Introduction

Microtubules are long polar tube-like polymers composed of α/β-tubulin dimers that provide mechanical stability to cells and act as tracks for motor-based transport. Tubulin dimers are preferentially added, and also lost, at microtubule plus ends, whereas microtubule minus ends are often anchored at microtubule organizing centers (MTOCs), structures of diverse composition that nucleate and anchor microtubules. In dividing cells, centrosomes are the main MTOCs, but many non-dividing or differentiated cells have less well characterized non-centrosomal (nc)MTOCs (Akhmanova & Kapitein, 2022).

Given their intrinsic polarity, microtubules predefine neuronal polarity. In axons, all microtubules are oriented with their plus ends towards the distal synapses, thus enabling axon-specific transport via kinesin-1. In contrast, a large fraction of dendritic microtubules are oriented with their plus ends towards the soma ("plus end-in"), enabling dendrite-specific transport via other motors including dynein. This orientation is particularly evident in invertebrates, where dendritic microtubules in primary dendrite branches almost uniformly adopt the plus end-in orientation. Despite the recognized importance of microtubule organization for neuronal function, it is still poorly understood how neuronal microtubule organization is achieved and maintained.

The differential organization of neuronal microtubules also underlies important developmental processes. The peripheral nociceptive class IV dendritic arborization (c4da) neurons of Drosophila larvae prune their long and branched sensory dendrites at the onset of metamorphosis. Pruning involves severing of proximal dendrite regions and subsequent degeneration of distal fragments (Rumpf et al, 2017; Furusawa & Emoto, 2021). The presumptive severing sites are marked by the local loss of microtubules, which predisposes them for mechanical tearing (Krämer et al, 2023). This local microtubule disassembly is initiated by an increase in microtubule dynamics at the onset of the pupal stage (Herzmann et al, 2017). Due to their uniform plus end- in orientation, dendritic microtubules can then shrink from their plus ends in a coordinated fashion, leading to their local loss in proximal dendrite regions (Herzmann et al, 2018; Rumpf et al, 2019). In support of this model, conditions that cause misorientation of dendritic microtubules bring about pruning defects. For example, pruning defects are associated with mutations in the kinesin-2 motor, which guides newly polymerized microtubules along preexisting ones, thus ensuring the identical orientation (Mattie et al, 2010; Herzmann et al, 2018). Microtubule orientation and pruning defects are also caused by loss of the minus end binding protein patronin (Wang et al, 2019), the microtubule polymerase/nucleation factor Minispindles (Msps)/XMAP-215 (Tang et al, 2020) and the serine/threonine phosphatase PP2A (Rui et al, 2020; Wolterhoff et al, 2020), but the underlying mechanisms are not always clear. An additional mechanism for dendritic microtubule orientation was described in C. elegans PVD neurons, where the localization of dendritic MTOCs is important for the establishment of uniform plus end-in organization. Here, a prominent ncMTOC resides at the tips of growing dendrites, such that microtubules can only grow towards the soma (Liang et al, 2020; Harterink et al, 2018). This ncMTOC was shown to reside on Rab11-positive recycling endosomes, and Rab11 was important for microtubule orientation (Liang et al, 2020).

Spectraplakins are a family of actin/microtubule cytoskeletal crosslinker proteins with important roles in microtubule organization in differentiated cells (Sanchez et al, 2021; Sun et al, 2019; Voelzmann et al, 2017). Anchored to the actin cortex via an N-terminal actin binding domain, they can recruit microtubules either directly via a GAS2-related domain, or via binding sites for end binding proteins (Noordstra et al, 2016; Ning et al, 2016; Nashchekin et al, 2016) (Slep et al, 2005). The Drosophila spectraplakin Short stop (Shot) is known to bundle and stabilize microtubules in axons (Alves-Silva et al, 2012, Okenve-Ramos et al, 2024).

Here, we describe an important role for Shot in dendrites. Loss of Shot causes pruning defects caused by mixed orientation of dendritic microtubules. During early dendrite growth stages, Shot localizes to dendrite tips in an actin-dependent manner where it recruits factors cooperating with an early-acting, Rab11-dependent MTOC. Our data provide a mechanism for MTOC tip recruitment for the establishment of oriented microtubule arrays in dendrites.

Results

Shot is required for c4da neuron dendrite pruning

C4da neurons have long and branched dendrites at the third instar larval (L3) stage (Fig. 1 A) that are pruned at the onset of the pupal stage, such that by 16 hours after puparium formation (h APF), they have lost all their dendrites (Fig. 1 A’, G - I). In order to identify new regulators of c4da neuron dendrite pruning, we expressed candidate RNAis in c4da neurons under pickpocket-GAL4 and assessed their effects on dendrite morphology and pruning. This approach identified a dsRNA line targeting Shot, which encodes the sole Drosophila member of the spectraplakins, a family of large adaptor proteins with binding sites for actin and microtubules. C4da neurons expressing shot dsRNA did not display major dendritic phenotypes at the larval stage (Fig. 1 B), even though their dendritic field coverage seemed somewhat reduced compared to control neurons. In contrast, dendrites could still be seen attached to the soma at 16 h APF in almost half the c4da neurons expressing shot dsRNA (Fig. 1 B’, G - I), indicative of pruning defects. In order to confirm this result, we used Mosaic Analysis with a Repressible Cell Marker (MARCM) to generate c4da neuron clones homozygous for the loss-of-function allele shot³ (Kolodziej et al, 1995). shot³ mutant c4da neurons had shorter dendrites at the L3 stage compared to control MARCM c4da neuron clones (Fig. 1 C, D). At 16 h APF, most shot³ mutant neurons still had dendrites attached to the soma (Fig. 1 C’ D’, G - I). We also generated an sgRNA line targeting a common exon of all Shot splice isoforms and found that it caused significant c4da neuron dendrite pruning defects when expressed in c4da neurons in a sensitized shot³/+ heterozygous background (Fig. 1 E’ -F’, G - I).

Figure 1.

Dendrite pruning is the first step of c4da neuron metamorphic remodeling and is followed by the regeneration of new dendrites adapted to the adult stage. We noted that c4da neurons lacking Shot also failed to initiate dendrite regrowth at 72 h APF, when control c4da neurons had regrown long and branched dendrites (Figure S1). We conclude that Shot is broadly required for c4da neuron dendrite remodeling.

Shot is required for uniform dendritic microtubule orientation

We and others have previously shown that c4da neuron dendrite pruning depends on local microtubule disassembly in proximal dendrites (Herzmann et al, 2017). As Shot is a microtubule regulator, we assessed the state of dendritic microtubules by immunofluorescence against the microtubule-associated protein futsch/MAP1B. At 5 h APF, continuous futsch staining could often still be detected in proximal dendrites of shot knockdown neurons, while control neurons had large gaps their futsch signal in these regions (Fig. S2). The uniform plus end-in orientation of dendritic microtubules is particularly important for dendrite pruning as microtubules must shrink away from the cell body in a coordinated fashion (Herzmann et al, 2018; Rumpf et al, 2019, Wang et al, 2019). We therefore asked whether loss of Shot affects dendritic microtubule orientation by using EB1::GFP to monitor the microtubule growth direction. In primary dendrites of third instar control c4da neurons, EB1::GFP comets (the plus ends of growing microtubules) moved exclusively retrogradely towards the soma (Fig. 2 A, C).

Figure 2.

Upon shot knockdown, however, a significant proportion of EB1::GFP comets could be seen moving anterogradely (Fig. 2 A, C). In dendrites of shot³ c4da neuron MARCM clones, we observed an even stronger increase in anterogradely moving comets in comparison to control MARCM clones (Fig. 2 B, C). In support of a broad role for Shot in dendritic microtubule organization, EB1 comets in shot³ mutant c4da neurons also moved at an increased speed compared to controls (Fig. 2 D). We conclude that Shot is required for the correct orientation of dendritic microtubules, and that this function likely underlies its role in pruning.

Actin and microtubule binding are crucial for Shot function in dendrites

In order to better understand the function of Shot in dendritic pruning and microtubule orientation, we next asked which domains of Shot are crucial for these functions. Of particular interest were the domains of Shot that bind actin and microtubules. Shot has two Calponin Homology (CH) domains at the N-terminus that mediate binding to specific actin structures, and a microtubule-binding GAS2 domain in its C-terminal region (Fig. 3 A). Shot also possesses functionally important binding sites for end binding proteins such as Patronin and EB1. We next expressed full length or truncated Shot variants in shot³ mutant c4da neurons using the GAL4/UAS system and asked whether they could rescue the pruning defects of the mutant. A full length Shot UAS transgene (Shot^FL) fully rescued the pruning defects of shot³ mutant neurons (Fig. 3 B, C, F, G). In contrast, Shot^βCH1::GFP, a Shot variant lacking the first CH domain that has decreased affinity for actin (Lee & Kolodziej, 2002), afforded only a partial rescue (Fig. 3 D, F, G). Interestingly, overexpression of a C-terminal Shot fragment containing the microtubule binding domain and adjacent regions, but not the CH or spectrin repeat domains (UAS-Cterm::YFP), also fully rescued the pruning defects of mutant cells (Fig. 3 E - G). In further support of non-functionality, overexpression of Shot^βCH1 caused strong pruning defects (Fig. S3).

Figure 3.

In order to assess the functionality of these Shot variants in dendritic microtubule organization, we made use of kinesin-lacZ (kin::lacZ), a fusion between the kinesin-1 motor domain and β-galacosidase. In control MARCM c4da neurons, kinesin-lacZ localizes exclusively to the soma and axon (Fig. 3 H), as the kinesin motor travels towards microtubule plus ends. In shot³ mutant c4da neurons, axonal kin::lacZ was reduced, and it could instead be found in distinct puncta in dendrites, indicating the presence of plus end-out microtubules in these dendrites (Fig. 3 I, M). Re-expression of full length Shot, but not of Shot^βCH1::GFP, significantly rescued the number of dendritic kin::lacZ puncta (Fig. 3 J, K, M). As with the pruning defects, overexpression of Shot-Cterm::YFP restored dendritic exclusion of kin::lacZ (Fig. 3 L, M).

It had previously been suggested that Shot and its mammalian homologues can bundle and/or stabilize microtubules. In order to assess the effects of Shot loss on dendritic microtubule structure, we labeled microtubules in c4da neurons by expression of mcherry-tagged α-tubulin and used 2D stimulated emission depletion (STED) microscopy for visualization. In primary dendrites of control c4da neurons, mCherry- tagged microtubules appeared as several long filaments (Fig. S4). In dendrites of c4da neurons lacking Shot, fewer filamentous structures were visible and the staining was more punctate, possibly indicating that these neurons have fewer and/or damaged microtubules (Fig. S4). Thus, both microtubule and actin binding activities of Shot are important for correct dendritic microtubule orientation and pruning.

Shot cooperates with actin to promote microtubule orientation

The above domain analysis demonstrated that the actin binding ability of Shot is important for its function. Spectraplakins often bind to cortical actin structures underneath the plasma membrane. Consistently, we found that endogenously GFP- tagged Shot localized to the rim of the dendrite in third instar larval c4da neuron dendrites (Fig. 4 A, A’) in structured illumination microscopy (SIM) imaging. To directly probe for a role for actin in dendritic microtubule organization, we sought to disassemble actin filaments in vivo. The actin-severing enzyme Mical is a pruning factor that is induced at the onset of metamorphosis (Kirilly et al, 2009). We used the GAL4/UAS system to overexpress Mical prematurely in larval c4da neurons. In dendrites of c4da neurons overexpressing Mical, a significant fraction of EB1::GFP comets moved anterogradely, indicating mixed microtubule orientation (Fig. 4 B, C). Mical overexpression combined with Shot knockdown did not lead to an additive increase in the percentage of anterograde comets, indicating epistasis between the manipulations (Fig. 4 B, C). However, these anterograde comets moved significantly faster than their retrograde counterparts (Fig. 4 D). Furthermore, movement of the anterograde comets in this sample was often irregular, and comets frequently slowed down and then sped up again (Fig. 4 B). Such phenomena are often seen when microtubules are not attached to specific anchoring sites such that they can be moved by microtubule motors.

Figure 4.

Mical overexpression suppresses the pruning defects of ecdysone signaling mutants (Kirilly et al, 2009; Rode et al, 2018). However, we observed a synergistic enhancement of the pruning defects caused by Shot knockdown when we overexpressed Mical (Fig. 4 E). Thus, Shot’s role in dendrite pruning might in part be related to microtubule anchoring.

Shot is recruited to dendritic tips during early dendrite growth

Our analysis of Shot domains showed that the Shot CH1 domain is important for dendrite pruning and microtubule orientation (Fig. 3). This domain is often required to recruit Shot to specific subcellular locations (Nashchekin et al, 2024). To test whether the CH1 domain is required for localization to the dendrite membrane (Fig. 4 A), we visualized transgenic Shot or Shot^ΔCH1 in third instar c4da neurons using SIM. Similar to endogenously GFP-tagged Shot, transgenic full length Shot::GFP localized in puncta and sometimes streaks (especially around dendritic branchpoints) close to the plasma membrane (Fig. 5 A). Shot^ΔCH1::GFP localized in a similar pattern to full-length Shot (Fig. 5 B), indicating that the CH1 domain is unlikely to be required for Shot cortical localization in dendritic shafts.

Figure 5.

Since dendritic microtubule organization is set up early during development, we investigated the localization of transgenic Shot::GFP in c4da neurons at the first larval instar stage, when c4da neuron dendrites are growing very fast. At this stage, Shot::GFP was highly enriched at dendrite tips, but could only be detected at low levels in the soma and the more proximal dendritic shafts (Fig. 5 C, E). Shot tip enrichment seemed transient, as it was not observed in mature c4da neuron dendrites at the third instar (Fig. S5). In contrast to full length Shot, Shot^βCH1::GFP showed significantly lower tip occupancy (Fig. 5 D, F) and often localized in dots that could be several micrometers away from the tips (Fig. 5 D, E). To independently test whether Shot tip localization was actin-dependent, we used UAS-Mical for enzymatic actin disruption. Mical overexpression did not abolish Shot at tips, but caused a significant broadening of the tip signal, again indicating disturbed Shot recruitment (Fig. 5 G, H). Thus, efficient dendrite tip recruitment of Shot during the early stages of c4da neuron dendrite growth requires the CH1 domain.

Rab11 is a component of a developmental dendritic MTOC

The fact that Shot variants lacking the CH1 domain cannot rescue the pruning defects of shot³ mutants suggested that dendrite tip localization of Shot was important for its function. The tips of growing dendrites have recently been linked to microtubule organization in C. elegans PVD sensory neurons. Here, tip-localized Rab11-positive vesicles recruit a ncMTOC (Liang et al, 2020; Harterink et al, 2018). The composition and localization of dendritic MTOCs in Drosophila sensory (c4da and c1da) neurons are debated. Microtubules have been observed originating in dendritic shafts, at branchpoints and at tips, and both Golgi outposts and Rab5 endosomes have been implicated as sites harbouring MTOCs (Ori-McKenney et al, 2012; Nguyen et al, 2014; Mukherjee et al, 2020; Yalgin et al, 2015; Weiner et al, 2020). Most studies in c4da neurons have focused on mature dendrites in third instar neurons. We therefore asked whether the same factors and/or a similar mechanism as in PVD neurons also act in early developing c4da neurons.

Rab11 is required for c4da neuron dendrite pruning (Krämer et al, 2019; Lin et al, 2020). Given the strong dependence of pruning on microtubule organization, it is interesting to speculate that it might perform a microtubule-related function in early c4da neuron dendrites as well. We first tested whether Rab11 colocalizes with the bona fide MTOC component γ-tubulin. To this end, we coexpressed Rab11::mCherry with GFP-tagged γ-tubulin in c4da neurons. At the first instar, a subset of Rab11::mCherry puncta - likely vesicles - could be seen overlapping with γ- tubulin::GFP in discrete dendritic puncta that could also be found at dendrite tips (Fig. 6 A). To test whether such puncta could nucleate microtubules, we observed Rab11::mCherry together with EB1::GFP. In first instar neurons, EB1::GFP comets could be observed originating from dendritic Rab11::mCherry puncta (6/9 neurons, 10/58 comets) (Fig. 6 B, C). These comets were mostly seen in dendritic shafts or at branchpoints, likely because we could only capture a subset of EB1::GFP comets in our experiments. No EB1::GFP comets could be observed from Rab11::mCherry puncta at the third instar (0/9 neurons, 27 comets) (Fig. 6 C). To test for a functional role for Rab11 in dendritic microtubule organization, we knocked it down and visualized growing microtubules at the third instar stage. Rab11 knockdown caused a significant fraction of EB1 comets (16 %) to move anterogradely, and overall comet speed was increased (Fig. 6 D - F). In contrast, a dsRNA targeting Rab5 did not cause a significant increase in anterograde comets in c4da neurons, and comet speed was normal (Fig. 6 D - F).

Figure 6.

The microtubule polymerase Msps/XMAP215 is another potential MTOC component linked to pruning and dendritic microtubule organization (Tang et al, 2020). In co- immunoprecipitations from transfected S2 cells, GFP-tagged Rab11, but not GTP- binding-deficient Rab11 S25N, could be co-precipitated with FLAG-tagged Msps (Fig. 6 G). Lastly, the minus end-binder Patronin is required for dendrite pruning and localizes to growing dendrite tips in PVD neurons (Wang et al, 2019; He et al, 2020). Transgenic Patronin::GFP was also enriched at dendrite tips - as well as in streaks along dendrite shafts - in first instar c4da neurons (Fig. 6 H). Taken together, our data suggest the existence of a developmentally regulated, Rab11-based MTOC in early growing c4da neuron dendrites that likely also acts at tips.

Shot recruits EB1 to tips and interacts with candidate MTOC components

The above observations opened up the possibility that Shot is part of an early developmental dendritic tip MTOC. In support of this notion, loss of Shot already caused altered microtubule orientation at the first instar stage (Fig. 7 A). To test whether Shot can recruit the microtubule nucleation machinery, we coexpressed Shot::GFP with red fluorescently tagged EB1 in c4da neurons. In the presence of Shot::GFP, EB1 also became recruited to dendrite tips at the first instar and colocalized with Shot::GFP (Fig. 7 B, C), furthermore, EB1 comets could be seen emanating from these Shot/EB1-rich tips (Fig. 7 D). To link Shot to the other microtubule regulators implicated in early dendritic microtubule nucleation, we tested for genetic interactions using pruning as a readout. To this end, we knocked down shot in c4da neurons and then crossed in heterozygous P element mutants in Patronin and Msps, or a dsRNA construct targeting EB1. While these manipulations on their own did not cause pruning defects, the number of secondary and higher order dendrite branches attached to the soma upon was strongly increased when combined with shot knockdown, indicating phenotypic enhancement (Fig. 7 E - I). To assess genetic interactions with Rab11, we knocked down rab11, which by itself already causes highly penetrant pruning defects, but most higher order dendrites are still pruned and only primary dendrites remain (Krämer et al, 2019) (Fig. 7 J, L). In contrast, rab11 knockdown in a shot³/+ heterozygous background also caused significantly more higher order dendrites to stay attached to the soma than with than rab11 knockdown alone (Fig. 7 K, L). Thus, EB1 imaging and genetic interactions link Shot to dendritic MTOC functions.

Figure 7.

Discussion

Microtubule organization in neurites is crucial for neuronal polarity, but the developmental mechanisms are still emerging. Here, we show that the spectraplakin Shot is required for the uniform plus end-in organization of dendritic microtubules. Our data indicate that the role of Shot is twofold. On the one hand, we provide evidence that Shot anchors microtubules via actin in mature neurons. If unanchored, a subset of plus end-out oriented microtubules in Shot loss-of-function neurons can undergo sliding, which has been linked to microtubule misorientation (He et al, 2020). The observation that only plus end-out microtubules show this behavior suggests that there might be other, orientation-dependent anchoring mechanisms that do not depend on Shot. Surprisingly, overexpression of the GAS2 microtubule binding domain was sufficient to rescue the phenotypes of shot mutant neurons. Microtubule length regulation and catastrophe rescue have recently been shown to be important for maintenance of dendritic microtubule orientation (Liang et al, 2024). It will be interesting to test whether the GAS2 domain can fulfil a similar function in c4da neurons.

The other function of Shot is as part of an early dendritic MTOC. Similar to the C. elegans PVD neuron, c4da neurons have a transient microtubule-linked structure at the tip of early growing dendrites, and they depend on Rab11 for proper microtubule organization. Shot interacts genetically with Rab11 and is a component of this tip structure. Early-stage tip MTOCs and Rab11 dependence may therefore be a general feature of dendritic development. The exact function of Shot in the tip MTOC is still unclear, but the observation that Shot can recruit EB1 to tips suggests that it may recruit several MTOC components. The dendritic tip MTOCs of C. elegans were suggested to be transported to dendrite tips on transient plus end-out microtubules (Liang et al, 2020). Plus end-out microtubules can be observed in terminal c4da neuron dendrites (Ori-McKenney et al, 2012; Hu et al, 2022). It is interesting to speculate that Shot might participate in such a transport function. Crucially, both proposed functions of Shot during microtubule organization depend on actin binding, linking the two main cytoskeletal components in neurite organization.

Materials and methods

Fly Strains

C4da neurons were labeled by UAS-CD8::GFP or UAS-tdtomato expression under ppk-GAL4 (Grueber et al, 2007). MARCM clones of shot³mutants (BL 5141) were induced with SOP-FLP (Matsubara et al, 2011) and labeled by tdtomato expression under nsyb-GAL4^R57C10. Endogenously tagged Shot::GFP was from (Voelzmann et al, 2024). EB1 analyses were carried out using either UAS-EB1::GFP (Zheng et al, 2008), ppk-EB1::GFP (Arthur et al, 2015), or UAS-EB1::mScarlet3. The following dsRNA lines were used: Orco (BL 31278) as control, Shot (BL 28336), Rab11 (VDRC 22198), Rab5 (VDRC 34096), EB1 (VDRC 24451). All dsRNAs were coexpressed with UAS-dcr2 (Dietzl et al, 2007). UAS transgenes were UAS-Shot^RE::GFP on the X chromosome (BL 29044), or mobilized to the third chromosome (this study), UAS-Shot^RC::GFP (1CH1, BL 29042), UAS-Shot Cterm::YFP (Nashchekin et al, 2024), UAS-kin::lacZ (Clark et al, 1997), UAS-Mical (Terman et al, 2002), UAS-αtub84B::mCherry (Villars et al, 2022), UAS-Rab11::GFP (BL 8506), UAS-tau^HA (Herzmann et al, 2017), UAS-γtub23C::GFP (Nguyen et al, 2011), UAS-Patronin::GFP (Derivery et al, 2015), UAS- cytoABKAR TA (Marzano et al, 2021) as inert UAS control. Additional mutant alleles were msps^MI14162 (BL 59478), patronin^k07433 (Kyoto 111217).

Cloning and transgenes

Shot sgRNAs GCTGCCCTCTCAGGCCGATT (target 1) and GAGTTCTCCAGAGTGGTCAC (target 2) were cloned into pCFD4w+ (gift from S. Schirmeier). For transgenesis, plasmids were injected into flies carrying the 86Fb acceptor site. C-terminally GFP-tagged Rab11 for biochemistry was cloned into pUAST attB carrying a C-terminal GFP tag using TOPO/TA cloning. The S25N mutation was inserted by PCR mutatgenesis. For Rab11::mCherry, the Rab11 entry clone was used to generate pUAST attB Rab11::mCherry, the insert was subcloned into pUAST, which was then injected into w¹¹¹⁸ flies using classical transposase- mediated transgenesis. To generate EB1::mScarlet3, a Drosophila codon optimized mScarlet3 (Gaudella et al, 2023, synthesized by Twist Bioscience) was added after the EB1 coding sequence separated by a CACACCTCCACTACCGCTAGCAGGCCGGCCACGCGTGGTACCTTCTGGTCCA linker and cloned into pUAST. The resulting plasmid was injected into wild type embryos by Bestgene. C-terminally FLAG-tagged Msps (isoform RB) was cloned into pUAST attB using standard cloning procedures.

Immunofluorescence

Briefly, appropriately staged pupal filets were fixed in 4% formaldehyde, blocked in PBS with 0.3 % Triton X-100 and 10% goat serum, and incubated with antibody in blocking buffer over night. C4da neurons labeled by ppk promotor fusions, endogenous or transgenic GFP-tagged Shot were visualized with chicken (1:500, Aves labs), rabbit (1:1000, Invitrogen A11122) or mouse (1:1000, Invitrogen A1120) anti-GFP. Tdtomato and mcherry-tagged transgenes were detected with rat anti-mCherry (1:1000, Invitrogen) or rabbit anti-DsRed antibodies (1:1000, Clontech). Kin::β-galactosidase fusion proteins were detected with rabbit anti-β-galactosidase (Cappel, preabsorbed in-house, 1:700). Secondary antibodies for regular immunofluorescence were conjugated to Alexa Fluor 488, 568 or 647. For STED microscopy of tagged tubulin, Abberior anti rabbit STAR RED were used as secondary antibodies.

Microscopy, time lapse imaging

For analyses of dendrite pruning phenotypes at 16 h APF, animals were dissected out of the pupal case and dorsal ddaC c4da neurons in segments A2 - A5 were imaged live on a Zeiss LSM710 confocal microscope with a 20x Plan Apochromat water objective (1.0 NA). For regrowth analyses at 72 h APF, animals were dissected out of the pupal case and ventral v’ada c4da neurons in segments A2 - A4 were imaged as described (Sanal et al, 2023). For live protein localization analyses, first instar larvae were immobilized on double sided tape, third instar larvae were briefly anaesthesized with ether before imaging. Neurons were then imaged live on Zeiss LSM710 or LSM880 microscopes with 63x Plan Apochromat DIC M27 oil objectives (1.4 NA). EB1::GFP imaging was performed on a Zeiss LSM880 microscope using a 40x Plan Apochromat FCS M27 (1.2 NA) oil objective. Consecutive images of a single plane were taken every second for 1 - 2 minutes.

Immunofluorescence images of kin::lacZ were obtained on a Zeiss LSM 710 microscope using a 40x Zeiss C Apochromat water objective (1.1 NA). Structured illumination microscopy (SIM) was performed on a Zeiss Elyra 7 microscope with a 63x Plan Apochromat DIC M27 oil objective (1.4 NA) with an Edge 4.2 SIM camera (50 ms exposure time), and deconvolution was done using SIM² software. All microscopes used ZEN Black software. Stimulated Emission Depletion Microscopy (STED) images were taken with an abberior STEDYCON equipped with an inverted IX83 microscope (Olympus), a 100x oil objective (Olympus UPLXAPO100XO, NA 1.45), using pulsed excitation lasers at 640 nm and a pulsed STED laser operating at 775 nm, continuous autofocus and gated detection with avalanche photodiode element detectors. All acquisition operations were controlled by STEDYCON Software (abberior Instruments), and images were deconvolved using Huygens deconvolution software. Microscopic images shown are maximum projections or (where indicated) single plane images. All processing was done in Fiji (Schindelin et al, 2012) using the plug-ins NeuronJ for dendrite length measurements, and Image Stabilizer for EB1::GFP comet analysis. Kymographs from different channels were overlayed using the KymoResliceWide plugin.

S2 cell culture, immunoprecipitation and Western blots

pUAST Rab11::GFP (wt or S25N) was cotransfected into S2 cells with pUAST- Msps^FLAG and Actin5C-GAL4 as described (Herzmann et al, 2017). After 72 hours, cells were harvested in ice-cold PBS, and lysed in cold lysis buffer (100 mM NaCl, 5 mM MgCl2, 50 mM Tris/HCl pH 7.6, 5 % glycerol, 1 % Triton X-100, 1x complete protease inhibitor). Cleared lysates were incubated with anti-FLAG beads (Sigma A- 2220) for 2 hours. After three washes with lysis buffer, bound proteins were eluted in SDS sample buffer. Samples were run on 8 % gels and blotted with JL-8 anti-GFP (Clontech, 1:1000) and FLAG M2 (Sigma F3165, 1:5000) antibodies. Bands were visualized on an Amersham Imager 680.

Quantification and statistical analyses

Phenotypic penetrance was assessed by counting the number of neurons with dendrites still attached to the soma. Here, significance was determined using a categorical two-tailed Fisher’s exact test (graphpad.com). Length of unpruned dendrites were measured using the Fiji NeuronJ plugin and compared using the Wilcoxon Mann Whitney test (Marx et al., 2016).

In order to quantify Shot dendrite tip localization, intensity profiles of the distal 10 μm of each dendrite were averaged. The resulting curves were then sectioned into 2 μm segments for statistical comparisons between genotypes using the Mann Whitney test.

Supplementary figures

Figure S1

Figure S2

Figure S3

Figure S4

Figure S5

Acknowledgements

We thank T. Zobel and S. Weischer at the Münster Imaging Network for expert advice on superresolution microscopy, D. Luchtman (abberior) for help with 2D STED, C. Klämbt for support, G. Tavosanis, A. Prokop, J. Wildonger, D. Nashchekin, the Bloomington, VDRC, and Kyoto stock centers and Addgene for fly lines and reagents. Requests for UAS-EB1::mScarlet3 should be directed to Melissa Rolls (mur22@psu.edu). We thank R. Hube for technical help in the inital stages of the project, T. Klein-Höing for help with cloning and A. Ziegler for advice on early larval c4da neuron imaging. MD, NS and NW are members of the CiM IMPRS and CRC1348 graduate schools, respectively. This work was supported by DFG grants RU1673/4-1 and RU1673/6-1 to SR. The authors declare no competing financial interests.

Additional information

Author Contributions

M. Davies designed experiments, performed phenotypic analyses and contributed reagents. N. Sanal and N. Wolterhoff performed phenotypic analyses. U. Gigengack performed biochemical experiments and contributed reagents. Y. Shen and I. Hahn contributed reagents. S. Rumpf designed experiments, performed biochemical analyses and contributed reagents. S. Rumpf wrote the manuscript with input from all other authors.