statistical significance.

Fig 12. Endosomal markers in Khc-73 mutant larvae.

(A) RAB5:YFP at muscle 4 NMJ terminal boutons stained with anti-GFP in control (VGlut-Gal4/+; UAS-RAB5:YFP/+) and Khc-73 (VGlut-Gal4/+; Khc-73¹⁴⁹/Khc-73¹⁴⁹; UAS-RAB5:YFP). (B) Quantification of RAB5:YFP fluorescence intensity at muscle 4 NMJ for genotypes in (A). n = 8, 12 NMJs. (C) RAB11:GFP at muscle 4 NMJ terminal boutons stained with anti-GFP in control (OK371-GAL4/ UAS-RAB11:GFP) and Khc-73 (OK371-GAL4, Khc-73¹⁴⁹/Khc-73¹⁴⁹; UAS-RAB11:GFP). (D) Quantification of RAB11:GFP puncta fluorescence intensity at muscle 4 NMJs for genotypes in (C). N = 10, 10 NMJs. (E) RAB7:GFP at muscle 4 NMJ terminal boutons stained with anti-GFP in control (OK371-GAL4/+; UAS-RAB7:GFP/+) and Khc-73 (OK371-GAL4, Khc-73¹⁴⁹/+, Khc-73¹⁴⁹; UAS-RAB7:GFP/+). (F) Quantification of RAB7:GFP puncta fluorescence intensity at muscle 4 NMJs for genotypes in (E). N = 20, 17 NMJs. Scale bar is 5μm. Error Bars are SEM. Student’s t-test. **P<0.01.

Fig 13. Live imaging of Rab7:GFP at synaptic terminals.

(A) Montage of RAB7:GFP retrograde movement at muscle 4 NMJ in control (OK371-GAL4/+; UAS-RAB7:GFP/+) and Khc-73 (OK371-GAL4, Khc-73¹⁴⁹/+, Khc-73¹⁴⁹; UAS-RAB7:GFP/+) larvae. Scale bar is 5μm. Time is seconds. (B) Histogram of Anterograde velocities of RAB7:GFP puncta in (A). N = 18(42), 12(36) NMJs(puncta). (C) Histogram of Retrograde velocities of RAB7:GFP puncta in (A). N = 18(42), 12(36) NMJs(puncta). (D) Average anterograde velocity of RAB7:GFP puncta for genotypes in (A). N = 18(42), 12(36) NMJs(puncta). (E) Average retrograde velocity of RAB7:GFP puncta for genotypes in (A). N = 18(42), 12(36) NMJs(puncta). (F) Average time RAB7:GFP puncta spent in each pause event. For genotypes in (A). N = 18(42), 12(36) NMJs(puncta). (G) Average number of pauses per RAB7:GFP puncta.for genotypes in (A). N = 18(42), 12(36) NMJs(puncta). (H) Total time RAB7:GFP puncta remained paused within proximal axon for genotypes in (A). N = 18(42), 12(36) NMJs(puncta). (I) TKV-YFP expression in control (BG380-Gal4/+; UAS-TKV-YFP/+) and Khc-73 (BG380-GAL4/+; Khc-73¹⁴⁹; UAS-TKV-YFP/+) larvae in the NMJ axon of muscle 4. (J) Quantification of the number of stationary puncta observed within the axon from time lapse movies of genotypes in (I). N = 6, 8 NMJs. Scale bar is 5 μm. Error Bars are SEM. Student’s t-test. *P<0.05, **P<0.01, ***P<0.001. ns-no statistical significance.

Fig 14. Model slide for Khc-73 function at the synaptic terminal.

(A) In wild-type neuromuscular junctions, Khc-73 mediates the efficient transit of retrograde bound endosomes to the retrograde transport machinery. Its function could include enhancement of sorting to retrograde pathways and/or exit of endosomes from the MVB. (B) In the absence of Khc-73, retrograde bound endosomes do not efficiently engage with the retrograde machinery and accumulate in late endosome compartments. This inefficient level of routing is sufficient for basal levels of BMP retrograde signaling. However, when BMP signaling is enhanced, Khc-73 function is needed to mediate the exit of retrograde endosomes from the synapse. In its absence, retrograde bound signaling complexes are slowed in their exit from the synapse, they accumulate and signaling to the cell body is inhibited.