The p150(Glued) CAP-Gly domain regulates initiation of retrograde transport at synaptic termini

Abstract

p150(Glued) is the major subunit of dynactin, a complex that functions with dynein in minus-end-directed microtubule transport. Mutations within the p150(Glued) CAP-Gly microtubule-binding domain cause neurodegenerative diseases through an unclear mechanism. A p150(Glued) motor neuron degenerative disease-associated mutation introduced into the Drosophila Glued locus generates a partial loss-of-function allele (Gl(G38S)) with impaired neurotransmitter release and adult-onset locomotor dysfunction. Disruption of the p150(Glued) CAP-Gly domain in neurons causes a specific disruption of vesicle trafficking at terminal boutons (TBs), the distal-most ends of synapses. Gl(G38S) larvae accumulate endosomes along with dynein and kinesin motor proteins within swollen TBs, and genetic analyses show that kinesin and p150(Glued) function cooperatively at TBs to coordinate transport. Therefore, the p150(Glued) CAP-Gly domain regulates dynein-mediated retrograde transport at synaptic termini, and this function of dynactin is disrupted by a mutation that causes motor neuron disease.

Figure 1. Gl^G38S Causes a Partial Loss of Glued Function

(A) p150^G38S-HA shows markedly reduced cosedimentation with microtubules in vivo. High-speed supernatants (HSS) from adult flies expressing WT or G38S HA-tagged p150 (p150^HA) under control of hs-GAL4 were incubated with taxol and GTP to stabilize microtubules, and microtubules and associated proteins were purified using ultracentriguation through a 15% sucrose cushion (see Figure S1C). (B) Toxicity of p150 overexpression in fly neurons. Shown schematically are the gross phenotypes observed with elav^C155-GAL4-mediated overexpression of UAS-p150 transgenic lines (number of lines with given phenotype in box; see also Figures S1A and S1B). The color scheme indicates severity of phenotype, with blue representing no observable phenotype and red signifying larval lethality. (C) Targeting strategy for generating the G38S “knock-in” allele of Glued. P{donor-Gl^G38S} denotes incorporation of the targeting construct within a P-element onto the second chromosome. Dp(Gl, P{w⁺, Gl^G38S}) indicates a line in which the endogenous Glued locus (red) has incorporated the mutated Glued targeting construct (blue) to form a tandem duplication flanked by the w⁺ gene. The bottom line indicates the resolution of the duplication following induction of Cre-I, leading to generation of the Gl^G38S allele (see Supplemental Methods). (D) Western analysis using an antibody directed against the p150 C-terminus shows that protein levels are reduced in Gl^G38S animals. The antibody recognizes a doublet in whole adult flies and bodies, but primarily recognizes a single band of 150 kD in adult heads. (E) Complementation analyses demonstrate that the Gl^G38S allele is a partial loss-of-function (LOF) Glued allele. Gl^1-3 and Gl ^Δ22 are null alleles. T80-GAL4 (a ubiquitous driver) in trans with UAS-p150^WT rescues the first-instar larval lethality of Gl^G38S/Gl^1-3 and Gl ^Δ22/Gl^1-3 flies, though rescued adults are male and female sterile. An ~23 kb BAC (CH322-82J07) containing the entire Glued locus (BAC {Gl+}) fully rescues lethality and sterility of Glued animals. (F) Gl^G38S larvae have normal locomotor activity, as indicated by the number of gridlines crossed per minute. N=18 animals for each genotype. (G–H) Gl^G38S adults exhibit progressive climbing deficits (G) and early lethality (H); this phenotype is rescued by ubiquitous expression of p150^WT (p150^WT represents a transgenic UAS-p150^WT (line #3) that drives constitutive expression). N > 50 flies per genotype.

Figure 2. Terminal NMJ Bouton rather than Axonal Transport Phenotypes in Gl^G38S Animals

(A) Representative kymographs of axonal transport of late endosomes in larval segmental nerves revealed by Rab7:GFP expressed in motoneurons. CNS is to the left and the NMJ to the right. Time is on the Y axis (scale bar = 10 sec) and proceeds from top to bottom. (B) Population analysis of Rab7:GFP particles in Gl^G38S animals demonstrates a reduced frequency of stationary endosomes along axons. (C) The number of endosomes that move along axons per minute (flux) is not altered in Gl^G38S animals. (D) Analysis of Rab7:GFP kinetics demonstrates no difference in the velocity or processivity (pause frequency, duration, or reversal frequency) in Gl^G38S animals. N=7 animals; 24–48 vesicles per animal. (E) NMJs develop normally in Gl^G38S/Gl^Δ22 third-instar larvae, with normal appearance of active zones (labeled with Brp); however, TBs are swollen (arrows). Segment A6 is shown. (F) (top) Quantification of type Ib bouton number per hemisegment in proximal (A2–A3) and distal (A5–A6) segments in wild type and Gl^G38S/Gl^Δ22 animals. n = 10 animals each. (bottom) TB volume is increased in Gl^G38S/Gl^Δ22 animals compared with wild type, and the TB volume of distal segments is increased relative to proximal segments (p=0.004). n=17 synapses from 5 animals of each genotype. (G) p150^HA, when expressed in motor neurons under control of OK371-GAL4, is enriched at microtubule (+) ends at NMJ TBs (arrow) and throughout axons (asterisk). A muscle 6–7 type, segment A3, NMJ is shown; dotted line outlines a type Is NMJ, arrowhead labels a type Ib NMJ. p150 is also enriched in microtubule loops (arrowhead, Figure S3D). (H) TBs are enriched in microtubule (+) ends. Expression of the motor domain of Khc fused to GFP (Khc^Head:GFP) leads to expression of GFP specifically within the NMJ TB (see also Figure S3B and enlarged in inset). Scale bar = 5μm. * p < 0.05 ** p =0.0003

Figure 3. Endosomes Accumulate at Terminal NMJ Boutons in Gl^G38S Mutant Larvae

(A) Neuronal membranes (labeled with anti-HRP) accumulate within TBs (arrows) of proximal (A2) and distal (A6) segments in Gl^G38S/Gl^Δ22 larvae. (B) Co-overexpression of Rab7:GFP with p150^G38S in motor neurons using D42-GAL4 demonstrates colocalization of anti-HRP staining with Rab7:GFP. (C) TB accumulation of anti-HRP partially colocalizes with mCD8:GFP (labeling membranes), synaptotagmin:GFP (Syt:GFP), and APP:YFP and is also seen with RNAi-mediated knockdown of Glued (enhanced with coexpression of UAS-Dcr2). APP-YFP is often present in small anti-HRP+ puncta at TBs of wild type animals (arrows). (D–E) Quantitation of the TB anti-HRP accumulation phenotype in different genotypes (D) and with RNAi-mediated knockdown (E) of dynactin subunits (Glued (Gl), capping protein alpha (cpa), p62 (CG12042)) and dynein (Dynein heavy chain 64C (dhc), Dynein intermediate chain (dic), and Dynein light intermediate chain (dlic)). N > 6 animals (24 synapses) for each genotype. In Gl^G38S animals and with Gl-RNAi, the TB anti-HRP accumulation is more severe in distal segments. Scale bar = 10 μm.

Figure 4. Dynein is Mislocalized to the Terminal NMJ Bouton in Gl^G38S Mutants

(A) Dynein heavy chain (Dhc) is mislocalized to NMJ TBs (labeled with anti-HRP) in Gl^G38S mutant animals. (B) Quantification of Dhc fluorescence intensity in the TB of proximal (A2 and A3) and distal (A6) segments. (C and D) Quantification of Dhc TB accumulation indicated as mean % of Dhc (+) synapses. N > 40 synapses from 5 animals for each genotype. (E) Gl-RNAi larvae are of the genotype UAS-Dcr2/+; OK371-GAL4/UAS-Gl-RNAi; MHC:Sh-GFP. Sh:GFP expressed in muscles labels the post-synaptic density. Scale bar = 10 μm. * p < 0.01 ** p < 0.0001 n.s. = not significant

Figure 5. Kinesin Heavy Chain Functions Synergistically with Glued at NMJ Terminal Boutons

(A) A 50% reduction in gene dosage of the anterograde microtubule motor kinesin heavy chain (khc) markedly enhances lethality of Gl^G38S animals. (B) Gl^G38S/Gl^Δ22 animals rarely survive to pupal stages when heterozygous for null khc (khc⁸/+). (C) Overexpression of p150^WT in the eye with GMR-Gal4 (GMR > p150^WT) causes a severe rough eye phenotype that is suppressed with a reduction in khc dosage. SEM images of the eye are shown with enlargements below. Note the severe fusion of ommatidia and bristle disorganization in GMR > p150^WT/+ animals that is suppressed in GMR > p150^WT/khc⁸. (D) Khc is expressed diffusely in axons (asterisk) and the NMJ in wild-type animals, whereas it accumulates at TBs (arrows) of Gl^G38S animals. Inset shows magnification of boxed area (Scale bar = 1 μm) with Khc accumulation at the periphery of the TB. (E) Kinesin and Dynein colocalize at TBs of Gl^G38S/Gl^Δ22 animals. (F–G) Genetic interaction between khc and Glued at TBs. Graph shows percentage of NMJs from segments A2-A3 with TB accumulations of anti-HRP. N > 24 synapses from > 6 animals of each genotype. Scale bar = 10 μm (D) and 5 μm (E–F). * p = 0.003 ** p<0.0001

Figure 6. p150^G38S Disrupts Retrograde Transport at TBs and Neurotransmitter Release

(A and B) Retrograde synaptic dense core vesicle (DCV) transport from TBs is impaired by overexpression of p150^G38S. (A) Top panels: D42 control (left, elav^C155-GAL4/Y; UAS-ANF:GFP/D42-GAL4) larval NMJs distribute DCVs (marked by ANF:GFP) throughout synaptic boutons, whereas D42>p150^G38S (right, elav^C155-GAL4/Y; UAS-ANF:GFP/D42-GAL4, UAS-p150^G38S) larval NMJs accumulate DCVs at TBs. After adjusting contrast to visualize individual DCVs (bottom three panels), all boutons except for the TB were bleached (t=0) and single DCVs were imaged exiting the TB. Scale bar = 10 μm. (B) Quantitation of ANF:GFP intensity reveals a ~10x increase in DCVs in TBs of D42>p150^G38S animals. N=6 animals. (C) Quantitation of DCV transport from TBs (retrograde synaptic flux) demonstrates a significant impairment in retrograde DCV flux from the TB with disruption of dynactin. Left panel shows data from FRAP experiments whereas right panel shows SPAIM results. N>7 animals for (−) control (D42 alone), G38S (D42>p150^G38S), WT (elav^C155> p150^WT), and dominant negative “DN” (elavGS>p150^ΔC). *p < 0.05 ** p<0.01. (D and E) Electrophysiology at the third-instar larval NMJ demonstrates that EJP amplitude is reduced in Gl^G38S mutant animals (N=12 animals; * p=1x10⁻⁴; ** p=1x10⁻⁵). (F–I) Spontaneous neurotransmitter release (mEJP amplitude and frequency) is not affected in Gl^G38S larvae. (I) The reduction in Gl^G38S EJP amplitude is due to reduced quantal content (* p=0.03).

Figure 7. Aggregates within Motor Neurons and Dynein Mislocalization is Specific to p150 Mutations that cause HMN7B but not Perry Syndrome

(A) In S2 cells, HA-tagged p150^WT is diffusely present in the cytoplasm and partially colocalizes with microtubules (labeled with anti-tubulin). Mutant p150 proteins (G38S and G50R) form large cytoplasmic aggregates in cells. (B) human p150^WT localizes diffusely throughout the cytoplasm in Drosophila motor neurons when expressed with OK371-GAL4, whereas human p150^G59S forms large cytoplasmic aggregates. mCD8:GFP (mGFP) labels motor neuron membranes. (C) mRNA and protein expression of HA-tagged p150 in fly heads of transgenic lines using elav-GAL4. qRT-PCR shows that WT, G38S, G50A, and G50R transgenic lines all express similar increased levels of glued mRNA compared with drive alone, whereas the G38S^Hi line expresses much higher levels. Western shows that p150^G38S-HA is expressed at lower levels than p150^G50A and p150^G50R, which are expressed at lower levels than p150^WT. A p150^G38S-HA transgenic line with a duplication of the P-element insertion and a robust level of p150 expression (G38S^Hi) was used to control for protein levels. (D) HA-tagged p150^WT and p150^G50R are diffusely expressed in the cytoplasm of motor neurons when expressed with OK371-GAL4, whereas p150^G38S forms cytoplasmic aggregates. (E–F) Overexpression of Perry mutant forms of p150 (G50A and G50R) does not cause the dynein TB phenotype observed with overexpression of p150^G38S. Furthermore, Perry mutant forms of p150 can rescue the dynein mislocalization phenotype seen in Gl^G38S/Gl^Δ22 animals. Quantitation performed as in Figure 5C–D. Scale bar = 5 μm. * p < 0.01 ** p < 0.0001

Figure 8. Model for p150^Glued Function at Synaptic Termini

(1) Kinesin (Khc) is required for delivery of p150^Glued and dynein (Dhc) to microtubule (+) ends at NMJ Terminal Boutons (TBs). Dynamically elongating microtubules may explore the TB and become stabilized through interactions between MT (+) ends and the CAP-Gly domain of p150^Glued. (2) This interaction allows dynein to initiate (−) end-directed retrograde endosomal transport, possibly by capturing small endosomal vesicles that have recently undergone endocytosis. This function of p150 is specifically disrupted by mutations in the CAP-Gly domain that cause motor neuron disease (G59S). In Gl^G38S larvae, the microtubule-binding function of the CAP-Gly domain of p150 is disrupted at synaptic termini, leading to an accumulation of enlarged endosomal vesicles and the kinesin and dynein motors.