ABSTRACT

Pruning that selectively removes unnecessary neurites without causing neuronal death is essential for sculpting the mature nervous system during development. In Drosophila, ddaC sensory neurons specifically prune their larval dendrites with intact axons during metamorphosis. However, the important role of endoplasmic reticulum (ER)-to-Golgi transport in dendrite pruning remains unknown. Here, in a clonal screen, we have identified Yif1, an uncharacterized Drosophila homolog of Yif1p that is known to be a regulator of ER-to-Golgi transport in yeast. We show that Yif1 is required for dendrite pruning of ddaC neurons but not for apoptosis of ddaF neurons. We further identify that the Yif1-binding partner Yip1 is also crucial for dendrite pruning. Yif1 forms a protein complex with Yip1 in S2 cells and ddaC neurons. Yip1 and Yif1 colocalize on ER/Golgi and are required for the integrity of Golgi apparatus and outposts. Moreover, we show that two GTPases, Rab1 and Sar1, which are known to regulate ER-to-Golgi transport, are essential for dendrite pruning of ddaC neurons. Finally, our data reveal that ER-to-Golgi transport promotes endocytosis and downregulation of the cell-adhesion molecule Neuroglian and thereby dendrite pruning.

INTRODUCTION

During animal development, neurons generate excessive cellular processes and connections at an earlier stage, and subsequently achieve accurate wiring via several regressive strategies (Schuldiner and Yaron, 2015). One of such strategies is pruning that eliminates unnecessary neurites and connections, without causing neuronal death at a later developmental stage (Luo and O'Leary, 2005; Yaniv and Schuldiner, 2016; Yu and Schuldiner, 2014). Pruning is a highly conserved process during the development of nervous systems in both vertebrates and invertebrates. In vertebrates, neurons first develop exuberant connections with many targets at embryonic stages and eliminate improper ones postnatally to establish functional connectivity (Schuldiner and Yaron, 2015). A typical example is stereotyped removal of axonal branches in layer 5 cortical neurons at an early postnatal stage in rats (Stanfield et al., 1982). In holometabolous insects, such as Drosophila, the nervous system undergoes dramatic remodeling to establish an adult-specific nervous system during metamorphosis: a transitional stage from larva to adult (Kanamori et al., 2015a; Truman, 1990). Two stereotyped pruning events occur in the central nervous system (CNS) or the peripheral nervous system (PNS) (Yu and Schuldiner, 2014). In the CNS, mushroom body γ neurons eliminate their larval axons and dendrites, and later re-extend their neurites to form adult-specific circuits (Lee et al., 1999). In the PNS, class I (ddaD/E) and class IV (ddaC) neurons, a subset of dorsal dendrite arborization (da) neurons, prune their larval dendrite arbors without affecting their axonal branches and subsequently regrow the adult-specific dendrites before eclosion (Kuo et al., 2005; Williams and Truman, 2005). Concomitantly, class II (ddaB) and class III (ddaA/ddaF) neurons are removed via apoptosis (Williams and Truman, 2005).

In Drosophila, ddaC sensory neurons have been established as a powerful paradigm for dissecting the molecular and cellular mechanisms underlying dendrite-specific pruning. In response to a late larval pulse of the steroid-molting hormone 20-hydroxyecdysone (ecdysone), ddaC neurons first sever the proximal region of their dendrites (severing step), then rapidly fragment the detached dendrites (fragmentation step) and clear the dendritic debris by phygocytosis (clearance step) (Fig. 1A) (Kuo et al., 2005; Williams and Truman, 2005). The dendrite-specific pruning events involve the formation of swellings and retracting bulbs, resembling the axon/dendrite degenerative process associated with brain injury and neurodegenerative diseases. Thus, understanding the mechanisms of developmental pruning would provide some insights into neurodegeneration in neuronal injury or diseases.

Extensive studies have revealed that ecdysone signaling pathway is the major gatekeeper of dendrite pruning. Upon the release at the late larval stage, ecdysone associates with the neuronal Ecdysone Receptor (EcR-B1) and its co-receptor Ultraspiracle (Usp) to activate downstream targets to initiate dendrite pruning (Yu and Schuldiner, 2014). We and others have previously reported that endocytosis is required to induce dendrite pruning (Kanamori et al., 2015b; Zhang et al., 2014). Via a Rab5/ESCRT-dependent endolysosomal degradation pathway, the L1-type cell-adhesion molecule Neuroglian (Nrg) is internalized from the cell surface to early endosomes and drastically downregulated to induce the onset of dendrite pruning (Zhang et al., 2014). More recently, we reported that Arf1/Sec71-dependent post-Golgi trafficking machinery governs dendrite pruning via Nrg endocytosis and downregulation (Wang et al., 2017).

The secretory pathway involves three elementary organelles: the endoplasmic reticulum (ER), the Golgi apparatus and the trans-Golgi network (TGN) (Rothman and Orci, 1992). Via the secretory pathway, protein and lipid supplies are provided to facilitate the specification and outgrowth of dendrites and axons during neuronal development (Valenzuela and Perez, 2015). In hippocampal neurons, the disruption of the secretory pathway through suppressing the post-Golgi trafficking or normal Arf1 function leads to a dendrite-outgrowth defect (Horton et al., 2005). Disruption of ER-to-Golgi transport by inhibiting Sar1 activity leads to shortened axons in mammalian neurons (Aridor and Fish, 2009). In Drosophila, the disruption of ER-to-Golgi transport by mutating Sar1 or Rab1 dramatically inhibits dendrite arbor elaboration with normal axonal elongation in sensory neurons (Ye et al., 2007). In addition to facilitating dendrites/axon outgrowth, the secretory pathway is also essential for the maintenance of dendrite arbors after neuron maturation (Horton et al., 2005). In stark contrast to its roles in neuronal growth and maintenance, we recently reported that Arf1/Sec71-mediated post-Golgi trafficking process is also crucial for dendrite pruning: a regressive event (Wang et al., 2017). However, a potential role for ER-to-Golgi transport in dendrite pruning remained elusive.

Here, we identify two novel genes, Yif1 and Yip1, that play important roles in dendrite pruning in ddaC sensory neurons during early metamorphosis. Their respective homologs are known to regulate ER-to-Golgi transport in yeast and mammals. We show that Yif1 associates with Yip1 in S2 cells and ddaC neurons. Moreover, Yip1 and Yif1 colocalize on the ER and cis-Golgi, and both are required for the integrity of the Golgi apparatus and outposts. We further demonstrate that the small GTPases Rab1 and Sar1, two key regulators of ER-to-Golgi transport, are also crucial for dendrite pruning of ddaC neurons. Importantly, our data indicate that the ER-to-Golgi transport promotes dendrite pruning partly via endocytosis and downregulation of the cell-adhesion molecule Neuroglian (Nrg). Thus, our data argue that some yet to be-identified molecules might be secreted into the dendrites to trigger Nrg internalization and dendrite pruning.

RESULTS

Yif1 is cell-autonomously required for dendrite pruning

To identify novel genes involved in dendrite pruning, we performed a large-scale clonal screen on the EMS-mutagenized third chromosome via Mosaic analysis with a repressible cell marker (MARCM) (Lee and Luo, 1999). From this screen, we isolated a mutant line, 10-46, that showed strong pruning defects by 16 h APF (n=16; Fig. 1E,J). An average of 14 primary and secondary dendrites remained attached to the soma in the homozygous ddaC clones of 10-46 (Fig. 1J). In contrast, wild-type ddaC neurons pruned all their dendrites (n=9; Fig. 1D,J). Overall dendrite termini of mutant ddaC neurons were significantly reduced in number at the wandering third instar larvae stage (wL3) compared with the wild type (n=3; Fig. S1). Using deficiency mapping and complementation tests with various lethal P-element insertion lines, we mapped 10-46 to a novel gene CG5484 (Fig. 1B). The following DNA sequencing analysis revealed a nonsense mutation in the coding region of CG5484 (Fig. S2A), producing a truncated protein with the first 71 deduced amino acids (Fig. 1C). CG5484 encodes an evolutionarily conserved transmembrane protein with 393 amino acid residues and shares the amino acid identity with yeast Yif1p and mammalian Yif1B (Fig. S2B). The Yif1 proteins contain five predicted transmembrane domains in its C-terminal region as well as a cytoplasmic domain in its N-terminal region (Matern et al., 2000) (Fig. 1C and Fig. S2B). Hence, we named the CG5484 gene as Yif1 and the 10-46 allele as Yif110-46.

Fig. 1.

Drosophila Yif1 is cell-autonomously required for dendrite pruning. (A) A schematic representation of dendrite pruning in ddaC neurons. (B) Yif110-46 failed to complement with Df(3R)BSC524 and Df(3R)BSC752, narrowing down the mutation between 97C3 and 97C5. P{EP}CG5484G19213, a P-element insertion in CG5484, failed to complement with Yif110-46. (C) The protein structure of Yif1 and the molecular lesion present in the Yif110-46 allele. (D-I) Live confocal images of ddaC neurons, visualized with the expression of UAS-mCD8GFP driven by ppk-Gal4. ddaC somas are indicated by red arrowheads. Whereas the wild-type ddaC neurons eliminated all the dendrites by 16 h APF (D), Yif110-46 MARCM (E), Yif1Δ25 MARCM (G) and Yif1Δ56 MARCM (H) ddaC clones exhibited strong dendrite pruning defects at 16 h APF. The pruning defects of Yif110-46 MARCM (F) and Yif1Δ56 MARCM (I) ddaC clones were rescued by the overexpression of a full-length Yif1 transgene. (J) Quantification of the number of primary and secondary dendrites attached to ddaC neurons. Number in each bar represents the sample number. Data are mean±s.e.m. N.S., not significant, ***P<0.001. Scale bar: 50 μm.

Fig. 1.

Drosophila Yif1 is cell-autonomously required for dendrite pruning. (A) A schematic representation of dendrite pruning in ddaC neurons. (B) Yif110-46 failed to complement with Df(3R)BSC524 and Df(3R)BSC752, narrowing down the mutation between 97C3 and 97C5. P{EP}CG5484G19213, a P-element insertion in CG5484, failed to complement with Yif110-46. (C) The protein structure of Yif1 and the molecular lesion present in the Yif110-46 allele. (D-I) Live confocal images of ddaC neurons, visualized with the expression of UAS-mCD8GFP driven by ppk-Gal4. ddaC somas are indicated by red arrowheads. Whereas the wild-type ddaC neurons eliminated all the dendrites by 16 h APF (D), Yif110-46 MARCM (E), Yif1Δ25 MARCM (G) and Yif1Δ56 MARCM (H) ddaC clones exhibited strong dendrite pruning defects at 16 h APF. The pruning defects of Yif110-46 MARCM (F) and Yif1Δ56 MARCM (I) ddaC clones were rescued by the overexpression of a full-length Yif1 transgene. (J) Quantification of the number of primary and secondary dendrites attached to ddaC neurons. Number in each bar represents the sample number. Data are mean±s.e.m. N.S., not significant, ***P<0.001. Scale bar: 50 μm.

To determine whether Yif1 is indispensable for dendrite pruning, we generated two additional Yif1 mutant alleles, Yif1Δ25 and Yif1Δ56, by imprecise excision of a P-element insertion line, P{EP}CG5484G19213. These two alleles uncover the majority of the Yif1-coding region, as well as N-terminal parts of its neighboring gene (Fig. S2C). Consistently, homozygous ddaC neurons of either Yif1Δ25 (n=9; Fig. 1G) or Yif1Δ56 (n=11; Fig. 1H) also displayed strong pruning defects to an extent similar to that in Yif110-46 clones. Yif1Δ25 and Yif1Δ56 mutant neurons retained an average of 13.4 and 12.1 primary and secondary dendrites attached to the soma, respectively (Fig. 1J). Moreover, the expression of a Yif1-Venus transgene into Yif110-46 (n=13; Fig. 1F) or Yif1Δ56 (n=8; Fig. 1I) ddaC neurons fully rescued their pruning defects.

Similar to ddaC neurons, wild-type ddaD/E sensory neurons pruned away their dendrites by 19 h APF (n=4; Fig. S3A). Yif110-46 ddaD/ddaE neurons failed to completely remove their dendrites at the same time point (n=13; Fig. S3A), suggesting that Yif1 is important for dendrite pruning of different sensory neurons. Wild-type ddaF neurons were eliminated by apoptosis at the early pupal stage (n=3; Fig. S3B). However, Yif110-46 mutant ddaF neurons were also eliminated (n=5; Fig. S3B), suggesting that Yif1 is dispensable for apoptosis of ddaF neurons. Collectively, these data demonstrate a cell-autonomous and essential role of Yif1 in dendrite pruning of da sensory neurons.

Yif1-Venus localizes on ER/Golgi compartments and Yif1 regulates the integrity of Golgi apparatus

To investigate the subcellular localization of Yif1, we generated several antibodies against four different regions of Yif1. However, none of the antibodies was able to detect any specific subcellular localization of the endogenous Yif1 protein in da sensory neurons. We therefore generated transgenes expressing Yif1 fused with Venus fluorescent protein at its C terminus (Yif1-Venus). The expression of Yif1-Venus fully rescued the pruning defects in Yif110-46 and Yif1Δ56 mutant ddaC neurons (Fig. 1F,I,J), suggesting that Yif1-Venus functionally substitutes for endogenous Yif1. In wild-type ddaC neurons, Yif1-Venus distributed as numerous punctate structures and partially overlapped with Sec31-mCherry, an ER exit site marker (n=14; Fig. 2A, Table 1). Importantly, the Yif1-Venus puncta more closely localized with GM130, a cis-Golgi marker (n=10; Fig. 2B, Table 1). Likewise, Yif1-Venus also colocalized with Rab1 (n=7; Fig. 2C, Table 1), the mammalian counterpart of which was reported to interact and colocalize with GM130 on the cis-Golgi compartments (Weide et al., 2001). Moreover, Yif1-Venus puncta were juxtaposed with two trans-Golgi markers, such as ADP-ribosylation factor 1-GFP (Arf1-GFP) (n=12; Fig. 2D, Table 1) (Wang et al., 2017) and galactosyltransferase-GFP (GalT-GFP) (n=11; Fig. S4A,E). However, the Yif1-Venus punctate pattern was different from those of the early endosomal marker GFP-Rab5 (n=8; Fig. S4B,E), the mitochondrial marker mito-GFP (n=9; Fig. S4C,E) or the recycling endosomal marker Rab4-mRFP (n=10; Fig. S4D,E). Thus, these findings suggest that Yif1-Venus localizes mainly on cis-Golgi and partly on ER.

Fig. 2.

Yif1 localizes on ER/Golgi compartments and regulates the integrity of Golgi apparatus in ddaC neurons. (A-D,E-G) Confocal images of wild-type and mutant ddaC neurons expressing UAS-Yif1-Venus, UAS-Sec31-mCherry, UAS-Arf1-GFP or UAS-mCD8GFP driven by ppk-Gal4 at late wL3 stage. ddaC somas are outlined with dashed lines. Insets show the magnified view of punctate structures. Line profiles show the arbitrary fluorescence intensity along the white lines. a.u. denotes arbitrary unit. Yif1-Venus puncta partially overlapped with Sec31-mCherry (A). Most of the Yif1-Venus puncta were colocalized with GM130 (B) or Rab1 (C) staining but juxtaposed with Arf1-GFP (D). Yif1-Venus signals were detected in live larvae (A,D) and in fixed tissues using an GFP antibody (B,C). Similar to the pattern in wild-type ddaC neurons (A), Sec31-mCherry still distributed as many discrete puncta in Yif110-46 MARCM ddaC clones (E). Unlike those in the wild type, the GM130 (F) and Arf1 (G) signals were absent in Yif110-46 MARCM ddaC clones. (H-I′) Confocal images of wild-type or mutant ddaC neurons expressing both UAS-mCD8GFP and UAS-ManII-Venus driven by ppk-Gal4. Although ManII-Venus exhibited punctate pattern in wild type (H), it became diffuse and showed no punctate localization in Yif110-46 neurons (I). The Golgi outposts, labelled by ManII-Venus, displayed discrete punctate pattern in the dendrites of wild-type neurons (H′); they became dim and diffuse in the dendrites of Yif110-46 neurons (I′). Scale bars: 10 μm.

Fig. 2.

Yif1 localizes on ER/Golgi compartments and regulates the integrity of Golgi apparatus in ddaC neurons. (A-D,E-G) Confocal images of wild-type and mutant ddaC neurons expressing UAS-Yif1-Venus, UAS-Sec31-mCherry, UAS-Arf1-GFP or UAS-mCD8GFP driven by ppk-Gal4 at late wL3 stage. ddaC somas are outlined with dashed lines. Insets show the magnified view of punctate structures. Line profiles show the arbitrary fluorescence intensity along the white lines. a.u. denotes arbitrary unit. Yif1-Venus puncta partially overlapped with Sec31-mCherry (A). Most of the Yif1-Venus puncta were colocalized with GM130 (B) or Rab1 (C) staining but juxtaposed with Arf1-GFP (D). Yif1-Venus signals were detected in live larvae (A,D) and in fixed tissues using an GFP antibody (B,C). Similar to the pattern in wild-type ddaC neurons (A), Sec31-mCherry still distributed as many discrete puncta in Yif110-46 MARCM ddaC clones (E). Unlike those in the wild type, the GM130 (F) and Arf1 (G) signals were absent in Yif110-46 MARCM ddaC clones. (H-I′) Confocal images of wild-type or mutant ddaC neurons expressing both UAS-mCD8GFP and UAS-ManII-Venus driven by ppk-Gal4. Although ManII-Venus exhibited punctate pattern in wild type (H), it became diffuse and showed no punctate localization in Yif110-46 neurons (I). The Golgi outposts, labelled by ManII-Venus, displayed discrete punctate pattern in the dendrites of wild-type neurons (H′); they became dim and diffuse in the dendrites of Yif110-46 neurons (I′). Scale bars: 10 μm.

Table 1.

Colocalization percentage of Yif1-Venus with other cellular markers

Colocalization percentage of Yif1-Venus with other cellular markers
Colocalization percentage of Yif1-Venus with other cellular markers

The subcellular localization of Yif1 is consistent with the role of its yeast and mammalian counterparts (Yif1p and Yif1B, respectively) as regulators of ER-to-Golgi transport (Alterio et al., 2015; Matern et al., 2000). We next investigated whether Yif1 is important for the integrity of ER or Golgi, or both. Interestingly, Sec31-mCherry puncta remained, although the number of puncta was slightly reduced (n=5; Fig. 2E). Notably, GM130 and Arf1 puncta were largely disrupted in Yif110-46 homozygous ddaC clones (n=7; Fig. 2F,G). Moreover, another Golgi marker ManII-Venus became diffuse in the soma of Yif110-46 homozygous ddaC clones (n=8; Fig. 2I), compared with its discrete puncta in wild-type ddaC neurons (n=10; Fig. 2H). Dendritic Golgi outposts that are labelled by ManII-Venus were also disrupted in Yif110-46 neurons (n=9; Fig. 2I′), compared with those in the wild type (n=10; Fig. 2H′). Taken together, Yif1-Venus localizes on ER/cis-Golgi compartments and Yif1 is primarily required for the integrity of the Golgi apparatus.

The N-terminal 90 amino acid fragment of Yif1 is dispensable for its function in dendrite pruning

A previous study reported that mammalian Yif1B, via its N-terminal 50 amino acid region, interacts with the serotonin G-protein-coupled 5-HT1A receptor and targets it to dendrites in hippocampal neurons (Al Awabdh et al., 2012). We then investigated whether the N-terminal region or other domains of Drosophila Yif1 are required for its function during dendrite pruning. To achieve this, we generated transgenes expressing various truncated Yif1 proteins (Fig. 3A).

Fig. 3.

The N-terminal fragment of Yif1 (1-90 amino acids) is dispensable for its function in dendrite pruning. (A) The schematic diagram of four truncated Yif1 variants. (B-F) Live confocal images of ddaC neurons visualized by ppk-Gal4-driven mCD8GFP. ddaC somas are indicated by red arrowheads. Yif110-46 MARCM ddaC clones showed strong dendrite pruning defect at 16 h APF (B). The overexpression of Yif11-195-Venus (C), Yif1196-393-Venus (D) or Yif11-323-Venus (E) failed to rescue the pruning defects in Yif110-46 MARCM ddaC clones. On the other hand, the overexpression of Yif191-393-Venus fully rescued the pruning defects in Yif110-46 ddaC neurons at 16 h APF (F). (G) Quantification of the number of primary and secondary dendrites attached to ddaC neurons. Number in each bar represents the sample number. Data are mean±s.e.m. Scale bar: 50 μm.

Fig. 3.

The N-terminal fragment of Yif1 (1-90 amino acids) is dispensable for its function in dendrite pruning. (A) The schematic diagram of four truncated Yif1 variants. (B-F) Live confocal images of ddaC neurons visualized by ppk-Gal4-driven mCD8GFP. ddaC somas are indicated by red arrowheads. Yif110-46 MARCM ddaC clones showed strong dendrite pruning defect at 16 h APF (B). The overexpression of Yif11-195-Venus (C), Yif1196-393-Venus (D) or Yif11-323-Venus (E) failed to rescue the pruning defects in Yif110-46 MARCM ddaC clones. On the other hand, the overexpression of Yif191-393-Venus fully rescued the pruning defects in Yif110-46 ddaC neurons at 16 h APF (F). (G) Quantification of the number of primary and secondary dendrites attached to ddaC neurons. Number in each bar represents the sample number. Data are mean±s.e.m. Scale bar: 50 μm.

The expression of the N-terminal half (1-195 amino acids) of Yif1 failed to rescue the pruning defect in Yif110-46 mutant ddaC neurons (n=15; Fig. 3B,C,G). The expression of the C-terminal region (196-393 amino acids) or the fifth transmembrane domain-deleted Yif1 variant (1-323 amino acids) did not rescue the pruning defect in Yif110-46 mutant neurons (n=16 and 12, respectively; Fig. 3D,E,G). These data suggest that both N-terminal and C-terminal regions are required for Yif1 to achieve its full function. When we expressed, in Yif110-46 mutant neurons, the truncated Yif1 lacking the first 90 amino acids, which is a region similar to the N-terminal 50 amino acids of mammalian Yif1B (Al Awabdh et al., 2012), dendrite pruning was fully restored (n=10; Fig. 3F,G), suggesting that Drosophila Yif1 regulates dendrite pruning independently of a potential 5-HT1A receptor-binding domain (see Discussion). As controls, these truncated proteins did not disturb dendrite pruning when expressed in wild-type neurons (data not shown).

Moreover, we assessed the colocalization of these Venus-tagged Yif1 variants with GM130 in ddaC neurons. Neither the N-terminal half nor the C-terminal half of Yif1 colocalized with GM130 on cis-Golgi, instead they accumulated in the nucleus or the cytoplasm, respectively (Fig. S5A,B). The Yif1 variant (1-323 amino acids) that lacks the fifth transmembrane domain still exhibited puncta structure and largely colocalized with GM130 (Fig. S5C); nevertheless, it failed to rescue Yif110-46 mutant defects (Fig. 3E). The Yif1 variant deleting the first 90 amino acids almost fully colocalized with GM130 (Fig. S5D), indicating that this region is dispensable for Yif1 localization and function.

The rescue experiments and localization assays indicate that the N-terminal 90 amino acid fragment of Yif1, which was reported to be crucial for the binding and dendritic target of the 5-HT1A receptor in mammalian neurons (Al Awabdh et al., 2012), is not important for dendrite pruning in Drosophila sensory neurons.

Yip1 is essential for dendrite pruning but dispensable for apoptosis of da sensory neurons

Yeast Yif1p was initially identified as a Yip1p-interacting factor (Matern et al., 2000). Yip1p localizes on Golgi apparatus and interacts with Ypt1p/Rab1 and Ypt31p/Rab11 to facilitate ER-to-Golgi transport (Yang et al., 1998). We identified CG12404, a Drosophila homolog of Yip1p, which shares high amino acid identity with yeast Yip1p and mammalian Yip1A (Fig. S6A). Thus, we named CG12404 as Yip1. Drosophila Yip1 encodes a conserved protein with four possible transmembrane domains at its C terminus (Fig. S6A). To examine whether Yip1 is important for dendrite pruning, we generated Yip1 mutants using CRISPR/Cas9 technology (Port et al., 2014). Two guide RNAs (gRNAs) were designed to target the first exon of Yip1 (Fig. 4A). Two independent mutants, Yip11-21 and Yip12-20, were recovered, which harbor small deletions at the expected positions (Fig. 4A). Notably, homozygous ddaC clones of either the Yip11-21 or Yip12-20 mutant exhibited strong pruning defects with an average of 11.5 and 12 primary and secondary dendrites attached to the soma by 16 h APF, respectively (n=14 and 12, respectively; Fig. 4C,D,G). In the controls, all dendrites were pruned away at this time point (n=11; Fig. 4B,G). When the CYFP-tagged full-length Yip1 protein was expressed in Yip11-21 or Yip12-20 mutant ddaC clones, the pruning defects were completely rescued (n=7 and 8, respectively; Fig. 4E-G), highlighting that loss of Yip1 function is responsible for the dendrite pruning defects in these Yip1 mutants.

Fig. 4.

Yip1 is required for dendrite pruning in sensory neurons. (A) The schematic diagram of the genome sequence of Yip1 loci, positions of gRNA-Yip1-targeting sites and deleted regions in two CRISPR/Cas9-induced Yip1 mutants. (B-F) Live confocal images of ddaC neurons labelled with mCD8GFP driven by ppk-Gal4. ddaC somas are indicated by red arrowheads. Unlike the wild-type neurons that completely pruned all dendrites (B), the Yip11-21 MARCM (C) or Yip12-20 MARCM (D) ddaC clones failed to prune the dendrites by 16 h APF. The overexpression of a full-length Yip1 transgene rescued the pruning defect in Yip11-21 MARCM (E) and Yip12-20 MARCM (F) ddaC clones. (G) Quantification of the number of primary and secondary dendrites attached to ddaC neurons. Number in each bar represents the sample number. Data are mean±s.e.m. N.S., not significant, *P<0.05, ***P<0.001. Scale bar: 50 μm.

Fig. 4.

Yip1 is required for dendrite pruning in sensory neurons. (A) The schematic diagram of the genome sequence of Yip1 loci, positions of gRNA-Yip1-targeting sites and deleted regions in two CRISPR/Cas9-induced Yip1 mutants. (B-F) Live confocal images of ddaC neurons labelled with mCD8GFP driven by ppk-Gal4. ddaC somas are indicated by red arrowheads. Unlike the wild-type neurons that completely pruned all dendrites (B), the Yip11-21 MARCM (C) or Yip12-20 MARCM (D) ddaC clones failed to prune the dendrites by 16 h APF. The overexpression of a full-length Yip1 transgene rescued the pruning defect in Yip11-21 MARCM (E) and Yip12-20 MARCM (F) ddaC clones. (G) Quantification of the number of primary and secondary dendrites attached to ddaC neurons. Number in each bar represents the sample number. Data are mean±s.e.m. N.S., not significant, *P<0.05, ***P<0.001. Scale bar: 50 μm.

Initial dendrite arborization was also affected in Yip1 mutant neurons, similar to that in Yif1 mutants. The dendrite arbors were significantly simplified in Yip11-21 mutant neurons at the wL3 larvae stage (n=3), compared with those in the wild type (n=3; Fig. S6B). Furthermore, the ddaD/ddaE neurons failed to prune away the dendrites by 19 h APF in the Yip11-21 mutant (n=10; Fig. S3A). Yip11-21 ddaF neurons, like wild-type neurons (n=3), died via apoptosis (n=3; Fig. S3B). Taken together, Yip1, like Yif1, governs dendrite pruning but not apoptosis of sensory neurons.

Yip1 associates with Yif1 on Golgi

We next demonstrated whether Drosophila Yif1 associates with Yip1, as do their yeast homologs. First, we carried out co-immunoprecipitation (co-IP) experiments in S2 cells transfected with Flag-tagged Yif1 and Myc-tagged Yip1. Yif1 was specifically detected in the immune complex when Myc-Yip1 was immunoprecipitated from the protein extracts of S2 cells using an anti-Myc antibody (n=3; Fig. 5A), suggesting that Yif1 and Yip1 form the same complex in Drosophila. Yif1 and Yip1 were also present in the same complex in S2 cells expressing Arf1DN (Fig. S7A). Second, we took advantage of the Bimolecular Fluorescence Complementation (BiFC) tool (Gohl et al., 2010) to examine their in vivo association. In this assay, Yif1 and Yip1 were fused to the non-fluorescent N-terminal half (NYFP) and C-terminal half (CYFP) of a yellow fluorescent protein (YFP), respectively. When NYFP-Yif1 and CYFP-Yip1 were co-transfected and expressed in S2 cells, strong BiFC signals were observed (Fig. S7B). In contrast, no fluorescent signal was detectable in three control experiments (Fig. S7B). These data suggest a specific in vivo association in S2 cells. To further investigate whether the Yif1-Yip1 association occurs in ddaC neurons, we generated transgenes expressing either NYFP-Yif1 or CYFP-Yip1. Notably, we also detected strong YFP fluorescent signals when NYFP-Yif1 and CYFP-Yip1 were simultaneously expressed in ddaC neurons via the ppk-Gal4 driver (Fig. 5E, Fig. S7C). No fluorescent signal was detectable in the control experiments where NYFP-Yif1 and CYFP, CYFP-Yip1 and NYFP, or CYFP and NYFP were co-expressed in ddaC neurons (Fig. 5B-D, Fig. S7C). The BiFC signals also distributed as discrete punctate structures in the dendrites (Fig. 5F). These discrete BiFC signals largely overlapped with the cis-Golgi marker GM130 (colocalization percentage, 71.30±14.8%, n=8; Fig. 5G). These data indicate that the Yif1-Yip1 association can occur on Golgi in ddaC neurons. Thus, Yif1 and Yip1 form a protein complex in S2 cells and ddaC neurons, and the association occurs on the Golgi.

Fig. 5.

Yip1 associates with Yif1 in S2 cells and ddaC neurons. (A) Yip1 associates with Yif1. S2 cells were co-transfected with Myc-Yip1 and Flag-Yif1. Anti-Myc antibody was used in immunoprecipitation followed by immunoblotting with anti-Flag antibody. (B-F) Confocal images of wild-type ddaC neurons expressing different Split-YFP fusion protein combinations and UAS-mCD8mCherry driven by ppk-Gal4 at late wL3 stage. ddaC somas are outlined with dashed lines. Whereas the co-expression of NYFP and CYFP (B), NYFP-Yif1 and CYFP (C) or NYFP and CYFP-Yip1 (D) failed to show any BiFC signal, the co-expression of NYFP-Yif1 and CYFP-Yip1 resulted in the reconstitution of YFP and showed strong punctate YFP fluorescence signals in both soma (E) and dendrites (F). (G) Confocal images of wild-type ddaC neurons expressing NYFP-Yif1 and CYFP-Yip1 driven by ppk-Gal4 at late wL3 stage. The majority of the BiFC signals (71.30±14.8%) colocalized with GM130 staining in ddaC neuron (n=8). Inset shows a magnified view of punctate structures. Line profiles show the arbitrary fluorescence intensity along the white lines. a.u. denotes arbitrary unit. Scale bar: 10 μm.

Fig. 5.

Yip1 associates with Yif1 in S2 cells and ddaC neurons. (A) Yip1 associates with Yif1. S2 cells were co-transfected with Myc-Yip1 and Flag-Yif1. Anti-Myc antibody was used in immunoprecipitation followed by immunoblotting with anti-Flag antibody. (B-F) Confocal images of wild-type ddaC neurons expressing different Split-YFP fusion protein combinations and UAS-mCD8mCherry driven by ppk-Gal4 at late wL3 stage. ddaC somas are outlined with dashed lines. Whereas the co-expression of NYFP and CYFP (B), NYFP-Yif1 and CYFP (C) or NYFP and CYFP-Yip1 (D) failed to show any BiFC signal, the co-expression of NYFP-Yif1 and CYFP-Yip1 resulted in the reconstitution of YFP and showed strong punctate YFP fluorescence signals in both soma (E) and dendrites (F). (G) Confocal images of wild-type ddaC neurons expressing NYFP-Yif1 and CYFP-Yip1 driven by ppk-Gal4 at late wL3 stage. The majority of the BiFC signals (71.30±14.8%) colocalized with GM130 staining in ddaC neuron (n=8). Inset shows a magnified view of punctate structures. Line profiles show the arbitrary fluorescence intensity along the white lines. a.u. denotes arbitrary unit. Scale bar: 10 μm.

CYFP-Yip1 colocalizes with Yif1-Venus on ER/Golgi compartments, and their localizations are co-dependent

To investigate the localization of Yip1 in ddaC neurons, we produced several antibodies against Yip1. Similar to anti-Yif1 antibodies, Yip1 antibodies could not detect endogenous protein but exhibited prominent punctate signals in CYFP-Yip1-overexpressing neurons. The CYFP-Yip1 protein is functional, as its expression fully rescued the pruning defects in Yip11-21 and Yip12-20 (Fig. 4E,F). The immunostaining with anti-Yip1 indicated that CYFP-Yip1 distributed as many punctate signals in ddaC neurons (Fig. 6). Those puncta colocalized partly with Sec31-mCherry (n=12; Fig. 6A, Table 2) and primarily with GM130 (n=14; Fig. 6B, Table 2); however, were adjacent to Arf1-GFP (n=13; Fig. 6C, Table 2). Moreover, CYFP-Yip1 overlapped with Yif1-Venus when both were co-expressed (n=13; Fig. 6D, Table 2). Therefore, CYFP-Yip1 is colocalized with Yif1-Venus on ER/cis-Golgi compartments.

Fig. 6.

Yip1 colocalizes with Yif1 on ER/Golgi compartments and their localizations are co-dependent. (A-D,E-G) Confocal images of wild-type and mutant ddaC neurons expressing UAS-CYFP-Yip1, UAS-Sec31-mCherry, UAS-Arf1-GFP, UAS-Yif1-Venus or UAS-mCD8GFP driven by ppk-Gal4 at late wL3 stage. ddaC somas are outlined with dashed lines. Insets show magnified views of punctate structures. Line profiles show the arbitrary fluorescence intensity along the white lines. a.u. denotes arbitrary unit. (A) CYFP-Yip1 partially colocalized with Sec31-mCherry. CYFP-Yip1 puncta largely colocalized with GM130 staining (B), but they were adjacent to Arf1-GFP signals (C). (D) The CYFP-Yip1 signals that were visualized with an GFP antibody overlapped with Yif1-Venus. Similar to the pattern in wild type (A), Sec31-mCherry expression remained punctate in Yip11-21 MARCM ddaC clones (E). Unlike those in the wild type, GM130 (F) or Arf1 (G) punctate staining were disrupted in Yip11-21 MARCM ddaC clones. (H-K) Confocal images of wild-type or mutant ddaC neurons at late wL3 stage immunostained using anti-GFP (in green) or anti-Yip1 (in red). Although Yif1-Venus exhibited a punctate pattern in wild-type neurons (H), it became diffuse in Yip11-21 MARCM ddaC clones (I). CYFP-Yip1, which exhibited punctate pattern in the wild type (J), became delocalized in Yif110-46 MARCM ddaC clones (K). Scale bar: 10 μm.

Fig. 6.

Yip1 colocalizes with Yif1 on ER/Golgi compartments and their localizations are co-dependent. (A-D,E-G) Confocal images of wild-type and mutant ddaC neurons expressing UAS-CYFP-Yip1, UAS-Sec31-mCherry, UAS-Arf1-GFP, UAS-Yif1-Venus or UAS-mCD8GFP driven by ppk-Gal4 at late wL3 stage. ddaC somas are outlined with dashed lines. Insets show magnified views of punctate structures. Line profiles show the arbitrary fluorescence intensity along the white lines. a.u. denotes arbitrary unit. (A) CYFP-Yip1 partially colocalized with Sec31-mCherry. CYFP-Yip1 puncta largely colocalized with GM130 staining (B), but they were adjacent to Arf1-GFP signals (C). (D) The CYFP-Yip1 signals that were visualized with an GFP antibody overlapped with Yif1-Venus. Similar to the pattern in wild type (A), Sec31-mCherry expression remained punctate in Yip11-21 MARCM ddaC clones (E). Unlike those in the wild type, GM130 (F) or Arf1 (G) punctate staining were disrupted in Yip11-21 MARCM ddaC clones. (H-K) Confocal images of wild-type or mutant ddaC neurons at late wL3 stage immunostained using anti-GFP (in green) or anti-Yip1 (in red). Although Yif1-Venus exhibited a punctate pattern in wild-type neurons (H), it became diffuse in Yip11-21 MARCM ddaC clones (I). CYFP-Yip1, which exhibited punctate pattern in the wild type (J), became delocalized in Yif110-46 MARCM ddaC clones (K). Scale bar: 10 μm.

Table 2.

Colocalization percentage of CYFP-Yip1 with other cellular markers

Colocalization percentage of CYFP-Yip1 with other cellular markers
Colocalization percentage of CYFP-Yip1 with other cellular markers

We next assessed whether Yip1, like Yif1, regulates Golgi integrity. The ER marker Sec31-mCherry largely remained punctate in the Yip11-21 mutant clones (n=5; Fig. 6E), similar to that in wild-type neurons (Fig. 6A). Importantly, the Golgi markers GM130 (n=9; Fig. 6F) and Arf1 (n=7; Fig. 6G) were disrupted in Yip11-21 mutant neurons. Thus, Yip1, like Yif1, is required for Golgi integrity. Moreover, in Yip11-21 neurons, Yif1-Venus became diffuse (n=4; Fig. 6I), compared with the discrete puncta in wild-type neurons (n=10; Fig. 6H). Similarly, whereas CYFP-Yip1 was localized as numerous punctate structures in wild-type neurons (n=8; Fig. 6J), these puncta dispersed in Yif110-46 mutant neurons (n=10; Fig. 6K). Thus, the subcellular localizations of Yif1 and Yip1 are co-dependent. However, we cannot exclude the possibility that their localization co-dependence is secondary to the disruption of Golgi structure caused by loss of either gene.

To examine whether Yif1 or Yip1 regulate the secretory pathway, we used a GFP-tagged version of Sec15 (Sec15-GFP), a conserved component of the exocyst, to visualize a subset of secretory vesicles in sensory neurons (Jafar-Nejad et al., 2005; Wang et al., 2017). In wild-type ddaC neurons, Sec15-GFP showed punctate signals that were separable from either Yif1 (n=11; Fig. S8A) or Yip1 signals (n=12; Fig. S8B). Notably, Sec15-GFP puncta were dramatically reduced in number and size in Yif110-46 mutant clones (n=8; Fig. S8C). Thus, Yif1 is important for proper formation of secretory vesicles in sensory neurons. We next investigated whether Yif1 and Yip1 regulate protein transport to the plasma membrane. To achieve this, we carried out a trafficking assay in which we measured the levels of extracellular mCD8 epitope at wL3 larval stage in a detergent-free condition (Murthy et al., 2003). In wild-type ddaC neurons, the extracellular mCD8 signals were prominent and the mCD8/GFP ratio reached more than 0.7 (n=13; Fig. S9A,E). Similar to that in Sec71DN mutant neurons (n=5; Fig. S9B,E) (Wang et al., 2017), the extracellular mCD8 epitope intensity became significantly weaker in Yif110-46 (n=12; Fig. S9C,E) or Yip11-21 (n=10; Fig. S9D,E) MARCM clones with the mCD8/GFP ratio reduced to ∼0.4. Collectively, CYFP-Yip1 and Yif1-Venus colocalize on the ER/Golgi; Yip1 and Yif1 regulate both Golgi integrity and protein secretion in ddaC sensory neurons.

Rab1 and Sar1 are required for dendrite pruning in ddaC neurons

It has been reported that yeast Yif1p and Yip1p interact with Ypt1p/Rab1 and facilitate the ER-to-Golgi transport (Matern et al., 2000; Yang et al., 1998). We show here that, in ddaC neurons, Yif1 colocalized with Rab1 (Fig. 2C), a key regulator of the COPII-mediated ER-to-Golgi transport (Lee et al., 2004). Importantly, punctate Rab1 signals were abolished in Yif1 or Yip1 mutants (n=15 and 13, respectively; Fig. S10A). We next tested whether Rab1 is involved in dendrite pruning. We first made use of a lethal P-element insertion line, Rab1e01287, which is inserted to the first exon of Rab1. Rab1e01287 mutant ddaC neurons exhibited notable pruning defects at 16 h APF (n=8; Fig. S10C,E). Second, the expression of Rab1 RNAi construct in ddaC neurons showed consistent pruning defects (n=18; Fig. S10D,E). Moreover, via the gene-switch system, we induced the expression of Rab1S25N, a Rab1 dominant-negative form (Nuoffer et al., 1994), at the middle third instar larval stage. Induced Rab1S25N expression led to normal formation of elaborate dendrite arbors at the WP (white prepupal) stage but inhibited dendrite pruning at 16 h APF (n=20; Fig. 7B). On average, 8.6 primary and secondary dendrites remained attached to the soma in the mutant neurons (Fig. 7I). In contrast, the ddaC neurons pruned their dendrites normally in non-induced controls (n=16; Fig. 7A,I). The effect of Rab1S25N is probably due to its being locked in a GDP-bound state. It has been reported that overexpression of wild-type mammalian Rab1 rescues the inhibitory effect of the Rab1S25N mutant on ER-to-Golgi transport (Nuoffer et al., 1994). Importantly, the pruning defects in Rab1S25N-overexpressing ddaC neurons were fully rescued by the co-expression of wild-type Rab1 (n=16; Fig. 7D,I) but not the expression of a control transgene (n=15; Fig. 7C,I), suggesting a specific effect of Rab1S25N in dendrite pruning. Thus, these data demonstrate that Rab1 is crucial for dendrite pruning of ddaC neurons.

Fig. 7.

Rab1 and Sar1 are required for dendrite pruning in ddaC neurons. (A-H) Live confocal images of ddaC neurons visualized by ppk-tdGFP. ddaC somas are indicated by red arrowheads. Although the wild-type ddaC neurons removed all dendrites at 16 h APF (A), induced expression of Rab1S25N via the RU486-inducible gene-switch system led to strong pruning defects by 16 h APF (B). When Rab1S25N and UAS-control were co-overexpressed via the gene-switch system, ddaC neurons displayed strong pruning defects and the phenotype was comparable with Rab1S25N single induction (C). These pruning defects caused by Rab1S25N overexpression were fully rescued by Rab1 co-overexpression (D). Similarly, unlike the control neurons (E), the induction of Sar1T34N overexpression (F) or Sar1T34N and UAS-control overexpression (G) via the gene-switch system caused strong pruning defects. These pruning defects were rescued by the reintroduction of the Sar1 protein (H). (I) Quantification of the number of primary and secondary dendrites attached to ddaC neurons. Number in each bar represents the number of neurons examined. Data are mean±s.e.m. N.S., not significant, ***P<0.001. Scale bar: 50 μm.

Fig. 7.

Rab1 and Sar1 are required for dendrite pruning in ddaC neurons. (A-H) Live confocal images of ddaC neurons visualized by ppk-tdGFP. ddaC somas are indicated by red arrowheads. Although the wild-type ddaC neurons removed all dendrites at 16 h APF (A), induced expression of Rab1S25N via the RU486-inducible gene-switch system led to strong pruning defects by 16 h APF (B). When Rab1S25N and UAS-control were co-overexpressed via the gene-switch system, ddaC neurons displayed strong pruning defects and the phenotype was comparable with Rab1S25N single induction (C). These pruning defects caused by Rab1S25N overexpression were fully rescued by Rab1 co-overexpression (D). Similarly, unlike the control neurons (E), the induction of Sar1T34N overexpression (F) or Sar1T34N and UAS-control overexpression (G) via the gene-switch system caused strong pruning defects. These pruning defects were rescued by the reintroduction of the Sar1 protein (H). (I) Quantification of the number of primary and secondary dendrites attached to ddaC neurons. Number in each bar represents the number of neurons examined. Data are mean±s.e.m. N.S., not significant, ***P<0.001. Scale bar: 50 μm.

We further analyzed potential involvement of Sar1, another regulator of the COPII-mediated ER-to-Golgi transport (Lee et al., 2004), in dendrite pruning. Sar111-3-63 mutant ddaC clones formed their simplified dendrite arbors at the WP stage; moreover, they exhibited pruning defects by 16 h APF (n=15; Fig. S11B,E). Consistently, the expression of Sar1 RNAi #1 or #2 in ddaC neurons also led to the defects in dendrite pruning (n=18 and 18, respectively; Fig. S11C-E). Moreover, via the gene-switch system, we pulse induced the expression of Sar1T34N, which is a GDP-locked form of Sar1 and can constitutively inhibit wild-type Sar1 function (Ivan et al., 2008), in ddaC neurons. Although the non-induced ddaC neurons eliminate their dendrites (n=12; Fig. 7E,I), the induction of Sar1T34N did not affect normal dendrite morphology at the WP stage but led to strong pruning defects by 16 h APF (n=17; Fig. 7F). On average, Sar1T34N neurons retained ∼10.6 primary and secondary dendrites that were attached to the soma (Fig. 7I). Similar to Rab1S25N, the pruning defects caused by Sar1T34N overexpression in ddaC neurons were fully restored by the co-expression of wild-type Sar1 (n=26; Fig. 7H,I) but not by the expression of a control transgene (n=14; Fig. 7G,I). Thus, two key regulators of the COPII-mediated ER-to-Golgi transport are also crucial for dendrite pruning.

Yif1/Yip1-mediated ER-Golgi transport is crucial for the downregulation of the cell-adhesion molecule Nrg prior to dendrite pruning

Previously, we reported that the downregulation of L1-CAM Nrg through Rab5/ESCRT-dependent endocytic pathway is a prerequisite step to initiate dendrite pruning in ddaC neurons (Zhang et al., 2014). In wild-type ddaC neurons, Nrg protein was redistributed to early endosomes at 6 h APF, and the overall protein levels were significantly decreased in the somas, dendrites and axons (Fig. 8A,G, Fig. S12A) (Zhang et al., 2014). However, in Rab5DN-expressing ddaC neurons, the Nrg protein accumulated in the somas, dendrites and axons (Fig. 8B,G), compared with the wild-type control (Fig. 8A) (Zhang et al., 2014). We examined whether the Yif1/Yip1-mediated ER-to-Golgi transport is required for Nrg downregulation prior to dendrite pruning. Intriguingly, the Nrg protein accumulated dramatically in the somas, dendrites and axons of Yif110-46 (Fig. 8C,G) or Yip11-21 (Fig. 8D,G) ddaC neurons, similar to those in Rab5DN-expressing neurons (Fig. 8B). Moreover, the Nrg protein levels also increased in the somas, dendrites and axons of Rab1S25N (Fig. 8E,G) or Sar1T34N (Fig. 8F,G) gene-switch ddaC neurons. These data indicate that ER-to-Golgi transport facilitates Nrg downregulation prior to dendrite pruning. We then investigated whether normal ER-to-Golgi transport is important for endosomal localization of Nrg at the onset of dendrite pruning. In wild-type ddaC neurons, Nrg colocalized with the early endosomal marker GFP-2×FYVE at 6 h APF (n=15; Fig. S12A). However, Nrg was rarely redistributed as discrete puncta and did not overlap with GFP-2×FYVE signals in Rab1S25N (n=6; Fig. S12B) or Sar1T34N gene-switch ddaC neurons (n=9; Fig. S12C). Thus, these data suggest that the ER-to-Golgi transport promotes endocytosis and downregulation of Nrg protein prior to dendrite pruning.

Fig. 8.

ER-Golgi protein transport is crucial for the downregulation of the cell-adhesion molecule Nrg prior to dendrite pruning. (A-F) Confocal images of wild-type and mutant ddaC neurons at 6 h APF immunostained using anti-Nrg antibody (in red). ddaC somas are outlined with dashed lines, the proximal dendrites are marked by curly brackets and the ddaE somas are marked with asterisks. Although Nrg exhibited a punctate expression pattern in wild-type neurons (A), it accumulated significantly in the somas, dendrites and axons of Rab5DN-overexpressing (B), Yif110-46 MARCM (C), Yip11-21 MARCM (D), Rab1S25N-overexpressing (E) and Sar1T34N-overexpressing (F) ddaC neurons. (G) The quantification of Nrg immunostaining intensities in somas, dendrites and axons of ddaC neurons. (H-J,L-N) Live confocal images of ddaC neurons labelled with mCD8GFP driven by ppk-Gal4. Although the Yif110-46 MARCM ddaC clones showed a strong pruning defect by 16 h APF (H), knocking down Nrg in Yif110-46 ddaC clones via Nrg RNAi #1 (I) or Nrg RNAi #2 (J) significantly attenuated pruning defects by 16 h APF. Compared with those defects in Yip11-21 MARCM ddaC clones (L), knocking down Nrg in Yip11-21 ddaC clones via Nrg RNAi #1 (M) or Nrg RNAi #2 (N) significantly suppressed pruning defects by 16 h APF. (K,O) Quantification of the number of primary and secondary dendrites attached to ddaC neurons. Number in each bar represent the sample number. Data are mean±s.e.m. ***P<0.001. Scale bars: 10 μm in A; 50 μm in H.

Fig. 8.

ER-Golgi protein transport is crucial for the downregulation of the cell-adhesion molecule Nrg prior to dendrite pruning. (A-F) Confocal images of wild-type and mutant ddaC neurons at 6 h APF immunostained using anti-Nrg antibody (in red). ddaC somas are outlined with dashed lines, the proximal dendrites are marked by curly brackets and the ddaE somas are marked with asterisks. Although Nrg exhibited a punctate expression pattern in wild-type neurons (A), it accumulated significantly in the somas, dendrites and axons of Rab5DN-overexpressing (B), Yif110-46 MARCM (C), Yip11-21 MARCM (D), Rab1S25N-overexpressing (E) and Sar1T34N-overexpressing (F) ddaC neurons. (G) The quantification of Nrg immunostaining intensities in somas, dendrites and axons of ddaC neurons. (H-J,L-N) Live confocal images of ddaC neurons labelled with mCD8GFP driven by ppk-Gal4. Although the Yif110-46 MARCM ddaC clones showed a strong pruning defect by 16 h APF (H), knocking down Nrg in Yif110-46 ddaC clones via Nrg RNAi #1 (I) or Nrg RNAi #2 (J) significantly attenuated pruning defects by 16 h APF. Compared with those defects in Yip11-21 MARCM ddaC clones (L), knocking down Nrg in Yip11-21 ddaC clones via Nrg RNAi #1 (M) or Nrg RNAi #2 (N) significantly suppressed pruning defects by 16 h APF. (K,O) Quantification of the number of primary and secondary dendrites attached to ddaC neurons. Number in each bar represent the sample number. Data are mean±s.e.m. ***P<0.001. Scale bars: 10 μm in A; 50 μm in H.

To ascertain whether ER-to-Golgi transport promotes dendrite pruning at least partly through Nrg downregulation, we knocked down Nrg, via two distinct RNAi lines, in Yif110-46 or Yip11-21 mutants. The expression of these two RNAi lines, which has been shown to efficiently knock down Nrg protein (Zhang et al., 2014), significantly suppressed the pruning defects of Yif110-46 (n=17 and 12; Fig. 8H-K) or Yip11-21 (n=17 and 21; Fig. 8L-O) mutant ddaC neurons. Thus, our results suggest that Yif1 and Yip1 act to facilitate Nrg endocytosis and downregulation to promote dendrite pruning likely in an indirect manner. To further strengthen it, we conducted double mutant analyses for Rab5 and Rab1/Sar1. Induced expression of Rab5DN at the late larval stage via the gene-switch system led to notable accumulation of ubiquitylated protein deposits on two or three enlarged endosomes, which were positive for ubiquitin (Fig. S13A) and the early endosomal marker Avl (Fig. S13F) at 6 h APF, suggesting a defect in endosomal fusion and maturation. These data are consistent with the known functions of Rab5 because Rab5 is dispensable for initial formation of endocytic vesicles but crucial for subsequent endosomal fusion and maturation. Neither ubiquitin nor Avl appeared to accumulate in Rab1S25N- (Fig. S13B,G) or Sar1T34N-expressing (Fig. S13D,I) ddaC neurons at 6 h APF. Importantly, in contrast to the Rab5DN single mutant, Rab1S25N and Rab5DN double mutants did not exhibit any ubiquitylated aggregates (Fig. S13C) and enlarged Avl-positive endosomes (Fig. S13H) in 6h-APF ddaC neurons, resembling Rab1S25N single-mutant neurons. These double-mutant phenotypes imply that Rab1 may facilitate initial formation of endocytic vesicles and promote endocytosis likely upstream of Rab5. Similar to Rab1S25N and Rab5DN double mutants, Sar1T34N and Rab5DN double-mutant neurons did not form ubiquitylated aggregates (Fig. S13E) or Avl-positive endosomes (Fig. S13J), resembling the Sar1T34N single mutant. Taken together, these genetic data further suggest that ER-to-Golgi transport may generally facilitate Rab5-dependent endocytosis prior to dendrite pruning in ddaC neurons.

DISCUSSION

Drosophila Yif1 and Yip1 are evolutionarily conserved transmembrane proteins that regulate the Golgi integrity

In the present study, we have identified for the first time two evolutionarily conserved Yif1 and Yip1 proteins in Drosophila, and demonstrate their important roles in dendrite pruning during metamorphosis. Drosophila Yif1 and Yip1 belong to the same protein family as yeast Yif1p/Yip1p. Yeast Yip1p was first identified as a Ypt1p/Rab1-interacting protein (Yang et al., 1998) and Yif1p was further discovered as a Yif1p-interacting factor (Matern et al., 2000). Via their C-terminal regions, Yif1p and Yip1p form a heteromeric integral membrane complex. These two proteins localize on the Golgi membranes and regulate ER-to-Golgi protein trafficking and secretion (Matern et al., 2000; Yang et al., 1998). Yip1p-Yif1p was also reported to regulate the fusion process between ER-derived vesicles and the Golgi apparatus (Barrowman et al., 2003). This complex was proposed to serve as a receptor on the Golgi membranes for vesicle docking and fusion. In mammals, Yif1B was discovered as a binding partner of the 5-HT1A serotonin receptor and localized on intermediate compartments of the Golgi (Carrel et al., 2008). Partial knockdown of Yif1B specifically disturbs the targeting of the 5-HT1A receptor to the distal dendrites (Carrel et al., 2008), whereas complete knockout of Yif1B leads to the disruption of Golgi integrity (Alterio et al., 2015). Likewise, mammalian Yip1 proteins were also discovered as trafficking regulators between ER exit sites, intermediate compartment and cis-Golgi (Tang et al., 2001; Yoshida et al., 2008).

Here, we show that Drosophila Yif1 and Yip form a protein complex in vivo and are functionally relevant during dendrite pruning. First, these two proteins associate in both S2 cells and ddaC sensory neurons, as revealed in co-IP experiments and BiFC assays. Second, they colocalize on ER and cis-Golgi, and importantly their localizations are mutually dependent. Third, both of them are required for the integrity of Golgi apparatus, similar to the COPII-mediated ER-to-Golgi trafficking regulators, such as Rab1 or Sar1 (Storrie et al., 1998; Wilson et al., 1994). Finally, removal of either of them caused the same phenotypes in terms of dendrite pruning, identical to those observed in Rab1 or Sar1 mutant neurons. Thus, it is plausible to suggest that Drosophila Yif1 and Yip1, like their homologs, can regulate the ER-to-Golgi transport to promote dendrite pruning. Consistent with our findings, previous genome-wide RNAi screens reported Yif1 as a potential component of the secretory pathway (D'Ambrosio and Vale, 2010) that might be involved in neural outgrowth and morphology (Sepp et al., 2008).

The ER-to-Golgi protein transport is prerequisite for dendrite pruning in Drosophila

Growing evidence indicates that the secretory pathway has been shown to regulate dendrite growth and maintenance in both Drosophila and mammals (Kennedy and Ehlers, 2011). The ER-to-Golgi transport is an early step of the canonical secretory pathway (Rothman and Orci, 1992). Suppression of the ER-Golgi transport inhibits dendrite outgrowth in the developing neurons (Horton et al., 2005; Ye et al., 2007). Either disrupting Golgi apparatus or attenuating Golgi outposts alters dendrite morphology and outgrowth in mature neurons (Horton et al., 2005; Ye et al., 2007). Our previous study illustrated that the small GTPas Arf1, which is known to regulate post-Golgi trafficking in mammals, is important for proper dendrite pruning of ddaC sensory neurons (Wang et al., 2017). Moreover, this present study describes Yif1 and Yip1 as two new regulators of dendrite pruning in sensory neurons. Yif1 and Yip1 appear to localize on ER and cis-Golgi, which is compatible with their roles in the ER-to-Golgi protein transport like their yeast and mammalian homologs (Alterio et al., 2015; Matern et al., 2000; Yang et al., 1998). Yif1 and Yip1 can facilitate the biogenesis of secretory vesicles and affect the proper Golgi structure in ddaC neurons. Consistently, we illustrate that Rab1 and Sar1, which are two key components of the COPII vesicles, are also essential for dendrite pruning. Thus, it is tempting to hypothesize that some yet to be-identified molecules might be specifically secreted into the dendrites to promote dendrite pruning. Interestingly, mammalian Yif1B functions as a scaffold protein to recruit the 5-HT1A receptor together with Yip1A and Rab6, probably in the same trafficking vesicles and specifically targets the receptor to the distal dendrites in rat neurons (Al Awabdh et al., 2012). However, our results suggest that this machinery might not be required for dendrite pruning of ddaC neurons. First, although the N-terminal 50 amino acid region of mammalian Yif1B is required for targeting the 5-HT1A receptor (Al Awabdh et al., 2012), we found that a similar N-terminal fragment of Yif1 is dispensable for dendrite pruning because the N-terminal deleted Yif1 transgenes completely rescued the pruning defects in Yif110-46 mutant neurons. In addition, the ddaC neurons underwent dendrite pruning normally in mutant ddaC neurons derived from a null Rab6 mutant or RNAi expression (n=6 and 12, respectively; data not shown). Future work would focus on the identification of secreted molecules that trigger dendrite pruning.

The ER-Golgi transport facilitates endocytosis and downregulation of Nrg during pruning

In this study, we show that ER-Golgi transport facilitates Nrg endocytosis prior to dendrite pruning at prepupal stage. It is also possible that ER-Golgi transport directly regulates the secretion of Nrg towards the plasma membrane at larval stages. Because of a lack of an anti-Nrg antibody against its extracellular domain, we were not able to examine this possibility. To bypass its early role in protein secretion, Rab1DN or Sar1DN expression was induced to temporally inhibit the EG-to-Golgi transport at late larval stage, when Nrg exocytosis was completed under normal ER-to-Golgi transport. We provided multiple lines of evidence indicating that the ER-Golgi transport facilitates Nrg endocytosis and downregulation before the onset of dendrite pruning. First, the blockade of ER-Golgi transport caused by loss of Yif1/Yip1 function or Rab1DN/Sar1DN induction leads to a significant increase of Nrg proteins in the somas, dendrites and axons, similar to that in Rab5 mutant neurons. Moreover, when the ER-Golgi transport was inhibited upon Rab1DN/Sar1DN induction, Nrg was no longer redistributed to endosomes, indicating defective Nrg endocytosis. Finally, the pruning defects caused by loss of Yif1 or Yip1 function were significantly suppressed by Nrg knockdown. We hypothesize that, in response to activation of ecdysone signalling, some unknown cell-surface or secreted molecules might be secreted via the canonical secretory pathway to facilitate Nrg endocytosis and thereby promote dendrite pruning. Over 1000 cell-surface or secreted molecules exist in Drosophila (Kurusu et al., 2008). Future studies will be necessary to identify such secreted molecules as potential ‘eat me’ signals for initiating Nrg endocytosis and dendrite pruning.

L1-CAM Nrg is the only known transmembrane protein that is endocytosed during dendrite pruning in ddaC neurons (Zhang et al., 2014). It is conceivable that ER-to-Golgi transport likely generally affects endocytosis of various transmembrane proteins. A previous study has elegantly showed that an increase of general endocytosis occurs prior to dendrite pruning; loss of Rab5 function results in blockade of general endocytosis and thereby a dendrite pruning defect (Kanamori et al., 2015b). Loss of Rab5 function leads to general endocytosis defects, including the formation of enlarged Avl-positive endosomes in ddaC neurons. These Rab5 mutant phenotypes were suppressed by the blockade of ER-to-Golgi transport, suggesting that aberrant ER-to-Golgi transport generally affects endocytosis, in addition to Nrg endocytosis. Thus, ER-to-Golgi transport may also generally facilitate Rab5-dependent endocytosis during dendrite pruning of ddaC sensory neurons.

MATERIALS AND METHODS

Fly strains

The following fly stocks were used in this study: UAS-GFP-2×FYVE, UAS-GFP-Rab5 and UAS-Rab5DN (M. Gonzalez-Gaitan, Geneva University, Switzerland) (Wucherpfennig et al., 2003), UAS-MicalN-ter (A. Kolodkin, Johns Hopkins University, Baltimore, MD, USA) (Terman et al., 2002), ppk-Gal4 on II and III (Y. Jan, University of California, San Francisco, CA, USA) (Grueber et al., 2003), SOP-flp (a kind gift from T. Uemura, Kyoto University, Japan), UAS-Sec31-mCherry (S. Luschig, University of Münster, Germany) (Förster et al., 2010), UAS-Arf1-GFP (T. J. Harris, University of Toronto, Canada) (Shao et al., 2010), UAS-NYFP-myc and UAS-CYFP-HA (S. Bogdan, University of Münster, Germany) (Gohl et al., 2010), UAS-Sec15-GFP (H. Bellen, Baylor College of Medicine, Houston, TX, USA) (Jafar-Nejad et al., 2005), and UAS-ManII-Venus and UAS-Sar1T34N (F.Y.’s lab).

The stocks from Bloomington Stock Center (BSC) are listed below: Df(3R)ED6232 (BL #8105), Df(3R)BSC524 (BL #25052), Df(3R)BSC752 (BL #26850), P{EP}CG5484G19213 (BL #31845), Gal4109(2)80 (BL #8769), UAS-GalT-GFP (BL #30902), UAS-Mito-HA-GFP (BL #8442), UAS-Rab4-mRFP (BL #8505), nanos-Cas9 (BL #54591), elav-Gal4, GsG2295-Gal4 (BL #40266), ppk-CD4-tdGFP (BL #35842 and BL #35843), UAS-YFP-Rab1S25N (BL #23236 and BL #9757), Rab1e01287 (BL #17936), Rab1 RNAi #1 (BL #27299), Sar111-3-63 (BL #53710), Nrg RNAi #1 (BL #38215) and Nrg RNAi #1 (BL #37496).

The stocks from Vienna Drosophila RNAi Center (VDRC) were Sar1 RNAi #1 (v34192) and Sar1 RNAi #2 (v108458).

The stock from FlyORF centre used was UAS-Sar1-ORF (#F001450).

For genotypes of the fly strains, please see the supplementary information for further details.

EMS mutagenesis

We used the standard procedure (Lewis and Bacher, 1968) to achieve the EMS-inducible mutagenesis with a concentration of 25 mM on isogenized third chromosome. Approximately 550 stable stocks carrying lethal mutations on the third chromosome were isolated and screened.

Generation of Yif1 mutants

The P{EP}CG5484G19213 transposon line was crossed with a Δ2-3 transposase expressing fly strain to induce imprecise excision. Thirty lethal lines were recovered and the genomic deletions were examined by PCR and DNA sequencing. Two imprecise excision lines Yif1Δ25 and Yif1Δ56 carry 2814 bp and 2306 bp deletions, respectively (Fig. S2C).

Generation of Yip1 mutants via CRISPR/Cas9 system

We used a standard procedure (Port et al., 2014) to clone the guide RNA sequences into the pCFD3 vector. The cloning primers used were: Yip11-21, 5′-GTCGCTTTTACGGCTCCGCGCCCT-3′ and 5′-AAACAGGGCGCGGAGCCGTAAAAG-3′; Yip12-20, 5′-GTCGGTTGTAGCTGGCATCGGCCG-3′ and 5′-AAACCGGCCGATGCCAGCTACAAC-3′. The pCFD3-gRNA vectors were inserted into the attP2 landing site via φC31 integration (BestGene). The resultant transgenic flies were crossed with a nanos-Cas9 fly strain (BL #54591) to generate Yip1 mutants. Their molecular lesions were determined by PCR and DNA sequencing.

Generation of transgenes and transgenic flies

UAS-Yif1

The full-length cDNA fragment of Yif1 was amplified from FI08032 (Drosophila Genomics Resource Center) with the following primers: 5′-ATGAACTACAATCCAAATCCG-3′ and 5′-GAAGGTCTTGGGCACCGTCAG-3′. The PCR product was cloned into pEntry (Invitrogen) and then inserted into pTWV (the Gateway System) or pUAST-NYFP vectors (Gohl et al., 2010) via the LR reaction (Invitrogen). The pUAST-NYFP-Yif1 was inserted into the attP2 landing site via φC31 integration.

UAS-Yip1

The full-length cDNA of Yip1 was amplified from RH67967 (Drosophila Genomics Resource Center) with the following primers: 5′-ATGTCTCAGTTCGGCGGACCC-3′ and 5′-GTAAATGGTAATCAACGCAAA-3′. The PCR product was cloned into pEntry and inserted into the pUAST-CYFP vector (Gohl et al., 2010) via the LR reaction. The pUAST-CYFP-Yip1 was inserted into the attP2 landing site via φC31 integration.

UAS-Yif1 truncated forms

The cDNA fragments corresponding to amino acids 1-195, 196-393, 1-323 and 91-393 were amplified from FI08032 and cloned into pTWV via the LR reaction. Transgenic lines were established by Bestgene.

Expression vectors in S2 cells

pAFW-Flag-Yif1 and pUAST-NYFP-Yif1 were constructed from pEntry-Yif1 via LR reaction. pAMW-myc-Yip1 and pUAST-CYFP-Yip1 were constructed from pEntry-Yip1 via LR reaction.

Generation and imaging of ddaC MARCM clones

We performed MARCM clonal analysis, dendrite imaging and quantification as before (Kirilly et al., 2009). Briefly, ddaC clones were screened and imaged at the WP stage for the dendrite arbor morphology. Then the same neurons were checked for dendrite pruning defects at 16 h APF.

RU486/mifepristone treatment for the gene-switch system

We used the gene-switch system to pulse induce expression of transgenes. Embryos with different genotypes were collected at 6 h intervals and cultivated on standard food until third instar larval stage. The larvae were then collected and transferred into mifepristone-containing medium (240 μg/ml, Sigma-Aldrich, M8046).

Yip1 antibody production

The cDNA fragment corresponding to the 5-110 amino acid fragment of Yip1 was amplified from RH67967 and cloned into the MBP expression vector (pMAL, NEB). Yip1 antibodies were generated by immunizing guinea pigs and mice with the MBP-Yip1 fusion protein.

Immunohistochemistry and antibodies

Rabbit monoclonal anti-GM130 (1:100, Abcam, ab52649), rabbit anti-Rab1 (1:250, F.Y.’s lab), guinea pig anti-Arf1 (1:250, F.Y.’s lab), rat anti-HA (1:200, Roche, 11867423001), rabbit anti-Myc (1:1000, Santa Cruz Biotechnology, A-14), guinea pig anti-GFP (1:1000, F.Y.’s lab), mouse anti-Yip1 (1:200, F.Y.’s lab), mouse monoclonal anti-Nrg (1:25, BP104, Developmental Studies Hybridoma Bank) and rat monoclonal anti-mCD8 (1:300, CATLOG Laboratories) were used. FITC, Cy3- or Cy5-conjugated secondary antibodies were used at 1:500 (Jackson Laboratories). For immunostaining, larvae or pupae were dissected in ice-cold PBS and fixed in 4% formaldehyde for 15 min. The samples within the same group of experiments were stained in the same tube and mounted in VectaShield mounting medium and imaged with Leica SPEII or SP8 using the same microscopy setting and processed in parallel. Data analysis and statistics were performed using Excel (Microsoft) and Imaris software.

Live imaging

For each genotype, multiple larvae were imaged. Larvae were washed briefly in 1× PBS before mounting in halocarbon oil for live imaging of ddaC neurons. Larvae were imaged for average of 5-15 min. The images within the same group of experiments were acquired using a Leica SP8 confocal microscope at the same setting. Data were processed in parallel and statistics was performed using Excel and Imaris software.

Cell culture, transfection and co-immunoprecipitation (co-IP)

Drosophila S2 cells were maintained in the Express Five serum free medium (SFM) with 10% L-glutamate and grown at 25°C. For co-IP, S2 cells were transfected with pAFW-Flag-Yif1 and/or pAMW-myc-Yip1 using Effectene Transfection Reagent. 48 h after transfection, cells were homogenized by the lysis buffer [25 mM Tris pH 8, 27.5 mM NaCl, 20 mM KCl, 25 mM sucrose, 1 mM DTT, 10%(v/v) glycerol, 0.5% NP40 and protease inhibitors]. The cell extracts were incubated with anti-Myc (1: 1000, ab32, Abcam) overnight at 4°C, followed by incubation of Protein A/G beads (Pierce Chemical Co.). A/G beads were washed three times with cold PBS. Protein samples were resolved by SDS-PAGE gels and analyzed with anti-Flag, anti-Myc HRP-conjugated antibody in western blotting. Each co-IP assay was repeated three times.

BiFC assay

For BiFC assays in S2 cells, cells were co-transfected with pUAST-NYFP and pUAST-CYFP, pUAST-NYFP-Yif1 and pUAST-CYFP, pUAST-NYFP and pUAST-CYFP-Yip1, or pUAST-NYFP-Yif1 and pUAST-CYFP-Yip1 using Effectene Transfection Reagent (Qiagen). Forty-eight hours after transfection, cells were fixed in 4% formaldehyde for 15 min and followed by immunostaining. The samples were mounted in VectaShield mounting medium and directly imaged with a Leica SP8 confocal microscope with the same setting.

For BiFC assays in fixed larval tissues and live larvae, transgenic animals carrying pUAST-NYFP and pUAST-CYFP, pUAST-NYFP-Yif1 and pUAST-CYFP, pUAST-NYFP and pUAST-CYFP-Yip1, and pUAST-NYFP-Yif1 and pUAST-CYFP-Yip1 were collected and fixed in 4% formaldehyde for 15 min and followed by immunostaining with GM130. The samples were mounted in VectaShield mounting medium and directly imaged with the confocal microscopy Leica SP8 with the same setting. Live transgenic larvae were collected and washed for direct detection of the BiFC YFP fluorescence using a Leica SP8 confocal microscope with the same setting.

Trafficking assay in ddaC neurons

The trafficking assays were conducted according to a previous report (Murthy et al., 2003). wL3 larvae of wild type or mutant were dissected and fixed in 4% formaldehyde for 15 min. The fillets were incubated with rat anti-CD8 (1:300, CALTAG Laboratories) overnight and washed three times with PBS in a detergent-free condition. The fillets were then subject to incubation with secondary antibody for 3 h. The samples were mounted in VectaShield mounting medium. The intensity of immunofluorescence was measured with the same confocal setting for both control and mutants.

Quantification and statistical analysis

For ddaC dendrites, live confocal images of ddaC neurons labelled by UAS-mCD8-GFP driven by ppk-Gal4 were captured at WP and 16 h APF stages. Both wild-type and mutant ddaC neurons were counted for the average numbers of primary and secondary dendrites attached to the soma. The number of neurons (n) in every group is shown in the bars. We used either two-tailed Student's t-test (two samples) or one-way ANOVA and Bonferroni test (multiple samples) to determine the statistical significance (*P<0.05, **P<0.01, ***P<0.001, N.S., not significant). Error bar represents s.e.m. Dorsal is upwards in all images.

For Nrg immunostaining, the intensity of Nrg was quantified at 6 h APF as shown previously (Zhang et al., 2014). Briefly, the boundaries of the soma/dendrite/axon of ddaC and ddaE neurons were outlined according to the GFP channel in ImageJ. The mean intensity values of soma/dendrites/axons in ddaC and ddaE neurons were measured after background subtraction (rolling ball radius=30). The intensity index in ddaC neuron is calculated using the ddaE neuron as an internal control.

For trafficking assay quantification, the boundaries of the primary dendrites were outlined in the mCD8 staining channel according to the GFP channel in ImageJ. The mean intensity values of mCD8 and GFP were determined after background reduction (rolling ball radius=30). The fluorescent intensity ratio in wild-type or mutant ddaC neurons was determined by calculating the ratio of mCD8 and GFP. The number of samples (n) in every group is shown in the bars. We used one-way ANOVA and Bonferroni test to determine the statistical significance (*P<0.05, **P<0.01, ***P<0.001, N.S., not significant). Error bars represent s.e.m.

Acknowledgements

We thank H. Bellen, S. Bogdan, M. Gonzalez-Gaitan, T. Harris, Y. Jan, A. Kolodkin, S. Luschnig, T. Uemura, the Bloomington Stock Center (BSC), DSHB (University of Iowa), Kyoto Stock Center (Japan) and VDRC (Austria) for generously providing antibodies and fly stocks. We thank members of the Yu laboratory for helpful assistance and discussion.

Footnotes

Author contributions

Conceptualization: F.Y.; Methodology: Q.W., Y.W., F.Y.; Software: Q.W.; Validation: Q.W., F.Y.; Formal analysis: Q.W., Y.W., F.Y.; Investigation: Q.W., Y.W.; Resources: Q.W., F.Y.; Data curation: Q.W., Y.W.; Writing - original draft: Q.W., F.Y.; Writing - review & editing: F.Y.; Visualization: Q.W., Y.W.; Supervision: F.Y.; Project administration: F.Y.; Funding acquisition: F.Y.

Funding

This work was supported by Temasek Life Sciences Laboratory (TLL-2040), Singapore (F.Y.).

References

Al Awabdh
,
S.
,
Miserey-Lenkei
,
S.
,
Bouceba
,
T.
,
Masson
,
J.
,
Kano
,
F.
,
Marinach-Patrice
,
C.
,
Hamon
,
M.
,
Emerit
,
M. B.
and
Darmon
,
M.
(
2012
).
A new vesicular scaffolding complex mediates the G-protein-coupled 5-HT1A receptor targeting to neuronal dendrites
.
J. Neurosci.
32
,
14227
-
14241
.
Alterio
,
J.
,
Masson
,
J.
,
Diaz
,
J.
,
Chachlaki
,
K.
,
Salman
,
H.
,
Areias
,
J.
,
Al Awabdh
,
S.
,
Emerit
,
M. B.
and
Darmon
,
M.
(
2015
).
Yif1B is involved in the anterograde traffic pathway and the Golgi architecture
.
Traffic
16
,
978
-
993
.
Aridor
,
M.
and
Fish
,
K. N.
(
2009
).
Selective targeting of ER exit sites supports axon development
.
Traffic
10
,
1669
-
1684
.
Barrowman
,
J.
,
Wang
,
W.
,
Zhang
,
Y.
and
Ferro-Novick
,
S.
(
2003
).
The Yip1p.Yif1p complex is required for the fusion competence of endoplasmic reticulum-derived vesicles
.
J. Biol. Chem.
278
,
19878
-
19884
.
Carrel
,
D.
,
Masson
,
J.
,
Al Awabdh
,
S.
,
Capra
,
C. B.
,
Lenkei
,
Z.
,
Hamon
,
M.
,
Emerit
,
M. B.
and
Darmon
,
M.
(
2008
).
Targeting of the 5-HT1A serotonin receptor to neuronal dendrites is mediated by Yif1B
.
J. Neurosci.
28
,
8063
-
8073
.
D'Ambrosio
,
M. V.
and
Vale
,
R. D.
(
2010
).
A whole genome RNAi screen of Drosophila S2 cell spreading performed using automated computational image analysis
.
J. Cell Biol.
191
,
471
-
478
.
Förster
,
D.
,
Armbruster
,
K.
and
Luschnig
,
S.
(
2010
).
Sec24-dependent secretion drives cell-autonomous expansion of tracheal tubes in Drosophila
.
Curr. Biol.
20
,
62
-
68
.
Gohl
,
C.
,
Banovic
,
D.
,
Grevelhörster
,
A.
and
Bogdan
,
S.
(
2010
).
WAVE forms hetero- and homo-oligomeric complexes at integrin junctions in Drosophila visualized by bimolecular fluorescence complementation
.
J. Biol. Chem.
285
,
40171
-
40179
.
Grueber
,
W. B.
,
Jan
,
L. Y.
and
Jan
,
Y. N.
(
2003
).
Different levels of the homeodomain protein cut regulate distinct dendrite branching patterns of Drosophila multidendritic neurons
.
Cell
112
,
805
-
818
.
Horton
,
A. C.
,
Rácz
,
B.
,
Monson
,
E. E.
,
Lin
,
A. L.
,
Weinberg
,
R. J.
and
Ehlers
,
M. D.
(
2005
).
Polarized secretory trafficking directs cargo for asymmetric dendrite growth and morphogenesis
.
Neuron
48
,
757
-
771
.
Ivan
,
V.
,
de Voer
,
G.
,
Xanthakis
,
D.
,
Spoorendonk
,
K. M.
,
Kondylis
,
V.
and
Rabouille
,
C.
(
2008
).
Drosophila Sec16 mediates the biogenesis of tER sites upstream of Sar1 through an arginine-rich motif
.
Mol. Biol. Cell
19
,
4352
-
4365
.
Jafar-Nejad
,
H.
,
Andrews
,
H. K.
,
Acar
,
M.
,
Bayat
,
V.
,
Wirtz-Peitz
,
F.
,
Mehta
,
S. Q.
,
Knoblich
,
J. A.
and
Bellen
,
H. J.
(
2005
).
Sec15, a component of the exocyst, promotes notch signaling during the asymmetric division of Drosophila sensory organ precursors
.
Dev. Cell
9
,
351
-
363
.
Kanamori
,
T.
,
Togashi
,
K.
,
Koizumi
,
H.
and
Emoto
,
K.
(
2015a
).
Dendritic remodeling: lessons from invertebrate model systems
.
Int. Rev. Cell Mol. Biol.
318
,
1
-
25
.
Kanamori
,
T.
,
Yoshino
,
J.
,
Yasunaga
,
K.-I.
,
Dairyo
,
Y.
and
Emoto
,
K.
(
2015b
).
Local endocytosis triggers dendritic thinning and pruning in Drosophila sensory neurons
.
Nat. Commun.
6
,
6515
.
Kennedy
,
M. J.
and
Ehlers
,
M. D.
(
2011
).
Mechanisms and function of dendritic exocytosis
.
Neuron
69
,
856
-
875
.
Kirilly
,
D.
,
Gu
,
Y.
,
Huang
,
Y.
,
Wu
,
Z.
,
Bashirullah
,
A.
,
Low
,
B. C.
,
Kolodkin
,
A. L.
,
Wang
,
H.
and
Yu
,
F.
(
2009
).
A genetic pathway composed of Sox14 and Mical governs severing of dendrites during pruning
.
Nat. Neurosci.
12
,
1497
-
1505
.
Kuo
,
C. T.
,
Jan
,
L. Y.
and
Jan
,
Y. N.
(
2005
).
Dendrite-specific remodeling of Drosophila sensory neurons requires matrix metalloproteases, ubiquitin-proteasome, and ecdysone signaling
.
Proc. Natl. Acad. Sci. USA
102
,
15230
-
15235
.
Kurusu
,
M.
,
Cording
,
A.
,
Taniguchi
,
M.
,
Menon
,
K.
,
Suzuki
,
E.
and
Zinn
,
K.
(
2008
).
A screen of cell-surface molecules identifies leucine-rich repeat proteins as key mediators of synaptic target selection
.
Neuron
59
,
972
-
985
.
Lee
,
T.
and
Luo
,
L.
(
1999
).
Mosaic analysis with a repressible cell marker for studies of gene function in neuronal morphogenesis
.
Neuron
22
,
451
-
461
.
Lee
,
T.
,
Lee
,
A.
and
Luo
,
L.
(
1999
).
Development of the Drosophila mushroom bodies: sequential generation of three distinct types of neurons from a neuroblast
.
Development
126
,
4065
-
4076
.
Lee
,
M. C. S.
,
Miller
,
E. A.
,
Goldberg
,
J.
,
Orci
,
L.
and
Schekman
,
R.
(
2004
).
Bi-directional protein transport between the ER and Golgi
.
Annu. Rev. Cell Dev. Biol.
20
,
87
-
123
.
Lewis
,
E. B.
and
Bacher
,
F.
(
1968
).
Method of feeding ethyl methanesulfonate (EMS) to Drosophila males
.
Dros. Inf. Serv.
43
,
193
.
Luo
,
L.
and
O'Leary
,
D. D. M.
(
2005
).
Axon retraction and degeneration in development and disease
.
Annu. Rev. Neurosci.
28
,
127
-
156
.
Matern
,
H.
,
Yang
,
X.
,
Andrulis
,
E.
,
Sternglanz
,
R.
,
Trepte
,
H. H.
and
Gallwitz
,
D.
(
2000
).
A novel Golgi membrane protein is part of a GTPase-binding protein complex involved in vesicle targeting
.
EMBO J.
19
,
4485
-
4492
.
Murthy
,
M.
,
Garza
,
D.
,
Scheller
,
R. H.
and
Schwarz
,
T. L.
(
2003
).
Mutations in the exocyst component Sec5 disrupt neuronal membrane traffic, but neurotransmitter release persists
.
Neuron
37
,
433
-
447
.
Nuoffer
,
C.
,
Davidson
,
H. W.
,
Matteson
,
J.
,
Meinkoth
,
J.
and
Balch
,
W. E.
(
1994
).
A GDP-bound of rab1 inhibits protein export from the endoplasmic reticulum and transport between Golgi compartments
.
J. Cell Biol.
125
,
225
-
237
.
Port
,
F.
,
Chen
,
H.-M.
,
Lee
,
T.
and
Bullock
,
S. L.
(
2014
).
Optimized CRISPR/Cas tools for efficient germline and somatic genome engineering in Drosophila
.
Proc. Natl. Acad. Sci. USA
111
,
E2967
-
E2976
.
Rothman
,
J. E.
and
Orci
,
L.
(
1992
).
Molecular dissection of the secretory pathway
.
Nature
355
,
409
-
415
.
Schuldiner
,
O.
and
Yaron
,
A.
(
2015
).
Mechanisms of developmental neurite pruning
.
Cell. Mol. Life Sci.
72
,
101
-
119
.
Sepp
,
K. J.
,
Hong
,
P.
,
Lizarraga
,
S. B.
,
Liu
,
J. S.
,
Mejia
,
L. A.
,
Walsh
,
C. A.
and
Perrimon
,
N.
(
2008
).
Identification of neural outgrowth genes using genome-wide RNAi
.
PLoS Genet.
4
,
e1000111
.
Shao
,
W.
,
Wu
,
J.
,
Chen
,
J.
,
Lee
,
D. M.
,
Tishkina
,
A.
and
Harris
,
T. J. C.
(
2010
).
A modifier screen for Bazooka/PAR-3 interacting genes in the Drosophila embryo epithelium
.
PLoS ONE
5
,
e9938
.
Stanfield
,
B. B.
,
O'Leary
,
D. D. M.
and
Fricks
,
C.
(
1982
).
Selective collateral elimination in early postnatal development restricts cortical distribution of rat pyramidal tract neurones
.
Nature
298
,
371
-
373
.
Storrie
,
B.
,
White
,
J.
,
Röttger
,
S.
,
Stelzer
,
E. H. K.
,
Suganuma
,
T.
and
Nilsson
,
T.
(
1998
).
Recycling of golgi-resident glycosyltransferases through the ER reveals a novel pathway and provides an explanation for nocodazole-induced Golgi scattering
.
J. Cell Biol.
143
,
1505
-
1521
.
Tang
,
B. L.
,
Ong
,
Y. S.
,
Huang
,
B.
,
Wei
,
S.
,
Wong
,
E. T.
,
Qi
,
R.
,
Horstmann
,
H.
and
Hong
,
W.
(
2001
).
A membrane protein enriched in endoplasmic reticulum exit sites interacts with COPII
.
J. Biol. Chem.
276
,
40008
-
40017
.
Terman
,
J. R.
,
Mao
,
T.
,
Pasterkamp
,
R. J.
,
Yu
,
H.-H.
and
Kolodkin
,
A. L.
(
2002
).
MICALs, a family of conserved flavoprotein oxidoreductases, function in plexin-mediated axonal repulsion
.
Cell
109
,
887
-
900
.
Truman
,
J. W.
(
1990
).
Metamorphosis of the central nervous system of Drosophila
.
J. Neurobiol.
21
,
1072
-
1084
.
Valenzuela
,
J. I.
and
Perez
,
F.
(
2015
).
Diversifying the secretory routes in neurons
.
Front. Neurosci.
9
,
358
.
Wang
,
Y.
,
Zhang
,
H.
,
Shi
,
M.
,
Liou
,
Y.-C.
,
Lu
,
L.
and
Yu
,
F.
(
2017
).
Sec71 functions as a GEF for the small GTPase Arf1 to govern dendrite pruning of Drosophila sensory neurons
.
Development
144
,
1851
-
1862
.
Weide
,
T.
,
Bayer
,
M.
,
Köster
,
M.
,
Siebrasse
,
J.-P.
,
Peters
,
R.
and
Barnekow
,
A.
(
2001
).
The Golgi matrix protein GM130: a specific interacting partner of the small GTPase rab1b
.
EMBO Rep.
2
,
336
-
341
.
Williams
,
D. W.
and
Truman
,
J. W.
(
2005
).
Cellular mechanisms of dendrite pruning in Drosophila: insights from in vivo time-lapse of remodeling dendritic arborizing sensory neurons
.
Development
132
,
3631
-
3642
.
Wilson
,
B. S.
,
Nuoffer
,
C.
,
Meinkoth
,
J. L.
,
McCaffery
,
M.
,
Feramisco
,
J. R.
,
Balch
,
W. E.
and
Farquhar
,
M. G.
(
1994
).
A Rab1 mutant affecting guanine nucleotide exchange promotes disassembly of the Golgi apparatus
.
J. Cell Biol.
125
,
557
-
571
.
Wucherpfennig
,
T.
,
Wilsch-Bräuninger
,
M.
and
González-Gaitán
,
M.
(
2003
).
Role of Drosophila Rab5 during endosomal trafficking at the synapse and evoked neurotransmitter release
.
J. Cell Biol.
161
,
609
-
624
.
Yang
,
X.
,
Matern
,
H. T.
and
Gallwitz
,
D.
(
1998
).
Specific binding to a novel and essential Golgi membrane protein (Yip1p) functionally links the transport GTPases Ypt1p and Ypt31p
.
EMBO J.
17
,
4954
-
4963
.
Yaniv
,
S. P.
and
Schuldiner
,
O.
(
2016
).
A fly's view of neuronal remodeling
.
Wiley Interdiscip. Rev. Dev. Biol.
5
,
618
-
635
.
Ye
,
B.
,
Zhang
,
Y.
,
Song
,
W.
,
Younger
,
S. H.
,
Jan
,
L. Y.
and
Jan
,
Y. N.
(
2007
).
Growing dendrites and axons differ in their reliance on the secretory pathway
.
Cell
130
,
717
-
729
.
Yoshida
,
Y.
,
Suzuki
,
K.
,
Yamamoto
,
A.
,
Sakai
,
N.
,
Bando
,
M.
,
Tanimoto
,
K.
,
Yamaguchi
,
Y.
,
Sakaguchi
,
T.
,
Akhter
,
H.
,
Fujii
,
G.
, et al. 
(
2008
).
YIPF5 and YIF1A recycle between the ER and the Golgi apparatus and are involved in the maintenance of the Golgi structure
.
Exp. Cell Res.
314
,
3427
-
3443
.
Yu
,
F.
and
Schuldiner
,
O.
(
2014
).
Axon and dendrite pruning in Drosophila
.
Curr. Opin. Neurobiol.
27
,
192
-
198
.
Zhang
,
H.
,
Wang
,
Y.
,
Wong
,
J. J. L.
,
Lim
,
K.-L.
,
Liou
,
Y.-C.
,
Wang
,
H.
and
Yu
,
F.
(
2014
).
Endocytic pathways downregulate the L1-type cell adhesion molecule neuroglian to promote dendrite pruning in Drosophila
.
Dev. Cell
30
,
463
-
478
.

Competing interests

The authors declare no competing or financial interests.

Supplementary information